Medical Radionuclide Imaging 1980 л P R O C E E D I N G S O F O R G A N I Z E D
BY
IA E A
HEIDELBERG.
A S Y M P O S I U M
Ф )
IN C O - O P E R A T I O N W I T H W H O 1-5 S E P T E M B E R
I9S0
W
'
Vol. I INT ER NA TIO NAL A T O M IC ENERGY AGENCY, VIENN A, 1981
MEDICAL RADIONUCLIDE IMAGING 1980
VOL. I
PROCEEDINGS SERIES
MEDICAL RADIONUCLIDE IMAGING 1980 /
PROCEEDINGS OF AN INTERNATIONAL SYMPOSIUM ON MEDICAL RADIONUCLIDE IMAGING ORGANIZED BY THE INTERNATIONAL ATOMIC ENERGY AGENCY IN CO-OPERATION WITH THE WORLD HEALTH ORGANIZATION AND HELD IN HEIDELBERG, 1 -5 SEPTEMBER 1980
In two volumes
VOL. I
INTERNATIONAL ATOMIC ENERGY AGENCY VIENNA, 1981
MEDICAL RADIONUCLIDE IMAGING 1980 IAEA, VIENNA, 1981 STI/PUB/564 ISBN 9 2 -0 - 0 1 0 0 8 1 -3
©
IA EA, 1981
Permission to reproduce o r translate the inform ation contained in this publication may be obtained by writing to the International Atomic Energy Agency, Wagramerstrasse 5, P.O. Box 100, A -1400 V ienna, Austria.
P rin ted by th e IA EÀ in A ustria Ju ly 1981
FOREWORD
The International Symposium on Medical Radionuclide Imaging organized by the International Atomic Energy Agency (IAEA) in co-operation with the World Health Organization (WHO) in Heidelberg, Federal Republic of Germany, from 1 to 5 September 1980, was the sixth of the IAEA’s symposia in the subject field, its predecessors having been held in Vienna (1959), Athens (1964), Salzburg (1968), Monte Carlo (1972) and Los Angeles (1976). The Agency is indebted to the Government of the Federal Republic of Germany for the invita tion to host the meeting, and is particularly grateful to the staff of the Institute for Nuclear Medicine, German Cancer Research Center, Heidelberg, for their substantial support, hospitality and co-operation. The symposium was attended by 332 participants and 61 observers from 36 countries and six international organizations. In the course of ten scientific sessions, a total of 92 papers were presented orally, these including 20 invited reviews covering all major aspects of the subject. A further 33 papers were presented in poster form. The programme also included two special events, a Joint IAEA/WHO Round Table on Quality Assurance in Radionuclide Imaging and a Round Table on Radionuclide Imaging in Relation to Other Imaging Modalities. Medical radionuclide imaging has made fresh advances during the four years that have elapsed since the Los Angeles meeting. Indeed, this modality no w d o m in a te s d iagnostic n u clea r m edicine. Its scope has b een f u rth e r ex ten d e d
through the continued elaboration of methodologies based on gamma camera/data-processing system combinations, both for dynamic studies and for tomographic sectional imaging. Additional progress has resulted from the development of new radiopharmaceuticals for imaging particular organs and tissues. There has been growing recognition that the optimum utilization of radionuclide imaging facilities requires appropriate quality assurance procedures. The position of radionuclide imaging has been increasingly challenged through the emergence o f other imaging modalities, notably computed tomo graphy (CT), using X-rays, and ultrasonography. To these must now be added nuclear magnetic resonance (NMR) imaging. It is becoming appreciated, how ever, that radionuclide imaging, depending as it does on the accumulation of the radiopharmaceutical in the organ or tissue of interest, provides particular insight into the biochemical, metabolic and physio-pathological processes within.
The symposium provided opportunities for exchange of information on all these aspects. The proceedings now published contain the full texts of all papers presented orally at the meeting, including invited reviews, summaries of papers presented in poster form and an edited record of the discussions. Volume I covers those parts of the meeting dealing with fundamental aspects, instrum entation and techniques, information processing and display, and radio pharmaceuticals. Volume II covers those dealing with quality assurance and all clinical applications, and includes the two round tables. It is hoped that the two volumes will constitute a useful guide to the present status of the subject.
EDITORIAL N O T E The papers and discussions have been edited by the e d ito ria l s ta ff o f the In te rn a tio n a l A to m ic Energy Agency to the extent considered necessary f o r the reader's assistance. The views expressed and the general style adopted remain, however, the responsibility o f the named authors o r participants. In a ddition, the views are n o t necessarily those o f the governments o f the n o m inating M em ber States o r o f the nom inating organizations. Where papers have been incorporated in to these Proceedings w ith o u t resetting by the Agency, this has been done w ith the knowledge o f the authors and th e ir governm ent authorities, and th e ir cooperation is g ra te fu lly acknowledged. The Proceedings have been p rin te d by com position typ in g and p h o to -o ffse t lithography. W ithin the lim ita tio n s imposed by this method, every e ffo rt has been made to m aintain a high e d ito ria l standard, in p a rticu la r to achieve, wherever practicable, consistency o f units and sym bols and co n fo rm ity to the standards recom m ended by com petent in te rn a tio n a l bodies. The use in these Proceedings o f p a rticu la r designations o f countries o r te rrito ries does n o t im p ly any judgem ent by the publisher, the IA E A , as to the legal status o f such countries o r territories, o f th e ir authorities and in stitu tio n s o r o f the d e lim ita tio n o f th e ir boundaries. The m ention o f specific companies o r o f th e ir products o r brand names does n o t im p ly any endorsement o r recom m endation on the p a rt o f the IA E A . A u th o rs are themselves responsible f o r ob taining the necessary permission to reproduce c o p yrig h t m aterial fro m o th er sources.
/
CONTENTS OF VOL. I FUNDAMENTAL ASPECTS (Session 1) Invited review paper: New perspectives in nuclear imaging (IAEA-SM -247/200)
............................
3
.............................................................................................................
26
H. N. Wagner, Jr. Discussion
INSTRUMENTATION AND TECHNIQUES (Sessions 1 (cont.)and 2) A multicrystal positron scanner for quantitative studies with positronemitting radionuclides (IAEA-SM-247/37) ....................................................
29
H. Ostertag, W. Kübler, R. Kubesch, W.J. Lorenz A hybrid scanner for positron imaging (IAEA-SM-247/4)
............................
41
A. Markkula, E. Vauramo, A. Virjo, P. Ahonen, H. Vilén, J.C.W. Crawley, N. Veall, M. Keinänen, J. Kulmala, V. Näntö, H. Wendelin A large-area stationary positron camera using wire chambers (IAEA-SM-247 /65 ) ...............................................................................................
49
A. Jeavons, B. Schorr, K. Kuli, D. Townsend, P. Frey, A. Donath A radioisotope scanner for use in developing countries (I AEA-SM-247/167) ..........................................................................................
73
J.C.W. Crawley, A.B. Ajdukiewicz, N. Bassett, A.G. Cronquist, N. Veall Investigation o f long-term in vivo tracer distribution patterns using an ultra-high-sensitivity scanning system (IAEA-SM -247/114) ..................
83
D. Gvozdanovic, K. V. Ettinger, D.B. Smith, C.G. Taylor, S. Gvozdanovic, J.R. Mallard Discussion ............................................................................................................. A germanium rotating laminar emission camera (ROLEC) (IAEA-SM-247/84) ...............................................................................................
105 107
W. Mauderli, M.M. Uñe, L.T. Fitzgerald Discussion ............................................................................................................. Invited review paper: Medical imaging by nuclear magnetic resonance: A review o f the Aberdeen physical and biological programme(IAEA-SM-247/201) ....
116
117
JR . Mallard, J.M.S. Hutchison, M.A. Foster, W.A. Edelstein, C.R. Ling, F.W. Smith, A. Reid, R. Selbie, G. Johnson, T.W. Redpath Discussion
.............................................................................................................
142
P oster presentations:
X-ray fluorescence tomography (IAEA-SM-247/78)
.....................................
145
J.A. Patton, R.R. Price, D.R. Prickens, F.D. Rollo, C.L. Partain, A.B. Brill La tomographie pulmonaire par diffusion Compton (IAEA-SM-247/150) ..
146
J.L. Moretti, E. Mathieu, J.-F. Cavellier, G. Roux Gamma-gamma-coincidence scintigraphy: Tomography without computerized image reconstruction (IAEA-SM-2 47/40) ...........................
147
H. von Boetticher, H. Helmers, E.-M. Muschol, P. Schreiber, I. Schmitz-Feuerhake Theoretical and experimental investigations o f 3-D imaging with complex coded apertures (IAEA-SM -247/137) .............................................................
148
E.R. Reinhardt, G. Laub, W. Müller-Schauenburg Discussion ............................................................................................................. Evaluation o f imaging properties o f planar stochastic coded apertures (IAEA-SM '247/44) ...............................................................................................
150 150
J. Kouris, N.M. Spyrou, D.F. Jackson HEADTOME: A hybrid emission tomograph for brain: Design concepts and preliminary results (IAEA-SM -247/13) ...............................................
153
I. Kanno, K. Uemura, Y. Miura, S. Miura, Y. Hirose, K. Koga, H. Hattori Rotational positron computed tomographs (IAEA-SM-24 7 /54) ..................
165
E. Tanaka, N. Nohara, M. Yamamoto, T. Tomitani, H. Murayama, Y. Tateno, K. Ishimatsu, K. Takami Invited review paper: Single photon imaging: New instrumentation and techniques (IAEA-SM-247/202) ..........................................................................................
173
G. Muehllehner, J. Colsher Discussion ............................................................................................................. Invited review paper: Positron computed tomography: Present and future design alternatives (IAEA-SM-24 7/203) ..........................................................................................
197
199
M.E. Phelps, E.J. Hoffman, Sung-Cheng Huang, D.E. Kuhl Discussion ............................................................................................................. Potential and limits o f quantitative studies in emission tomography (IAEA-SM-247/139) ..........................................................................................
F. Soussaline, A.E. Todd-Pokropek, D. Comar, C. Raynaud, C. Kellershohn
229 231
SPET I: Figures o f merit for two multiple detector (single slice) and one area detector (multiple slice) single photon emission tomographic instruments (IAEA-SM -247/31) .......................................................................
243
PH. Jarritt, I.D. Cullum, P.J. Ell Discussion ............................................................................................................. SPET II: The clinical role o f single photon emission tomography (IAEA-SM -243/32) ...............................................................................................
253 255
P.J. Ell, O. Khan, P.H. Jarritt, R.G. Radia Discussion ............................................................................................................. Clinical results o f quantitative single photon emission tomography (IAEA-SM-247/25) ..............................................................................................
261 263
K.E. Britton, B. Shapiro, A. T. Elliott Discussion ............................................................................................................. Scintillation camera radionuclide imaging with low-energy isotopes .......................................... (iodine-125, caesium-131) (IAEA-SM-247/61)
269 271
G. Endert, E. Klose, E. Schumann Poster presentations: Design o f a fast positron camera head (IAEA-SM -247/134) ....................... C. Nahmias, D.B. Kenyon, E.S. Garnett, K. Kouris Interdependence o f dead-time losses and camera inhomogeneity (IAEA-SM-247/42) ...............................................................................................
281
282
D. Lange, H.J. Hermann Discussion ............................................................................................................. Automatic alignment o f radionuclide images (IAEA-SM-247/7) ..................
283 283
D.C. Barber INFORMATION PROCESSING AND DISPLAY (Session 3) Invited review paper: New developments in techniques for information processing in radionuclide imaging (IAEA-SM -247/204) ....................................................
287
R. Di Paola, A.E. Todd-Pokropek Maximum entropy reconstructions in emission tomography (IAEA-SM -247/128) ..........................................................................................
313
M.C. Kemp Discussion ............................................................................................................. Longitudinal and transverse digital image reconstruction with a tomographic scanner (IAEA-SM -247/76) ....................................................
323 325
D R. Pickens, R.R. Price, J.J. Erickson, J.A. Patton, C.L. Partain, F.D. Rollo Discussion
.............................................................................................................
333
Development o f the computerized Anger tomographic scanner for positron imaging (IAEA-SM -247/127) ........................................................
335
E.A. Silverstein, E.W. Fordham, A. Chung-Bin, T. Wachtor, R. W. Atcher Emission-computerized tomography for visualization o f organ volumes (IAEA-SM-247/103) ..........................................................................................
345
K.J. Vikterlöf В. Söder borg Discussion ............................................................................................................. Construction and clinical application o f complex utility programs in the SEGAMS-80 system (IAEA-SM-247/43) ...............................................
349 351
E. Máté, J. Csirik, L. Csernay, A. Makay An electronic image processing device featuring continuously selectable two-dimensional bipolar filter functions and real-time operation (IAEA-SM-247/100) ....................................................................................... 357
B.D. Charleston, F.H. Beckman, M.J. Franco, D.B. Charleston Discussion ............................................................................................................. Practical application o f deconvolution techniques to dynamic studies (IAEA-SM-247/26) ........................................................................................... .
365 367
C.C. Nimmon, T. Y. Lee, K.E. Britton, M. Granowska, S. Gruenewald Discussion ............................................................................................................. Experience with a mobile data storage device for transfer o f studies from the critical care unit to a central nuclear medicine computer (IAEA-SM -247/11) ...............................................................................................
387
389
T.D. Cradduck, A.A. Driedger Discussion ...................................................................................... ...................... The future o f functional imaging: Success or failure? (IAEA-SM-247/86) ...................................... ........................................................
399 401
W.J. MacIntyre, S.A. Cook, R.T. Go Discussion
.............................................................................................................
417
Poster presentations: Optimized display o f digitized scintigraphic data (IAEA-SM-247/51) ........
419
B.F. Hutton, J. Cormack Discussion ............................................................................................................. The effect o f display matrix size on the quality o f digitized images (IAEA-SM -247/115) ..........................................................................................
420 420
P. Sharp, R. Chesser Discussion ............................................................................................................. 421 Non-uniformity and non-stationarity in emission tomography (IAEA-SM -247/140) ........................................................................................... ' 422
A. Todd-Pokropek, S. Zurowski, F. Soussaline
Influence o f non-uniform resolution on image quality and quantitation in positron-emission computed tomography (PCT) (IAEA-SM-247/ 92) ...............................................................................................
423
E.J. Hoffman, Sung-Cheng Huang, D.L. Plummer, M.E. Phelps, D.E. Kuhl Restauration d’images scintigraphiques par filtres adaptatifs à support borné (IAEA-SM -247/144) ................................................................................
425
Y. Bizais, Ph. de Larminat, J.-P. Lemort, A.E. Todd-Pokropek
RADIOPHARMACEUTICALS (Sessions 4 and 4a) Invited review paper: New developments in radiopharmaceuticals for imaging: A review (IAEA-SM-247/205) .......................................................................
429
G. Subramanian, J. G. McAfee Discussion ............................................................................................................. Solid-phase labelling: An improved method for the preparation o f technetium-labelled pharmaceuticals (IAEA-SM-247/49) .......................
452 453
P H. Cox Discussion ............................................................................................................. Technetium-99m-BIDA: A potential cholescintigraphic agent for icteric patients (IAEA-SM -247/152) .............................................................
456 459
J. Weininger, J. Trumper, E. Lubin, M. Cohen, T. Sadeh, M. Juszinsky Cholescintigraphy in jaundiced patients: Comparison o f diethyl iminodiacetic acid (IDA) with p-butyl-IDA (IAEA-SM -247/153) ...........
465
R. Dudczak, P. Angelberger, M. Wagner-Löffler, К. Kletter, P. Ferenci,
H. Frischauf Discussion ............................................................................................................... Clinical applications o f indium-1 1 1-acetylacetone-labelled blood cells (IAEA-SM -247/39) ...............................................................................................
475 477
P. Georgi, H. Sinn, H. Wellman, J.H. Clorius, W. Becker Kinetics and migration o f indium-11 1-labelled human lymphocytes (IAEA-SM-247/9 5 ) ...............................................................................................
487
D.A. Goodwin, J.R. Heckman, L.F. Fajardo, A. Calin, S.J. Propst, C. I. Diaman ti Use o f indium-111-oxinate-labelled granulocytes and thrombocytes in kidney transplantation (IAEA-SM -247/129) ...............................................
499
E.A. van Royen, J.B. van der Schoot, M.R. Hardeman, S. Surachno, J.H. ten Veen, J. Vreeken, J.M. Wilmink Discussion
..................................................................................................................
509
Radioimmunodetection o f various human carcinomata with radiolabelled antibodies to tumour-associated antigens (IAEA-SM-247/70) ................
511
E.E. Kim, F.H. Deland, J.R. Salyer, S.J. Bennett, D.M. Goldenberg Discussion ............................................................................................................. Evaluation of new radiopharmaceuticals for tumour localization: The value o f the human tumour xenograft (IAEA-SM-247/6) .........
518 519
D.M. Taylor Invited review paper: Radiation dose to the patient in radionuclide studies (IAEA-SM-247/206) ..........................................................................................
527
H.D. Roedler Discussion
.............................................................................................................
542
Poster presentation: Use of cell-type specific antibodies for radioimmunodetection o f breast métastasés with a high-purity germanium camera (IAEA-SM-247/77) ....
543
J.A. Peterson, T. Wilbanks, S. Miller, L. Kaufman, D. Ordendahl, R.L. Ceriani A new germanium-68/gallium-68 radioisotope generator system for production o f gallium-68 in dilute HCl (IAEA-SM-247/38) ..................
545
J. Schuhmacher, W. Maier-Borst Discussion ............................................................................................................. Indium-113m-phytate liver scanning agent (IAEA-SM -247/131 ) .............
550 551
E. Lachnik, W. Zulczyk, J. Wiza, I. Liciñska, W. Jakubowski, W. Graban Discussion ............................................................................................................. Scintigraphic evaluation o f chronic cholecystitis with cholelithiasis using technetium-99m-labelled pyridoxylidene isoleucine (IAEA-SM -247/106) ..........................................................................................
557
559
E. Alp, C.F. Bekdik, Y. Laleli, M. T. Ercan Discussion ............................................................................................................. Tin-117m-labelled radiopharmaceuticals: Effects o f structural modifications on the adrenal uptake o f steroids labelled in the sidechain with tin-117m (IAEA-SM -247/89) ...............................................
566
567
F.F. Knapp, Jr. Discussion ............................................................................................................. Ytterbium-169 citrate for detection o f inflammatory lesions (IAEA-SM -247/94) ...............................................................................................
577 579
R.L. Hayes, B.L. Byrd, J.J. Rafter, J.E. Carlton, J.L. Coffey Discussion
.............................................................................................................
585
Spallation-produced thulium-167 for medical applications (IAEA-SM-247/60) ........................................................................................................... :............ 587
G.-J. Beyer, R. Muenze, W.D. Fromm, W.G. Franke, H. Henke, V.A. Khalkin, N.A. Lebedev Discussion ............................................................................................................. Study o f the dose and distribution o f technetium-99m-MDP in various bones and tissues (IAEA-SM -247/113) .........................................................
I.
598 599
Othman, N.M. Spyrou, R.A.A. Khan
Structure-activity relationship o f technetium-99m-and indium-111labelled N-diphosphonates (IAEA-SM -247/111 ) .....................................
613
P. Andreou, E. Chiotellis, A. Varvarigou, C. Koutoulidis Some aspects o f the mechanism o f uptake o f technetium-99mdiphosphonates (IAEA-SM -247/122) .............................................................
623
R.A.M.J. Claessens, Z. Kolar, J.E.J. Schmitz, J.G.M. van den Linden, J.J. Steggerda, I. Kazem Discussion
.............................................................................................................
List of Chairmen o f Sessions and Secretariat
....................................................
633 635
IAEA-SM -247/200
Invited Review Paper NEW PERSPECTIVES IN NUCLEAR IMAGING H.N. WAGNER, Jr. The Johns Hopkins Medical Institutions, Baltimore, Maryland, United States o f America Abstract NEW PERSPECTIVES IN NUCLEAR IMAGING. As the science of nuclear medicine has evolved, it is becoming increasingly clear that it is concerned with biodistribution, the science of how the chemical constituents of the body move from one place to another. Perhaps more than any other branch of clinical medicine, it rests on a foundation of chemistry, physics and biology. In nuclear medicine, biodistribution is studied by quantitative nuclear imaging; chemistry is translated into radio pharmacology; physics into instrumentation; and biology into physiology. Nuclear medicine can be thought of as applied physiology and physiological chemistry. The hierarchy of nuclear medicine extends from its foundation on basic sciences all the way up to the study of the human being in health and disease, to a concern with patients’ problems and physicians’ decisions. A landmark event occurred in 1941 when Hamilton in California and Hertz in Boston first measured the accumulation of radioiodine in the human thyroid with a GeigerMüller tube, thus beginning the era of in vivo biochemistry in man. It is clear that the orienta tion of these pioneers of nuclear medicine was that of chemistry. Today, forty years later, the orientation of the field has returned to chemistry, but chemistry of a different sort. Where the pioneers concentrated on the biodistribution of elements, the modern nuclear chemist is concerned with molecules, with relating chemical structure to biological distributions. An example is the case of the iminodiacetate compounds. Using a new technique known as mixed ligand analysis, Bums and his colleagues were able to determine that the hepatobiliary agent, HID A, developed by Loberg, was an anionic bis-complex with a charge of minus one. Further study by Smith and by Depamphilis added the important additional information that the co-ordination number of technetium was five and that an oxo-oxygen was in the apical position. Having determined the chem ical s tru ctu re of this class of compounds, they were able to study a series of analogues and found that the amount of the complexes excreted in the bile was directly proportional to the natural log of the molecular weight to charge ratio. Agents of this type are widely used today in diagnostic medicine and illustrate the direction of the field, towards the synthesis of labelled compounds whose biodistribution depends on their chemical interaction with structurally specific binding sites, i.e. receptors, enzymes or binding sites of active transport systems. Other examples include amino acids for pancreatic imaging; deoxyglucose for studies of regional brain and heart metabolism; fatty acids for studies of the heart; steroid hormones for breast tumours; and muscarinic compounds for study of the cholinergic system of the heart. Most of these compounds are labelled with UC, 18F or 13 N, available only from cyclotrons, but the extension of the work to the more widely available nuclides, iodine-123 and technetium99m, is also proceeding at a rapid pace, stimulated chiefly by the success with the positron- ■ emitting compounds. Advances in nuclear imaging include the development of both positron and single photon emission tomography which, combined with the new radioactive tracers,
3
4
W AGNER
are providing new insights in physiology and biochemistry. Biodistribution studies in man permit measurement of regional as well as global function. Improved perception is being augmented by improved quantification and automation. Functional imaging is becoming commonplace. In the author’s opinion, the field of nuclear imaging has reached a high state of development, can now be considered as the science of biodistributions, and is likely to continue its outstanding rate of growth.
A navigator must constantly ask: 1. How have we gotten where we are? 2. Where do we want to go? 3. Are we likely to be able to get there? The philosopher, Alfred Whitehead, said 'It is a wellfounded historical generalization that the last thing to be discovered in any science is what the science is really about.' For decades, the field of nuclear medicine has been searching for its identity. Various definitions have been suggested in the past. In the I960 s nuclear medicine was defined as 'the application of radioactive materials to the diagnosis and treatment of patients and the study of human disease'. The field justified its existence in that it was a reasonable way for some persons to spend the greatest part of their profes sional interests and activities. A more recent statement is 'that it is concerned with the study of the motion and change of body structures and functions, with regional biochemistry and physiology'. A major theme of nuclear medicine is its concern with biodistribution; it is a science concerned with what goes where within the human body and how these movements are brought about. Nuclear medicine can perhaps be thought of as the 'applied science of biodistribution', resting on a solid found ation of chemistry, physics and biology. These biodistribu tions are studied by quantitative nuclear imaging. Chemistry is translated into radiopharmacology; physics into instrument ation; and biology into physiology. The field can be thought of as a hierarchy, with chemistry, physics and biology at the base and extending up to a concern with the human being in health and disease, a concern w i t h .human problems and medical décisions. To one who has viewed the nuclear medicine scene for over two decades, it seems clear that advances in nuclear imaging with radioactive tracers have provided and will continue to provide new insights in physiology and biochemistry. Perhaps their greatest contribution is that they permit measurement of regional as well as overall organ function. This helps over come one of the major limitations in the' past in the applica tion of physiological measurements, i.e. biological variabil ity. Certain body functions may vary several hundred percent
IA E A -S M -2 4 7 /2 0 0
F IG .l.
D r. Joseph H a m ilto n o f the U niversity o f C alifornia a t Berkeley who, a t the same
tim e as H e rtz in Boston, in tro d u ce d measurement o f the uptake o f ra d io io d in e by the th y ro id in the diagnosis o f th y ro id diseases.
even among normal people, whereas the variability of a given function within an organ or between paired organs is much less. This leads to one of the basic tenets of nuclear medicine, namely, that an abnormality in regional function may be detected before the overall function of the organ is outside of the range of values in normal people. This is called the 'homogeneity principle'. Over the past two decades, perception of regional function has improved tremendously, and has been augmented by improved quantification and automation. The latter also improves performance by greatly decreasing the variability inherent in subjective interpretation of nuclear images. In essence, the major advances in nuclear medicine in recent years can be summarized as: better chemistry, better quantification and better physiology.
5
6
WAGNER
Ln
F IG .2.
I MW/Z I
A n example o f the relatio n sh ip between chem ical stru ctu re and b io d is trib u tio n . The
b ilia ry excre tio n o f im in o ace tic acid com pounds is p ro p o rtio n a l to th e na tura l log o f the m ole cular w eight divid e d by the m olecular charge.
When we consider the work over 50 years ago of pioneers such as Blumgart and his colleagues who studied the pathophysi ology of the circulation with radon solutions and cloud chambers, we can better appreciate the advances that have been made [1]. Few people realize that these first studies of the heart with radioactive tracers and external detectors were performed nearly 15 years before the first measurement of the uptake of radioactive iodine by Hamilton [2] in California and Hertz [3] in Boston. It has always interested me that the 1941 photograph of Hamilton shows him seated in front of the periodic table, symbolizing the chemical orientation of the field (Fig. 1). The renaissance of nuclear medicine today also involves chemistry--but chemistry with a difference. Hamilton and the other pioneers of the tracer method were studying elements. They hoped that a whole host of diseases would be found that would be analogous to what they had discovered in the case of the thyroid, namely, that abnormal metabolism of an element was involved, i.e. iodine, and that disease resulted from hyperactivity or underactivity of the gland. Their disappointment that similar advances were not possible in the case of other elements can be gleaned from a reading of their research reports. Today the chemistry of nuclear medicine involves compounds, particularly those bound by membrane and intracellular receptors. Major accomplishments are being made in our understanding of
IAEA-SM -247/200
F IG .3.
Increased deoxyglucose m etabolism in the o c c ip ita l co rte x (a rro w ) in association w ith
visual s tim u la tio n o f a human being by w h ite lig h t and com plex visual patterns. These studies were perfo rm e d by Phelps and others a t the U niversity o f C a lifornia, Los Angeles, and are published by perm ission.
7
8
WAGNER WM
¡ÍÉiSSIs *■ % :t:ïïV: :::
seiswi
WtKSSSBSÊ
ярю ир « т
вИИ -
т ^ щ
шШ ЯИШ
Я Т т , 5в1в¥Ш -
М
||||8 Щ
ï l f i S i Î l  ï Ш Ё 1 Ш Ё 9 Ш Я в й в Я Ж - ¥*mm m
W ï ii
W
||! й |§ |§ ||||
«им
ш FIG .4.
1ШШёв#Й One o f the earliest brain scans perfo rm e d a t Johns H o p kin s H o sp ita l in B a ltim o re w ith
a re c tilin e a r scanner and iodine-131 album in.
the relationship between chemical structure and the biodistribution of a molecule after its injection. An example from our laboratory is the case of the iminoacetic acid (IDA) compounds, first introduced into nuclear medicine by Harvey and his associates [4]. Burns and others at Johns Hopkins developed a new analytical technique, called "mixed ligand" analysis that helps determine the structure of compounds in tracer quantities [5]. In the case of the IDA compounds, their work revealed the structure of technetium-99m labeled HIDA [6, 7]. Knowledge of the chemical structure made it possible to synthesize a whole group of analogues and determine the rela tive affinity of these compounds for the liver and biliary system. It was found that the biliary excretion could be pre dicted from the natural log of the molecular weight of the com pound divided by its charge (Fig. 2). This represented one of the first examples of how biodistribution was based on chemical structure. Similar results have been obtained in the case of steroid hormones by Katzenellenbogen and his associates in Illinois [8]. It is predictable that a host of structurebiodistribution relationships of this sort will be discovered. The discovery of structurally specific binding sites over the past decade has been a great stimulus to research in
IAEA-SM -247/200
9
Homonymous Hemianopsia
F IG .5. R e la tio n sh ip between regional glucose m etabolism and degree o f stim u la tio n o f the p rim a ry (PVC) and associative (A VC) visual co rte x in man.
nuclear medicine. These include: (1) membrane and intra cellular receptors, (2) enzymes, and (3) the binding substance in active transport systems. Examples include the l'minoacetic acid compounds, deoxyglucose uptake in the brain and heart, fatty acid uptake in the myocardium, steroid hormone uptake in the uterus and mammary tumors, and uptake of muscarinic cholinergic compounds at the neuromuscular junction. Let me illustrate these points with a few examples: The B rain: Figure 3 illustrates the increased metabolic activity seen in the occipital cortex in the regions concerned with vision when the subject is stimulated by looking at white light or a complex visual pattern. The images were obtained by Phelps and his associates in Los Angeles, California, using a positron-emission tomography device and fluorine-18 deoxyglucose to measure regional metabolism [9]. For the first time, we can visualize the chemical processes associated with mental function in the living human being. What a difference from the brain scans performed in the late 1950 s with iodine131 labeled human serum albumin and a 3-inch rectilinear scanner (Fig. 4)! The most recent work of the University of California at Los Angeles (UCLA) group has shown that there
F IG . 6 .
R e g io n a l cerebral b lo o d flo w measured w ith , 3 N H 3 and re g io n a l cerebral m etabolism
measured w ith 1SF D G in a p a tie n t w ith te m p o ra l lobe epilepsy. Betw een seizures (in te ric ta lj, there was decreased deoxyglucose m etabolism in the le ft te m p o ra l lobe. D u rin g tw o separate fo c a l seizures, the area th a t had previously been h yp o a ctive became hyperactive. These studies were p e rfo rm e d by K u h la n d his associates a t the U nive rsity o f C a lifo rn ia , Los Angeles, and are p u b lish e d w ith th e ir permission.
11
IAEA-SM-24 7 /2 0 0
FIG .
7. Three patients w ith te m p o ra l lobe resections in trea tm e nt o f fo c a l seizures. The
arrows indicate regions o f decreased deoxyglucose m etabolism .
These studies were perform ed
by K u h la n d his associates a t the U niversity o f C alifornia, L o s Angeles, a n d a ré published w ith th e ir perm ission.
FIG . 8 .
N o rm a l X -ray ( transm ission) co m p u te d tom ography and absent deoxyglucose
m etabolism and cerebral b lo o d flo w in a dead person. S tu d y by Phelps et al. (U C L A ) and published w ith permission.
12
F IG .9.
WAGNER
Increased deoxyglucose m etabolism in the rig h t cerebral co rte x in man associated
w ith liste n in g to a sto ry being to ld in the le ft ear. S tu d y by A la v i and associates a t the U niversity o f Pennsylvania; published w ith permission.
is a quantitative relationship between the degree of visual stimulation and the percentage increase in glucose metabolism in the visual cortex, both the primary (PVC) and associated (AVC) visual cortex (Fig. 5) [10]. The translation of these fundamental observations to the care of patients is proceeding more rapidly than we might have expected. Kuhl and others have begun to examine regional cerebral blood flow with 13N H 3 and regional glucose metabolism with 18F deoxyglucose in patients with temporal lobe epilepsy [11-13]. They have observed that regional glucose metabolism is below normal in the involved regions between convulsions, and far above normal in these same regions during convulsions (Fig. 6). Regional increases in cerebral blood flow were also observed, but to a lesser degree than glucose metabolism. In most of the subjects, no structural abnormalities were observed in transmission computed tomograms. The post-operative studies after removal of part of the temporal lobes are seen in Figure 7.
IA EA-SM -247/200
FIG . 10.
Ernest Lawrence, in ve n to r o f the cyclo tro n , standing beside his fo u rth cyclo tro n ,
the fir s t devoted to m edical uses.
Perhaps the best illustration of the relative information content of transmission computed tomography (TCT) and emission computed tomography (ECT) is the study by Phelps shown in Figure 8. The image with TCT is normal while the ECT images show absent glucose metabolism and cerebral blood flow. The patient in this case was dead. Another example of measurable chemical activity associ ated with mental activity is shown in Figure 9. This study by Alavi and others at the University of Pennsylvania shows the increased metabolic activity in the right cerebral hemisphere that resulted from the subject's listening to a story being played through an earphone over either the right or left ear. These investigators have also found decreased glucose metabolism in the frontal lobes in aged persons compared to young persons, and especially reduced frontal activity in patients with senile dementia [14]. These studies illustrate the use of positron-emitting radionuclides, especially carbon-11, nitrogen-13 and fluorine18, the first two being available only through the use of a cyclotron. Just as the renaissance of nuclear medicine involves the rebirth of chemistry, so also it involves the
13
14
WAGNER
/ MC16F SHIELDED CYCLOTRON
FIG. I I.
Diagram of a small, self-shielded cyclotron designed for hospital use.
FIG.12 . A positron-emitting tomography device, usually called a PET scanner.
IA EA-SM -247/200
15
VERTEX VIEW O F HEAD (90 min post injection)
L LATERAL VIEW O F HEAD
W H O LE BODY
(110 min post injection)
(165 min post injectionI
F IG .13. D is trib u tio n o f iodine-123 N -iso p ro p yl-p-io d o a m p h e ta m in e in a m onkey. S tudy by W inchell et al. and published w ith permission.
rebirth of the cyclotron--but again with a difference. In Figure 10, we see Nobellaureate Ernest Lawrence and his first 'medical cyclotron'. Among the first nuclides that he produced with this instrument were radioactive phosphorus and sodium. In those days, cyclotrons were complex devices that required a team of skilled physicists to keep them operating. Today cyclotrons have been simplified in design to the point that they can be located right in the heart of nuclear medicine departments in hospitals (Fig. 11). The evolution of the university hospital cyclotron has been the result of a series of happenings: (1) the renewed chemical orientation of nuclear medicine, (2) competition in other areas of noninvasive imaging, (3) increasing emphasis on quantification, (4) the development of two cyclotron-produced radionuclides (gallium-67 and thallium-201), (5) improvement in cyclotron design, (6) advances in computers and automation, (7) the effect of transmission tomography on emission tomography, (8) the acceptance of expensive technology, (9) advances in receptor chemistry, and (10) the slowing of production of technetium-99m tracers. For best results, positron-emitting radionuclides require the use of a positron-emitting tomography device, similar to that shown in Figure 12.
16
WAGNER
F IG .14.
Single p h o to n tom ography w ith a ro ta tin g A nger camera.
It is unlikely that cyclotrons and positron-emission tomography will spread beyond university medical centers or Other research centers within the next decade. But it need not take that long for the fruits of research with positronemitting radiochemicals to be translated into medical practice The advances made with positron-emitting compounds have stimulated increasing efforts to produce analogous compounds labeled either with iodine-123 or technetium-99m. An example is the case of iodine-123 N-isopropyl p-iodoamphetamine (Fig. 13), produced by Winchell and his colleagues in California [15]. This tracer is believed to be bound by serotonin and norepinephrine receptors and is able to cross the blood-brain barrier. It can be studied by conventional Anger-type scintillation cameras or by single photon tomographic systems, such as that illustrated in Figure 14. Before leaving the brain, I would like to illustrate how the advance of nuclear medicine depends on advances in basic
IAEA-SM -247/200
F IG . 15.
A c c u m u la tio n o f
С deoxyglucose in the brain o f a m onkey. The dark areas
correspond to the areas o f m etabolism . In the upper p ictu re (A ) the m onkey was seeing; in the m id dle (B), the m onkey was b lin d e d ; and in the lo w e r image (C) the m onkey was blinded in one eye. These studies were perfo rm e d by S o k o lo ff et al. and are published w ith th e ir perm ission.
18
FIG -16.
WAGNER
E a rly heart scan a t Johns H o p kin s in 1958 p e rfo rm e d w ith iodine-131 a lb u m in
and a re c tilin e a r scanner. The p a tie n t has a p e rica rd ia l effusion.
science. Figure 15 is the original study by Sokoloff and others at the National Institutes of Health in Bethesda, Maryland, U.S.A. [16]. It shows the areas of uptake of carbon 14 deoxyglucose in the brain of a monkey when the monkey is seeing with both eyes (upper picture, A); blinded in both eyes (middle picture, B); or blinded in one eye only (lower picture, C). This work led to the human studies with fluorine 18 deoxyglucose. More recently, Sokoloff has reported with Carolyn Smith that regional protein synthesis associated with regeneration of nerves results in an increased uptake of carbon-14 leucine in the appropriate regions of the brain of
19
IAEA-SM -247/200
VEN TRICULAR
F IG .18.
PRESSURE - VOLUME - TIM E
CURVE
C ontinuous m o n ito rin g o f le ft ve n tricu la r fu n c tio n (upper tracing) together w ith
the electrocardiogram in a p a tie n t w ith a single prem ature ve n tricu la r beat fo llo w e d by a com pensatory pause.
WAGNER
20
F IG .19.
A device fo r m o n ito rin g le ft ve n tricu la r fu n c tio n , called a nuclear stethoscope.
experimental animals [17]. Such findings are also likely to be extended to man by means of nitrogen-13 tracers. The Heart: In 1958, thirty years after Blumgart's pioneering work, Rejali and others in Cleveland, Ohio, U.S.A., introduced the first nuclear imaging technique that was to achieve widespread use in clinical cardiology, namely, the use of radioiodinated albumin and the rectilinear scanner in the diagnosis of pericardial effusion [18]. An example of an early study of this type is shown in Figure 16. Again we can see the relatively poor quality of the image, but it was the forerunner of gated blood pool imaging, now widely performed.
IAEA-SM -247/200
RR (msec)-. 1050
820
§40
21
1090
IffiO
/ \
F IG .20.
16/32S 49
115/305 95/30 IS/32 38 32 47
1000
£90 J
V
= ;.£ Г т щ ; íh -.V;
T30
14/323 43
/ v
И/305 36
\ / v
л.
4 /
У
15/315 105/30 45 35
■
\ V \ ./ 16/32 50
7/28 25
C ontinuous m o n ito rin g o f le ft ve n tricu la r fu n c tio n in a p a tie n t w ith a tria l fib rilla tio n .
Indeed, perhaps the greatest contribution of nuclear medicine to cardiology is that it makes it possible to monitor the changes in the volume of blood within the various chambers of the heart throughout the cardiac cycle. For decades, cardi ology has been concerned primarily with the changes in pressure within the cardiac chambers as a function of time. If we can monitor both volume and pressure changes, we can characterize completely the thermodynamic or energy state of the heart [19]. This can be portrayed as.a so-called 'pressurevolume loop', as shown in Figure 17. Today, it is common to monitor the left ventricular time-actiyity curve. In the future, we will be able to monitor right and left ventricular work as well. An example of continuous monitoring of left ventricular activity on a beat-to-beat basis is shown in Figure 18. The upper curve is a different type of image from the visual image. Instead of being a spatial image or a functional image, it is a third type--a temporal image, analogous to tne electrocardiogram. In the illustration, the upper curve shows the changes in activity within the left ventricle as a function of time. The activity is proportional to volume so that we can see the end-diastolic and end-systolic volume as well as the rates of emptying and filling of the ventricle. In the patient, whose tracing is shown, we can see the hemodynamic events asso ciated with a premature ventricular beat followed by a compensa tory pause. In our department we have investigated ectopic beats in this way, using the device shown in Figure 19, a device we have called a 'nuclear stethoscope'. Its principal use is in continuous beat-to-beat monitoring of spontaneous beats and those resulting from perturbations, such as exercise, cold pressor or isometric stress, or the administration of drugs [20-22]. An example of the changes in ventricular volume in patients with atrial fibrillation is shown in Figure 20.
22
W AGNER
PACiÑO í NDUCED ISCHEMIA
F IG .21. M easurement o f regional m yoca rd ia l perfusion w ith n N H 3 and m yoca rd ia l m etabolism w ith 18F deoxyglucose
IAEA-SM -247/200
23
At present it is necessary to use a blood pressure cuff or intracardiac catheters to monitor pressure simultaneously with the volume measurements. Schuler and his colleagues in Heidelberg are using similar approaches to try to examine the contractile state of the heart independent of changes in enddiastolic volume and aortic pressure [23]. Extending the physi ological studies of Sagawa [24], they are examining the rela tionship between end-systolic volume and peak systolic pressure, measured with a blood pressure cuff. They have been able to show that the slope of this relationship is related to the state of the ventricular muscle. Just as positron-emitting nuclides have revolutionized the study of the brain, so also have they begun to have an equally important impact in cardiology. As in the case of the brain, using 13Из to measure regional myocardial blood flow and 18F deoxyglucose to measure regional glucose metabolism, Schelbert and his colleagues in California have been able to show that under certain conditions, regional blood flow and glucose metabolism are uncoupled [25]. For example, ischemia in dogs reduces glucose metabolism whereas hypoxia increases it [26]. In patients with idiopathic cardiomyopathy, myocardial blood flow and glucose metabolism are dissociated in certain regions of the heart (Fig, 21). As in the case of the brain, both single photon and positron-emitting radio-compounds are under development. These include technetium-99m arsene compounds introduced by Deutsch in Ohio [27]; beta adrenergic compounds labeled with iodine-123, introduced by Wieland in Michigan [28]; cholinergic compounds, introduced by the Hopkins group [29]; tellurium-123mlabeled fatty acids, introduced by Knapp and associates from Oak Ridge [30]; iodine-123 and bromine-75 labeled fatty acids, i n t r o d u c e d by S t o c k ! in e t a l . of J ü l i c h , F.R.,Germany [31]; and carbon-ll'labeled glucose, introduced by Vyska [32]. Other Organ Systems : Time will not permit consideration of other important perspectives, including the renewed interest in the study of platelets; the. increasing clinical study of the biliary system; the increasing number of studies of the skeletal system; the use of inhaled radioactive aerosols to study the distribution and clearance of inhaled substances in the lungs, especially as a point of entry of toxic substances into the body; and the general interest in quantification of most nuclear imaging techniques. What Lies Ahead? In the early days of nuclear medicine, pioneers visualized a molecular orientation instead of a cellular approach to our understanding of the way the body functions in health and disease. What were they able to accomplish? Did they identify new diseases, help explain the mechanisms of old diseases, or improve our ability to reverse or arrest disease processes? To some degree, the answer to
24
WAGNER
these questions is yes. Largely as a result of the use of radiolabeled compounds, we have learned that changes in the amino acid structure of proteins, or the production of defi cient or excessive quantities of an enzyme will result in disease. Examples include defects in thyroid hormone produc tion; pulmonary emphysema due to deficiency in the enzyme, a , -antitrypsin; lactose intolerance due to a congenital or acquired deficiency in the enzyme lactase; galactosemia; alkaptonuria; and many others. Most of these types of diseases are genetically determined, and are relatively uncom mon, being autosomal or x-linked recessive inborn errors of metabolism. For the most part, however, medical diagnosis is unfor tunately still dominated by an anatomical orientation, supple mented by the information provided by measurement of the chemical constituents of the blood. Surgery remains a major form of treatment. With the tremendous growth in interest in pharmacology, it is likely that nuclear medicine will continue to thrive for the following reasons: 1. Nuclear medicine is applied human biochemistry and physiology. 2. We can study the dynamic state of body constituents. 3. We can measure both regional and overall organ function. 4. We can monitor the effects of drug therapy. 5. We can study the most important characteristic of living systems: motion and change.
REFERENCES [1] BLUMGART, H.L., WEISS, S., Studies on the velocity of blood flow: VII. The pulmonary circulation time in normal resting individuals, J. Clin. Invest. 4 (1927) 399. [2] HAMILTON, J.G., Absorption of the radioactive isotopes of sodium, potassium, chlorine, bromine, and iodine in normal human subjects, Am. J. Physiol. 127 (1938) 667. [3] HERTZ, S., Radioactive iodine in the study of thyroid physiology, Proc. Soc. Exp. Biol. Med. 38 (1938) 510. [4] HARVEY, E., LOBERG, M., COOPER, M., Tc-99m HIDA: A new radiopharmaceutical for hepato-biliary imaging, J. Nucl. Med. 16 (1975) 533. [5] BURNS, H.D., WORLEY, P., WAGNER, H.N., Jr., MARZILLI, L., RISCH, V., “Design of technetium radiopharmaceuticals” , The Chemistry of Radiopharmaceuticals (HEINDEL, N.D., BURNS, H.D., HONDA, T., BRADY, L.W., Eds), Masson Publishing, New York (1978) 269. [ 6 ] MARZILLI, L.G., WORLEY, P., BURNS, H.D., A new electrophoretic method for determining ligand: technetium stoichiometry in carrier free 9 9 mTc-radiopharmaceuticals, J. Nucl. Med. 20 (1979) 871.
IAEA-SM -247/200
25
LOBERG, M.D., FIE L D S, A .T., C hem ical stru c tu re o f technetium -99m -labeled N -(2,6-dim ethylphenylcarbam oylm ethyl) im inodiacetic acid (Tc-H ID A ), Int. J. Appl. R adiat. Isot. 29 (19 7 8 ) 167. K A TZEN ELLEN BO G EN , J.A ., CARLSO N, K.E., HEIM AN, D .F., LLOYD, J.E ., R adiopharm aceuticals: S tru ctu re A ctivity R elationships (SPEN CER, R.P., Ed.), G ruñe and S tra tto n , New Y ork (in press). PHELPS, M.E., M AZZIOTTA, J., M ILLER, J., KUHL, D.E., C om parison o f quantitative transaxial and coronal p ositron tom ography o f the brain, J. Nucl. Med. 21 (1 9 8 0 ) P15 (abstract). PHELPS, M.E., KUHL, D.E., M A ZZIOTTA , J .C., T om ographic m apping o f the m etabolic changes in th e visual cortex during visual stim ulation o f v olunteers and p atien ts w ith visual deficits, J. Nucl. Med. 21 (1 9 8 0 ) P21 (abstract). KUHL, D.E., A LAV I, A., REIV ICH , M .,“ C om puterized em ission tom ography and d e term in atio n of local brain fu n c tio n ” , Non-Invasive Brain Im aging (DeBLANC, H., SO REN SON , J., Eds), Society o f N uclear M edicine, New Y ork (1975). K U HL, D.E., PH ELPS, M .E., ENGEL, J.P ., e t al., R elationship o f local cerebral glucose utilizatio n and relative perfusion in stroke and epilepsy, J. C om put. Assist. Tom ogr. 2 (1 9 7 8 )6 5 5 . KUHL, D.E., ENGEL, J., Jr., PHELPS, M.E., Em ission com p u ted tom ography o f F - l 8 fluorodeoxyglucose and N-13 am m onia in partial epilepsy, J. Nucl. Med. 21 (1980) P21 (abstract). ALAVI, A., RINTELM ANN , W., REIV ICH , M., et al., R esponse to aud ito ry stim ulation as m apped by p o sitro n em ission to m ography (PET) and F - l 8 deoxyglucose (FD G ), J. Nucl. Med. 21 (1 9 8 0 ) P21 (abstact). W INCHELL, H.S., HO RST, W.D., BRAU N, L., O L D E N D O R F, W.H., Single-pass brain u p take and w ashout o f 1-123 N -isopropyl-p-iodoam phetam ine and its binding to brain cortical synaptosom es, J. Nucl. Med. 21 (1 9 8 0 ) P22 (abstract). SO K O LO FF, L., REIV ICH , M., KENN ED Y, C., et al., T he ( 14C) deoxyglucose m ethod for the m easurem ent o f local cerebral glucose utilization: th eo ry , procedure and norm al values in the conscious and anesthetized albino rat, J. N eurochem . 28 (1 9 7 7 ) 897. SO K O LO FF, L., SMITH, C., personal com m unication. R E JA L I, A.M., M acIN TY RE, W .J., FR IE D E L L , H.L., A radioisotope m ethod o f visualization o f blood pools, Am. J. R oentgenol. 79 (1 9 5 8 ) 129. BOU RGUIG NON , M.H., W AGNER, H.N., Jr., Noninvasive m easurem ent o f ventricular pressure th ro u g h o u t systole, Am. J. Cardiol. 44 (1 9 7 9 ) 466. W AGNER, H.N., Jr., WAKE, R., N IC K O L O FF, E., N A TA R A JA N , Т .К ., The nuclear stethoscope: a sim ple device for generation o f left ventricular volum e curves, Am. J. Cardiol. 38 (19 7 6 ) 747. W AGNER, H.N., Jr., RIGO , P., BA X TER, R .H ., e t al., M onitoring ventricular function a t rest and during exercise w ith a non-im aging nuclear d e te c to r, Am. J. Cardiol. 43 (1 9 7 9 ) 975. CAM ARGO, E .E ., HA RRISO N , K.S., W AGNER, H.N., Jr., et al., Noninvasive beat to beat m onitoring o f left ventricular fu n c tio n by a nonim aging nuclear detecto r during p rem atu re ventricular co ntractions, Am. J. Cardiol. 45 (1 9 8 0 ) 1219. SCH ULER, G., OLSH AUSEN , K.V., MEHMEL, H., et al., Noninvasive d eterm ination o f left ventricular end systolic pressure volum e relation by radionuclide angiography, J. Nucl. Med. 21 (1 9 8 0 ) P64 (abstract). SAGAWA, K., The ventricular pressure-volum e diagram revisited, Circ. Res. 43 (1978) 677.
26
WAGNER
[25] SCHELBERT, H., PHELPS, M., HUANG, H., et al., N-13 am m onia as an in d ic a to r of flow: factors influencing its u p tak e and re te n tio n in m yocardium , J. Nucl. Med. 21 (1980) P89. [26] PHELPS, M.E., HOFFM AN, E .J., SELIN, C., et al., Investigation o f 18F-2 deoxyglucose for m easurem ent o f m yocardial glucose m etabolism , J. Nucl. Med. 19 (1 9 7 8 ) 1311. [27] DEUTSCH, E., GLAVAN, K.A ., FER G U SO N , D.L., D evelopm ent o f a Tc-99m m yocardial imaging agent to replace T l-201, J. Nucl. Med. 21 (1 9 8 0 ) P56 (abstract). [28] W IELAND, D.M., BROWN, L.E., W ORTHING TON , K.C., et al., H eart imaging w ith an I23I-noradrenaline storage analog, J. Nucl. Med. 21 (1 9 8 0 ) P90 (abstract). [29]
BURNS, H.D., M A RZILLI, L.G., DANNALS, R .F., e t al., 125I-4-iodophenyltrim ethylam m onium ion, an iodinated acetylcholinesterase in h ib ito r w ith p o ten tial as a m yocardial imaging agent, J. Nucl. Med. (in press). [30] KNAPP, F .F ., Jr., BU TLER, T .A ., AM BROSE, K .R ., et al., “ M yocardial up tak e o f 123mTe-labeled long-chain fa tty acids: effects o f h e te ro ato m position and to ta l chain length” , T hird Intern atio n al Sym posium o n R adiopharm aceutical C hem istry, 1 6 -2 0 . June 1980, W ashington U niversity, St. Louis, Missouri, 175. [31] STÖCKLIN, G., COENEN, H.H., HARM AND, M .F., et al., 15-(p-brom ophenyl-) pentadecanoic acid: a new p o ten tial agent for m yocardial imaging, J. Nucl. Med. 21 (1 9 8 0 ) P58 (abstract). [32]
DEUTSCH, E., GLAVAN, K.A., FER G U SO N , D.L., D evelopm ent o f a Tc-99m m yocardial imaging agent to replace T l-201, J. Nucl. Med. 21 (1 9 8 0 ) P56 (abstract).
DISCUSSION
» E. PETURSSON: I understood from your oral presentation that cyclotrons will increasingly constitute part of the equipment o f nuclear medicine departments in the near future and that their costs will be about the same as for CT scanners. Could you tell us how much space is needed for a cyclotron and the number of staff necessary to run it and prepare the radiopharmaceuticals. H. WAGNER, Jr.: Our cyclotron will be installed in our nuclear medicine department in April 1981. The cost will be approximately one million US dollars. A positron emission tomography imaging system currently costs about US $750 000. We will have one cyclotron operator, supervised by a chemist and physicist.
\
INSTRUMENTATION AND TECHNIQUES Sessions 1 (cont.) and 2
IA EA-SM -247/37
A MULTICRYSTAL POSITRON SCANNER FOR QUANTITATIVE STUDIES WITH POSITRON-EMITTING RADIONUCLIDES H. OSTERTAG, W. KUBLER, R. KUBESCH, W.J. LORENZ Institute o f Nuclear Medicine, German Cancer Research Center, Heidelberg, Federal Republic o f Germany
Abstract A M U LTICRY STA L PO SITRO N SCANNER FO R Q U A N TITA TIV E STUDIES W ITH PO SITRO N -EM ITTIN G RAD IONU CLID ES. A non-tom ographic m ulticrystal w hole-body scanner for qu an titativ e po sitro n imaging has been designed. The d e te cto r system consists o f 64 coincidence d e te c to r pairs arranged in tw o opposing d e te cto r banks. N al crystals o f 38-m m d iam eter and 76-m m length are used. The p a tie n t moves linearly betw een th e statio n ary transverse d e te c to r banks. The scanning area is 64-cm wide and up to 192-cm long. The spatial resolution is 2 cm at a sam pling distance o f 1 cm. The plane sensitivity am ounts to 6400 co u n ts/s for a pure posi tro n e m itte r o f 1 /¿Ci/cm 2 . The accuracy o f q u an titativ e activity m easurem ents is b e tte r th an ± 15% for activities up to a few ßC i/с т 2 . The design o f th e in stru m e n t, and its capabilities and lim itations, are discussed. Initial experim ental and clinical results are presented.
,
1.
INTRODUCTION
In recent years institutes and industry have developed excellent instrumenta tion for positron-emission-computed tomography. While this method is very capable in regional imaging and quantitation, conventional positron scanning continues to play an important role where total organ or total body quantitation is significant, e.g. in studying the kinetics o f positron-labelled compounds by sequential whole-body imaging for dosimetric calculations and for slow metabolic studies. Only few single photon instruments with high efficiency for 5 1 1-keV gamma rays have been designed and are described in the literature. In 1966 Anger [ 1 ] presented his multicrystal whole-body scanner MARK II. A linear crystal was used by Crawley and co-workers [2 ,3 ] in their whole-body scanner. A similar 29
30
OSTERTAG et al.
instrument has been built in Finland [4]. 1 Of the tomographic instruments, to our knowledge only the ECAT [5] is capable o f a rectilinear scan mode in addi tion to its tomographic mode. Of course, rectilinear scanners and Anger’s tom o graphic scanner [6] have been used for partial- or total-body positron imaging. Some o f these instruments have been modified to include coincidence techniques. We have built an instrument specifically for positron coincidence measurement and eventually will extend it to single photons.
2.
DESIGN CONSIDERATIONS AND CHOICE OF PARAMETERS
Since the instrument should be capable o f imaging the total body as well as individual organs, a scanning system was chosen. A dual scanning camera was ruled out because o f the low efficiency o f the camera crystals for 5 1 1-keV quanta, and because o f the very severe count-rate problems with large unshielded crystals. A multicrystal system was selected, as high sensitivity is most important in whole-body imaging. If there is enough room, it is easier to move the patient than the heavy detector block. Because o f speed and comfort to the patient a linear movement without lateral scanning was chosen. Figure 1 shows the principle o f the positron scanner. The coincidence detector pairs are packed into two detector banks, one above and one below the patient. The detectors are assembled in horizontal lines displaced against each other. The lateral sampling distance is defined by the fixed displacement o f the detector lines. The lateral scanning width is determined by the full width o f a patient. The width o f the scanning area and the necessary sampling distance — which should be less or equal to one half o f the resolution distance — define the minimum number o f detectors. The plane sensitivity o f a single positron detector pair is given by
where R is the radius o f the round detector crystals, 2a is the distance between the two detectors o f the pair and e is the detection efficiency for the 5 1 1-keV annihilation quanta. The FWHMs o f the spread functions for a point source and a line source in the symmetry plane between the two crystals are 0.81 R and
1 M ARKKULA, A., VA URAM O, E., V IR JO , A., AH ONEN, P., V ILËN, H., CRAW LEY, J., V EALL, N., KEIN A N EN , M., KULM ALA, J., NÄ NTÖ, V., W ENDELIN, H., “ A hybrid scanner for po sitro n im aging” , these Proceedings, IAEA-SM -247/4.
31
IAEA-SM -247/37
F IG .l.
Principle o f the positron scanner.
0.91 R respectively. Thus the plane sensitivity o f a single pair detector is pro portional to the fourth power o f resolution:
S ~ (FWHM)4
(2)
If the resolution is improved by a factor o f 2, sensitivity is reduced by a factor o f 16. In fact, by halving the radius, a fourfold number o f detectors could be assembled in a given detector area (this is not quite true, as the shielding between the defector crystals cannot simply be halved) but there would still remain a loss o f a factor o f 4 in sensitivity, not to speak o f the tremendous increase in costs. Increasing the detector diameter not only decreases resolution but also creates new problems, since the geometry of small sources leads to differences in activity determinations. The count rates o f a central point source, a line source and a plane source o f the same activity in the field o f view o f a pair detector are in the ratio o f 1 : 4/ 37Г : 1/4. A depth-independent mean resolution o f 2 cm seemed sufficient for the purpose o f the instrument. Therefore a sampling distance o f 1 cm was chosen and 64 standard Nal detectors o f 38-mm diameter X 76 mm are used. The total efficiency o f such a detector for 5 1 1-keV gamma quanta is about 0.75 for an axial point source at 25-cm distance.
32
OSTERTAG et al.
FIG.2.
Sim plified block diagram o f the positron scanner.
The detector crystals must be embedded in a leád block. The thick frontal shield contains an aperture o f 35-mm diameter for each crystal. By this aperture the theoretical resolution for a line source in the midplane is determined as 16 mm FWHM. The thickness L o f the aperture shield is very important since it determines the singles count rate. For round apertures the ratio o f the coincidence count rate to singles count rate is proportional to L2 , everything else being constant. For an infinite plane source this ratio is o f the order o f 0.5% (L = 5 cm, 2a = 46 cm). The thickness o f the frontal shield is even more important for the ratio o f true to random coincidences. This ratio is proportional to L4 , i.e. an increase o f the thick ness o f the shielding by 20% reduces the random coincidences by a factor o f 2 .
3.
SYSTEM DESIGN
Each o f the two opposing detector banks (Fig. 1) contains the 64 Nal crystals (38-mm ф X 76 mm) with their pertinent PM tubes and preamplifiers. The detec tor arrangement is similar to that in Anger’s whole-body scanner MARK II [ 1]. The detectors are mounted in five parallel lines that are displaced in successive 1-cm steps. The distance between the centre lines is 5 cm. Each detector has a free aperture o f 35-mm diameter in a 5-cm-thick lead shield. The lateral shielding between individual detectors is at least 4 mm o f lead. The patient bed moves linearly — either continuously or stepwise — between the stationary detectors. The longitudinal and transverse sampling distances are 1 cm each, the transverse sampling distance being defined by the detector arrangement and the longitudinal sampling by the position o f the patient bed. The scanning area is 64-cm wide. Its length is variable up to 192 cm.
IAEA-SM -247/37
33
Only directly opposing detectors are coupled in coincidence. The coincidence time resolution is 2 т = 32 ns. The resolving time o f the detectors and the elec tronics is 1 jus. In Fig.2 a simplified block diagram o f the electronics o f the posi tron scanner is shown. Each detector pair has its own independent counting channel. Either coincidences or singles can be counted. The thresholds o f the constant fraction discriminators are variable between about 60 and 600 keV, typically 100 keV being used. The system is controlled by a small computer (DEC LSI-11). The raw data are continuously shown on a modified TV display with an image matrix o f 128 X 192 pixels. Two whole-body scans can be presented simultaneously. Data are stored on a floppy disc. The instrument is supplemented by an independent image processing system that is controlled by a second LSI-11. Data transfer is done via floppy disc. The image processing device offers an interpolated presentation o f the scans on a finer display matrix o f 256 X 384 pixels. Further image processing permits optimization o f display parameters, colour or grey scales, and the use of regions o f interest and profiles. The final images are recorded on-X-ray film.
4.
MEASUREMENTS
4.1. Count rates Image quality and accuracy o f quantitative activity measurements depend strongly on the count-rate capabilities o f the instrument. The region o f useful count rates is determined by the resolving time p o f the electronics including detectors, and o f the coincidence resolving time 2 r. Figure 3 shows the expected rates o f t h e singles, the t r u e c o in c i d e n c e s a n d the r a n d o m c o in c i d e n c e s a s a f u n c tion o f activity density for a single detector pair o f the positron scanner. An infinite plane source o f constant activity density placed in the symmetry plane without absorbing material was assumed. The most critical point, the rapid increase o f the random coincidences with the activity, is obvious. As an infinite plane source was assumed this figure represents something like a worst case. In clinical applications the activity distributions are smaller, and the apparent activity densities are usually less than 1.5 pCi/cm2 due to attenuation. 4.2. Calibration, adjustment o f sensitivity For the purpose o f calibration, a homogeneous flood phantom o f 27 X 70 cm2 , filled with 68Ge solution (giving about 1.4 ¿iCi/cm2 o f 68Ga) is successively placed directly in front o f each detector bank.2 The thresholds o f 2 1Ci= 3.70X 1010Bq.
34
OSTERTAG et al.
FIG.3. E xpected singles count rate (N s), true ( N j) and random coincidences (Np_) as a fun ction o f a ctivity density fo r a single d etecto r pair. Parameters: p = Ißs, 2 t = 32 ns, d e te cto r distance 2a = 46 cm, aperture diam eter 2 R = 3.5 cm, efficiency e = 0.6, effective thickness o f shielding L eff = L —2 ß l = 5 —1.19 = 3.81 cm. D o tte d line: ideal singles and true coincidences.
the constant fraction discriminators are adjusted by the computer until the single count rates o f all detectors are within ± 1%. The remaining sensitivity differences in the coincidence channels are corrected for by a coincidence measurement with the flood phantom. 4.3. Emission and transmission scans Either simple emission scans can be carried out or combined emission/ transmission scans for absolute activity determinations. In the second case there are two possibilities: 1.
One starts with the transmission scan followed by injection o f the radio nuclide and sequential emission scans (e.g. when measuring the kinetics o f a labelled compound).
2.
One starts with the injection, and after a time interval necessary for sufficient uptake, consecutive emission and transmission scans are taken (e.g. bone scans).
Both methods are equivalent. Generally the statistical errors o f the first method are smaller, but most often the second method will be dictated by the type o f investigation. In both methods the radioactive decay o f the nuclide can be corrected for exactly.
35
IA EA-SM -247/37 rel. units
fiü-cm ‘
-
Г — 1 T I 1--- 1-1- —I-»-12 4 6 8 10 12 Absorber thickness
/cm
FIG.4. Calculated a c tiv ity densities fo r variable activities in a 12-cm 'thick water phantom (left) and variable thickness o f water absorber (right).
The transmission scans are taken by placing the rectangular 68Ge flood phantom upon the lower detector bank during the scan. The total activity in the phantom is about 2.6 mCi. The gamma dose rate at the surface o f the phantom is 40 mrad/h. The additional dose to the patient during the trans mission scan is less than 10 mrad.
5.
RESULTS
5.1. Sensitivity and resolution The plane source sensitivity o f the positron scanner is 6400 counts/s for a pure positron emitter o f 1 juCi/cm2 and a detector distance o f 46 cm, without absorber. The FWHM o f the measured line spread functions for a single detector pair with the same detector distance is 16.5 mm in the symmetry plane and 21.0 mm at 11 cm above and below the symmetry plane. The corresponding figures for the LSF with 12 cm o f water as absorbing medium are 17.5 mm and 21.0 mm. This shows that the sampling theorem is approximately fulfilled by using a sampling distance o f 1 cm, and that misinformation will not occur. 5.2. Activity measurement The results o f a phantom experiment are demonstrated in Fig.4. In the left part o f the image the activity in a 1 2 -cm-thick water phantom was varied, in the right part a constant activity was covered by water o f increasing thickness.
FIG.5. Exam ples o f bone scans with lsF. L eft: W hole-body scan o f a 4 7-year-old man w ith right knee trauma; R ight: A nterior and lateral views o f th e skull o f a p a tie n t w ith a m eningiom a.
IA EA-SM -247/37
37
FIG. 6. Gallium-68 liver scans o f a 38-year-old man. L eft: emission scan; M iddle: transmission scan; R ight: attenu ation -corrected emission scan.
In both cases one observes the correct relation between calculated and real activity density. This measurement was done with a 36-cm detector distance, and the random coincidences were taken into account. In work with patients, a detector distance o f 46 cm is generally used. Under these circumstances and without considering the random coincidences, the accuracy o f the calculated activity density is about ± 15% for activities up to about 1 juCi/cm2 for extended sources and 5 MCi/cm2 for small sources o f about 4 cm diameter. 5.3. Clinical examples As examples o f whole- and partial-body images Fig.5 shows some 18F bone scans. The whole-body scan on the left side was taken 3.5 h after IV injection of 5 mCi o f 18 F. The total activity at the time o f measurement was about 600 pCi. The scan time was 20 min. Typical whole-body scan times are 6—10 minutes for bone scans taken two hours after injection o f 5 mCi o f 18F. For short-lived radionuclides, half-life correction is provided. The system has been used for totalbody imaging even with the very short-lived 13N.
38
OSTERTAG et al.
At the right o f Fig.5 frontal and lateral scans o f the skull o f a patient with a meningioma are shown. The patient was imaged 2.3 h after injection o f 10 mCi o f 18F. Each scan took 4 min. Figure 6 shows an example o f an activity determination. Emission and transmission scans were measured, starting 8 min after injection o f 616 pCi o f a 68Ga-labelled liver imaging compound. The whole-body scan times were 20 min each. On the right side the attenuation-corrected image o f the activity distribution is presented. The activity in the ROI was found to be 55% o f the total-body activity, calculated for the starting time o f the measurement.
6.
DISCUSSION
The examples show that the positron scanner is useful for clinical investigation and that it represents an appropriate compromise between the required high sensi tivity and the necessary resolution. The sensitivity o f the positron scanner is comparable with the sensitivity o f a large field gamma camera for 99Tcm. Assuming a collimator efficiency o f 2 X 10'4 , with 38-cm field o f view, and a plane source with 1 pCi/cm2 , such a camera delivers about 6800 counts/s. The sensitivity could be increased by including oblique coincidence lines. This may be carried out in the future, but only for the direct neighbours o f opposing detectors as otherwise the depth-independent resolution will be lost. Obviously, the 2-cm resolution o f the positron scanner is much worse than that o f the gamma camera. But if one compares 99Tcm-whole-body scans taken with a gamma camera and whole-body installation, there is not much difference in resolution. Nevertheless the images look quite different, especially in the thoracic region, because o f the depth independence o f sensitivity and resolution, whereas in the camera image, mainly those parts that are close to the collimator are well imaged, others being blurred by absorption and depth-dependent resolu tion. Furthermore, only one scan is necessary with the positron scanner in contrast to the anterior and posterior scans with a gamma camera. To improve the accuracy o f activity determination it is essential to measure simultaneously the random coincidences, especially when higher activities are used than have been used up to now. Calculations have shown that the theoretical accuracy o f quantitation will be better than ±5%, independent o f the source geometry, for activity densities up to 10 juCi/cm2 if the resolving time p is reduced from 1 jus to 100 ns, e.g. by the pulse shortening method proposed by Muehllehner [7J. Appropriate modification o f the electronics is planned. A second set o f counters will be added in order to be able to measure the random coincidences ■ and to correct the measured count rate exactly. This, furthermore, offers the possibility o f using the instrument with collimators as a high-sensitivity double multicrystal scanner for any gamma-ray emitter.
IAEA-SM -247/37
39
ACKNOW LEDGEM ENTS
The authors would like to thank P. Wörner, W. Obers (mechanics) and H. Kohler (electronics) for their invaluable technical assistance in the design o f the instrument.
REFERENCES [1]
A N GER, H.O., W hole-body scanner M ARK II, J. Nucl. Med. 7 (1 9 6 6 ) 331.
[2]
CRAW LEY, J.C.W ., V E A L L , N., “ The design and som e clinical applications o f a hybrid scanner” , Medical R adioisotope S cintigraphy 1972 (Proc. Sym p. M onte Carlo, 1972) 1, IAEA, V ienna (1 9 7 3 ) 105. CRAW LEY, J.C.W ., GIBBS, G.P., “ Q uantitative w hole-body scanning using a hybrid scanner” , Medical R adionuclide Im aging (Proc. Sym p. Los Angeles, 1976) 1, IAEA, V ienna (1 9 7 7 ) 369. AHONEN, P., “ H ybrid scanner: principles o f fu n ctioning” , M edical A pplications of C yclotrons (Proc. Sym p. T urku, 1977), A nn. Univ. T urku. Ser. D 8 (1 9 7 8 ) 129. PH ELPS, M.E., HO FFM AN N, E .J., HUANG, S.C., KUHL, D .E., ECAT: A n e w com puterized tom ographic imaging system for p ositron-em itting radiopharm aceuticals, J. Nucl. Med. 1 9 (1 9 7 8 ) 635. AN GER, H.O., “ M ultiplane tom ographic gam m a-ray scanner” , M edical R adioisotope Scintigraphy (Proc. Sym p. Salzburg, 1968) 1, IAEA, V ienna (1 9 6 9 ) 203. M U EH LLEH N ER, G., CO LSH ER, J.G ., Use o f position sensitive d etectors in positron
[3]
[4] [5]
[6] [7]
imaging, IE EE Trans. Nucl. Sei. NS-27 N o .l (1 9 8 0 ) 569.
IAEA-SM -247/4
A HYBRID SCANNER FOR POSITRON IMAGING A. MARKKULA, E. VAURAMO, A. VIRJO Institute o f Biomedical Sciences, University o f Tampere, Tampere, Finland P. AHONEN, H. VILEN Technical Research Centre o f Finland J.C.W. CRAWLEY, N. VEALL MRC Clinical Research Centre, Harrow, Middlesex, United Kingdom M. KEINÄNEN, J. KULMALA, V. NÄNTÖ, H. WENDELIN Cyclotron Laboratory, University o f Turku, Turku, Finland
Abstract A H Y BRID SCANNER F O R PO SITRO N IMAGING. A h ybrid scanner has been built, w hich is specially designed fo r to ta l b o d y m etabolic studies w ith short-lived labelled com pounds. T he scanner consists o f tw o parallel detectors, 40-cm long, one above and one below the p atien t, w hich m ove along the body. T he positions of scintillations along the d e te cto rs are detected by pho to m u ltip liers at the ends o f the crystals, and the p ositions o f scintillations along th e p a tie n t’s body are derived from th e po sitio n of the m echanism carrying th e detectors. B oth single p h o to n and coincidence m odes are possible. The collim ation was optim ized for sem i-dynam ic total-body m etabolic studies. T h e sensitivity o f the in stru m en t to a po in t source is a bout 400 counts/juCi per m in ute in the single p h o to n m ode and 60 counts/piCi per m inute in the coincidence m ode w ith po sitro n energy. T he resolution is 28 m m versus 22 m m (FW HM ), and the m axim al co u n t rates w ith 25% loss are 2 2 0 0 0 versus 1 700 counts/s. T he scanning speed can be varied from 0.5 to 20 cm /s and the in stru m en t is designed to scan five tim es per m in u te over one m etre o f the body. T his m akes relatively accurate m easurem ents possible on p h enom ena th a t occur over 2 to 20 m inutes. As prelim inary applications the accum ulation o f 1SF in a ra b b it and a study o f e thanol m etabolism in a hum an subject are described.
1. ORGANIZATION The work reported here was begun at the initiative o f N. Veall. The developed equipment — the hybrid scanner — has been built in Tampere by the University 41
MARKKULA et al.
FIG.l. Schematic diagram o f the hybrid scanner.
•,i A O ■*»*'" ■ямйвм»
Ш КмВт
k -п , ;
яштйшш
Xy__
N . V
ядвар ,
* i
-imyllJs!
‘•
FIG.2. The hybrid scanner.
.t.* ,r.'-
43
IAEA-SM -247/4
TABLE I. SPECIFICATIONS OF THE HYBRID SCANNER Single d e te cto r R esolution fo r positrons w ith collim ators (FWHM, m m ) A pparent c o u n t rates w ith 25% loss (co u n ts/s) Sensitivity to a p o in t source (counts//LiCi per s)
P o sitro n m easurem ent
28
22
22000
1700
400
60
Scanning area (cm 2)
40X 200
Scanning speed (cm /s)
0 .5 -2 0
o f Tampere and the Technical Research Centre o f Finland. After preliminary testing the scanner has been transferred to the Cyclotron Laboratory o f the University o f Turku, to be used in various applications with positron-emitting radionuclides.
2. AIMS OF THE PROJECT The project was started in 1974 as a university research and development project, so the aims have been partly educational. This also explains the relatively long time span o f the project. The primary aim, however, was to construct a low-cost, total-body imaging device for semi-dynamic imaging with positron-emitting radionuclides. There was a need to study metabolic phenomena in humans, especially with nC-labelled compounds such as alcohol and sugar. Here the kinetics o f the compound are unknown and the time spans o f the phenomena are from one minute upwards. In 1974, commercial positron cameras were not available, and now that they have become available, they have turned out to be expensive, with a typical exposure time o f about four minutes. The emphasis in this work has been on low cost and on speed, which has inevitably led to relatively modest resolution.
3. CONSTRUCTION AND SPECIFICATIONS OF THE SCANNER The scanner consists o f two parallel detectors, one above and one below the patient, which move along the body (Figs 1 and 2). Both detectors are
44
MARKKULA et al.
FIG.3. E xperim ent with a rabbit; active isotope 1SF, coincidence m ode, 7 0000 counts per scan, scanning tim e 8 minutes. Times below scan indicate tim e after injection.
IAEA-SM -247/4
45
40-crn long and 5-cm in diameter and are built from Nal(Tl) crystals. One is a stack o f 40 crystals each 1-cm thick, and the other is a single crystal. The positions o f scintillations along the crystal assemblies are detected by two photo multipliers, one at each end [ 1 ], and the positions along the patient’s body are derived from the position o f the mechanism carrying the detectors. The electronic circuits include facilities for both single photon and coincidence modes. An ordinary oscilloscope with a Polaroid camera is used as display medium. The collimators have been designed to maximize sensitivity for given resolution [2]. Collimators exist at present for three energies: " T c m, 131I, and positron-emitting isotopes. Only straight slit collimators are used. The optimal length o f the collimators turns out to be surprisingly high: about 12 cm for " T cm , and 20 cm for 131I and positrons. This is because there is a different dependence o f sensitivity on resolution compared with scanners or gamma cameras. On the other hand, the long collimators ensure more even response as a function o f depth. The shielding o f the collimators has been designed for energies up to 840 MeV, and the total weight with shielding and supporting case is about 150 kg. The main specifications o f the scanner are given in Table I. The single detector resolution is for the upper detector, which consists o f 40 pieces. For the lower crystal, which consists o f a single crystal, the resolution is poorer, about 60 mm FWHM with positrons. This is caused by much lower attenuation o f the light quanta along the axis o f the crystal compared with the upper detector. However, the energy resolution is better in the lower detector for the same reason. In coincidence measurements with positrons the resulting total resolution is, of course, better than the resolution o f either o f the detectors, so that in positron measurements the poor resolution o f the lower detector is not a serious drawback. Although an oscilloscope with a Polaroid camera is used for display, the electronic x-, y- and z-signals coming from the scanner allow coupling to any standard data handling equipment ordinarily used with gamma cameras.
4. APPLICATIONS Many possible applications o f the hybrid scanner have already been described elsewhere [3, 4]. Here we only report two preliminary experiments with positronemitting radionuclides. 4.1. Experiments with animals The accumulation o f 18F in the skeleton o f a rabbit was followed after the intravenous administration o f 18F (NaF, activity 2.25 mCi).1 The images show 1 1 C i= 3 .7 X 1010Bq.
46
MARKKULA et al.
5MIN
30 MIN
40 MIN
50 MIN
60 MIN
70 MIN
FIG.4. E xperim ent w ith a human subject; nC-ethanol, single m ode, a bou t 1 2 0 0 0 0 counts per scan, scanning tim e 4 minutes. Times below scans indicate tim e after the peroral adm inistration.
(see Fig.3) that 60 minutes after the tracer injection there was an accumulation o f 18F in bones, and within 90 minutes the uptake o f 18F in the vertebral spine, skull bones and the joints o f the lower extremities became distinct. It has thus been demonstrated that the resolution o f the the hybrid scanner is good enough for animal studies. 4.2. Pharmacokinetics o f ethanol in a human subject The experiments with nC-ethanol are part o f a project to study the pharmaco kinetics and metabolism o f ethanol with chronic alcoholics and controls. With
IAEA-SM -247/4
47
nC-labelled ethanol it is possible to conduct human experiments and measure the turnover rates in different organs. A method o f measuring the kinetics o f the ethanol metabolism is under test. As part o f this project, total body distributions after an oral administration are needed. The distribution o f ethanol was followed in one human subject after the peroral administration o f 1.5 mCi nC-ethanol. The experiment was carried out in the fasting state. There was a large amount o f nC radioactivity in the stomach during the experiment. Carbon-11 carried to the small intestine could be shown to be absorbed from the proximal parts o f the small intestine (Fig.4b). There was no large amount o f uC-ethanol in the medial or distal parts o f the small intestine, demonstrating the rapid diffusion o f ethanol across the intestinal wall to the portal circulation (Fig.4c). The proportion o f nC accumulating in the liver was found to be high (Fig.4e), compared with the UC fraction seen elsewhere in the body, which confirms the main role o f the liver in the metabolism o f ethanol. Carbon-11 radioactivity was seen to be equally distributed in the body, due to the distribution o f ethanol in the total water volume o f the body. There was no distinct accumulation of nC in the brain.
5.
DISCUSSION
The hybrid scanner described here has proved to be useful for semi-dynamic measurements with positron-emitting radionuclides. The main drawback is the low count rate capability in the coincidence mode (Table I). With existing electronic components it should be easy to reduce the dead times o f the apparatus by an order o f magnitude at low cost. The relatively poor resolution compared with commercial positron cameras is not a serious drawback in studies similar to that reported in Section 4.2.
REFERENCES [1] CRAW LEY, J.C.W ., VEALL, N., “ The design and som e clinical applications o f a hybrid scanner” , M edical R adioisotope Scintigraphy (Proc. Sym p. M onte C arlo, 1972) 1, IAEA, V ienna (1 9 7 3 ) 105. [2] VAURAM O, E., V IR JO , A., T he design o f the d e te cto r and collim ators for a hybrid scanner, Br. J. R adiol. 50 599 (1 9 7 7 ) 808. [3] CRAW LEY, J.C.W ., GIBBS, G.P., “ Q uantitative w hole-body scanning using a hybrid scanner” , M edical R adionuclide Im aging (Proc. Sym p. Los Angeles, 1976) 1, IAEA, V ienna (19 7 7 ) 369. [4] A H O N EN , P., “ H ybrid scanner, principles o f functioning” , M edical A pplications of C yclotrons (Proc. Sym p. T urku, 1977), A nn. Univ. T urku ., Ser. D 8 (1 9 7 8 ) 129.
IAEA-SM -247/65
A LARGE-AREA STATIONARY POSITRON CAMERA USING WIRE CHAMBERS A. JEAVONS, B. SCHORR, K. KULL CERN, Geneva D. TOWNSEND, P. FREY, A. DONATH Department o f Nuclear Medicine, Cantonal Hospital, Geneva, Switzerland
Abstract A L A RG E-A REA STA TIO N A RY PO SITRO N CAM ERA USING WIRE CHAMBERS. A t CERN and the Geneva H ospital investigations are being u n d e rta k en on the application of wire cham bers to po sitro n imaging. Wire cham bers are used extensively in high-energy physics research fo r th e follow ing reasons: (a) H igh spatial resolution ( ~ 1 m m ); (b) F ast resolving tim e ( ~ 20 ns); (c) High rate capability ( ~ 106p a rticle s, s"1 cm ’ 2 ); and (d) Large area for low cost. Clearly these are valuable a ttrib u te s fo r a p o sitro n cam era. T he difficulty is to retain these pro p erties while efficiently converting the 0.5 MeV gam m a rays resulting from positron annihila tion in to an ionizing form , suitable for d e te ctio n by a wire cham ber. A t C ERN this problem has been solved by th e ad dition o f a high-density d rift space to a p ro p o rtio n a l cham ber. A cam era is now in o p eratio n for po sitro n em ission tom ography. I t consists o f tw o 2 0 X 20 cm cham bers, 30 cm apart. E ach cham ber is 8.5% efficient for detecting 0.5 MeV p h otons. A coincidence tim e gate o f 100 ns is em ployed. R ecently, a new m ode of cham ber op eratio n has been achieved. W ith a suitable gas m ixture and a strong electric field, electron avalanche m ultiplication in the holes o f the converter is obtained. A n um ber o f advantages result: th e m ain one is a reduction in the coincidence tim e from 100 ns to 20 ns. T he spatial resolution is lim ited by th e range of the positrons in tissue. F o r a 1-mm line source of 22N a in plastic, we o b tain 2-mm FWHM, and for 68Ga, 3.5-mrn. W ith a 300 ßC i source o f 68G a inside a 20-cm diam eter plastic scattering bolus, the c o u n t rate obtain ed is 4000 coincidences s’"1, o f w hich 50% are random . W ith the new m ode o f o p eratio n and a tim e gate o f 20 ns, the random coincidence ra te will drop to less th an 20%. In vitro, three-dim ensional images containing m ore th an one m illion true, unscattered coincidences can be obtain ed in 20 m inutes. Lim ited-angle tom ographic reco n stru ctio n by F o u rie r techniques o f such images provides 16 slices o f 1-cm thickness. T hus the imaging tim e is 1—2 m in p e r slice o f ~ 10s events. T he cam era has been em ployed to image an isolated hum an h e art filled w ith 68Ga-EDTA, and y o ung rabbits injected w ith a bone-labelling tracer 68Ga-DTPMP.
1.
IN T RODUCTION M u l t i w i r e p ro p or ti on al chambers are u s e d e x t en si v el y as two-
dimensional, p ar ti c l e d etectors in h ig h - e n e r g y physics research for the following reasons:
49
50
JEAVONS et al. Lead Insulator
/ Photons i
Electric Field
FIG. 1. The principle o f the high-density drift space. P hotons are sto p p ed by the lead and produce fa st electrons that can escape to an adjacent hole. The electric field extracts the resulting ionization. N o te that ionization can be produ ced ahead o f the p o in t o f im pact o f the photon : this shortens the tim e resolution.
i)
H i g h spatial r es ol u t i o n
ii)
Fast r esolving ti m e
iii)
H i g h rate c ap ab il it y
iv)
1 mm). 20 n s ) . 1 0 6 p a r t i c l e s * s -1 c m - 2),
Large area for low cost.
Cle ar ly these are v al ua bl e attributes for a po si t r o n camera. T h e diffic ul t y is to r e t a i n these properties, w h i l s t e ff iciently c on ve rt in g the 0.5 M e V gamma-rays r esulting f r om p o s i t r o n a n n i h i l a t i o n into a n ionizing form, suitable for de t e c t i o n b y the w i r e chamber.
Ear ly attempts [1, 2 ] h av e no t re al iz e d the full
pot en ti al o f the w i r e chamber.
A t CERN, the p r o b l e m has b e en
s ol ve d b y the addition o f a high-de ns i ty d rift space [3]] to the w i r e chamber.
S u ch a drift space is il lu strated in Fig. 1.
It
consists o f a s an dw ic h o f lead a n d e p o xy -r es in fibre-glass sheets p er fo r a t e d w i t h a large nu mb er of holes close together.
Photons
interact in the lead bars a n d p r o du ce fast electrons w h i c h can e s cape to a n adjacent hole.
The free electrons, resulting from gas
i onization in a hole, m a y be extracted b y an el e ctric drift field
IAEA-SM -247/65
FIG.2.
51
A general view o f the CERN positron camera.
and d et e c t e d b y the w i r e chamber.
T h e top o f the converter is
c o ve re d w i t h a sheet o f le ad g l u e d to a flat glass p la t e to e l e c tr ically "c l o s e 1 the c o nverter a n d p r e ve nt electrons drifting b a c k w ar d s to the c h a m be r window, w h i c h is at e ar t h potential. A t CERN a n d the G e n e v a Ho sp it a l the ap pl i c a t i o n o f such h i g h - d e n s i t y propor ti o na l chambers to m e d ic a l p o s i t r o n imaging is u n d er investigation, f ollowing the successful a p p li ca ti o n of such a s y s t e m to the m ea su r e m e n t of the angular c o rr e la ti on of p o s i t r o n a n n i h i la ti on r a di at io n for soli d- st at e physics research [4,5].
In this p a p e r the p e r fo r ma nc e of the latest ca me ra is
presented, togeth er w i t h the t om o graphic r e c o n st ru ct io n t ec h niq u e an d some imaging results.
Details m a y be foun d elsewhere
of chamber c o n s t r uc ti o n [6,7J a n d electronics [8], and p r e l i m i n ar y m e d i c a l imaging results [9].
52
JEAVONS et al.
FIG.3.
The spatial resolution o f the camera fo r various line sources.
2.
C A M E R A PERFORMANCE
2.1
Imaging m et h o d T he c am er a consists o f two 20 x 20-cm chambers, 3 0 - c m apart:
see Fig. 2.
Each c h a mb er is e qu ipped w i t h a 1-cm thick c onverter
d r i l le d w i t h 0. 8- mm di a me te r holes o n a h exagonal p a t t e r n w i t h 1 m m b e t w e e n hole centres.
T he object to b e imaged is p l a c e d b e
tw e e n the chambers. B y ba ck - pr o j e c t i n g all the coincident gammar a y pairs de te c t e d b y the chambers, a b l u r r e d three-dimensional image o f the object is obtained.
Because of the limi te d solid-
angle acceptance of the camera the spatial r es ol u t i o n i n the d i r e c t i o n perp en di cu l ar to the chambers is degraded, and so the image is treated as a series of thick slices p ar allel to the chambers.
A future c amera w i l l comprise four chambers to p r o
vide equal r es ol u t i o n in a ll d i m e n s i o n s .
A m at h e m a t i c a l simu
l ation o f this p r o b l e m is prese nt ed in the next section.
The
b lu rr i n g in the b a c k - pr o je ct ed image is rem o ve d b y frequency space filtering, taking into account the limited acceptance of the camera.
53
IAEA-SM -247/6S
FREQUENCY(cm-1) FIG.4. The m odulation transfer fun ction (Fourier transformJ o f the data in Fig.3 com pared with other cameras [11, 12].
2.2
Spatial r es ol ut io n The over-all spatial r e s o l u t i o n o f the camera is de te rm in e d
b y several factors:
the chamber resolution, the p a r a l l a x error
due to the c onverter thickness, the d eparture f ro m c o llinearity of the two ga mma rays, a n d the distance trav e ll ed b y the p os it r o n before annihilation.
E a c h of the factors c a n b e ^ 1 m m [9,10"!.
We h av e m e a s u r e d the over-all re s ol ut io n for a 1 -m m line source located ce ntrally b e t w e e n the chambers, and find that, as e x p e c ted, the p o s i t r o n range err or dominates (Fig. 3).
For 22Na
(maximum p o s i t r o n en e r g y 0.5 MeV) in p l a st ic w e ob t a i n 2-m m FWHM, a r esult sl i gh tl y b e tt er t ha n that of earlier w o r k [9], p ro b ab ly due to the thi nn er converter. degrades to 3.5 mm.
For 68Ga (1.9 MeV) the r es ol u ti on
Repl ac in g the p l a s ti c tube w i t h tu ngsten
eliminates the p o s i t r o n range error, a n d the result for 22N a is ne a r l y reproduced. B y taking the Four ie r tr a n s f o r m o f the line -s p re ad functions of Fig. 3, the spatial freq ue nc y sp e ct ru m or m o d u l a t i o n transfer fu nc t i o n o f the ca me ra is obtained.
This is c om pa re d w i t h other
54
JEAVONS et al.
cameras 1_11,12J in Fig. 4.
It m a y b e se en that m u c h b e t t e r c o n
trast is achiev ed at all spatial frequencies. 2.3
E n e r g y resolut io n A l t h o u g h there is no correspondence b e t w e e n p h o t o n e ne rg y
a n d chamber p ulse height, the c am er a does have a n im portant in t rinsic sens it iv it y to p h o t o n energy.
A s the p h o t o n en er gy falls,
the inter ac ti on p ro b a b i l i t y rises as the ph o to el e c t r i c effect dominates.
This w o u l d increase the d e t e ct i on efficiency, b ut it
is m o r e than offset b y the fact that the lower e n er gy p h o t o e l e c trons ha ve lower probabilities o f e sc aping f ro m the lead.
Down
to ab out 200 k e V d e t e c t i o n ef f ic ie nc y falls slowly, after w h i c h it drops rapid l y [3J.
Thus the c a me ra is insensitive to a s i g
n if ic an t amount o f the C o mp to n- s ca tt er ed r a d i at io n that exists in a c li nical imaging situation.
The r e j e c t i o n should b e as good
as w i t h Nal w i t h a w i d e - e n e r g y window, and comparable w i t h BGO [10]. 2.4
C ou nting rate T h e m a x i m u m useful count rate o f the c amera is limi te d b y
the occurrence o f accidental coincidences.
A fi gure-of-merit
exists [9j, w h i c h is (chamber d e t e c t i o n e f f i c i e n c y ) 2/coinc i de nc e time resolution.
Currently, the d et e c t i o n ef f ic ie nc y o f ea ch
c h a mb er is 8.5°s, and, u n t il recently, the time re so lu ti o n w as 100 ns.
Increasing the d e t e ct io n e ff ic ie nc y requires a thicker
converter, to stop mo r e photons, w h e re a s d ec r ea si ng the time r e so lu t i o n demands a thinner co nverter as the time r es ol u ti on is d o m i n a t e d b y the dr if t- ti m e o f the free e lectrons through the converter. converters.
T h e so lu t io n to this d i l em m a is to u s e several t h i n A camera has re c en tl y b e e n annou nc ed [13,14] that
takes this pri nc ip l e to the limit:
w i r e planes are interleaved
w i t h e ve r y sheet o f the converter.
T h e time re s ol ut io n o f the
b a s i c propor ti on al c h a m be r is ac hi e v e d (20 n s ) , bu t the spatial
55
IAEA-SM -247/65
TOTAL COINCIDENCE RATE (s'1) FIG.5. The dependence o f the accidental coincidence rate on the to ta l counting rate in a sim ulated clinical imaging situation.
r es ol ut io n degrades to ^ 6 m m as there are n o holes in the c o n ve r t e r to trap the fast e l e c t r o n s . The dri f t time w i t h i n a conv er te r m a y b e r e d u c e d b y o p t i m i z i n g the gas m i x t u r e in the chamber and the dri ft volt a ge app li ed to the converter.
A time re s ol u t i o n of 100 ns is achieved
u s i n g a m i x t u r e o f 70% m e t h a n e an d 30% n eo p en ta ne (2-2 dimethylpropane).
M e t h a n e pr o vides a hi g h e l e c t r o n dri f t ve lo c i t y
(107 .cm-s- 1) at m o d e s t d ri f t fields (1 kV* c m ”1) a nd the n e o p e n tane improves the speci fi c ioniza ti on o f the gas m i x t u r e a nd stabilizes the cham be r at h i g h g a i n o pe r a t i o n so that full e f f ic iency m a y be achieved.
The co unting rate o f the c am er a usi ng
this gas m i x t u r e is g i v e n in Fig. 5.
A ra p i d b ui l d - u p of a c c i
de nt al coincidences w i t h increasing source st re n g t h is evident.
FIG. 6. The tim e spectrum fo r a 4-m m-thick converter in coincidence w ith a fa st scintillator. Electric field in the converter: 10 k V c m ' 1. Gas
D rift velocity (c m 's'x)
FWHM (ns)
(aI Isobutane (b) Carbon dioxide ' (c) N eon + 7% carbon dioxide
~5X106 ~ 101
72 36
~ 2 X 101
12
N o te the lower fla t (accidental coincidence) background in (c).
57
IAEA-SM -247/65
W i t h a 300-yCi source of 68G a inside a 20-cm d i am et er p l as ti c s ca tt er in g bolus, a coincidence rate o f 4000 s-1 is obtained, of w h i c h 50% are accidental coin c id en ce s. 1 In an effort to decrease the time re s o l u t i o n further, the u se of neon, w h i c h at h i g h d rift fields is a f aster gas than m e th an e [ 1 5 J , has b e e n investigated.
A small amount o f quenching
gas, s u ch as ca r b o n diox i de w h i c h also has a h i g h d rift ve lo ci t y at h i g h fields [l5j, is a d ded to stabilize the c h a mb er operation. This type o f gas m i x t u r e is k n o w n as a Pen ni ng mixture, as it exhibits a c on siderable increase in ionizat io n at m o d e s t fields 5 k V c m - 1 ) owin g to the P e n n i ng effect [ 1 6 , 1 7 j.
In this e f
fect, e le c t r o n collisions excite gas atoms to m e t a s t a b l e states w h i c h c a nnot d e c a y b y p h o t o n emission; instead the atoms ionize the sec o nd ar y gas b y collision.
Thus a n e w m o d e o f chamber
o p e ra t io n is p o s s i b l e Г18~! : e le c tr on ampl if ic at io n in the holes 9 ~~ “ o f the conv er t er w i t h further a m p li fi c at io n in the w i r e chamber. This o p er at io n is s i m i la r to ot her w o r k [19] w i t h argon-acetone mixtures, b u t there the effect is as cr i be d to photo-ionization. T her e are b en efits of overco mi ng e le c tr on losses in the c onverter a n d ha v i n g c h am be r op e r a t i o n w i t h a safe, n o n p o ly m e r i z i n g gas mixture.
But it is the e ffect o n the time
r es ol ut io n that is important:
w i t h a n o p t i m u m amount o f c a rbon
dioxide (y 71) an el ec t r o n d r if t v e l o c i t y 'v 2 x 1 0 7 c m *s _1 is o bt ai n e d for a dri f t f ield of 10 k V c m - 1 .
Further, the a m p li f i
c at i o n renders the cham b er se nsitive to the first electrons to emerge f r o m the converter.
T h e se c a n be w e l l forw ar d o f the
poi nt of e m e r ge nc e o f the fast el ec t r o n fr om the lead (see Fig. 1). in Fig. 6:
T h e o ve r-all b e n ef i t of this m o de of o p e r a t i o n is shown a 4 - m m t h i ck con ve rt e r in c oi n cidence w i t h a fast
scint il la to r gives a time r e s o l u t i o n o f 12 ns FWHM.
This is
c om pa re d w i t h co nventional o pe r a t i o n in isobutane (72 ns) and 1 1Ci=3.70X 1010Bq.
58
JEAVONS et al.
c ar bo n dioxide (36 n s ) . the amplification:
Y e t another advantage m i g h t accrue from
a m e as ur e me nt of the p u l se height sh o ul d be
a n indication of the de p t h of the event in the converter, thus p r o vi di n g correc ti o n information for the time re so l ut io n and the p ar al la x error. In the next sec ti on the tomographic recons tr uc t io n techniques are d e s cr ib ed and in S e c ti o n 4 some imaging results are presented. T hese results w e r e take n w i t h me th an e/ n eo pe nt an e in the chambers and a 100 ns time resolution.
3.
IMAGE RE CO NSTRUCTION T he re co nstruction o f a three-dimensional image o f the a c
t iv it y distri bu ti on w i t h i n the object p l a c e d b e t w e e n the d e t e c tors is perf o rm ed b y F o u ri er deconvolution,
This,se ct i on
summarizes the re co ns t ru ct io n algorithm, a nd then presents a s im ul at io n stu d y that illustrates some difficulties o f limited angle tomography. 3.1
T he or y A n u n k n o w n d is t r i b u t i o n o f p o s i t r o n a ct iv it y f(x), a fter
i maging a nd b a c k p r o j e c t i o n of individual p o s i t r o n a nn ih i la ti on events, is tra ns fo r me d into a fu nc ti o n f(x) through c o n vo l ut io n w i t h the s p a t ia ll y in variant p o in t respo ns e fu nc ti on o f the i maging system, h ( x ) : f(x) = f(x) * h(x)
(1)
w h e r e * denotes a three-dim en si on a l convolution, a n d x = (x,y,z). Let к = (kx , ky, k z) b e a p o i nt in fr equency space and d efine the t hree-dimensional Fourier tra ns fo rm T for a ny function f(x) b y ° 0 CO CO T [f]Q 0
= f
f f f ( x ) e"
. —
dx = F(k)
(2)
59
IAEA-SM -247/65
T a k i n g F o u ri er transforms o f Eq. (1) a nd u s i n g t he c o n v o l u tion theorem: T [ f ]( k) = T[f](k) T[h](k)
(3)
S o l v in g for f(x), as suming T [ h ] (k) f 0 f(x) = r ' t T f f j M / T ' L h j M j w
(4)
w h e r e T _1 denotes the inverse Four ie r transform. e x pr es s i o n
for T [ h J ( k ) = H(k) m a y b e o bt ai ne d
b y as s uming
algebraic ex pr es s i o n for h(x) and evaluating the Fourier integral.
A clos e d- fo rm an
co r re sponding
T h e f o r m o f h(x) appropriate for a s y s t e m of
limited angular acceptance is h(x) = ——
if
x e R
2ïïr2 = 0
(5) x i R
if
w h e re r = |x| and R is the co ne - s h a p e d r e g i o n w i t h semi-angle y c orresponding to the c h a m be r a cc ep t an ce (Fig. 7).
Evaluat in g
H(k) w i t h h(x) g i v e n b y Eq. (5), the result, e x pr es se d in p o lar coordinates, is o f the form Н(к,©,Ф ) = ¿
С(0,Ф;ЧО
w h e r e the angl e- d ep en de nt f u nc t i o n С(0,Ф;Ч/) o riginates f r o m the n o n - u n i f o r m sa m pl in g o f Fouri er space.
Preci s e expressions for
С(0,Ф;ЧО m a y b e f o u n d e lsewhere [20], w h e r e it is sh o w n that for
f < тт/2, С(0,Ф;4О contains a cone of zeros.
T he semi-angle of
this cone increases w i t h d e cr ea s i n g Y, and the c on d i t i o n H(k) t 0, re qu ir ed for Eq. (4), cannot b e s a t is fi e d for all к w i t h i n the ba n d w i d t h o f the imaging system.
Instead, a f re quency-space
funct io n Fç(k) is obtained, s uc h that, ideally, Fc (k) = F(k) = 0
if
кe С
if
к фС
60
JEAVONS et al.
x//( = moximum acceptance angle in xz plane = maximum acceptance angle in yz plane
FIG. 7.
The acceptance cone fo r a centrally position ed p o in t source; fo r square detectors,
ф1=ф2= ф .
w h er e С is the r eg i on for w h i c h H(k) ф 0.
The c or re sponding
r eal-space fu nc ti on f^(x) is g i ve n b y fc (x) = r T F j C x ) a nd the p ro x i m i t y of this fu nc t i o n to the true fu nc t i o n f(x) d e p ends s tr ongly o n ¥. However, if f(x) has com pa c t support, F(k) is a n entire f u nction and F^Qc) m a y b e c o n ti nu e d into the un m e a s u r e d cone, t here by p r o vi di n g a n estimate o f F(k) at the frequencies for w h i c h H(k) = 0.
Suppose, for a r e g i o n Вin real space f(x) f 0 = 0
if
x
eВ
if
x фЪ
and f(x) > £(x) , for some g i ve n functi on £(x)
61
IAEA-SM -247/6S
Papoulis [21j sugg es t ed a n iterative a lg o r i t h m that c a n be u s e d to continue Fr (k); Lent an d T u y [22]: 1)
the v e r s i o n d es c r i b e d he re is due to *th for the n iteration,
Set Re f?(x) = Re fn - 1 (x) Im f?(x) = (l-A?) Ini fn _ 1 (x)
2) Set fz(x)
= f?(x)
if
x e В
f?(x) = (1-A?) f?(x) 3)
4)
Set f?(x)
= f 2 (x)
f a ®
= fz(x)
if
if
x фВ
f?(x) > A(x)
+ * 3 U ( x ) - f?(x)>
Set Fn (k) = T[ f? ]( k)
if
f?(x) < A(x)
if
кe
к ФС
Fn (k) = A ^Fc (k) + ( l - A ^ T [ f ? ] ( k ) 5)
if
С
fn (x) = T _ 1 [Fn ](x) T h e fu nc t i o n
(n = 1).
is u s e d for f°(x) in the first iteration
T h e r e l a x at i on pa ra me te r s A^1, A^, A?, a nd A ^ m a y be
c h os en to improve convergence, b u t for the simulations descr ib ed below, all four parame te rs are unity, w i t h £(x) = 0 negative densities.
to exclude
It has b e e n s h o wn [23] that, in the absence
o f noise, fn (x) converges to f(x) for large n. 3.2 S im ul a t i o n studies S om e s i m u l a t i o n studies w e r e p e r f o r m e d to e x am in e the image d is to r t i o n c au se d b y t he limited a n g ul ar acceptance, a n d the e ffectiveness o f the iterative a l g o r i t h m to corr ec t for it.
To
a v oi d additional d is tortions due to imperfect filtering, define the op er a to r
su ch that for a ny f un c ti on F(k) in frequency
space, Q c [ F ] ( k ) = F(k) = 0
if
k e C
if
k é С
FIG.8.
(a) Tw o-dim ensional p o in t source distribution. (b) L im ited-angle reconstruction o f the p o in t source in (a); 4 /= 3 0 . (c) Sam e as (b¡, b u t after 30 iterations o f the L en t-T uy algorithm. The o b jec t region o f su pport is a square o f side 2 2 pixels. (d) Sim ulated ph an tom consisting o f a uniform disc o f un it a m plitu de and 18 p ixels diam eter, w ith an off-centre h o t s p o t o f size 2 X 2 pixels and three tim es the am plitude o f the disc. (e) L im ited-angle reconstruction o f the phantom in fdj; ф = 30°. (f) Sam e as fo r (c) f o r the phantom in (d).
63
IAEA-SM -247/6S
w h e r e the r e g i o n С w a s d e f i n e d above.
For p e r f e c t filtering,
Fc (k) = Q c [F](k)
and th erefore
has the effect o f zeroing all frequ en cy a m pl i
tudes w i t h i n the u n m e a s u r e d cone.
A d i s t or te d real-s pa ce di s tr i
bution, fç(x),is o bt ai ne d from a si mulated p h a n t o m distribution, f(x), b y fc (x) = r ^ f J C x ) w h e r e the d is tortions are due only to the m i s s i n g cone.
A l s o of
interest is the f u n c t i o n fn ( x ) , w h i c h is o bt ai n ed f r o m f ^ Q O b y n iterations o f the L e nt - T u y algorithm. F or simplicity, the simulations are t wo -dimensional a nd the f unctions f(x), fç(x), a nd f 3 0 (x) are d is p l a y e d i sometrically w i t h the z axis, w h i c h is pe rp en d i c u l a r to the p a i r o f detectors, run ni ng f ro m b o t t o m left to top right.
The p h a n t o m is ge nerated
w i t h i n a 32 x 32 grid, and a 30° acceptance angle is assumed. F igures 8a to 8c s h o w the f un c ti on f(x), f ^ ( x ) , an d f 3 0 (x) for a cen tr al ly p o s i t i o n e d p o in t source, a n d Figs
8d to 8f show
the same functions for a m o r e c o m pl ic at ed p h a n t o m consis ti ng o f a u n i f o r m disc o f d ia me t er 18 pixels w i t h a 2 x 2 pixel, offcentre, hot spot co nt ai ni n g three times the a c ti v i t y in the disc. T h e m o s t obvious features in compa ri ng Figs
8a and 8b is the
a ppeara nc e o f o scillations along the edge o f the acceptance cone, a n e l o n g a t i o n o f the poi nt s ource in the z-direction, a nd a 761 r e d u c t i o n in amplitude.
T h e s e oscillations h a v e b e e n d i s c us se d
in de tail b y T a m a n d Perez- Me n de z [24].
Imposing a n o bject s u p
p o r t of a c en t r a l l y p o s i t i o n e d square of side 22 pixels a n d p e r forming 30 iterations o f the Le nt - Tu y a l g o ri th m improves the sharpness of the source, b u t some oscillations remain, as w e ll as a n ampli tu de r e d u c t i o n o f 6 6 % (Fig. 8c).
64
F IG .9.
JEAVONS et al.
(a) A fou r-detector reconstruction o f the phantom in Fig.8(d); ф= 30° fo r each pair o f detectors. (b) A s fo r (a), after 30 iterations o f the L ent-T uy algorithm. Same o b jec t su pport as fo r the p o in t source.
For the d i s c p h a n t o m (Fig. 8 d ) , the limited angle d i s t o r tions are considerable (Fig. 8 e ) , w i t h the it eration proc ed ur e pro du ci ng o nl y m a rg in al improvement, m o s t of w h i c h comes f ro m imposing the object support (same as for the p oint source) in the first iteration.
Th e h o t spot amplitude r ed u c t i o n before
and after i teration is 62% and 45%, respectively. T h e importance of a dding a s econd p a i r o f detectors at 90° to the first pair, i.e. pe r p e n d i c u l a r to the x-axis, is il l u s t ra te d in Fig. 9, (a) be f o r e and (b) a fter iteration, for the
65
IAEA-SM -247/65
FIG.10.
(a) A s fo r Fig.8(f), but with a circular region o f support o f diam eter 18 pixels, (b) A s fo r Fig.9(b), b u t with the same region o f support as fo r (a).
disc phantom.
The e lo ng a t i o n in z is eliminated, and the hot
spot amplitude is redu ce d b y on ly 20% after iteration. p a r i s o n of Figs
Com
8f an d 9b shows the c onsiderable improvement.
The sen si ti vi ty o f the i t eration m e t h o d to a g o od knowledge o f the object supp or t is d e m o n s tr a te d in Fig. 10, w h e r e the s quare support r e g i o n has b e e n re p l a c e d w i t h the exact circular sup po rt o f the original phantom, for a) two d e tectors an d b) four detectors. only 8%.
In Fig. 10b, the h o t spot amplitude is r e d u c e d b y
66
JEAVONS et al.
No. of Iterations FIG .I L The residual, d , comparing the reconstructed image with the real distribution, p lo tte d as a function o f the num ber o f iterations. Curves fo r both tw o and fo u r d e tecto rs w ith square and circular o b jec t support.
The s e results are s um ma ri ze d in Fig. 11 for the disc p h a n tom, w h e r e the residual d, d e f in e d b y (() denotes m e a n value)
d2 = { Z D f c ® - f w 3 2} / { I [ f w - № ) > ] 2} is p l o t t e d as a fu nc ti on o f the n u m b e r of iterations, for two and four detectors w i t h b o t h square a n d circular o bject support. It is cle a r that imposing the ob ject supp or t r e g i o n in the first i teration is the m o s t important step in reducing d, b u t since d is a global figure, a sma l le r v a l ue does n o t always indicate a b et t e r image.
67
IAEA-SM -247/65
T o co nclude this section, it s ho u ld b e n o t e d that in p r a c tice the di st ortions il lu strated in Figs
8b a nd 8e are o f te n
m a s k e d b y increasing the p i xe l d i m e n s i o n in the z-direction, i.e. b y t aking thicker sections.
This is n o t u nreasonable, since
w i t h a 30° a cceptance angle it cannot b e e xp ec t ed that the same spatial r e s o l u t i o n m a y b e achiev e d in b o t h x a n d z. E v e n th o ug h the simulations p r e s e n t e d in this s e ct io n w er e two di me nsional a n d nois e free, the importance o f at least two pairs o f detec to rs a nd g o o d k n o wl ed ge o f the o bj ec t support has b e e n cle a rl y demonstrated. 4.
IMAGING W I T H 68Ga To c om plete this paper, seme images, t a k en w i t h the p os i t r o n
c am er a d e s c ri b ed above a nd re c on st r u c t e d u s i n g the m e t h o d o u t lined in S e c t i on 3.1, are presented.
The p o s i t r o n emit te r was
68G a - E D T A (heart study) and 68G a- DT PM P (rabbit studies). A series o f i n vitro images of a n isolated h u m a n heart, the c avities fi ll ed w i t h 68Ga-EDTA, are s ho w n in Fig. 12 to d e m o n s trate the to mo graphic c a pa bi l i t y o f the camera.
S i x te e n se qu e n
tial sections, 1 - cm t h i ck are shown, as v i e w e d f r o m above the heart.
Sections (a) to (c) and sections (n) to (p) are located
p h y s i c a l l y b e l o w a n d above the heart, respectively.
Intermediate
slices sh ow ac ti vi t y in the r ight a n d left ventr ic le s and right a nd left atria, w i t h the interve nt ri cu l ar se p tu m clear ly seen in slices (f) and ( g ) .
T h e card ia c b o r d e r in slice (i) is f ormed
b y the co ro na ry sinus, the lower limit o f the left at r i u m and rig ht atrium, w i t h the entrant p o r t i o n of the inferior v e n a cava o n the p o s t e r i o r an d r ight lateral side.
The r ight and left a n
terior bor de rs are f ormed b y the rig ht a nd left ventricles.
In vivo imaging o f a series of y o u n g rabbits w a s p e r f or me d w i t h the bo ne - l a b e l l i n g tracer 68Ga-DTPMP, w h i c h w a s injected into an ear vein.
A b o u t 1\ hours af ter injection, the rabbit
'
68
JEAVONS et al.
FIG. 12. Sixteen transverse slices o f an isolated human heart. A b o u t 1 0 s counts in each central slice. Imaging tim e equivalent to 1 min per slice.
w as anaesthetized, a nd imaging b e g u n with, typically, less t ha n 100 y Ci o f activity w i t h i n the f ield of view.
Some r e p r e s e n t a
tive images ap p ea r in Fig. 13. T h e first image (a) shows the blurred, low-contrast, backp ro j e c t e d d is t ri bu ti on o f a single slice through the front p o r t io n o f a rabbit.
T he next image (b) shows the same slice after
Four ie r d e co nv ol u ti on o f the off-plane activity.
The skull,
cervical an d thoracic ve r t e b r a l column, s ho ulder joint and front leg are all clea r ly seen, as is the strong isotope absorp ti on in the individual vertebrae.
The h i g h spot o f ac tivity s ee n at top
left originates fr om residual a ct i vi ty at the site o f the intr a v enous i njection in the left ear.
A se c o n d slice (c) through
the same rabbit shows ac t iv it y in the ribs, a nd in the scapula.
IAEA-SM-247 /6 5
69
FIG. 13. In vivo imaging w ith a you n g rabbit injected w ith 100 ß C i o f the bone-seeking tracer №Ga-DTPMP. (a) The blurred, back-projected image o f a single longitudinal section. (bj The same section after Fourier deconvolution, showing the fro n t portion o f the rabbit. (с) A differen t section through the same rabbit. (dj A section showing the com plete rabbit. A ll sections are 1-cm thick and are displayed w ith a 2 mm X 2 mm p ix el size. The num ber o f counts per section is a b o u t 2 X 1 0 s . The grey scale assignm ent is w ith the black level corresponding to the minim um count density. The total num ber o f grey levels is 64.
Finally, a fo urth image (d) shows the s ke le t on of a small rabbit that is almost entir e ly w i t h i n the fie ld of v i e w o f the camera, w i t h the h e a d t uc k ed round, eliminating the c h a r a ct er is ti c c u r v a ture of the cervical ver te br al c olumn s ee n in the othe r images, The spine is clear ly visible, as are the individual vert eb ra e as far as the lower lumbar region.
The skull cap is outside the
field o f view, b u t a p o r t i o n o f the base, in cluding m a x i l l a and m andibula, can be seen.
H i g h e r activity, consist e nt w i t h in
cre as ed b o n e m e t a b o l i s m in y o u n g rabbits, c a n be s e e n at some
70
JEAVONS et al.
metaphysis - ep ip hy si s a r e a o f the front and rear legs, c o rr es p o n d ing to shoulder, elbo w 5.
and knee.
C ONCLUSION A p roportional cham be r p o s i t r o n camera has b e e n de m on st ra te d
w i t h the following characteristics: i) ii) iii) iv) v)
Spatial r es ol u t i o n 2-3 mm. Coincidence time re so l u t i o n ^ 100 ns. D e t e c t i o n e f f i ci en c y per cham be r ^ 10%. M o d e s t energy resolution, Large area, low cost.
Tomo gr ap hi c techniques have b e e n d e ve lo pe d to re construct images in three dimensions f ro m the limit ed set o f pr oj ections p r o v i d e d b y this statio na ry camera. 30 min.
Small animals c a n b e imaged in about
Devel op me nt is continuing to pro vi de a c amera compri si ng
four, larger (30-cm) chambers w i t h improved de t e c t i o n e ff ic i e n c y (20-300s) and time r e s o l u ti on (20 ns) for h um a n studies.
S u ch a
c amera sh ou l d m ak e a significant c o nt ri bu ti o n to the fi e l d o f n u c l e a r medicine. ACKNOWLEDGEMENTS W e w i s h to thank m a n y p e op le for their contributions: C. R i vo i r o n for chamber construction;
A. Gandi, D. Berthet,
J. G ue ri n an d A. Ba rba f or c onverter construction;
R. M a g n an in i
and G. L ee for solving ha rd wa re a nd software problems; an d G. Roub au d for the 68Ga radiopharmaceuticals; the p h ot ographs of Figs
12 and 13;
G. Ba nna
D. A d a m for
R. Nierhaus for the p r o g r a m
that gen er a te d the isometric plots o f Figs
8, 9
and 10.
Fina l ly
w e thank the CE R N Scientific T yping Service for ef fi ciently p r e p ar in g this paper, and Prof es so r L. V a n H o v e and Dr. E. Ga b at hu le r o f C E R N for their support a n d encouragement. This w o r k w a s s u pp or t ed in part b y the Fonds n at i on al suisse de la Recherche scientifique, gr ant No. 3.849-0.79.
IAEA-SM -247/6S
71
REFERENCES [1] REYNO LD S, R.A ., SN Y D ER, R .E., O V ERTO N , T .R ., A m ultiw ire p ro p o rtio n a l cham ber p ositron cam era: Initial results, Phys. Med. Biol. 20 (1 9 7 5 ) 136. [2] LIM, C.B., CHU, D., KAUFM AN, L., PEREZ-M EN D EZ, V., H A T T N E R , R., PRICE, D.C., Initial characterization of a m ultiw ire p ro p o rtio n a l cham ber po sitro n cam era, IEEE Trans. Nucl. Sei. NS-22 (1 9 7 5 ) 388. [3] JEA V O N S, A.P., CHA RPAK , G., STUBBS, R .J., T he high-density m ultiw ire drift cham ber, Nucl. Instrum . M ethods 124 (1 9 7 5 ) 491. [4] MANUEL, A.A., FISC H E R , O., PETER, M., JEA V O N S, A.P., An application o f pro p o r tional cham bers to the m easurem ent o f the electronic p roperties o f solids by positron annihilation, Nucl. Instrum . M ethods 156 (19 7 8 ) 67. [5] M ANUEL, A.A., SAM OILOV, S., FISC H E R , O., PE T ER , M., JEA V O N S, A.P., The use of high-density p ro p o rtio n a l cham bers for p o sitro n annihilation studies in alum inium and copper, Helv. Phys. A cta. 52 (1 9 7 9 ) 255. [6] JEA V O N S, A.P., T he high-density p ro p o rtio n a l cham ber and its applications, Nucl. Instrum . M ethods 156 (1 9 7 8 ) 41. [7] JEA V O N S, A.P., “ The CERN p ro p o rtio n al cham ber positron cam era” (Proc. 5 th Int. Conf. on P ositron A nnihilation, Lake Y am anaka, Japan), Japan In stitu te o f Metals, T okyo (1 9 7 9 ) 355. [8] JEA V O N S, A.P., FO R D , N.L., LIN D BERG , B„ PARKM AN, C., H A JD U K , Z., Twodim ensional p ro p o rtio n a l cham ber read o u t using digital techniques, IE E E Trans. Nucl. Sei. NS-23 (1 9 7 6 ) 259. [9] JEA V O N S, A.P., TOW NSEND, D.W., FO R D , N .L., KU LL, K., M A NUEL, A ., FISC H E R , O., PE T ER , M., A high-resolution p ro p o rtio n a l cham ber po sitro n cam era and its applications, IE EE Trans. Nucl. Sei. NS-25 (1 9 7 8 ) 164. [10] DOUGLAS, R .J., STEW ART, A.T., BIRD , L., “ Q ueen’s-Carleton m o difications to the Jeavons gam'ma cam eras — M edical imaging characteristics” (Proc. 5 th In t. C onf. on P ositron A nnihilation, Lake Y am anaka, Jap an ), Jap an In stitu te o f M etals, T o k y o (1979) 825. [11] H O FFM A N , E .J., HUANG, S.C., PHELPS, M .E., Q u an titatio n in p o sitro n em ission co m p u ted tom ography:
1. E ffe c t o f object size, J. C om p ut. A ssist. T om ogr. 3 (1 9 7 9 ) 299.
[ 12] BENNETT, G., B rookhaven N ational L ab o rato ry , private com m unication. [13] BATEMAN, J.E ., CON NOLLY, J.F ., A hybrid MWPC gam m a ray detecting system for applications in nuclear m edicine, Nucl. Instrum . M ethods 156 (1 9 7 8 ) 27. [14] BATEMAN, J.E ., CON NOLLY, J.F ., STEPHENSON, R., FL E SH E R , A.C., “T h e develop m ent of the R u th erfo rd L ab o rato ry MWPC p ositron cam era” (Wire C ham ber Conf. V ienna, 1980), to be published in Nucl. Instrum . M ethods. [15] BROWN, S.C., Basic D ata of Plasm a Physics, 1966, T he MIT Press, C am bridge, Mass. ( 1967). [16] LOEB, L.B., Basic Processes of G aseous E lectronics, U niversity of C alifornia Press, Berkeley and L os A ngeles (1961). [17] DRU Y V ESTEY N , M .J., PENNING, F.M ., T he m echanism of electrical discharges in gases of low pressure, Rev. M od. Phys. 12 (19 4 0 ) 88. [ 18] JEA V O N S, A., K U LL, K „ LINDBERG, B„ LEE, G „ TOW NSEND, D „ F R E Y , P., DONATH, A., A p ro p o rtio n a l cham ber po sitro n cam era for m edical imaging (Wire Cham ber Conf. V ienna, 1980), to be published in Nucl. Instrum . M ethods. [19] BRESKIN , A , CHARPAK, G., MAJEWSKI, S., M ELCHART, G „ PE T ER SEN , G „ SAULI, F., T he m ultistep avalanche cham ber: a new fam ily of fast, high-rate particle detectors, Nucl. In stru m . M ethods 161 (1 9 7 9 ) 19.
72
JEAVONS et al.
[20] SCH O RR, B., TOW NSEND, D., F ilters fo r lim ited angle tom ography, sub m itted to Phys. Med. Biol., 1980. [21] PAPOULIS, A., A new algorithm in spectral analysis and band-lim ited e xtrapolation, IE EE Trans. C ircuits and System s CAS-22 (19 7 5 ) 735. [22] LEN T, A., TUY, H., A n Iterative M ethod fo r the E x tra p o latio n o f Band-lim ited F unctions, M edical Im age Processing G roup Tech. Rep. MIPG 35, S tate U niversity o f New Y ork at B uffalo (1979). [23] ТАМ, K.C., PEREZ-M EN DEZ, V., M acDONALD, B., L im ited angle 3-D reconstructions from c ontinuous and pinhole projections, IEEE Trans. Nucl. Sei. NS-27 (1 9 8 0 ) 445. [24] ТАМ, K.C., PEREZ-M EN DEZ, V., L im ited-A ngle R e co n stru c tio n s using F o u rie r T ransform Itera tio n s and R adon T ransform Iterations, L aw rence Berkeley Rep. LB L-10207 (1979).
IA EA-SM -247/167
A RADIOISOTOPE SCANNER FOR USE IN DEVELOPING COUNTRIES J.C.W. CRAWLEY MRC Clinical Research Centre, Harrow, Middlesex, United Kingdom A.B. AJDUKIEWICZ, N. BASSETT Medical Research Council Laboratories, Fajara, The Gambia A.G. CRONQUIST, N. VEALL MRC Clinical Research Centre, Harrow, Middlesex, United Kingdom
Abstract A RAD IOISO TO PE SCA NNER FO R USE IN D EVELOPIN G COU NTRIES. A scanner was designed for use at the M edical R esearch C ouncil L aboratories in The G am bia, in a research p roject on the epidem iology and trea tm e n t o f hepatom a. T he d etector consists o f a 40-cm long by 5-cm d iam eter sodium iodide crystal assem bly w ith one p h o to m ultiplier at each end m oving sideways over th e area to be scanned. A fork lift tru ck form s th e m ain fram e o f the in stru m e n t w ith a yoke carrying the d e te cto r in place o f the forks. This assem bly can be raised and low ered by a m anually o perated hydraulic system and ro tated to scan from above, below or beside the p atien t. T he relatively thick d e te c to r can be used w ith in d iu m -113 as well as technetium -99m . A small cabinet m o u n ted on the fork lift tru ck houses the electronic circuits and carries the polaroid cam era for recording the scans. Rechargeable b atteries floating across a DC pow er supply b uffer the in stru m e n t against the wide fluctuations in m ains voltage and freq u e n t pow er cuts th a t occur at Fajara. T he in stru m e n t was designed and c o n stru cted at H arrow before shipm ent by sea to T he G am bia, w here it has been in use for a bout 18 m onths.
IN T R O DU C T I O N It c an b e a r g u e d that r e l a t i v e l y s o p h i s t i c a t e d a n d e x p e n sive te c h n i q u e s s u c h as r a d i o i s o t o p e s c i n t i g r a p h y s h o u l d n o t h a v e a h i g h p r i o r i t y in c o u n t r i e s whe r e there is a d e s p e r a t e shor t a g e o f m o r e b a sic m e d i c a l facilities. T h e r e are, however, s i t u a t i o n s i n w h i c h r a d i o i s o t o p e s c i n t i g raphy can he u s e d as a r e s e a r c h tool for epi d e m i o l o g i c a l stud i e s or to e v a l u a t e o b j e c t i v e l y t he r e s p o n s e to different forms of tre a t m e n t a n d t h u s
73
74
CRAWLEY et al.
h e l p to i n c r e a s e the e f f e c tiveness o f the l i m i t e d m e d i c a l faci li t i e s that m a y be available. Commercially available equip m e n t is oft e n u n s u i t a b l e for u se u n d e r the en v i r o n m e n t a l a n d o t h e r c o n d i t i o n s w h i c h pre v a i l in m a n y developing countries. T h e a d v e r s e c o n d i t i o n s that a r e frequently e n c o u n t e r e d include 1.
U n r e l i a b l e p o w e r supplies
2.
E x t r e m e s o f temp e r a t u r e a n d h u m i d i t y
3.
L o n g delays in o b t a i n i n g supplies or spa r e p a r t s .
if. S h o r t a g e o f s k i l l e d pers o n nel capable of o p e r a t i n g a n d m a i n t a i n i n g s o p h i s t icated equipment.
5- Shortage of funds for purchasing and maintaining expensive apparatus. 6.
P o o r d o c u m e n t a t i o n w i t h c ommercial apparatus l e a d i n g to m i s u n d e r s t a n d i n g w h e n o p erating or r e p a i r i n g t he apparatus.
T h i s p a p e r d e scribes a n i nstrument designed for us e at the M e d i c a l R e s e a r c h C o u n c i l Laboratories, Fajara, The Gambia, w h e r e t h e r e w a s a n e e d for a scan n e r in a research project on the in c i d e n c e a n d treatment of hepatoma.
DE S I G N O F T H E I N S T R U M E N T
The main considerations in designing the instrument were 1. Rugged construction for maximumreliability 2. Ease of maintenance 3. Ease of operation if. Minimum cost compatiblewith adequate performance 5. Independence of powersupplies 6 . Ability to use indium-113m (^^Inm ) as well as technetium-99m (^°Tcm). The advantage of ^^Inmover ^Tcmis that the generator has a longer half life ( 1 15 days compared with the 6 6 .7 h halflife of 9“rpcm)5 so delays in delivery can be tolerated. The high gamma energy (392 keV) is not ideal for the gamma camera because the thin crystal used in that instrument is relatively
IAE A -SM -247/167
75
F IG .l. The scanner. The electronic circuits are m oun ted on the fro n t o f the instrum ent with the swinging scanner head on the back in place o f the forks. The handwheel a t the side can rotate the d e te c to r assem bly to p o in t up, dow n or sideways.
insensitive to higher energy gamma rays and the rather thick septa required in the collimator degrade the resolution D J-
The rectilinear scanner may be used with ^^Inmbut it is relatively slow because it collects information from a small area of the subject at any one time. The hybrid scanner Q2, ma-J be a suitable compromise. It has already been used in a bedside scanner И . where it proved useful for liver scans.
76
CRAWLEY et al.
FIG.2. B lock diagram o f the electronic circuits. In pu t pulses are stretch ed to 2 ßs i f they are above the discrim inator threshold and to 8 ¡is if they are accepted by the single-channel analyser.
The shortage of spare parts and of skilled maintenance personnel in the developing countries makes it essential to keep the scanner as simple as possible whilst the shortage of skilled operators makes simplicity of operation an important factor. These requirements are interdependent because an instrument that is easy to use may require sophisticated cir cuits that replace the skill of the operator. The main frame of the instrument was a modified fork lift truck (Formula Model 0MS-20-69, Lift Stacker, Slough, Bucks., UK) (Fig. 1 ) with a yoke carrying the detector assembly mounted in place of the forks. A rotating system supplied with the truck enabled the scanner head to be operated from under the couch, above the couch or in a vertical position. Eaising and lowering the scanner head was by a manually operated hydraulic system. The electronic circuits were mounted in a small cabinet on the other side of the frame giving some counter balance (Fig. 1).
IAEA-SM -247/167
77
The entire system operated from a 2¿f volt DCsupply using two 'long life' lead acid batteries floating across a 10 amp charger (both batteries, rated at 5 years' useful life, and the charger, were supplied by Accumulatorenfabrik Bronneschein GmbH, Baedingen/Hessen, F. R. Germany). The motor driving the detector head operated directly from the 2 ¿fvolt supply, but the electronic circuits ran from an inverter giving a 2hß volt, 50 Hz sine wave at 200 VA (Valradio Ltd, Feltham, Middlesex, UK, type 02¿f/200S). Where possible, commercial nuclear instrumentation module (NIM) units were used. The circuits that were constructed at Harrow were built in the NIMunits and on two plug-in circuit boards mounted behind a control panel added to a commercial display oscil loscope (Electronic Visuals Ltd, Staines, Middlesex, UK, model No. EV8o6). The hybrid scanner has only two photomultipliers, so it is relatively simple to adjust. The motor moving the scanner head is operated via relays from the flash gun contacts of a Polaroid camera that records the scintigram. When the shutter is opened, the scanner head moves to and fro over the subject until a preset number of scintillations has been collected. A scaler then operates a relay which stops the motor at the end of the next sweep over the subject, so all the operator has to do is position the patient and press the camera cable release. The display oscilloscope has six low impedance inputs, X+, X-, Y+, Y-, Z1 and Z2, so it was possible to use the Y+ and Y- inputs to subtract the outputs of the logarithmic amplifiers (Fig. 2). A switch was provided so that the display oscilloscope could also be used to show the setting of the energy window and to check waveforms at a number of points in the circuit. Aposi tion on the CROswitch also allowed the display oscilloscope to be u s e d as a g e n eral purpose CR O for l o o k i n g at oth e r waveforms.
The circuit is similar to that described in an earlier paper DO, except that the delay lines are eliminated by using discri minators in the amplifiers to stretch the pulses and to continue stretching only if the single channel analyser accepts the out put pulses from the geometric mean circuit (Fig. 2). Ahigh speed (5 MHz bandwidth) analogue multiplier (Computing Techniques Ltd, M16) is used to multiply the two amplifier outputs in less than 2 us and pulses that fall within the window of the single channel analyser are stretched to 8 y.s in order to obtain the logarithms of the input pulses. A time base started by the first pulse to arrive is used for the circuit which displays the 'window' setting of the SCA. The same time base is used to display the most important pulses in the system when checking the performance of the various parts of the circuit and when adjusting the gains of the two amplifiers.
78
CRAWLEY et al. □ Ш
public supply standby working hours
■ standby out of working hours
FIG.3. A histogram o f mains voltage readings taken over a three-week period a t Fajara. The charger characteristic is superim posed to show that m axim um current was available fo r m ost o f the time. The mean voltage was 196.
I N S T A L L A T I O N A N D P R E L I M I N A R Y O P E RA TI ON
After the instrument had been constructed at the Clinical Research Centre it was used for a trial period of a few weeks at Harrow before transport by sea to Banjul. A ten-mile road journey to Fajara was followed by 11 months’ storage awaiting completion of a scanning room. The sodium iûdide crystal was retained at Harrow and taken by air just before the instrument was installed. Assembly and adjustment took less than two days. Aphantom was constructed with four straight line defects, 2 cmwide, to check the linearity and four circular defects, 2 , 3 , if and 5 cm diameter, to check the resolution. The phantomwas used for the initial adjustments and patient scanning started during the first week. There is a high incidence of hepatoma in The Gambia and about 5 patients per week present with symptoms that indicate a liver scan. Avoltmeter was used to check the mains supply during the 20 day installation period and a histogram of voltage readings is shown in Fig. 3» There were power failures almost every
IAEA-SM -247/167
FIG.5.
A scan o f the phantom .
80
CRAWLEY et al.
day and a histogram of the voltage from the standby generators is also shown in Fig. 3 . The battery voltage varied by about 1 volt from charge to discharge and a slight change in the speed of the motor could be heard, but any effects of changes in motor speed were avera ged out over several sweeps of the detector. Typically, it took about 10 sweeps to collect a scintigram with 200 000 counts. The batteries supplied enough power to continue scanning during the periods between failure of the public power supplies and manually starting the standby generators (5 to 20 minutes). A lateral scan of the liver of a patient is shown in Fig. if. PERFORMANCE THEFW HMalong the crystal was 2.2 cmand the collimator was designed to give a better resolution (1 cmFW HM) at right angles to the_crystal [j+J rather than equal resolution in both directions £3_| • The 2 cmdefect just resolved through 3 cmof scattering material (Fig. 5)* Uniformity was a major problem because the individual slices in the sodium iodide detector £3 , 5^] were not equally sensitive. Correction for uniformity was carried out using thin lead shims to increase the thickness of the septa of the collimator in regions of high sensitivity until measurements with a point source placed over each hole in turn gave a variation of less than 15 %DISCUSSION The design and construction of the instrument was adequate for the rather rough treatment that it received on the journey to Fajara, but it seems likely that it would have received more careful treatment had it not looked so robust. The charger, floating batteries and inverter system were almost immune to the wide fluctuations in mains voltage and even bridged the gaps between the failure of the public supply and the start of the standby generators. Scanning has been carried out for 18 months with only minor faults and one overhaul. Faults were dealt with by telephone from Harrow or by correspondence with the help of an engineer from the Cable and Wireless Ltd station in The Gambia. A compréhensive instruction book including circuits and component layouts as well as operating instructions was left with the instrument. Photographs of electrical waveforms, with the CR0 switched to a number of test points, were included in the instruction book. Electrical stability has been high and
IA EA-SM -247/167
81
only minor adjustments of the single channel analyser have been required to keep the electronic circuits at peak performance. Of the first 272 patient scans, 100 were diagnosed as hepa toma and 8 as hepatic abscess. One splenic abscess was first diagnosed on scanning. The scans provide useful information for biopsy or aspiration and, as most histological and biochemi cal investigations have been carried out in the United Kingdom, the scans have proved useful diagnostically. A comparison of scans with other clinical findings will be published elsewhere. The resolution would not be regarded as adequate for many purposes, but in view of the large hepatomas found in The Gambia j_6TTis sufficient for that purpose. An instrument based upon the sam'e principles but using a detector with 12 photomultipliers, forming a ’strip of gamma camera’ [j7_J , would have a reso lution approaching that of the gamma camera, but be capable of working with higher energy isotopes. It may even be possible to produce such an instrument at a price well below that of the mobile gamma camera. Nowthat microprocessors and memory de vices are becoming cheaper it should be possible to build in correction for uniformity and linearity distortions similar to that used in the double detector hybrid scanner at Harrow[jT3 • ACKNOWLEDGEMENTS The advice and encouragement of Dr R. A. Dudley is grate fully acknowledged, as well as helpful discussion with Dr R. Floyrac. The help and encouragement of several members of the staff at Fajara was greatly appreciated, as is the help of Mr D. Bay of Cable and Wireless Ltd. The work was supported in part by a grant from the Inter national Atomic Energy Agency, Vienna. REFERENCES [1] VAURAM O, E., V IR G O , A., The design o f the d etecto rs and collim ators for a hybrid scanner, Br. J. R adiol. SO (1 9 7 7 ) 808. [2] DAVIS, T.P., M A RTO NE, R .J., T he hybrid radioisotope scanner, J. N ucl. Med. 7 (1 9 6 6 ) 114. [3] CRAW LEY, J.C.W ., V EALL, N., “ T he design and som e clinical applications of a hybrid scanner” , M edical R adioisotope S cintigraphy 1972 (Proc. Sym p. M onte Carlo, 1972) 1, IA E A , V ienna (1 9 7 3 ) 105. [4] PLA N IO L, T h., FL O Y R A C , R „ IT TI, R., PO U RCELO T, L., «L a scintigraphie «au lit du m alade» — R ésultats o b ten u s a l’aide d ’u n scintigraphe hybride portatif», M edical R adioisotope S cintigraphy 1972 (Proc. Sym p. M onte Carlo, 1972)1, IA EA , V ienna (1 9 7 3 ) 113.
82
CRAWLEY et al.
[5] CRAW LEY, J.C.W ., GIBBS, G.P., “ Q uantitative w hole-body scanning using a hybrid scanner” , Medical R adionuclide Imaging (Proc. Sym p. Los Angeles, 1976) 1, IAEA, V ienna (1 9 7 7 ) 369. [6] JO HNSON, P .J., W ILLIAMS, R., Clinical and therapeutic aspects of hepatocellular carcinom a, Trans. R. Soc. T rop. Med. Hyg. 71 (1977) 461. [7] BOK, B„ M O RETTI, J.L ., FO N R O G ET, J., TH EBA U LT, B., A new u nidirectional w hole body scanner, Eur. J. Nucl. Med. 3 (1978) 55.
IA EA -SM -247/114
INVESTIGATION OF LONG-TERM IN VIVO TRACER DISTRIBUTION PATTERNS USING AN ULTRA-HIGH-SENSITIVITY SCANNING SYSTEM D. GVOZDANOVIC, K.V. ETTINGER, D.B. SMITH, C.G. TAYLOR, S. GVOZDANOVIC, J.R. MALLARD Department o f Bio-Medical Physics and Bio-Engineering, University o f Aberdeen, Aberdeen, United Kingdom
Abstract IN V ESTIG A TIO N OF LONG-TERM IN VIVO T R A C E R D ISTRIBU TIO N PA TTERN S USING AN ULTRA-HIG H-SENSITIVITY SCANNING SYSTEM. An ultra-high-sensitivity scanning system for the coarse imaging o f rad io tracer d istribution w ithin the hum an body is described. Slow dynam ic curves o f tracer d istrib u tio n in a n um ber of selected areas can be produced by sequential scanning and co m p u ter analysis of the images produced. The system was developed round a conventional shadow-shield w hole-body counter. The fo u r detecto rs are arranged in tw o vertical pairs, positioned laterally across the scanning direction. In the scanning m ode only the tw o b o tto m detecto rs are used; these are fitte d w ith very coarse focusing m ultih o le collim ators, giving a FWHM o f 100 m m fo r s lCr and 120 m m for 59Fe. A digitally co ntrolled bed moves longitudinally through the shadow -shield tunnel and steps laterally at the end o f each passage. Signals from each of the detecto rs, after pulseheight analysis, are recorded in the m em ory o f a m ultichannel analyser addressed by the digital co n tro l o f th e bed or directly o u tp u tte d on to a p unched tape. A pair o f scans is produced c oncurrently and th e lateral m ovem ent o f the bed is designed so th a t at th e end o f the scanning procedure th e tw o scans jo in in the m iddle. T he scan image m ay have any fo rm at, m ore or less, provided th a t this is w ithin 4 К pixels. The scan d a ta are transferred to a co m p u te r (PDP-8 or PD P-11) w here the images are assem bled and stored for fu rth e r processing. This includes m u lticolour images o f th e tracer d istrib u tio n and sequential dynam ic analysis o f selected areas o f the image. Effective images can be p roduced from total-body activity burdens o f the order o f 1—100 ßCi. T he system was used in a n u m b er o f ro u tin e clinical investigations (red cell sequestration sites, ferrokinetics) and research studies in trace elem ent d istrib u tio n (Cu, Zn) as well as in studies o f the d istrib u tio n of o th er labelled substances (vitam in B-12).
1.
INTRODUCTION
Radionuclide imaging has made steady progress in the few past years, a main tendency being the development o f imaging systems to suite " T c m and possibly a few other short-lived isotopes o f similarly low radiotoxicity. Ultra-high-sensitivity 83
84
GVOZDANOVIC et al.
imaging is separate from this main development, its aim being to enable use to be made o f small doses o f radionuclides o f longer half-life that have higher radio toxicity but facilitate tracer tests of much longer duration. Such systems have been described in the literature [ 1 - 4 ] and put to use in some conventional clinical tests that were originally developed with stationary counters over some areas o f the body that are o f interest. The number o f clinical tests that can make use of such ultra-high-sensitivity scanning systems is perhaps small at the moment; this is a field that was possibly neglected in past years, and there should be at least a number of trace element tests and research programmes to which such a system would ideally be suited. In this report such an ultra-high-sensitivity system is described. It has a sensitivity 1 0 0 -1 0 0 0 times higher than a conventional one. Benefits derived from recent developments in mainstream imaging and image processing were used in the system, particularly those relating to dynamic analysis o f areas o f interest in consecutive images. As well as conventional applications, the system was used in a study o f two trace elements: 64Cu and 65Zn.
2.
DESCRIPTION OF THE SYSTEM
The ultra-high-sensitivity imaging system was developed around a conventional shadow-shield ( 100-mm lead, 20-mm iron) scanning bed whole-body counter. The counter was built originally with four detectors arranged in two opposing pairs, the two detectors in each pair being positioned laterally across the scanning direction (F ig .l). The distance between the opposing pairs is 800 mm (face to face) and the distance between the detectors in lateral pairs is 250 mm (centre to centre). All detectors are 100 X 150 mm diameter Nal(Tl) crystals with single photo multipliers. The system uses for imaging at present the two bottom detectors only. These are fitted with coarse multihole focusing collimators which are 80-mm thick, each having 12 holes of 23 mm (max.) diameter and a focal point at 300 mm from the face (Fig.2). The focus of the collimators is chosen so as to fall in the middle o f the patient’s body when lying on the scanning bed. The scans produced in this arrangement are essentially posterior views but taken with a deep focus. The lateral separation o f the detectors o f 250 mm was decided during the con struction o f the whole-body counter as an optimum for best counting uniformity. The lateral movement of the bed is, however, 200 mm only, and collimators with asymmetrical focus had to be used to make the effective lateral separation o f the detectors equal to that o f the maximal bed movement. The bed is controlled digitally, moves longitudinally through the shadowtunnel and steps laterally at the end of each passage. With the two detectors a pair o f scans is produced concurrently; the two scans join in the middle at the end of the scanning procedure and cover the whole width o f the bed. The digital
IAEA-SM-247 /1 1 4
F IG .l. W hole-body counter. View through the shadow-shield tunnel. The bed is in the starting position, fu lly on one side o f the tunnel. During scanning it m oves in steps fu lly to the other side.
FIG.2.
Collim ators in position over the tw o b o tto m detectors.
86
GVOZDANOVIC et al.
FIG.3. E lectronic system used initially to input scanning data into a conventional m ultichannel pulse-height analyser with a m em ory capacity o f 1 К channels.
control o f the bed allows practically free choice o f the pixel size (from 1 X 1 to 100 X 100 mm). The pixel size determines the sensitivity in a similar way as resolution and a much narrower range o f these is practicable. A pixel size o f 25 X 25 mm was found suitable for most applications and was used as standard. The second factor limiting pixel size is the data storage capacity. A multichannel pulse-height analyser was used initially as the storage device. Signals from the detectors were amplified in separate charge-sensitive amplifiers and the energy selected in separate single-channel analysers which provided the two image content signals (one E.H.T. unit with dividers was, however, used). The position o f the bed provided address signals; by controlling externally the most significant digit o f the analyser’s memory two separate scans could be entered (Fig.3). Each longitudinal passage o f the bed was thus generating two strings o f information, the longitudinal passages being separated by a number o f channels with zero content. At the end o f a scan the data were outputted in an ASCII format paper tape, the tape was read into a computer (PDP-8 or PDP-11) and the image assembled. The total capacity o f the system was limited by the number o f channels available in the storage device (taking into account also some channels lost in the separation o f the individual profiles). A matrix o f 16 X 52 = 832 pixels could be stored comfortably in a 1 К memory, and the format was used in most examinations. The system described suffered from some non-linearity in positional addressing and required long time for data outputting. Furthermore, no data could be col lected in both channels simultaneously although this was not a serious shortcoming
IA EA-SM -247/114
87
FIG.4. Isom etric presentation o f the resolution o f the collim ators; p o in t sources at focal distance. Top: 51 Cr, b o tto m : 59Fe. The separation betw een poin ts and lines is 25 mm.
with the low count rates obtained from the low activities used. A system allowing direct output o f up to four channels o f data on to an ASCII paper tape was developed later, using a microprocessor-operated device. The matrix size is now limited only by the computer’s memory capacity, which in the practical con siderations o f low activity counting by far exceeds possible requirements. Once the image has been assembled in a matrix, standard computing facilities o f the BGAMMA program system can provide the following functions: (a) (b) (c) (d)
Image display: display range, background subtraction, colour range selection, grey scale. Image processing: interpolation, various smooth routines, selected transfer frequency filtration. Dynamic read o f selected area content: square/irregular area, area size, normalized. Graphical output o f dynamic curves.
Special computing programs were developed to cope with long-term studies: (a)
Time tag for each frame. There is not necessarily equal time lapse between sequential scans/frames.
88
GVOZDANOVIC et al.
TABLE I. DETECTION SENSITIVITY AND EFFICIENCY OF THE WHOLEBODY COUNTER SCANNING SYSTEM FOR POINT SOURCES IN THE COLLIMATOR’S FOCUS IN THE STATIONARY SITUATION
Sensitivity (Counts//iiCi-s)
Isotope
51 Cr
8.8
B ackground 51Cr region
4.0
59Fe
2.6 X 10~3
2.0 X 10"3
75
Background s9Fe region
(b) (c) (d)
3.
D etectio n efficiency (C ounts/disintegration)
2.8
Decay correction. Study duration may be only a fraction o f the halflife o f the tracer (e.g. 6SZn) or many half-lives (e.g. 64Cu or 47Ca). Physiological background subtraction, e.g. contribution o f blood back ground in red cell survival studies. Deconvolution o f overlapping images? The merits o f such a routine have to be assessed.
SYSTEM PERFORMANCE
For scanning purposes the single-channel analysers are tuned to include only the entire peak(s) o f the isotope used and no Compton region. With the relatively large Nal(Tl) crystals used in the system, the Compton region count is relatively small when sources in air are counted. In the real scan situation the Compton region count is therefore composed mainly o f scattered radiation, and inclusion o f these counts would lower the resolution still further. For photopeak energies only the spatial resolution o f the system is for sources in air at focal distance with the following FWHM (full width half maximum): 51Cr (0.32 MeV)
10 cm
59Fe (1.10 and 1.29 MeV)
12 cm
89
IA EA-SM -247/114
TABLE II. DETECTION SENSITIVITY AND EFFICIENCY OF THE WHOLEBODY COUNTER SCANNING SYSTEM FOR UNIFORM EXTENDED SOURCES
Sensitivity (C ounts- c m 2lp C i ■s)
D etectio n efficiency (C o unts/disintegration)
51Cr
1 519
0.446
59Fe
26 470
0.715
TABLE III. NUMBER OF COUNTS COLLECTED FROM A POINT SOURCE IN AN INTEGRATION AREA EXTENDING TO THE HALF-MAXIMUM LIMITS FOR A SCANNING SPEED OF 1 PIXEL/s AND A PIXEL SIZE OF 25 X 25 mm
Isotope
s l Cr
A pproxim ate size o f the integration area (pixels) Integrated background c ou nt Point spread fu n c tio n integral over the area (pixels (approxim ate)) Registered counts//uCi
s9Cr
24
36
100
101
15
24
132
1848
With the maximum diameter o f the area taken by the holes on the face of the collimator being only slightly larger than the FWHM, the latter does not change much when the source is closer to the face o f the collimator than the focal distance; the shape o f the transfer function changes, however. Some deterioration o f the resolution is observed when the source is farther than the focal distance or when it is submerged in a water phantom. A three-dimensional presentation o f the scan o f a point source is given in Fig.4. The sensitivity o f the system can be described in terms o f the number of counts collected in a unit time from a unit activity source positioned in the
90
GVOZDANOVIC et al.
З Ь л Л И
РГ . '1. ,y 'Ш .í
■ i . fl %
FIG.5. Ultra-high-sensitivity w hole-body scanner images. L eft: Original image o f 0.4 ¡ i d 58Co (vitam in В-12) in а 16 У. 52 m atrix. Centre: The same image after background subtraction and interpolation. R ight: A natom ical sites covered b y the images.
collimator’s focus in a stationary situation. If, instead o f activity, the number of emitted photons is used, the figure obtained represents the geometric efficiency o f the system with the collimator in place (Table I). Detection limits can be derived from these figures. Besides the factors given in Table I, the statistics o f the count number in each pixel depends on collection time. In practice this cannot be extended beyond a practical limit which is 1 or 2 s for a normal scan o f 832 pixels. With two standard deviations required to give an acceptable level o f significance, the detection limit was calculated to be 0.5 pCi for 51Cr and 0.05 /¿Ci for 59Fe when these are presented as point sources. For extended or area sources, the sensitivity can be expressed as a number o f counts collected in a unit time when presented with a source o f unit-specific activity (juCi/cm2); the collection then does not depend on whether the situation is stationary or the source or detector are moving (Table II).1
1 1 Ci = 3.70 X 1010 Bq.
IAEA-SM -247/114
91
FIG. 7. R e d cell survival stu dy: 70 ßCi 51 Cr were given as red cell label. Images taken on days 0, 2, 9 and 29, to ta l b o d y burden decreasing from the initial one dow n to 30 ¡ id . B lood p ool image seen a t the beginning disappears and splenic concentration becom es evident.
92
GVOZDANOVIC et al.
д а -7
C o rre la tio n
E s t i m . M e a n L ife D a y s
12 .7
C o r r e la t io n : - 0 . 6 9 0
H a lf- life
8-8
E s tim a te d
S u r v iv a l
D ays
D ays:
- O .S ^ S O
FIG .8. Disappearance o f 51 Cr-labelled red cells from circulation in the pa tien t whose scans were given in Fig. 7.
The detection efficiency figures for point sources are similar for 51Cr and
59Fe, a slight fall for the latter one being due to the lower photopeak efficiency o f the detector at this energy; the situation appears at first glance to be reversed for area sources. We must not forget, however, that the latter figures are an area integral o f the efficiencies under the point spread function o f the collimator’s performance for the particular energies. The apparent increase in the efficiency for detection o f s9Fe is caused by the lower resolution for this isotope; when the figures for extended sources are divided by the values of the integrals for spread functions (these were obtained by approximation), the results obtained are similar to those in Table I. Detection limits for extended sources o f humanoid shape in supine position (projection area o f 0.5 m2) can be obtained from the figures in Table II. By using the same statistical criteria as for the point source, we require a total body burden of 13 AiCiof 51Cr or
1.8 /¿Ci of 59Fe
IA EA-SM -247/114
93
FIG.9. D ynam ic curves o f integration over selected areas o f interest in red cell survival studies in the pa tien t whose scans were given in Fig. 7. N o te that the actual count rate is given for heart, and fo r other areas the difference betw een the actual count and the count e xpected from the blood pool.
to obtain a significantly distinguishable image at a scanning speed o f 1 pixel per second. In the practical situation o f following a tracer distribution in the human body, this means that satisfactory images can be obtained with, for instance, 50—70 juCi o f s lCr or 5 —7 цCi o f 59Fe in 15 minutes o f scanning in a standard format o f 16 X 52 pixels of 25 X 25 mm size. Useful distribution patterns of tracer concentration areas can be obtained from total-body burdens as small as 0.5 p d of, for ex a m p le,65Zn in scanning times o f 30 minutes. The performance described refers to individual images. A major interest of the investigations is, however, numerical analysis o f integrated counts in selected ‘areas o f interest’ in sequential scans. The analysis o f the system performance in such conditions is very complex. Relatively simple assessments can be made for the two extreme cases o f the integration over a part or the entire area o f the point-spread function for a point source and for the integration over an area for an extended source. The numerical values for the integration over a part o f the point-spread function with the integration limits selected at half o f
GVOZDANOVIC et al.
94
FIG. 10. Ferrokinetics stu dy: 10 p C i s9Fe were given IV (citrate). Images taken a t days О, 1, 7 and 12, total body burden remaining similar to th at injected. The tracer is transferred from the blood p o o l to the liver, no narrow uptake is seen. L ater only part o f the tracer is returned to the blood pool.
A .R .I. M e d ic a l IRON METABOLISM STUDIES TIPNAaTmIEeN: _______ 71• •
P h y s ic s is o t o p e c l in ic S u rfa ce Counting________________
___________ Unit N o : 4 9 1 3 1 3
W a r d : I A g T- D a t e : l S - ^ -43
FIG. 11. D ynam ic curves o f integration over selected areas o f interest in ferrokinetics studies in the pa tien t whose scans were presented in Fig.10.
95
IA EA -SM -247/114
A .R .I. M e d ic a l P h y s ic s RON METABOLISM STUDIES B lo o d I r o n PATIENT -, Name1.
PS
90 80 70
IS O T O P E
I М Я.
CLINIC
D a te : I g . ^ T Q
R ed C e lls Iro n
60 50
У / 7*—
40 £ 30
/
20
i—••}
)—
ÿ)/
/ и
10,
/ /
À /
A 2 HOU RS 3
5
DAYS
10
FIG .12. Disappearance o f Fe-tracer from plasma and appearance in red cells in the patient w hose scans were given in F ig.l 0. Thin lines show normal variation limits.
the maximum level are given in Table III. Extended source integration is a simple multiplication o f the sensitivity figures of Table II by the size o f the area of interest, provided the source extends beyond the integration limits. In the practical situation o f tracer distribution in the human body, the dimensions o f the accumulation site will rarely fulfil the requirements o f an extended source, at the same time these will be larger than a point source and a situation between the two described will apply. The analysis presented shows nevertheless that by integration over an area a larger number o f counts is collected, improving the statistics and allowing activities smaller than the detection limit to be followed. The integration over an average practical area o f interest yields up to a couple o f thousand counts for scans o f 51Cr and s9Fe at the activity levels already described. At the other end o f the scale, for 0.5 ^Ci 6sZn integration areas give more than 100 counts net with a similar background count. Evidently the significance o f changes in tracer count in the areas o f interest has to be assessed for each individual area separately. The actual images are presented by using a conventional colour range. A typical unprocessed image o f the distribution o f 0.4 ¡±Ci o f s 8Co-labelled vitamin B-12 is presented with its anatomical outlines in Fig.5. This image was collected in 1600 seconds, has a 16 X 52 matrix and is displayed in a standard 64 X 64 frame using the 12-colour range. The second image in the figure presents the same
96
GVOZDANOVIC et al.
FIG. 13. D istribution o f 47Ca tracer: 7 /i d were given orally, images were taken on days 2 and 7, total body burden was 3.5 and 1.7 pC i respectively. N orm al distribution in spine and cancellous bone was observed in both images.
FIG. 14. D istribution o f Ca tracer: 7 pC i were given orally, image taken on day 5, total b o d y burden 1.6 pCi. The accum ulation in both shoulders, particularly the left, is higher than that in the spine.
IA EA-SM -247/114
97
data after processing: background was subtracted and remaining data interpolated (display frame 128 X 128). The image is easier to follow although the amount of real information in the image has not changed. Smoothing routines can improve image perceptability still further; the choice of type and o f amount o f smoothing depends very much on image content. For dynamic analysis o f the area, integration data from original images were used only.
4.
CLINICAL APPLICATIONS
The imaging system has found routine use in a number o f clinical tests, both in diagnosis and in research. With vitamin B-12, the whole-body retention is used routinely as the measure of its uptake in patients with impaired kidney function but this figure may some times be misleading. Apart from variation in transit time, degradation o f the substance may cause non-specific intestinal retention. Concentration o f the label in the liver o f the patient demonstrated by imaging at the end o f the excretion period confirms the normal uptake pattern (Fig. 6). In haematological investigations the attainment o f numerical data is essential. The pictorial result of a red cell survival study is given in Fig. 7 in a patient with marked shortening o f red cell survival (Fig. 8). Although splenic involvement could be suggested from pictorial results, numerical analysis o f selected area integration shows clearly the magnitude o f splenic involvement (Fig.9). Counting over the heart is used for assessment o f blood background in conventional studies with fixed detectors. This appears to be to some extent misleading (proximity o f large areas o f bone marrow and the spleen) and large areas over head or thigh gave a better agreement with the blood counting curve. The other important tracer in haematology is iron; an example o f a ferrokinetics study is given in Figs 10 and 11. In this patient most o f the tracer was handled by the liver, and a large spleen with a large blood pool contained a fair amount o f tracer in various stages o f the study but did not release the tracer into the red cells (Fig. 12). Practically no bone marrow uptake was seen in this patient. Radioactive tracer studies for both major elements and trace elements in the body present a large field for the use o f the system. For example, distribution o f calcium was imaged. Very detailed images can nowadays be obtained on bone activity using " T c m-labelled polyphosphates and the 47Ca-images cannot compare with this. The 47Ca-tracer can be imaged with the system for much longer periods (up to 10 days using only the conventional dose o f 10 juCi used in calcium uptake studies) and thus give information on handling o f calcium in medium-term stores. An example o f normal calcium distribution throughout the spine and other cancellous bone is presented in Fig. 13. A case o f anomalous calcification o f the T ext cont. on p. 104.
VO
00
GVOZDANOVIC
et al.
FIG.15. Copper m etabolism stu dy: MCu was given. Top: 5 0 ßC i were given IV. Images taken on days 0, 1 and 2, to ta l b o d y burden 50, 10 and 2 p C i respectively. A lth ough som e blood p o o l is evident in all images, concentration is in the liver, w ith som e signs o f endogenous excretion. B o tto m : 100 p C i were given orally w ith 2 5 0 m l m ilk and 2 m g Cu carrier. Im ages taken on days 0, 1, 2 and 3, to ta l b o d y burden was 100, 1 0 ,2 and 0.5 p C i respectively. Passage o f the tracer through the gastro-intestinal tract is observed during the test, w ith only a small fraction o f the tracer being retained in the liver on the last day.
100
GVOZDANOVIC et al.
( 1)
(2)
> и ¡> M S
FIG .16. Zinc m etabolism stu d y : 0.5 pC i Zn were given IV. Average to ta l b o d y burden was 0.4 pCi. Top: Images o f single scans on days 2 and 31. B o tto m : C om bin ed images o f days 0, 2 and 3 (1), 7 and 9 (2) and 24, 31, 38 and 4 9 ¡3). Initial accum ulation in the liver w ith subsequent transfer into skeleton is eviden t from the single-scan images b u t m uch clearer in com bined images.
F IG .17. Z inc m etabolism stu dy: 1.5 ßC i 65Zn was given orally. Images taken on days 0, 3, 7 and 17. E x c ep t fo r day 0, to ta l b o d y burden was 0.4 f i d . R em aining background from previou s I V stu d y was substracted. A fter the initial passage o f the tracer through the gastro intestinal system , the accum ulation is fir s t in the liver, later in the skeleton and generally throu ghou t the body.
LONG-TERM ZINC METABOLIC STUDYr 0 . 5
yC i
6 5 Zn G IV EN J V
WITH 0 . 3
mg Zn CARRIER
AREA COUNTING: PEAK COUNTS I N BOTH AREAS NORMALIZED TO 1 0 0
L I V E R
S P I N E
100
CO
Ы о о oí Ui и
-
0 x
в—
-
«»*
.
's»
О Г
✓
60
" с
/
/ О )
40
IAEA-SM-247/114
CONCENTRATION
о Ы b tío < a
TRACER
< 2
RELATIVE
H ы > w H 20 J < Ы
ce
10 10 DAYS AFTER ADMINISTRATION PEAK A C T IV IT Y IN THE INTEGRATED AREA: 0 . 0 5U C Í
20 30 40 50 DAYS AFTER ADMINISTRATION
PEAK A C T IV IT Y IN THE INTEGRATED AREA: 0 . 0 3 u C i
FIG. 18. D ynam ic curves o f integration over selected areas o f interest in a zin c m etabolism study in the patien t whose scans were given in F ig .l 6. A tw o-expon en tial loss from liver is dem onstrated, the material being transferred to the skeleton. О
ui
GVOZDANOVIC et al.
104
connective tissue o f the left shoulder in a patient on dialysis who neglected to follow the prescribed regime is given in Fig. 14. Although the lesion was demonstrated first radiologically, calcium retention followed fast changes in the patient’s regime before any changes in radiological images could be demonstrated. In trace element investigations, an example o f the use of 64Cu is given for a patient suspected o f Wilson’s disease (Fig. 15). The sequence in the top row presents the accumulation o f the tracer after an IV injection, the main site o f concentration being liver. Later on there is some copper in the intestines; this appears to be due to endogenous excretion which is re-absorbed and the patient maintains almost all tracer throughout the test. In the bottom row the tracer was given orally to the same patient in milk, with some carrier copper added (2 mg). Note the passage through the intestines, some accumulation in the liver and some tracer found in the descending colon even on the third day after the dose was given. Another trace element study is that o f 65Zn tracer (Fig. 16). The top row presents a sequence after IV administration. Initial concentration o f the tracer is in the liver, from where it is later transferred to the cancellous bone. Combi nation images, representing a sum o f selected images from each phase, are given in the bottom row; these show the changes even better. A similar sequence after oral administration is presented in Fig. 17. The sequence starts with the passage o f the unabsorbed material through the gut. Useful numerical data can be obtained even from such low density images; integration results o f counts over liver and spine in the IV series o f scans are given in Fig. 18.
5.
DISCUSSION
Stationary detectors positioned over areas o f interest can generally gather more counts per area than a scanning system. With stationary detectors, even before a test has started there is a problem in deciding where to position such detectors, both in respect to organ position and to the prospective site o f accumula tion. The collection area may include sites that are not really wanted; the repositioning in successive counts is very critical. With the scanning system all positioning o f a patient can be related to a fixed anatomical site, e.g. vertex, and no marks on the body are required. Furthermore, all points o f the body are positionally interrelated in a scan and a digital shift o f the image, if the image is found to be out o f true position, moves all points with the image and their interrelationship does not change. At the end o f the investigation an unlimited number o f choices for selection of areas o f interest is possible, and m ost unforeseen interference can thus be eliminated.
IA EA -SM -247/114
105
ACKNOWLEDGEMENTS Thanks are due to W.I. Keyes, P.E. Undrill and R. Chesser for their help in writing the computer programs and for computer system maintenance. Electronic system maintenance was provided by K. Ross, and many patient scans were made in the initial stages o f the work by P. Stapleton.
REFERENCES [ 1 ] WATTS, J.R., et al., Design and evaluation of a scanning-bed whole b o d y c ounter for use in clinical studies, Health Phys. 23 (197 2) 63. [2] C OTTRALL, H.J., HODT, H.J., KENT, R.W., T R O T T , N.G., “ Investigations using a whole-body scanning system with digital co ntrol and read-out” , Radioaktive Isotope in Klinik und Forschung (Int. Symp. Gastein, 1972) 10 (FE L LIN G ER , K., H Ö F E R , R., Eds) Urban & Schwarzenberg, Munich-Berlin-Vienna (1 973) 45. [3] T R O T T , N.G., CO T T RA L L, M.F., WELLS, D.G., McCREADY, V.R., “ Investigations of sequential distributions using a low-background whole-body scanner with digital o u t p u t ” , Medical Radioisotope Scintigraphy 1972 (Proc. Symp. Monte Carlo, 1972) 1, IAEA, Vienna (1973) 61. [4] LILLICRAP, S.C., ST EERE, H., CLINK, H.M., “D istribution and dosim etry of 52F e and 59Fe in bone m arrow and o th e r organs” , Radioaktive Isotope in Klinik u n d Forschung (Int. Symp. Gastein, 1976) 12 (H Ö FE R , R., Ed.), H. Egermann ( 1976) 79.
DISCUSSION N.G. TROTT: Have you found any interesting clinical applications yet for long-term 59Fe measurements? D. GVOZDANOVIC: We measure the patients with 59Fe until a stable blood level o f tracer is reached in general. If I remember correctly, we have long-term iron distribution data for one patient only, and here the results were normal.
IA EA-SM -247/84
A G E R M A N I U M ROTATING L A M I N A R EMISSION C A M E R A (ROLEC) W. MAUDERLI, M.M. URIE, L.T. FITZGERALD Department o f Radiology, College o f Medicine, University o f Florida, Gainesville, Florida, United States o f America
Abstract A GERMANIUM R OTATING LAMINAR EMISSION CAMERA (ROLEC). Experimental and theore tical results of a p r o to ty p e rotating lam inar emission camera (ROLEC) for nuclear medicine imaging are re ported. A 1 1.5-mm-thick, 45 X 45-mm highpurity germ anium d etector is segm ented into thirty 1.47-mm-wide parallel channels and collimated with 39-mm-high paralleled tungsten plates. Projection data acquired at multiple angular orientations as the detector-collimator assembly is ro tated a b o u t its centre are mathematically re co nstructed to yield an image o f the activity distribution. For the above given geometry, the spatial re solution of th e ROLEC is at least twice as good, at all distances, as th at o f gamma cameras o f the Anger type with high resolution collimators. The better energy resolution of the germ anium enhances the detection and resolution of the ROLEC in com parison with gamma cameras with Nal(Tl) crystals, the relative superiority increasing with greater volumes and with greater depths. Adequate sensitivity is m aintained while achieving these i m provem ents in spatial resolution and, in practice, ROLEC images are acquired in less time th a n pinhole collimator images with gamma cameras.
PRINCIPLE
The principle of mathematically reconstructing a radioactive distribution from strip projections measured at multiple angular orientations has been employed in a germanium detector device designed for nuclear medicine imaging. This imaging principle1 obviates some of the inherent limitations of garnna cameras of the Anger type, allowing significantly better spatial resolution while maintaining adequate sensitivity. Garana cameras of the Anger type have an inherent limitation in their spatial resolving ability imposed by the imaging principle, whereas the spatial resolution of the ROLEC device depends mainly on the width of the detector strips, which theo retically can be made as small as desirable. 1
Which was independently conceived by Keyes and Tosswill [ 1 ], [2].
107
108
MAUDERLI et al.
F IG .l.
Cross-section o f the R O L E C imaging device.
Figure 1 is a cross-section of the head of ROLEC. A cold finger extends downwards from the liquid nitrogen dewar to the 45 x 45-irm germanium, below which are the collimator plates. These are enclosed in a standard cryostat and shielded with 2.5-nm lead, except for the 4.5-cm diameter circular viewing aperture at the bottom. The assembly is positioned on a circular platform which rotates about the detector center. A block diagram of the system is shown in Figure 2. The detector is biased at 1000 V and each individual strip is connected to a charge sensitive preamplifier, a pulse shaping amplifier, and a lower-level discriminator. Pulses of sufficient amplitude are stored at a unique address in a microcomputer (SOL 8080) memory as a count in a particular detector channel at that particular angular orientation. After a period of time, which is variable, the same microcomputer pulses a stepping motor, causing rotation of the detector-collimator. Acquisition at the new angular orientation begins. This process is repeated for a given number of equally spaced
IAEA-SM-2 4 7 /8 4
FIG.2.
109
Block diagram o f the R O LEC device.
projection angles through 180°. The data of each channel at each angular orientation are then transferred to an AMDAHL 470 computer where the activity distribution is mathematically reconstructed using a convolution back-projection technique with ramp filter. The resultant image is brought back to the micro computer and displayed on a CRT in a 30 x 30 matrix with 16 gray levels.
THEORETICAL CONSIDERATIONS Collimator
The collimator design is decisive in determining the spatial resolution of the ROLEC device. The main parameters are colli mator sheet thickness, spacing of collimator sheets and collimator height. The sheet thickness (0.18 nm W) has been optionalized to have a septal penetration effect of less than 0.5% and yet an open area collimator ratio of 88%, compared to about 70% for high resolution parallel hole collimators.
110
MAUDERLI et al.
£1=39 mm (Collimator Height) % H= 10 Heiaht above 40 Collimator 80 mm 7 o o --------- in— 1---------------- n --------------------- n -----;
FWHM-— :-
1.25*
i3.35
[гзб
Г
-
7 i Г2Т 0 77Г21
)Г,1 CHANNEL Ф ---- *-
-
i Г312 *0
1 г 3
100%: EVENTS REGISTERED IN CHANNEL 0
% H-
120
160
200mm /
гг
FWH
Mf
\
■h
,5.00 t
3 4.80 t
P
“
-i
1? 'rrr гг т Л, 1 i - i я 1 1?ч4 '5 CHANNEL*----- »* Inunit« of chonnel width FIG.3.
7М
' Л 7.2 \Ъ >
п
гг гЛ ’' 5я . тЛ ñ í Р ■> 1 ? з %5>6 « * 3 '» <и ? 3 4 5 в _
Calculated p o in t spread function.
The sheet separation (1.3 run) and collimator height (39.0 irni) have been selected to obtain a spatial resolution of about twice that of conmercially available cameras. Figure 3 shows the theoretical point spread functions for various heights above the collimator face, for the prototype RQLEC. The cross-hatched areas shows the effect of placing the detectors behind the colli mator assembly, rather than between the collimator sheets. DETECTOR
The laminar collimator design provides collimation in only one-dimension, exposing an entire detector strip to incoming photons and significantly increasing detector efficiency when conpared to a hole collimator.
IAEA-SM-247/84
1 11
о
L --------
FIG.4. Relative sensitivity fo r a p o in t source located above the m iddle o f the d e te cto r strip (referenced to a 50-m m -long detector). L = D etecto r length; D = H eight above d etecto r strip.
44
80 120 200 LENGTH OF DETECTOR
L in mm
300
FIG.5. R elative sen sitivity fo r a p o in t source com pared with hole collim ation as function o f d e te cto r strip length and source height above detector. Hole diam eter is assum ed equal to collim ator sh eet spacing.
112
MAUDERLI et al. в
FIG .6. E ffect o f prim ary absorption events in channel О on neighboring channels, as function o f several energy discrim inator levels.
Contrary to Anger-type cameras, the detector sensitivity increases with increase of detector length. Figure 4 shows the sensitivity increase relative to a 50-rmi long detector strip, as function of detector length L and point source distance D above detector. Figure 5 shows the relative strip detector response for a point source (as function o f detector length L and height D above detector) compared to a hole collimated detector. This shows that the overall gain in sensitivity including the effect of the increased open area ratio can be very appreciable. It has to be mentioned however that this gain in sensitivity is partly compensated for by the fact that the mathematical reconstruction procedure introduces additional noise, as has been shown for instance by Budinger et al. [3]. For identical signal-to noise ratios of the two types of imaging
113
IAE A-S M -2 47/84
3mm
6mm
ROLEC
HIGH RESOLUTION
4mm
PINHOLE
MINIMUM SEPARATION RESOLVED AT FACE OF COLLIMATORS
#
♦
6 mm ROLEC
13 mm HIGH RESOLUTION
6 mm PINHOLE
MINIMUM SEPARATION RESOLVED AT IOCM FROM COLLIMATORS FIG. 7. Images o f parallel line sources.
devices, the ROLEC device has therefore to collect more counts. It should however be pointed out that the ROLEC device was mainly designed to attain an improved spatial resolution rather than an increase in sensitivity. The energy resolution is 3.5 - 4 keV for most detectors except the two outside detectors with 6.5 keV. With an energy discrimi nation level of 135 keV (for the 140 keV line of Tc-99 m) the influence of channels neighboring the channel with the primary photon absorption is less than 1%. Figure 6 illustrates the calculated values of this cross-talk effect for different energy discriminator levels. EXPERIMENTAL RESULTS Imaging conditions
Activity distributions to be imaged are placed beneath the circular acceptance aperture. Since all activity must be within
114
MAUDERLI et al.
О HIGH
W
ROLEC
€ > С O L L I N 1A T 0 R
«
R E S C D L U T IO N
Ш
PINHOLE COLLIMATOR
2 .5 cm 0
2 .0 cm 0
I 5 cm 0
I 2cm 0
0 . 9 cm 0
FIG.8. Images o f spherical cold lesions at surface o f a 14-cm activity.
the field of view of all detectors at all angular orientations for the reconstruction process to be valid, objects are restricted to a maximum width of 4.5 cm. Figure 7 shows the images of parallel line sources as close as possible to the collimator face and 10 cm frcm collimator face. The ROLEC device shows an increase of resolution by about a factor of two, compared to a high resolution hole collimated Angertype camera. For clinical diagnosis, cold lesion detection is equal in importance to hot spot resolution. To investigate this capability, activity-free spheres (beeswax), ranging in diameter from 0.7 cm to 2.5 cm, were placed at different depths in 4.2-cm diameter cylinders of various heights. The results of one combination of depth of spheres and total depth of activity - at the surface of 14 cm activity - is shown in Figure 8. The ROLEC is able to detect smaller cold regions. The 1.2-cm diameter sphere is clearly evident and even the 0.9-cm diameter one is suggested. With the high resolution collimator of the gamma camera, there is a
IAE A-SM -247/84
115
FIG.9. Images o f the head o f a rat injected with 1.3 mCi Tc-99m-MDP. L e ft: Anger-type gamma camera. Center: R O L E C image. R ight: X-ray picture.
suggestion of a cool area with the 2.5 and 2.0-era diameter spheres, the 1.5-cm diameter sphere is very questionable, and anything smaller is not seen at all. The potential clinical significance of the improved hot spot resolution and cold lesion detection capability of the ROLEC is demonstrated in Figure 9. These images of the head of a rat injected with 1.3 mCi Tc-99m-MDP illustrate the superior detail obtained with the ROLEC.2 The gamma camera (left) shows no more than a general area of activity; the ROLEC (center) detects several distinct regions, which correspond ex ceedingly well with the bone densities in the radiograph (right) as would be expected for the bone-seeking radiopharmaceutical. Imaging times for the ROLEC are very acceptable. The ROLEC image of the rat head (Figure 8) was acquired in 390 seconds (185 000 counts). There was no attempt to minimize the time and, in fact, an image acquired in 190 seconds had the same detail (but it was slightly noisier). The ganma camera with the high resolution collimator took 196 seconds to acquire the pre-set 20,000 counts, which is the standard technique for small lesions. Other studies have shown that the ROLEC does not require unsatisfactorily long times, and, in fact, is usually faster than a ganma camera with a pinhole collimator. The fact that the entire object to be imaged must be at all times in the field of view renders the present device too small to be clinically useful. Computer simulations show however that it may be expected that a larger version would produce clinically superior images. 2
1 Ci = 3 . 7 0 X 1010Bq.
116
MAUDERLI et al.
REFERENCES [1] KEYES, W.I., The fan-beam gamma camera, Phys. Med. Biol. 20 3 (19 75) 489. [2] TOSSWILL, C.H., private com m unication to C.M. Williams, April 1975, Computerized rotating laminar collimation imaging system, U.S. Patent No. 4 ,0 90 ,0 80 (1978). [3] BUDINGER, T.F., DERENZO, S.E., GULLBERG, G.T., et al., Emission com puter assisted to m ogra p hy with single-photon and positron annihilation p h o to n emitters, J. Com put. Assist. Tomogr. (Com put. Tomogr.) 1 (19 7 7 ) 131.
DISCUSSION N.G. TROTT : Have you found that the performance o f your system is good enough to give the promised high resolution under the scattering conditions experienced in clinical work? W. MAUDERLI: Yes, we have produced images under scattering conditions as they exist in clinical work and have found the influence quite small.
IAEA-SM -247/201
I n v ite d R e v ie w P a p er
M E D IC A L IM A G IN G B Y N U C L E A R M A G N E T IC R E S O N A N C E A R e v ie w o f th e A b e r d e e n p h y s ic a l a n d b io lo g ic a l p r o g r a m m e J.R. MALLARD, J.M.S. HUTCHISON, M.A. FOSTER, W.A. EDELSTEIN*, C.R. LING Department o f Bio-Medical Physics and Bio-Engineering, University o f Aberdeen F.W. SMITH Aberdeen Royal Infirmary A. REID, R. SELBIE, G. JOHNSON, T.W. REDPATH Department o f Bio-Medical Physics and Bio-Engineering, University o f Aberdeen, Aberdeen, United Kingdom
Abstract MEDICAL IMAGING BY NU CLEAR MAGNETIC RESONANCE: A REVIEW O F THE ABERDEEN PHYSICAL AND BIOLOGICAL PROGRAMME. Nuclear m agnetic resonance imaging as used in the Aberdeen machine involves placing the subject in a static m agnetic field and applying a sequence o f radiofrequency pulses (1.7 MHz) and m agnetic field gradient pulses in three orthogonal directions. Images are formed o f any transverse section across the body, which display either the distribution of p roto ns in water (and fat) in tissues, or the spin-lattice relaxation time T i . Thus tw o new imaging para meters can be used to characterize norm al and diseased tissue. The m etho d has the advantage o f n o t using ionizing radiation, and it appears to be a safe, non-invasive procedure. A 128-s scan gives tw o arrays, one of p ro to n concentratio n, and one of T i. Each transverse section is 18.5-mm th ick with elements 7.5 m m by 7.5 m m (1 cm3). Simultaneously with the machine developm ent, a biological program me has been followed. In vitro measurements have been made of Ti for rab bit tissues at 24 MHz and 2.5 MHz: soft tissues range from 141 ms for liver to 463 ms for testis. Malignant tum ours generally have longer T t , and ra t thigh muscle
* Present address: General Electric Corporation Research and Development L aboratory, Schenectady, New York, United States of America.
117
MALLARD et al.
118
immediately surrounding implanted Yoshida sarcoma shows longer Tj due to reactions. Also, it is expected that conditions which affect the water conte nt of a tissue, e.g. oedem a, should increase its Tx. Tomographic sections from head, tho rax and abdom en (including the pelvis) have been examined on healthy volunteers. No adverse effect was experienced. The cerebral cortex, pineal gland, choroid plexes, sagittal sinus, cerebellum, fo u rth ventricle, brain stem, paranasal sinuses, orbits, eyes and ocular muscles have been seen. Major blood vessels in the tru n k , right and left ventricles o f the heart, lungs, breasts and chest wall, liver, spleen, kidneys, stomach, colon, vertebral canal, lum bar muscles, rectum and bladder have all been imaged. T he clinical role remains to be explored.
T h is p a p e r re v ie w s th e m ent s in c e 1973 w h ic h , in a E . G o r d o n u n t i l 1974 (now a S u t h e r l a n d u n t i l 1978 (now
1.
w o r k o f t h e NMR g r o u p o f t h i s d e p a r t d d i t i o n t o t h e a u t h o r s , i n c l u d e d R. t O x f o r d I n s t r u m e n t s L t d . ) a n d R. a t C u lh am L a b o r a t o r y , A b in g d o n ) .
INTRODUCTION
R a d i o n u c l i d e I m a g i n g o r i s o t o p e s c a n n i n g i s a way o f l e a r n in g so m e th in g a b o u t a p a t i e n t by d e t e r m i n i n g t h e d i s t r i b u t i o n o f a p a r t i c u l a r a to m ic n u c le u s i n t h e b o d y . I t i s do n e by d e t e c t in g t h e gam m a-ray a c tiv e decay.
p h o to n s
w h ich
th e
n u c le u s
em its
in
its
ra d io
A new a n d q u i t e d i f f e r e n t fo rm o f n u c l e a r im a g in g u s e s N u c l e a r M a g n e t i c R e s o n a n c e o r NMR. I t i s a n o t h e r w a y o f l o c a l i s in g a p a r t i c u l a r n u c l e u s i n t h e b o d y , a n d on e d o e s i t by d e t e c t i n g t h e r a d io f r e q u e n c y f i e l d s w h ich t h e n u c l e i r a d i a t e a f t e r t h e y h a v e b e e n e x c i t e d i n a p a r t i c u l a r w a y . No r a d i o a c t i v i t y i s in v o lv e d a t a l l - m e re ly t h e n a t u r a l p h y s i c a l p r o p e r t i e s o f p a r t i c u l a r a to m ic n u c l e i . NMR s p e c t r o s c o p y h a s e x i s t e d s i n c e 1 9 4 5 . I t h a s b e e n u s e d a lm o s t e n t i r e l y by p h y s i c i s t s , c h e m i s t s , b i o l o g i s t s a n d b i o c h e m i s t s t o s t u d y a to m ic a n d m o le c u la r s t r u c t u r e . The p o s s i b i l i t y o f u s i n g NMR f o r i m a g i n g w a s n o t p r o p o s e d u n t i l 1 9 7 3 a n d t h i s h as s tim u la te d c o n s id e ra b le a c t i v it y o v er th e p a s t s e v e ra l y e a rs [ 1 - 8 ] . Of p a r t i c u l a r i n t e r e s t a re th e p o t e n t i a l m ed ic al a p p li c a t i o n s , w h e r e o n e m i g h t m a k e i m a g e s u s i n g NMR m e a s u r a b l e p a ra m e te rs such as hy d ro g en p ro to n d e n s ity in th e w a te r o f body t i s s u e s o r NMR r e l a x a t i o n t i m e s o f t h e s e p r o t o n s . In
th is
p a p e r we show
such
im a g e s d e r iv e d
from
a NMR
to m o g r a p h i c s e c t i o n im a g e r fro m h e a d a n d t r u n k o f h e a l t h y hum an s u b j e c t s . The p r i n c i p l e o f t h e im a g in g m eth o d i s d e s c r i b e d . D e t a i l s o f th e i n s t r u m e n t a t i o n an d im a g in g t e c h n iq u e s c an be f o u n d e ls e w h e r e , [9 , 10 ], b u t a b r i e f d e s c r i p t i o n o f t h e A b e rd e e n m ac h in e i s g iv e n h e r e .
IAEA-SM -247/201
119
Applied Magnetic Field Direction of nuclear bar magnet
Vertical
I\
precession
precession
Nucleus
Spinning Top in Gravitational
Model of Spinning
Field
in Static Magnetic Field
F IG .l. L eft: A spinning top is shown pre cessing around the vertical fie ld o f gravity. R ight: A spinning nucleus is shown with the direction o f its m agnetization: its precession around the direction o f the applied m agnetic field is also shown.
2.
THE P R I N C I P L E
OF NMR IMAGING
I t h e l p s t o go b a c k t o c h i l d h o o d d a y s o f w h ip a n d t o p i ( F ig u re 1 ). The to p w as sp u n a b o u t i t s a x i s by t h e w h ip b u t i t a l s o w o b b le d a ro u n d t h e v e r t i c a l p a r t i c u l a r l y a s i t slo w e d down. T h is w o b b lin g i s c a l l e d a p r e c e s s i o n . The p r o to n s o f h y d ro g e n a l s o s p i n a n d h a v e an a s s o c i a t e d m a g n e tic f i e l d . They b e h a v e l i k e v e r y t i n y b a r m a g n e t s o f a d e f i n i t e s t r e n g t h o r m a g n e t i c m o m en t. I f
th e y
a re
p la c e d
in
a
m a g n e tic
f ie ld ,
th e y
w ill
lin e
up
m o re
o r le s s p a r a l l e l to th a t f i e l d and w ill p re c e s s aro u n d i t , as t h e to p d o e s a r o u n d t h e v e r t i c a l g r a v i t a t i o n a l f i e l d . The r a t e o r fre q u e n c y o f t h i s p r e c e s s io n i s p r o p o r tio n a l to th e m a g n e tic f i e l d s t r e n g t h i n w h ich th e y a r e p l a c e d . In n u c le a r m a g n e tic r e s o n a n c e , one m akes u se o f t h i s p r e c e s s io n to s tu d y th e a to m ic n u c le i and t h e i r s u rro u n d in g s by i r r a d i a t i n g th em w i t h e l e c t r o m a g n e t i c r a d i a t i o n o f e x a c t l y t h e sam e f r e q u e n c y a s t h e i r p r e c e s s i o n . At t h a t f r e q u e n c y th e y a b s o rb e n e rg y from th e r a d i a t i o n - a r e s o n a n c e a b s o r p t io n - and change t h e i r a lig n m e n t r e l a t i v e to th e a p p lie d I t i s e v e n p o s s i b l e , f o r e x a m p le , t o t u r n th em f l i p th em o v e r , i f t h e m a g n e tic f i e l d a n d t h e io n a r e a p p lie d s im u lta n e o u s ly a s a p u ls e - a i s a l s o p o s s i b l e t o t u r n th em t h r o u g h 90 - a f r e q u e n c y o f r a d i a t i o n w h ich h a s t o be u s e d i s f r e q u e n c y b a n d , n o r m a l l y i n t h e r e g i o n o f MHz
m a g n e tic f i e l d . th r o u g h 180 and a p p ro p ria te r a d ia t 180 p u lse . I t 90 p u lse .T h e in th e r a d io o r t e n s o f MHz.
120
MALLARD et al. UNIFORM FIELD
t o t a l s ig n a l from b o th tu b e s o f w a te r (a )
tu b e in w eaker field
tu b e in s tro n g e r field
( increasing frequency or x FIG.2. (a) Two tubes o f water are placed inside a uniform magnetic field. The NMR signal measured is the total signal from both tubes. (b) The same tw o tubes are n ow placed in a m agnetic field which has a gradient so that one tube is in a weaker field strength than the other. The tube, in the weaker field gives o u t a signal a t a low er frequency than the tube in the stronger field. The signals are therefore separated and the tw o tubes are iden tified in their appropriate position s related to the frequency o f their signal. The size o f the left-hand signal is larger because it is a bigger tube containing m ore water.
A f t e r a 9 0 ° p u l s e , t h e n u c le i h a v e s u r p l u s e n e rg y w h ich t h e y r a d i a t e t o t h e i r s u r r o u n d i n g s a t t h e sam e r e s o n a n t f r e q u e n c y . From a s a m p le c o n t a i n i n g a l a r g e n u m b e r o f s u c h n u c l e i , one by one th e y a r e th e n a b le to f a l l b a c k , o r r e l a x , t o t h e i r o r i g i n a l a lig n m e n t so t h a t th e sa m p le r e - e m i t s i t s r e s o n a n t f r e q u e n c y w h ich c a n be d e t e c t e d a n d m e a s u re d . A f t e r a 180 p u l s e , th e s u r p lu s n u c le a r e n e rg y i s a b s o rb e d by th e s u r r o u n d i n g m ed iu m . T he l e n g t h o f t i m e n e e d e d f o r t h i s i s a s s o c i a t e d w ith th e e n v iro n m en t o f th e p r o to n s b e c a u se th e e a s i e r i t i s f o r th em t o p a s s e n e r g y t o n e i g h b o u r i n g a to m s t h e q u i c k e r th ey can r e tu r n to t h e i r o r ig in a l s t a te ; th e s tr e n g th o f th e s ig n a l f a l l s e x p o n e n tia lly w ith tim e a t a r a t e c h a r a c t e r i s t i c o f t h e i r e n v iro n m e n t - a m e a su re m e n t known a s th e r e l a x a t i o n tim e . C o n s i d e r t h e s i m p l e e x a m p le o f tw o t u b e s o f w a t e r l y i n g s i d e b y s i d e i n a u n i f o r m m a g n e t i c f i e l d ( F i g u r e 2 ) . By i r r a d i a t in g th em w i t h e l e c t r o m a g n e t i c w a v e s o f t h e a p p r o p r i a t e p a r t o f t h e s p e c t r u m a n d v a r y i n g t h e f r e q u e n c y fro m lo w t o h i g h , o n e
121
IAEA-SM -247/201
c a n f i n d t h e r e s o n a n t f r e q u e n c y a t w h ich th e p r o t o n s o f th e h y d ro g e n a to m s a r e f l i p p e d o v e r . H a v in g fo u n d t h a t , o n e c an th e n m e a su re th e r a t e a t w h ich th e sy s te m r e l a x e s b a c k . H ow ever, i t i s n o t p o s s i b l e t o d i s t i n g u i s h b e t w e e n t h e tw o t u b e s o f w a t e r b e c a u s e b o t h h a v e b e e n s u b j e c t t o t h e sam e m a g n e tic f i e l d a n d h e n c e h a v e t h e sa m e r e s o n a n t f r e q u e n c y ; b u t i f o n e now a d d s a f i e l d g r a d ie n t to th e a p p lie d m a g n e tic f i e l d , so t h a t th e f i e l d i s s t r o n g e r o n o n e s i d e t h a n t h e o t h e r , o n e t u b e w i l l now b e in a s l i g h tl y w eaker f i e l d th an th e o th e r and w ill re s o n a te a t a s l i g h t l y lo w e r f r e q u e n c y . T h is tim e , w hen th e f r e q u e n c y o f r a d i a t i o n i s s w e p t fro m low t o h i g h , t h e t u b e p o s i t i o n e d i n t h e lo w e r f i e l d w i l l a b s o rb e n e rg y and th e n r e - e m i t i t a t a lo w e r f r e q u e n c y t h a n t h e o t h e r t u b e ( F i g u r e 2 ) . We h a v e t h u s o b t a i n e d som e i n f o r m a t i o n o n t h e p o s i t i o n o f t h e tw o t u b e s i n t h e f i e l d , m ea su re d from th e d i f f e r e n t f r e q u e n c i e s , and c a n d i s t i n g u i s h b e tw e e n th em s p a t i a l l y . I f , a s in F ig u re 2, one tu b e i s b ig g e r th a n th e o th e r , th e n i t w i l l c o n t a i n m ore w a t e r p r o t o n s a n d t h e s i z e o f i t s s i g n a l w i l l b e l a r g e r , s o t h a t from th e fr e q u e n c y o f th e l a r g e r s i g n a l i t i s p o s s i b l e to d e c id e w h ich o f t h e tu b e s i s th e b i g g e r o n e. To c a r r y t h i s e x a m p le f u r t h e r , i f o n e o f t h e t u b e s w e re t o be f i l l e d w ith w a te r c o n ta in in g a s m a ll am ount o f a s u b s ta n c e such a s c o p p e r s u l p h a t e , w h ich e n a b l e s th e p r o to n s t o r e l a x f a s t e r , t h e n by m e a s u r in g t h e tim e t a k e n f o r t h e s i g n a l t o d e c a y a t t h e tw o d i f f e r e n t f r e q u e n c i e s i t i s p o s s i b l e t o m e a s u re t h e r e l a x a t i o n tim e o f t h e p r o t o n s in e a c h tu b e s e p a r a t e l y a n d , from th e d i f f e r e n t decay tim e s , so i d e n t i f y th e tu b e c o n ta in in g copper su lp h a te . We h a v e , s p a t ia l
im a g e
th e re fo re , o f
e ith e r
th e th e
b a sic p ro to n
in fo rm a tio n
to
c o n c e n t r a t io n
produce
a
d is t r ib u t io n
o r t h e r e l a x a t i o n t im e d i s t r i b u t i o n i n a c o m p le x s a m p le . The sp e c tru m in (b) i s a o n e - d im e n s io n a l p r o f i l e o f p r o to n c o n c e n tr a t i o n i n f o r m a t i o n , a n d by s a m p lin g t h i s p r o f i l e w i t h t h e m a g n e tic f i e l d g r a d ie n t a p p lie d a t a s e r i e s o f a n g le s aro u n d th e tu b e s , a t w o - d i m e n s i o n a l m ap o f t h e p r o t o n c o n c e n t r a t i o n d i s t r i b u t i o n c a n b e r e c o n s t r u c t e d u s i n g t h e p r i n c i p l e now k n o w n g e n e r a l l y a s co m p u ted to m o g ra p h y [1 ]. T h e p r o g r a m m e d e s c r i b e d h e r e b e g a n by p r o d u c i n g i n 1974 a n im a g e o f a w h o le m ouse u s i n g t h i s m eth o d
[11 ].
3.
THE ABERDEEN NMR WHOLE-BODY IMAGING MACHINE
C o m p le te d e s c r i p t i o n s o f t h e p r e s e n t m a c h in e a n d i t s e l e c t r o n i c su b sy stem s a re g iv e n in H u tc h iso n e t a l . 1980. I t i s b a se d on a fo u r c o i l , a ir - c o r e d e le c tro m a g n e t m an u fa ctu re d by t h e O x f o r d I n s t r u m e n t C om pany ( F i g u r e 3 ) . I t p r o d u c e s a s t a t i c
122
MALLARD et al.
FIG.3.
The A berdeen NMR imaging machine (January 1979).
123
IAEA-SM-247 /201 m a g n e tic
fie ld
of
.0 4
T and
a
c o n s e q u e n t NMR f r e q u e n c y
of
1 .7
MHz f o r t h e h y d r o g e n p r o t o n s o f b o d y t i s s u e s . T h i s f i e l d i s a p p lie d in th e p o s te r io r - a n te r io r d ir e c tio n o f a su p in e ^ p a t i e n t . T h e m axim um f i e l d i n h o m o g e n e i t y i s a b o u t 6 x 1 0 at a r a d iu s o f 0 .2 3 m ., a p p ro x im a te ly tw ic e a tta in a b le w ith t h i s c o n f ig u r a tio n .
th e
am ount
th e o re tic a lly
T h e m a c h i n e u s e s t h e NMR s i g n a l s f r o m h y d r o g e n p r o t o n s i n w a t e r a n d i n f a t . I t p r o d u c e s tw o k i n d s o f im a g e s . The f i r s t u s e s p r o to n d e n s i ty a s th e im a g in g p a ra m e te r, and th e se c o n d u s e s p ro to n s p i n - l a t t i c e r e l a x a t i o n tim e (T ^ ). in w et tis s u e s d ep en d s e s s e n t i a l l y on th e r e l a t i v e p r o p o r tio n s o f f r e e and bound w a t e r w i t h i n t h e t i s s u e u n d e r e x a m i n a t i o n ; t i s s u e s w i t h m ore f r e e w a t e r h a v e l o n g e r r e l a x a t i o n t i m e s [ 1 2 ] . T^ v a l u e s o f n orm al and m a lig n an t a n im a l a n d hum an t i s s u e s h a v e b e e n p r e s e n t e d [1 3 , 14, 7 ]an d t h e i r i m p l i c a t i o n s f o r i n - v i v o c l i n i c a l NMR i m a g i n g h a v e b e e n d i s c u s s e d [ 7 ] . The im a g in g p r i n c i p l e em p lo y ed i s t h a t o f ' s e l e c t i v e e x c i t a t i o n 1 [ 1 5 , 1 6 ]. T h e p a r t i c u l a r f o r m u s e d [ 1 0 ] t o l e r a t e s a m u c h l a r g e r m a g n e tic f i e l d n o n -u n ifo rm ity th a n th e p r o j e c ti o n r e c o n s t r u c t i o n m e th o d [ 1 7 , 18 ] a n d a l s o t h e e f f e c t o f p a t i e n t m ovem ent i s l e s s d r a s t i c . I t u s e s a s p e c t r a l l y s h a p e d r a d i o f r e q u e n c y (R F) p u ls e in th e p re s e n c e o f a f i e l d g r a d ie n t to e x c ite s p in s in a t h i n s l i c e o f th e t h r e e - d im e n s i o n a l sa m p le . I n th e s i m p le s t form o f t h i s te c h n iq u e , th e f i e l d g ra d ie n t is th e n tu rn e d t h r o u g h 90 to a d i r e c t i o n p a r a l l e l to th e s e l e c te d s l i c e , and th e sp e c tru m , o r F o u r ie r tr a n s f o r m ,o f th e r e s u l t i n g f r e e in d u c tio n s ig n a l th e n r e p re s e n ts th e d i s t r ib u t i o n o f s p in s w ith in t h e s l i c e a lo n g th e se c o n d g r a d i e n t d i r e c t i o n . So f a r t h e r e i s no d i s c r im i n a t i o n in th e t h i r d o rth o g o n a l d i r e c t i o n , and any im a g e fo rm e d i s a p r o j e c t i o n o r sh a d o w g ra p h i n t h i s d i r e c t i o n . In o rd e r to d is c rim in a te in th e t h ir d d ir e c tio t r u e to m o g r a p h ic im a g e s , i t i s n e c e s s a r y t o em p lo y g r a d i e n t i n t h i s d i r e c t i o n a l o n g w i t h som e s e l e c t i v T he m o s t r e c e n t s e l e c t i v e p r o c e s s w h ic h we u s e i s w arp im a g in g w h ich i s d e s c r i b e d f u l l y in [ 1 0 ] . The p u l s e
sequence
used
fo r
th e
new
im a g in g
n a e c
and a ch iev e fie ld process. a lle d sp in -
m eth o d
is
show n
in f ig u r e 4 . T h is r e f e r s to a s e t o f c o o rd in a te a x e s w ith z v e r t i c a l a lo n g th e s t a t i c f i e l d from p o s t e r i o r t o a n t e r i o r o f a su p in e p a t i e n t : y h o r i z o n t a l a lo n g th e lo n g a x is o f th e p a t i e n t fro m h e a d t o f o o t ; a n d x h o r i z o n t a l l a t e r a l l y a c r o s s t h e p a t i e n t , fro m r i g h t t o l e f t . B e g in n in g a t i n t e r v a l 3 , a t h i n s la b o f s p in s t r a n s v e r s e l y across th e p a t i e n t a r e s e l e c t i v e l y e x c i t e d by t h e 90 G a u ssia n - s h a p e d , n a r r o w - b a n d RF p u l s e a p p l i e d i n t h e p r e s e n c e o f a
MALLARD et al.
124 ADIABATC FAST
PULSE SEQUENCE S P IN -WARP IMAGING FIG.4. The main features (diagrammatic) o f the sequences o f R F pulses and m agnetic field gradients, divided into six intervals.
y m a g n e t i c f i e l d g r a d i e n t (G + p e a k v a l u e t h i s e x c i t a t i o n ( a n d r e p h a s i X g G" d u r i n g
2 . 7 m T /m ). F o l l o w i n g in te r v a l 4 w ith
d e p h a s i n g G J ) a ' r e a d o u t ' f i e l d g r a d i e n t (G+ 0 . 5 m T / m ) in th e x d i r e c t i o n in i n t e r v a l 5 and th e sp e c tru m fm ore p r e c i s e l y , F o u r i e r tr a n s f o rm ) o f th e F re e I n d u c tio n s i g n a l (sa m p le d M tim e s ) i s , in e f f e c t, a p r o je c tio n o f th e p ro to n d e n s ity in th e s l i c e o n to th e x - a x i s ( a s i n F i g 2 ( b ) r i g h t ) . The f r e q u e n c y i s d i r e c t l y r e l a t e d to th e x - c o o rd in a te and a m p litu d e to th e t o t a l sp in s in one o f th e M c o lu m n s i n t h e z d i r e c t i o n (fro m b a c k t o f r o n t o f th e p a t i e n t ) : th e r e i s no d is c r im in a tio n in th e z - d i r e c t i o n as y e t. The a ll
o n to
z -d is c rim in a tio n
is
th e
T h is
sam e
a x is
x.
o b ta in e d w o u ld
by be
tak in g a silly
n p ro je c tio n s, th in g
to
do
in
is o to p e o r X -ray to m o g ra p h ic im a g in g b e c a u s e th e p r o j e c t i o n w o u ld b e t h e sam e e a c h tim e a n d t h u s p r o j e c t i o n s h a v e t o be t a k e n f r o m d i f f e r e n t a n g l e s . W h a t m a k e s NMR d i f f e r e n t i s t h e a b i l i t y to u se p h a se in f o r m a tio n in th e s i g n a l . Each freq u en cy com ponent o f th e sp e c tru m h as p h a se in f o r m a tio n a s w e ll a s a m p litu d e : a l t e r n a t i v e l y i t can be r e g a r d e d a s a co m p lex num ber w ith b o th a r e a l and an im a g in a ry c o m p o n e n t.
to
Suppose th e n , a p h a se t w is t, o r w arp, i s d e li b e r a t e l y added e a c h c o lu m n o f s p i n s , a s show n i n F i g . 5 . C o n s i d e r a lo n g
125
IAEA-SM -247/201
I \
I
I
\
\
— > z =o -1\ <—
*—
I
\
\
VECTOR SUM
I
1
\
4
COLUMN
1
COLUMN 2
COLUMN 3
FIG.S. The effect o f phase tw is t (or warp) upon colum ns o f spins (diagram matic) and their vectorial sums (see text).
u n i f o r m c o l u m n , a s i n c o l u m n 1 . W hen t h e c o n t r i b u t i o n f r o m e a c h s p i n i s a d d e d v e c t o r i a l l y , t h e v e c t o r sum i s q u i t e s m a l l ( t h e r e i s a b o u t 1% f u l l c y c l e s o f p h a s e t w i s t i n t h i s c o l u m n ) . C o l u m n 2 h a s f e w e r s p i n s i n i t b u t i s d i v i d e d i n t o tw o s e p a r a t e p o r t i o n s . T h e r e s u l t a n t v e c t o r sum i s q u i t e l a r g e . C o lu m n 3 h a s t h e sam e s t r u c t u r e a s i n c o lu m n 2 , b u t m oved d o w n w ard s a l i t t l e i n z . T h e v e c t o r sum h a s t h e sa m e a m p l i t u d e a s i n c o lu m n 2 b u t i s s h i f t e d i n p h a s e . W hat h a s m a x im iz e s th e r e s p o n s e t o th e o f th e phase tw is t.
happened is th a t th e ph ase s p a tia l frequency eq u al to
tw ist th a t
T h is p h a s e t w i s t i s i n t r o d u c e d by a d d in g a z f i e l d g r a d i e n t p u lse , G , a f t e r th e s e le c tiv e e x c ita tio n and b e fo re th e re a d o u t, i . e . d u rin g in te r v a l 4 (F ig . 4 . ) . In each su c e ssiv e p u lse sequence,
G^ h a s
th e
sam e
shape
but
its
a m p litu d e
changes
by
126
MALLARD et al.
TABLE I. NMR Ti RELAXATION VALUES FOR RABBIT TISSUES AT 24 MHz AND 2.5 MHz AND THE RATIO BETWEEN THESE VALUES (RESULTS EXPRESSED AS MEAN ± SD) Selected from Ref. [14] Tissue
T! at 24 MHz
Ti at 2.5 MHz
Ratio
Skeletal muscle (thigh)
554 + 32
182 ± 12
3.05 1 0.11
Liver
311 ± 15
141 ± 16
2.22 1 0.20
Spleen
5 0 9 + 11
258 ±
4
1.98 1 0.04
Grey cerebral tissue
644 ± 40
332 ± 22
2.00 + 0.01
White cerebral tissue
469 ± 13
264 ± 11
1.77 1 0.08
Kidney cortex
4 0 6 + 41
206±27
1.98 1 0.10
Kidney medulla Red fem ur marrow
801 1 35
426 ± 61
(202 ± 14)
(175 1 50)
1 .9 0 1 0 .2 4 (1.2
10.27)
Heart ventricle
637 ± 28
243 1
4
2.62 ± 0.07
Whole blood, heparinized
8 7 2 1 43
372 1 34
2 .1 3 1 0 .2 3
1078 ± 446
888 1 388
1 .2 2 1 0 .0 5
Bile
e q u a l s t e p s f r o m z e r o t o a m axim um v a l u e , e a c h o n e m a x i m i z i n g th e re sp o n se to a d if f e r e n t s p a t ia l freq u en cy in th e z - d ir e c tio n In p r a c t i c e , 64 o f t h e s e p u l s e s e q u e n c e s a r e u s e d fro m i n t e r v a l 3 onw ards and a tw o -d im e n s io n a l F o u r ie r tra n s fo rm o f th e m s i g n a l s p ro d u c e d a 64 x m im a g e . They a r e c a l l e d s ig n a ls and c o n ta in m o stly p ro to n d e n s ity in f o r m a tio n .
The a b o v e
sequence
is
sim ila r
in
p rin c ip le
to
one
proposed
by R ic h a r d E r n s t a n d c o l l e a g u e s [ 2 , 1 9 ] , in te r v a l o f G r a th e r th an a m p litu d e , as
who c h a n g e d t h e t i m e u se d h e r e . The p r e s e n t
m eth o d r e l a x e s f ie ld re q u ire d
of
th e s t r i n g e n t h o m o g e n e ity by E r n s t ' s m e th o d .
S p in -la ttic e
re la x a tio n
tim e
(T ^)
th e
sta tic
in fo rm a tio n
o b ta in e d by s ta r tin g th e sequence a t in te a re in v e rte d and one w a its d u rin g in te r v a l a t e l y e q u a l to th e a v e ra g e r e l a x a t i o n tim e m s) b e f o r e f o l l o w i n g w i t h t h e s e q u e n c e s t a r
can
m a g n e tic
be
r v a l 1. The s p i n s 2 f o r a tim e a p p ro x im i n t h e sa m p le (200 t i n g w ith in te r v a l
127
IAEA-SM -247/201 3. T h is i n i t i a l i n v e r s io n i s done h e r e by u s in g an a d f a s t p a s s a g e f o r w h i c h t h e RF m a g n e tic f i e l d in th e f r a m e i s a b o u t 2 0 yT a n d s w e e p s f r o m +8 kH z t o - 8 kH z t o t h e c e n t r a l L a r m o r f r e q u e n c y i n 1 0 ms g i v i n g m ore in v e rsio n . S ig n a ls d e r iv e d from th e p u ls e se q u e n c e i n t e r v a l 1 a r e c a l l e d S2 s i g n a l s a n d c o n t a i n and
T
re la x a tio n
tim e
in fo rm a tio n .
A 12 8 -second
scan
c o lle c ts
d a ta
for
b o th
ia b a tic ro ta tin g re la tiv e t h a n 90%
s t a r t i n g w ith b o th p ro to n d e n s ity
a p ro to n
d e n sity
and T im ag e. Such a s c a n c o n s i s t s o f 64 s i g n a l s o f ty p e SI a n d 64 s i g n a l s o f t y p e S 2 . One e c h o s i g n a l i s c o l l e c t e d e a c h se c o n d . I n i t i a l l y , th e SI d a ta and th e (S 1-S 2) d a ta a re F o u r ie r t r a n s f o r m e d i n t o tw o 6 4 x 6 4 e l e m e n t a r r a y s . T h e T^ v a l u e f o r e a c h i m a g i n g e l e m e n t i s c a l c u l a t e d f r o m t h e f o r m u l a T^ = 2 0 0 / l n ( 2 x S 1 / ( S 1 - S 2 ) ) . The 64 x 64 e le m e n t S I a n d T a r ra y s a re th en i n t e r p o l a t e d i n t o 2 5 6 x 2 5 6 a r r a y s a n d d i s p l a y e d o n a CONRAC C T - d i s p l a y u s i n g 16 c o l o u r - c o d e d l e v e l s , o r g r e y s c a l e . The to m o g ra p h ic s e c t i o n s h av e a G a u s s ia n p r o f i l e w ith t h i c k ness ( e q u i v a l e n t r e c t a n g u l a r w i d t h ) 1 8 . 5 mm. E a c h i m a g i n g e l e m e n t i s 7 . 5 mg w i d e x 7 . 5 mm. h i g h x 1 8 . 5 mm t h i c k , c o m p r i s i n g a v o l u m e o f 1 cm . I n e a c h im a g in g e le m e n t, th e u n c e r t a i n t y i n p r o t o n d e n s i t y f o r m u s c le t i s s u e i s a b o u t 2 .5 % a n d t h e u n c e r ta in ty in T i s a b o u t 3 .5 % f o r T^ a b o u t 2 0 0 m s , w i t h som e w o rsen in g as T d e p a r t s f ro m t h a t o p tim u m v a l u e . F o r t h e r a n g e o f T^ v a l u e s o f t i s s u e s r e p o r t e d i n t h e n e x t S e c t i o n (1 4 1 t o 4 6 3 ms a t 2 . 5 m H z) a b o u t 4 0 l e v e l s c o u l d b e d i s p l a y e d .
4.
.
SU PP O RT IN G PROGRAMME OF B IO L O G IC A L NMR MEASUREMENTS
IN V I T R O 4 .1 .
N orm al
tissu e s
T he hum an b o d y i s e x t r e m e l y c o m p le x . W hen, i n a d d i t i o n , we f i r s t a t t e m p t t o im a g e a new p r o p e r t y s u c h a s p r o t o n m a g n e t i c re s o n a n c e , th e r e i s bound to be d i f f i c u l t y in i n t e r p r e t i n g th e r e s u l t s . To t h e s e d i f f i c u l t i e s , f u r t h e r c o n f u s i o n may b e a d d e d i f e a r l y NMR i m a g e s a r e c o m p a r e d w i t h o t h e r i m a g e s d e r i v e d f r o m c o m p le te ly d i f f e r e n t t i s s u e p r o p e r t i e s , e .g . X -ray a b s o r p tio n , ra d io p h a rm a c e u tic a l c o n c e n tra tio n , o r u ltra so u n d c h a r a c te r is tic i m p e d a n c e . W ith t h i s i n m in d , we h a v e c a r r i e d o u t a b i o l o g i c a l b a c k - u p program m e on n o rm a l a n d p a t h o l o g i c a l t i s s u e s t o e a s e t h e p r o b le m o f im ag e i n t e r p r e t a t i o n a n d t o f i n d p o i n t e r s t o w a r d s t h e m o s t f r u i t f u l f i e l d s f o r t h e a p p l i c a t i o n o f NMR i m a g i n g . I t h a s a lw a y s b e en o u r i n t e n t i o n s e p a r a t e l y from p r o t o n c o n c e n t r a t i o n .
t o a t t e m p t t o im a g e T^ As a r e s u l t , i t i s v e r y
128
MALLARD et al.
r e le v a n t to m easure T v a lu e s o f t i s s u e sa m p le s in v i t r o , in comm on w i t h s e v e r a l o t h e r t e a m s , e . g . [ 1 3 ] . I t c a n b e n o t e d h e r e t h a t o t h e r im a g in g te a m s, so f a r , h av e n o t c o n c e n tr a t e d on t r y i n g t o im ag e s e p a r a t e l y , b u t d i s p l a y a co m p lex f u n c t i o n o f and p ro to n c o n c e n tra tio n . M any s t u d i e s
have
been
m ade
of
p ro to n
re la x a tio n
tim e
in
b i o l o g i c a l t i s s u e s [20] b u t m ost o f th e s e h av e b e e n a t f r e q u e n c i e s a b o v e 1 0 MHz. T h e s t u d i e s m a d e h e r e o f a w i d e r a n g e o f n o r m a l r a b b i t t i s s u e s a t 2 4 MHz a n d 2 . 5 MHz h a v e h e l p e d t o r e l a t e t h e T^ v a l u e s t o t h e l o w e r f r e q u e n c i e s o f t h e im a g in g sy s te m s [1 4 ]. 39 d i f f e r e n t t i s s u e s w ere e x am in ed a n d T a b le I show s a se le c tio n o f T v a l u e s fro m [ 1 4 ] , w h ich r a n g e fro m th e lo w e s t o f 141 m s e c f o r l i v e r to 463 m sec f o r t e s t i s , w ith th e body f l u i d s , b i l e a n d b l o o d s e r u m b e i n g m uch h i g h e r a t o v e r 800 m se c ( a t 2 . 5 M H z), t h e f r e q u e n c y c l o s e s t t o o u r i m a g i n g f r e q u e n c y . F o r f r e e w a t e r t h e r e l a x a t i o n tim e T^ i s a b o u t 3 s e c and i t i s b e l i e v e d t h a t t h e v e r y m uch r e d u c e d T v a lu e s fo r tis s u e s a r e a s s o c i a t e d w i t h t h e way i n w h ic h t h e w a t e r i s b o u n d t o p r o t e i n s i n t h e c e l l s a n d t h e t i g h t n e s s o f t h a t b i n d i n g . The c lo s e r th e b in d in g , th e s h o r t e r th e r e l a x a t i o n tim e . For th e s o f t tis s u e s , th e re is a th re e fo ld range of T v a lu e s w h ich t o g e t h e r w i t h t h e 3 .5% u n c e r t a i n t y o f f e r s good s o f t t i s s u e d i s c r i m i n a t i o n . The w a te r c o n t e n t o f s o f t t i s s u e s r a n g e s o n l y f r o m a m i n i m u m o f 69% f o r s k i n t h r o u g h 79% f o r s k e l e t a l m u sc le , s p le e n and b r a i n t i s s u e (72% f o r w h i t e
lu n g , b ra in
t o a m axim um o f tis su e ).
83% f o r
grey
T h e r a t i o b e t w e e n T ^ r e l a x a t i o n t i m e s a t 2 4 MHz a n d 2 . 5 MHz f o r m o s t o f t h e r a b b i t t i s s u e s l i e i n t h e r a n g e 1 . 9 t o 2 . 2 . The v a lu e s f o r s t r i a t e d and c a r d i a c m u sc le a r e , h o w ev er, h ig h e r , w h i l s t th e n e rv o u s t i s s u e s a r e , in g e n e r a l, lo w e r, d e c re a s e in t h e r a t i o f r o m t h e m o r e comm on v a l u e p e r h a p s r e f l e c t i n g t h e i n c r e a s e in t h e am ount o f m y e lin m em branes in th e t i s s u e sa m p le . W h i l s t t h e i n v i t r o m e a s u r e m e n t o f T ^ a t 2 . 5 MHz s h o u l d g i v e a g o o d g u i d e t o t h e T^ v a l u e s t o b e e x p e c t e d f o r t h e e q u i v a l e n t t i s s u e i n v i v o a t 1 . 7 MHz i n t h e i m a g i n g m a c h i n e ( s e e b e l o w ) , f u r t h e r t i s s u e m e a s u r e m e n t s a r e p l a n n e d a t 1 . 7 MHz i n a new i n v i t r o s p e c t r o m e t e r t o g i v e u s a b e t t e r e q u i v a l e n c e .
4 .2 .
P a th o lo g ic a l
tissu e s
M any g r o u p s h a v e m e a s u r e d t h e r e l a x a t i o n t i m e o f w a t e r p ro to n s in tu m o u rs. R e s u lts have in d ic a te d t h a t m a lig n a n t t is s u e s
129
IAEA-SM -247/201
g e n e r a lly have lo n g e r r e l a x a t i o n tim e s th a n th e e q u iv a le n t t i s s u e s [ 1 3 , 2 1 ] , a l t h o u g h i t i s now r e c o g n i z e d t h a t t h e r e a c o n s id e ra b le o v e rla p o f v a lu e s [2 2 , 2 3 ]. In v i t r o r e s u l t s o b ta in e d in t h i s l a b o r a t o r y [2 4 , 25 ]h a v e b e en .re v ie w e d in a n d w ere in a c c o r d w ith t h o s e o f o t h e r w o r k e r s . V ery s m a ll s t u d i e s s u g g e s t t h a t p e r h a p s hum an b r e a s t tu m o u rs a n d l i v e r m é t a s t a s é s m ig h t h av e an i n c r e a s e d s u f f ic ie n tly above t h su rro u n d in g s fo r d e te c tio n .
norm al is ' [7] ? sc a le e ir
The e f f e c t o n T^ c a u s e d by b i o l o g i c a l p r o c e s s e s s u c h a s i n f l a m m a t i o n , l o c a l im m unal r e a c t i o n s a n d i n v a s i o n ö f n e i g h b o u r in g t i s s u e by tu m o u r w e re e x a m in e d i n [25]-. I t w as f o u n d t h a t r a t m u sc le im m e d ia te ly a d j a c e n t t o im p la n te d Y o s h id a sa rc o m a s h o w e d i n c r e a s e d T^ r e l a x a t i o n t i m e , w h i c h g r a d u a l l y f e l l t o t h e n o rm a l m u s c le v a l u e a s o n e p r o g r e s s e d f u r t h e r aw ay fro m t h e tu m o u r. A lth o u g h t h i s s h o u ld h a v e th e e f f e c t o f b l u r r i n g th e . e d g e o f a tu m o u r a n d d e c r e a s i n g t h e r e s o l u t i o n o f t h e im a g e , t h e r e m a y , p e r h a p s , b e a g a i n i n t h a t s m a l l t u m o u r s may. a p p e a r o n T^ im ag es t o be l a r g e r th a n th e y á r e : t h i s m ig h t h e lp to d e t e c t them e a r l i e r . In v e s tig a tio n s a re in p ro g re ss to T^ a s s o c i a t e d w i t h p u r e l y i n f l a m m a t o r y in je c tin g tu rp e n tin e u n d er th e s k in o f L in g
and F o s te r
[26]
have
m easure th e ch an g es r e a c t i o n s in m u sc le r a t th ig h .
a tte m p te d
to
fin d
th e
sta g e
in by
of
c a rc in o g e n e s is a t.w h ic h th e T r e l a x a t i o n tim e in c r e a s e above n o rm a l o c c u r s , by e x a m in in g c h a n g e s i n T . o f l i v e r a n d s p l e e n t i s s u e o f r a t s b e i n g f e d o n h e p a t i c c a r c i n o g e n ; ( DA B) . A l t h o u g h t h e r e w as o n ly a s m a ll c h a n g e in l i v e r T , a d r a m a tic f a l l i n T^ r e l a x a t i o n t i m e o f t h e s p l e e n w a s f o u n d fro m 6 0 0 m s e c (n o rm al) to 350 m s e c . T h is a p p e a re d t o be a s s o c i a t e d w ith a tim e - r e la te d f a l l in w a te r c o n te n t o f th e s p le e n and an in c re a s e in p a r a m a g n e tic i r o n , p o s s i b l y from i n c r e a s e d b reak d o w n o f r e d b l o o d c e l l s . S i n c e ' t h i s f a l l i n s p l e e n T^ v a l u e s i s a s s o c i a t e d w i t h ä r a p i d l y p r o g r e s s i v e h a e m o l y t i c a n a e m i a , i t m ay b e t h a t . in c l i n i c a l c ir c u m s ta n c e s w h ich g e n e r a te a s i m i l a r . c o n d i t i o n i n t h e h u m a n , a p r o g r e s s i v e r e d u c t i o n i n h u m a n s p l e e n T ' '• m i g h t be fo u n d ., " • I t i s i m p o r t a n t t o r e a l i z e t h a t m any c l i n i c a l c o n d i t i o n s o t h e r t h a n m a l i g n a n c y m ay c h a n g e , t h e T ^ o f b o d y t i s s . u e s . A л . c o n d i t i o n w h i c h g i v e s r i s e t o a c h a n g e i n w a t e r c o n t e n t o f at i s s u e , e i t h e r i n t r a c e l l u l a r o r e x t r a c e l l u l a r , e . g . o ed em a,' s h o u l d i n c r e a s e T^ i n t h e a f f e c t e d r e g i o n . A c o n s i d e r a b l e n u m b e r o f s u g g e s t i o n s h a v e b e e n m ade [ 7 ] o f p o t e n t i a l , d i a g n o s t i c o r i n v e s t i g a t i v e c l i n i c a l < u s e s o f NMR i m a g i n g :' b u t t h e y h a v e n o t y e t b e e n e x a m i n e d , a n d m u s t a w a i t c l i n i c â l t r i a l s o f ' t h e NMR im a g in g
sy ste m s.
130
M A L L A R D et ai.
(a )
IAEA-SM -247/201
131
(с) Brain
FIG. 6. A x ia l section o f the head a t the level o f the pineal gland in 2 7-year-old fem ale volunteer, (a) proton concentration, (bj T \, (c) anatom ical guide.
5.
SA FETY A S PE C TS
T h e p o t e n t i a l h a z a r d s o f NMR i m a g i n g n e e d t o b e i n v e s t i g a t e d m o r e p r e c i s e l y , a l t h o u g h i n t h e ■p r e s e n t , ' g e n e r a t i o n o f m a c h i n e s t h e r e seem s to be l i t t l e c h an c e o f s e r i o u s e f f e c t U p o n -th e' . . p a t i e n t . A W o r k i n g P a r t y o n t h e s u b j e c t s è t u p b y t h e U -.K .,: V N a t i o n a l R a d i o l o g i c a l P r o t e c t i o n B o a r d h a s r e p o r t e d ■r e c e n t l y [ 2 7 ] . T h e p o s s i b l e h a z a r d s i n c l u d e t i s s u e h e a t i n g b y t h e r a d i o ’t freq u e n c y f i e l d s and th e e f f e c t o f th e in d u ced c u r re n ts cau sed by t h e p u l s e s o f m a g n e tic f i e l d g r a d i e n t . -T he e f f e c t o f . t h e s t a t i c m a g n e tic f i e l d i s b e lie v e d to be l e s s s i g n i f i c a n t .b u t' s h o u ld b e k e p t b e lo w 2 .5 T o r 25 000 g a u s s [ 2 7 ] . The A b e rd ee n m ac h in e o p e r a t e s a t 0 .0 4 T.
The l i m i t s e t upon th e a llo w e d r a d io f r e q u e n c y h e a tin g e f f è c t h a s b e e n b a s e d on n o t e x c e e d i n g t h e b a s a l , m e t a b o l i c r a t é , w h ich " i s a p p r o x i m a t e l y 1 W /k g o r 7 0 w a t t s t o t a l . T h e A b e r d e e n ¡ l e v e l i s o n ly up t o 0 .2 5 W t o th e w h o le b o d y .- A ls o , v a ry in g , f i e l d s ': , o f 20 Т/ s e c o u r m ac h in e near
th e
are th e
ends .o f
b e lie v e d n o t peak r a te o f th e
p a tie n t
t o c a u s e b i o l o g i c a l e f f e c t s .',’ I n ' ■' c h a n g e o f - f i e l d j u s t . ' ‘r e a c h e s . 3 T / s e c tu b e.
■ I t i s w o r t h s t r e s s i n g t h a t V o l u n t e e r s ' a r e . q u i t e 'ü r i à w a i 'e , o f th e p re s e n c e o f any o f th e f i e l d s , o r any s u b je c tiv e e f f e c t s due
to
th em .
132
MALLARD et al.
(b)
133
IAEA-SM -247/20I
„ Right, ventricle Inter ventricular septum Left ventricle Left pulmonary vessels Lung ^Descending aorta
RIGHT
LEFT
FIG. 7. A xial section through the thorax o f a 33-year-old fem ale volunteer a t the level o f the 6th dorsal vertebra, (a) proton concentration, (b) T \, (cj anatom ical guide.
C o n d i t i o n s w h i c h a r e a d v i s a b l e f o r h u m a n NMR i m a g i n g e x p erim e n ts d u rin g th e very e a rly s ta g e o f i t s u se have been su g g e ste d [2 7 ]. H e alth y v o lu n te e rs sh o u ld n o t be e p i l e p t i c s o r h a v e h e a r t d i s e a s e o r , a t th e p r e s e n t s ta g e o f k n o w led g e, be p re g n a n t. T here sh o u ld be a m ed ic al e x a m in a tio n b e fo re h a n d , a few d a y s a f t e r a n d 6 m o n th s l a t e r , a n d d u rin g , t h e e x p o s u r e a p e rs o n sh o u ld be p r e s e n t w ith r e s u s c i t a t i o n e q u ip m e n t. F or p a t i e n t s , t h e NMR i m a g e m u s t b e a d d i t i o n a l t o . a l l n o r m a l c l i n i c a l p ro c e d u re s and a d o c to r sh o u ld be p r e s e n t w ith r e s u s c i t a t i o n eq u ip m e n t. F o r b o th , r e c o r d s s h o u ld be k e p t an d p u ls e r a t e and r e a c t i o n tim e s o b s e r v e d d u r in g e x p o s u r e . T hese re c o m m e n d a tio n s w i l l b e r e v ie w e d a f t e r m ore e x p e r i e n c e h a s b e e n g a in e d . T h i s new i m a g i n g t e c h n i q u e h a s t h e a d v a n t a g e o f .n o t u s i n g io n iz in g r a d i a t i o n , and i t a p p e a rs to be a s a f e , n o ri-in v a s iv e pro ced u re.
6.
HUMAN I N VIV O NMR IMAGES
6 .1 .
V o lu n tee rs
To e v v o lu n te e rs th o ra x and No a d v e r s e
a l u a t e o u r s y s t e m we h a v e e x a m i n e d 11 h e a l t h y a g e d 20 t o 55 y e a r s . T o m o g ra p h ic s e c t i o n s fro m h e a d , abdom en (in c lu d in g , th e p e l v i s ) h av e b e e n ,o b ta i n e d . e f f e c t w as e x p e r i e n c e d by a n y o f t h e v o l u n t e e r s e i t h e r
d u rin g
a fte r
or
th e
e x a m in a tio n .
134
MALLARD et al.
(a )
(b)
135
IAEA-SM -247/201 Abdominal fat and muscles
(C )
Gas in stomach Stomach Aorta Upper pole left kidney Spleen Left forearm
Liver Splenic vein Inferior vena cava Upper pole right kidney
Spinal canal LEFT
RIGHT
FIG.8. A xial section through the upper abdom en o f a 33-year-old fem ale volunteer a t the level o f the 12th dorsal vertebra, (a) proton concentration, (b) T¡, fc) anatom ical guide.
TABLE II. NMR T, RELAXATION VALUES FOR TISSUES OF HUMAN VOLUNTEERS IMAGED BY THE MACHINE DESCRIBED. THEY ARE COMPARED WITH VALUES FROM RABBIT TISSUES.MEASURED IN VITRO AT 2.5 MHz [14] T l relax atio n tim e (m illiseconds) 1.7 MHz
2.5 MHz
Fem ale in vivo
Male in vivo
R abbit in vitro [ 14]
H eart ventricle
274
236
243
Liver
166
143
141 258
Spleen
269
264
Skeletal muscle
182
177
182
K idney cortex K idney m edulla
325
331
206 /4 2 6
Brain, white
245
N ot taken
264
Brain, grey
292
N o t tak en
332
MALLARD et al.
(a )
(b) FIG. 9.
(caption on p. 138).
IAEA-SM -247/201
(С)
(d)
137
138
MALLARD et al.
(f) FIG.9. P atient G.McK. Age 64. Inoperable oesophageal carcinoma and hepatic métastasés. (a) NMR Ту section at DIO showing carcinoma behind the heart. Elevated R diaphragm. N ote bilateral sm all pleural effusion, (b) and (c) NM R T¡ section through liver dem onstrating prim ary tum our and hepatic métastasés, (d) technetium sulphur colloid liver scan (AP) o f same patient, fej E -C A T scanner section o f same patient, (fj Bone scan o f same p atien t showing spine m étastasés n o t su spected until NMR Fig. (c) was taken and suggested their presence.
IAEA-SM-247/201
139
In the head, most of the important structures have been demonstrated. Figure 6 shows the cerebral cortex, pineal gland, the choroid plexes and the saggital sinus in a 27-year-old female volunteer. Lower sections have demonstrated the cerebellum, fourth ventricle and brain stem. The paranasal sinuses, orbits, eyes and ocular muscles have also been seen.
Sections through the thorax have demonstrated very clearly the major blood vessels, heart and lungs; these are best seen in the images. The chest wall and breast tissue are best demonstrated in the proton density images. Figure 7 is a section at the level of the sixth dorsal vertebra in a 33-year-old female and shows clearly the right and left ventricles of the heart, the left main pulmonary artery and the descending aorta. The lungs, breasts and arms are also clearly seen.
Abdominal sections have again demonstrated the major blood vessels and organs. The liver, spleen and kidneys are especially well seen in the T^ images. The stomach and colon are seen in both the proton density and T^ images, especially when a gas/fluid level is present. Figure 8, a section at the level of the twelfth dorsal vertebra/first lumbar vertebra, shows clearly the vertebral canal and lumbar muscles in the proton density image in the same female as in Fig. 7. The T^ image shows the liver with the splenic vein draining into the porta hepatis together with the spleen and the upper pole of the left kidney. A fluid level is seen in the stomach. We were not able to demon strate the pancreas in any of the volunteers. The small bowel is seen as a conglomerate of different densities which is more clearly demonstrated when it is surrounded by fat.
In the pelvis the descending colon, rectum and bladder are seen easily as are the testes in the male.
A comparison has been made in Table II between T^ relaxation times of some tissues when imaged in vivo (at 1.7 MHz) in these volunteers and those reported [14] for rabbit tissues in^ vitro at 2.5 MHz. The agreement is good. In the case of the kidney, in vivo it is entire and functioning and so it appears as a combination of the two in vitro values. Also, for the in vivo value measured for grey brain, it is likely that some white brain tissue will have been imaged in the same region of the display.
140
MALLARD et al.
6.2. Patients As a b e g i n n i n g of the e x p l o r a t i o n o f its p o s s i b l e c l i n i c a l roles, we intend to study a number of patients who have primary carcinoma, plus evidence to suggest hepatic or pulmonary m é t a s t a s é s .
the p r e s e n c e
of either
s e c t i o n s o f t h e f i r s t p a t i e n t to be e x a m i n e d in the m a c h i n e d e s c r i b e d h e r e a r e s h o w n i n F i g . 9. T h e i n o p e r a b l e oeso p h a g e a l c a r c i n o m a is s e e n b e h i n d t h e h e a r t in F ig. 9(a), together with elevated right diaphragm and small bilateral pleural effusion. Figs 9 ( b ) a n d (c) a r e sections through the liver showing multiple hepatic métastasés and primary 9 ( d ) (e) a n d (f) a r e s h o w n demonstrate confirm tumour. Figs a t i o n o f t h e s e f i n d i n g s , (d) b e i n g a n A P Тс- s u l p h u r c o l l o i d g a m m a c a m e r a i m a g e a n d (e) a s i n g l e p h o t o n e m i s s i o n t o m o g r a p h t h rough the liver p e r f o r m e d on the A b ^ g d e e n Section Scanner [ 2 8 J . F i g . 9 ( f ) i s a n A P b o n e i m a g e ( ШТс ( M D P ) ; g a m m a c a m e r a ) t a k e n a f t e r Fig. 9(c) h a d s u g g e s t e d the p r e s e n c e of s pinal méta s t a s é s : there h a d b e e n no e v i d e n c e of b one m é t a s t a s é s unt i l Fig. 9(c) was taken.
7.
CONCLUSION
We b e lieve our equi p m e n t p rovides a p o t e n tially useful d i a g n o s t i c i m a g i n g s y s t e m w h i c h i t is h o p e d w i l l p r o v i d e v a l u a b l e i n f o r m a t i o n in the i n v e s t i g a t i o n a n d d i a g n o s i s o f a large number of conditions. I mages are f o r m e d o f any t r a n s v e r s e s e c t i o n a c r o s s the b o d y , w h i c h d i s p l a y e i t h e r t h e d i s t r i b u t i o n o f p r o t o n s in w a t e r a n d fat, or the s p i n - l a t t i c e r e l a x a t i o n time, T , of t h o s e w a t e r p r o tons. Thus two c o m p l e t e l y n e w i m a g i n g p a r a m e t e r s can now be used to c h aracterize ation and diagnosis.
of
normal
and
diseased
tissue
for
investig
T h i s n e w t e c h n i q u e s h o u l d b e u s e f u l in t h e i n v e s t i g a t i o n t h e h e a d a n d n e c k , t h o r a x a n d a b d o m e n . I t s u s e in the
i n v e s t i g a t i o n o f p e l v i c d i s e a s e in b o t h the m a l e a n d f e m a l e should be pursued because of the absence of ionising radiation, a n d its p o t e n t i a l as a b r e a s t s c a n n i n g a g e n t s h o u l d be e x p lored. W h i l s t it w i l l be u s e d to p r o v i d e i n f o r m a t i o n o n t h o s e c o n d i t ions w h i c h can a lso be s t u d i e d by oth e r i m a g i n g techniques, this new procedure may provide further information not yet readily available.
IAEA-SM -247/201
141
It a l s o s e e m s v e r y l i k e l y t h a t it m a y p r o v i d e u s e f u l c l i n i c a l i n f o r m a t i o n o n t h o s e c o n d i t i o n s w h i c h h a v e not, so far, y i e l d e d to o t h e r i m a g i n g t e c hniques, for example, i n f l a m m a t o r y or oedematous states. The full clinical role of this new imaging t e c h n i q u e r e m a i n s to be explored.
ACKNOWLEDGEMENTS We are grateful to the Medical Research Council for grant support o f the project and o f W.A. Edelstein. G. Johnson and T.W. Redpath are in receipt of MRC Research Studentships. C. R. Ling is supported by the Research Fund of the Grampian Health Board. We are deeply indebted to Technicare Inc. for the gift o f a computer and display system in May 1980. We are deeply indebted to the normal and patient volunteers. Patents are held on the imaging machine for the University o f Aberdeen and the Medical Research Council by the National Research Development Corporation.
REFERENCES [1]
L AUTERBUR, P., Image form ation by induced local interactions, examples employing nuclear m agnetic resonance, N ature (L o n d o n ) 242 (1973) 190. [2] KUMAR, A., WELTI, D., ERNST, R.R., NM R Fourier zeugm atography, J. Magn. Resonance 18 (1 975 ) 68. [3] HINSHAW, W.S., Image form ation by nuclear magnetic resonance: the sensitive point [4] [5] [6] [7]
[8]
m ethod , J. Appl. Phys. 47 (1 976 ) 3709. DAMADIAN, R., GOLDSMITH, M., M IN KOFF, L., NMR in cancer: F o n a r image o f the live h u m an body, Physiol. Chem. Phys. 8 (19 7 7 ) 97. MANSFIELD, P., PYKETT, I.L., Biological and medical imaging by NMR, J. Magn. Resonance 29 (19 7 8 ) 69. HOULT, D.I., R otating frame zeugmatography, J. Magn. Resonance 33 (1 979 ) 183. MALLARD, J.R., HUTCHISON, J.M.S., EDELSTEIN, W.A., LING, C.R., FO ST ER, M.A., JO HNSON, G., In-vivo NM R imaging in medicine: the Aberdeen approach, b o th physical and biological, Philos. Trans. R. Soc. L ondon, Ser. В 289 (19 8 0 ) 519. HOLLAND, G.N., HAWKES, R.D., MOORE, W.S., NM R tom o gra ph y of the brain,
J. C om put. Assist. Tomography 4 (19 8 0 ) 429. [9] HUTCHINSO N, J.M.S., EDELSTEIN, W.A., JOHNSON, G., A whole-body NM R imaging machine, J. Phys. E (L o ndo n) Sei. Instrum. 13 (1980). [10] EDELSTEIN, W.A., HU TCHISON, J.M.S., JOHNSON, G., REDPATH, T.W., Spin warp NMR imaging and applications to h u m a n whole-body imaging, Phys. Med. Biol. 25 (1 980 ) 751. [11] HU TCHISON, J.M.S., MALLARD, J.R., GOLL, C.C., “ In-vivo imaging of b ody structures using p r o to n resonance” , Proc. 18th Ampere Congress, N ottin gham (ALLEN, P.S., ANDREW, E.R., BATES,C.A., Eds), University of N ottingham ( 197 4) 283.
142
MALLARD et al.
[12] PACKER, K.J., The dynamics of watei in heterogeneous systems, Philos. Trans. R. Soc. L o ndon, Ser. В. 278 (1 977 ) 59. [13] DAMAD1AN, R., ZANER, K., HÖ R, D., DIMAIO, T., MLNKOFF, L., GO LDSM ITH, M., Nuclear magnetic resonance as a new tool in cancer research: T um ours by NM R, Ann. N.Y. Acad. Sei. 222 (1 973 ) 1048. [14] LING, C.R., FO ST ER, M.A., HUTCHISON, J.M.S., Comparison of NMR water p ro to n T, relaxation times o f rabbit tissues at 24 MHz and 2.5 MHz, Phys. Med. Biol. 25 (19 8 0 ) 748. [15] MANSFIELD, P., MAUDSLEY, A.A., BAINES, T., Fast scan p ro to n density imaging by NMR, J. Phys. E ( L on don ) Sei. Instrum. 9 (19 7 6 ) 271. [16] HU TCHISON, J.M.S., “ Imaging by nuclear m agnetic resonance” , Proc. 7 th L.H. Gray Conf. on Medical Images, Leeds (HAY, G.A., Ed.), Bristol, Inst, o f Physics, Wiley (19 76) 135. [17] HU TCHISON, J.M.S., SU TH ERLAN D, R.J., MALLARD, J.R., NMR imaging: image recovery und er m agnetic fields with large non-uniformities, J. Phys. E (L o ndo n) Sei. Instrum . 11 (1 978 ) 217. [18] SU THERLAND, R.J., HUTCHISON, J.M.S., Three-dimensional NMR imaging using selective excitation, J. Phys. E (L on don ) Sei. Instrum . 11 (19 7 8 ) 79. [19] ERNST, R.R ., Gyromagnetic resonance Fourier transform Zeugm atography, U.S. Pa tent 4,0 70,6 11, 1978. [20] GADIAN, D.G., Nuclear m agnetic resonance in living tissue, Contem p. Phys. 13 (1 977 ) 351. [21] HOLLIS, D.P., ECONOMOU, J.S., PA RKS, L.C., EGGLESTON, J.C., SARYAN, L.A., CZEISLER, J.L., Nuclear magnetic resonance studies o f several experimental and h u m an malignant tumors, Cancer Res. 33 (1973) 2156. [22]
EGGLESTON, J.C., SA RYAN, L.A., HOLLIS, D.P., Nuclear magnetic resonance investigations of h u m an neoplastic and abnormal non-neoplastic tissues, Cancer Res. 35 (1 9 7 5 ) 1326.
[23] KOUTCHER, J.A ., GOLDSMITH, M., DAMADIAN, R., NMR in cancer: X. A malignancy index t o discriminate norm al and cancerous tissue, Cancer 41 (197 8) 174. [24] GORDON, R.E., MALLARD, J.R., PHILIP, J.F., Chapter 16 in Nuclear Magnetic Resonance Effect in Cancer (DAMADIAN, R., Ed.), Pacific Publishing Co., Portland, Oregon (1978). [25] LING, C.R., FO ST ER , M.A., MALLARD, J.R., Changes in NM R relaxation times o f adjacent muscle after im plan tation o f malignant and no rm al tissue, Br. J. Cancer 40 (1 9 7 9 ) 898. [26]
LING, C.R., FO ST ER , M.A., DAB-induced changes in NMR relaxation times, water and iron c o n te n t of rat tissue, Br. J. Cancer 42 (1980). [27] NATIO NAL RADIOLOGICAL PROTECTION BOARD, Outline of Advice o n Exposure to Nuclear Magnetic R esonance Clinical Imaging, NMR 2 /8 0 /3 (1980). [28] BOWLEY, A.R., TAYLOR, G.G., CAUSER, D.A., BARBER, D.C., KEYES, W.I., UN DRIL L, P.E., CO RFIEL D , J.R ., MALLÁRD, J.R., A radioisotope scanner for rectilinear, arc, transverse section and longitudinal section scanning: (ASS — the Aberdeen Section Scanner), Br. J. Radiol. 4 6 (1973) 262.
DISCUSSION J.S. LAUGHLIN: I should like to congratulate you on this demonstration o f total body scanning with T¡ spin relaxation and proton location in NMR. What is the length o f time required for a cross-section scan? And would even better resolution require longer scan times?
IAEA-SM -247/201
143
J.R. MALLARD: With the Aberdeen system, we need 128 pulse sequences to provide the data from which two transverse-section images can be constructed. One is a 64 X 64 array o f proton concentration information: the other is a 64 X 64 array o f Tj relaxation time information. Each pulse sequence takes 1 second and so the two images need 128 seconds. Regarding your second question, I take it that by better resolution you mean better spatial resolution. For this we would have to use a smaller resolution volume. Our present resolution volume is 1 cm 3 , with a pixel o f 7.5 mm X 7.5 mm, which is 18.5-mm thick. For this volume, proton concentration can be determined with an uncertainty o f 2.5%, and Tj can be determined with 3.5% uncertainty. It is a signal-to-noise ratio problem. Reducing the resolution volume by half, and thereby imaging with better spatial resolution, would double these uncertainties. To improve spatial resolution without losing accuracy o f the imaged parameter would require the re-design o f the machine with a better excitation and signal acquisition system, which would improve the signal-to-noise ratio. This would require either a different pulse sequence, or different magnetic field gradients and, perhaps, static field and RF, or both, depending upon the improvement to be achieved. In general, the signal-to-noise ratio is related to (frequency ) 372 and hence to (magnetic field)372. To halve the linear resolution (i.e. volume element reduced to 1/ 8) leads, therefore, to the need for a four-fold rise in magnetic field strength, which, in turn, requires a new magnet design. A.E. TODD-POKROPEK: Mr. Mallard, o f the speakers this morning, three have been seeking to improve resolution at the expense, presumably, o f sensitivity, and the others, to some extent, exactly the opposite. Could you: (a) Comment on the resolution/sensitivity compromises that exist in NMR? (b) Comment on the imaging o f ‘heavier nuclei’, e.g. 31P, where, I believe, much less resolution can be obtained, and the systems proposed have been called ‘systems for regional “in vivo” biochemistry’? J.R. MALLARD: The spatial resolution o f a given system is dependent, among other things, upon the resolution volume for which the system is designed. Decreasing the resolution volume increases the uncertainty o f the measurement o f either the proton concentration or the proton T t relaxation time, and hence reduces the chance o f detecting a small change in them from volume to volume. You mention heavier nuclei. I should repeat that the machine which has been described in my paper images the protons o f water (and fat). It is important to think o f its potential uses in terms o f water, and not in terms o f elements or compounds injected into the patient. Since approximately 75% o f the human body is water, this new method o f imaging seems likely to have an impact in medicine, in some way yet to be explored, at the resolution and accuracy achieved here. I am amazed by the general lack o f knowledge about the water inside us,
144
MALLARD et al.
even though one quickly thinks o f common (and less common) ailments that obviously influence water balance in a region. The inherent sensitivity o f the method depends upon the spin (and magnetic) properties o f the nucleus concerned, and its occurrence. Phosphorus-31 is a potential candidate. The natural phosphorus takes part in the chain concerned with the flow o f energy through a tissue. But to localize the 31P requires much higher static magnetic fields than are needed for protons. To provide this field, a super conducting magnet is necessary, which will be expensive to build if it is to be large enough and uniform enough for a human trunk. Based upon current experience with smaller animal systems, the team at Oxford Instruments Ltd. in the United Kingdom hopes to achieve a meaningful measurement o f 31P in a volume o f about 10-cm diameter. This application might be useful for the heart, liver, kidneys and spleen.
P o s te r P r e s e n ta tio n s X-RAY FLUORESCENCE TOMOGRAPHY* J.A. PATTON, R.R. PRICE, D.R. PRICKENS, F.D. ROLLO, C.L. PARTAIN Vanderbilt University Medical Center, Nashville, Tennessee A.B. BRILL Brookhaven National Laboratory, Upton, New York, United States o f America
X-ray fluorescence scanning has been used in the laboratory o f the Vanderbilt University Medical Center for more than eight years as a routine clinical procedure for imaging the stable iodine distribution within the thyroid. Of particular interest has been the applicability o f the technique in differentiating between malignant and benign nodules that are ‘cold’ in radioisotope scans on the basis o f their iodine content. The current system is calibrated to provide regional as well as total gland iodine content measurements using computer techniques. Methods have been developed to compare the iodine content o f a nodule with that o f a corresponding normal area by calculating a ratio o f nodule to normal iodine content. In a study o f 107 patients who subsequently went to surgery after • fluorescent scans, 98% o f those patients with an iodine content ratio greater than 0.6 had benign nodules whereas 54% o f those with ratios less than 0.6 had malignant nodules. To improve nodule delineation and permit the simultaneous mapping of " T c m-pertechnetate and stable iodine distributions, a new X-ray fluorescence system is being developed around an array o f high purity germanium detectors which should provide significant improvements over currently available techniques. A source and detector collimator system has been fabricated consisting o f eight detector collimators focused to a common point. A parallel hole source collimator has been placed in the centre o f the apparatus which currently allows a single 1 Ci source o f 241Am to irradiate uniformly a 1.9-cm diameter field .1 The collimation system provides a FWHM resolution o f 0.56 cm in the focal plane at 6.4 cm and a separation between planes o f 1 cm. Longitudinal section images are reconstructed
* Work supported in part by NIH G ran t No. 5 ROI GM 23621-3.
1 1 Ci = 3.70 X 1010 Bq. 145
146
POSTER PRESENTATIONS
by backprojection. Phantom studies have demonstrated that high quality tomographic images o f both radioactive ( 99Tcm) and stable ( 127I) isotope distribu tions can be obtained with this system. A multi-source system is currently being constructed to provide the sensitivity necessary for patient studies. This new system should improve the ability to discriminate nodules with low iodine content from benign tissues, and to aid the surgeon in deciding which nodules need to be removed surgically.
LA TOMOGRAPHIE PULMONAIRE PAR DIFFUSION COMPTON J.-L. MORETTI, E. MATHIEU, J.-F. CAVELLIER, G. ROUX Service de médecine nucléaire, Hôpital Henri-Mondor, Créteil, France L’intensité d’un rayonnement diffusé par effet Compton dans la matière est fonction du nombre d’électrons par unité de volume du matériau diffusant. Pour les tissus présentant un rapport Z/A constant, l’intensité du faisceau diffusé est proportionnelle à la densité. L’application de ce principe permet de mesurer la densité tissulaire à l’aide d’un rayonnement monochromatique collimaté et d’un détecteur focalisé associé à un système d’analyse. Cette technique a permis de mesurer une densité tissulaire ponctuelle (Kaufman, Garnett, Reiss) ou d’obtenir la distribution des densités d’une coupe anatomique sous la forme d’une image analogique (Okuyama, Guzzardi). Ces systèmes tomographiques d’imagerie comportaient généralement une irradiation en nappe et une collimation droite pour laquelle la contribution du rayonnement polydiffusé ne permettait pas une mesure précise de la densité tissulaire. Le dispositif d’imagerie que nous avons employé utilise une irradiation en pinceau fin et une détection focalisée limitant la contribution de la polydiffusion.
METHODE ET PATIENTS Le système de détection informatisé est un scintigraphe à barreau corps entier, équipé d’irradiateurs solidaires de la tête détectrice, placés face à face, contenant chacun une source de 55 GBq (1,5 Ci) d’iridium 192 (Tphys = 74 jours).
POSTER PRESENTATIONS
147
Le faisceau est finement collimaté; la détection du rayonnement diffusé est effectuée à 90° grâce à un collimateur focalisé sur la ligne d’irradiation aux énergies de 195 à 244 keV. Le plan tomographique est construit par le déplace ment synchrone du détecteur et du pinceau d’irradiation au niveau d’un plan déterminé chez le patient par le positionnement de l’ensemble. Après des essais sur fantômes, l’application clinique fut réalisée chez 40 patients présentant des affections pulmonaires diverses.
RESULTATS La résolution spatiale est de 6 mm (largeur à mi-hauteur) pour des coupes de 10 mm d’épaisseur moyenne et la résolution en densité est de 2%. Les images tomographiques obtenues permettent une bonne analyse de la distribution des densités pulmonaires normale et pathologique, même en l’absence de compensation des phénomènes d’atténuation.
GAMMA-GAMM A-COIN CIDEN CE SCINTIGRAPHY : TOMOGRAPHY WITHOUT COMPUTERIZED IMAGE RECONSTRUCTION H. von BOETTICHER, H. HELMERS, E.-M. MUSCHOL, P. SCHREIBER, I. SCHMITZ-FEUERHAKE Department o f Physics, University o f Bremen, Bremen, Federal Republic o f Germany
Radionuclides emitting two or more gamma rays per decay such as 43K, 48Cr, 73Se, 75Se, 178Ta, 192Ir, 196Au and 202T1 allow direct three-dimensional scintigraphy by gamma-gamma-coincidence measurements without computerized image reconstruction. Two systems based on this method were designed. A scintillation camera was coupled to two horizontal detectors in opposite arrangements. The latter consist o f two 15X15X5 cm 3 Nal(Tl) crystals, each with a focusing tungsten slit collimator at a distance o f 24 cm. The collimator parameters were optimized by computer programs. At the horizontal centre of the system, the FWHM o f the point spread function in vertical Z-direction amounts to 2.5 and 3.2 cm in water, with and without energy discrimination
148
POSTER PRESENTATIONS
respectively. A maximum coincidence count rate o f 4.4% was obtained, compared with the normal camera mode, for a 7SSe point source in the centre of a 30X17X15 cm 3 water phantom. In order to compensate for the reduction o f efficiency in this system, the use o f pinhole collimators for the camera will be indicated. The applicability o f the camera system may be limited for imaging large organs because the distance between the two horizontal detectors is too large. Therefore a 3-D scanner was developed to overcome these limitations. Seven 12.7 cm ф X 5 cm Nal(Tl) detectors are coupled to long focusing hole collimators and inclined towards each other so that they focus on a common point. Using every combination o f two o f these seven detectors in coincidence (( 2) pairs), a coincidence count rate of 4.8% was reached, compared with the normal scanner mode, for a 75Se point source in 6.5 cm water. In addition to the vertical depth discrimination (FWHM o f the point spread function 2.0 cm in water) there is a resolution improvement in horizontal direction (FWHM o f the point spread function in water 1.2 cm instead o f 1.7 cm). Several phantom studies were undertaken with both systems.
THEORETICAL AND EXPERIMENTAL INVESTIGATIONS OF 3-D IMAGING WITH COMPLEX CODED APERTURES* E.R. REINHARDT, G. LAUB Institute o f Physical Electronics, University o f Stuttgart, Stuttgart W. MÜLLER-SCHAUENBURG Department o f Nuclear Medicine, Institute o f Radiological Medicine, University o f Tübingen, Tübingen, Federal Republic o f Germany
The feasibility and performance o f imaging gamma-emitting extended objects by means o f coded apertures have been investigated both theoretically and experimentally [1—3]. This imaging technique offers the following advantages: 1. Formation o f 3-D images. 2. No correlation between geometric efficiency and spatial resolution. * Work supported by the Bundesministerium für Forschung und Technologie, Federal Republic of Germany.
POSTER PRESENTATIONS
149
In many ways the coded image is similar to a hologram and has to be decoded or reconstructed in a second step, using autocorrelation or deconvolution procedures. For autocorrelation the imaging process is said to be optimal when the autocorrelation o f the aperture с (x, y) can be described by a delta function:
c ( x ,y ) © c ( x , y ) = 5 ( x ,y )
(1)
Transforming Eq. (1) into the Fourier domain yields Eq. (2): С (u, v) • C* (u, v) = 1
(2)
С (u, v) denotes the Fourier transform o f с (x, y) and C* (u, v) denotes the conjugate complex function o f С (u, v). It is evident that all functions given by Eq. (3) satisfy Eq. (2): С (u, v) = exp [j ф (u, v)]
( 3)
Using Eq. (3) a large number o f coded apertures can be determined. In this way it is possible to compute optimized apertures for special applications. Some potential aperture functions will be described. The reconstructed image can be focused sequentially on different planes o f the 3-D object. But the evaluation o f the 3-D object information is limited by the superposition o f defocused images corresponding to other object planes. By implementing a priori information on the object, this restrictive influence can be reduced. In general it is possible to overcome these difficulties with a special imaging procedure. Using N linearly independent coded apertures, the intensity distribu tions o f N different object planes can be computed, whereas the influence o f defocused images is eliminated (emission tomography with coded apertures). The linearly independent apertures can be determined systematically [4]. For imaging gamma-emitting objects in nuclear medicine a complex Fresnel zone plate collimator has been adapted to a conventional Anger camera. The efficiency o f imaging with coded apertures will be demonstrated by experimental results obtained by in vivo imaging various human organs as well as by imaging 3-D phantoms. Discussion o f the results concentrates on the following aspects: dose reduction, spatial and depth resolution, and signal-to-noise ratio. In particular the properties o f conventional coded apertures and emission tomography with coded apertures are discussed.
150
POSTER PRESENTATIONS
REFERENCES [1] [2] [3] [4]
BARRETT, H.H., HORRIGAN, F.A., Appl. Opt. 12 (1973) 2686. BRUNOL, J., FONROGET, J., Opt. Commun. 25 (1978) 35. REINHARDT, E.R., LAUB, G., BLOSS, W.H., Biomed. Tech. 24 (1979) 225. REINHARDT, E.R., Abbildungssysteme m it codierenden Aperturen und ihre Anwendung in der Radiologie, Dissertation, University of Stuttgart, 1979.
DISCUSSION J.-C. ROUCAYROL: Have you used other combinations o f coded apertures apart from the one shown on your interesting poster? E.R. REINHARDT: We have performed experimental investigations o f complex Fresnel zone plates and annular apertures. Other apertures have also been used in numerical simulation experiments. A large number o f suitable apertures can be determined from a formula presented in the paper. For example, it is possible to use to advantage complex annular apertures or complex pinhole distributions.
EVALUATION OF IMAGING PROPERTIES OF PLANAR STOCHASTIC CODED APERTURES K. KOURIS, N.M. SPYROU, D.F. JACKSON Medical and Environmental Physics Group, Department o f Physics, University o f Surrey, Guildford, Surrey, United Kingdom
A major drawback o f gamma-ray imaging systems that use collimators is that only a small fraction of the emitted photons are detected. Systems that use coded apertures increase the detection efficiency and offer tomography but do not always improve image quality. An evaluation is presented, by computer simulations, o f the imaging properties o f planar stochastic coded apertures. At the source depth, the reconstruction yields a spread function that exhibits a strong and sharp peak and no sidelobe structure. Reconstructions at underlying and overlying tomographic planes yield spread functions which are smaller in magnitude and more broad in shape without any systematic sidelobes; their
POSTER PRESENTATIONS
151
presence, however, would be expected to cause artifacts. The magnitude o f the in-focus peak response varies with source location on the tomographic plane; a post reconstruction solid angle factor improves the uniformity. Out-of-focus reconstructions are associated with greater fluctuations than in-focus ones. For equally out-of-focus planes, the peak o f the overlying reconstruction is greater than that o f the underlying one. Symmetry can be imposed by using the factor: (depth o f focus)-2 . Both the in-plane and depth resolutions improve with decreasing source depth and with decreasing aperture size. The accuracy associated with a recon struction decreases as the strength o f the background increases. Stochastic coded apertures are expected to be most useful in applications involving small anatomical regions, near the surface, that exhibit a weak background.
IAEA-SM-247/13
HEADTOM E:
A H Y B R ID E M IS S IO N
T O M O G R A P H F O R B R A IN D e s ig n c o n c e p ts a n d p r e lim in a r y r e s u lts I. KANNO, K. UEMURA, Y. MIURA, S. MI UR A Division o f Radiology and Nuclear Medicine, Research Institute o f Brain and Blood Vessels, Akita, Akita City Y. HIROSE, K. KOGA, H. HATTORI Shimadzu Corporation, Nakagyo-ku, K yoto, Japan
Abstract HEADTOME: A HYBRID EMISSION TOMOGRAPH FOR BRAIN: DESIGN CONCEPTS AND PRELIMINARY RESULTS. A new emission tomograph, HEADTOME (hybrid emission advanced dynamic tomograph), combining a single-photon emission tomograph (SET) and a positron emission tomograph (PET) in one system, has been developed. The Headtome has 64 Nal (TI) detectors arrayed on a 42 cm ф circular ring. Two collimator systems slide axially on the inner face of the ring according to SET and PET imaging. A collimator of SET consists of 64 sets of 2-mm-thick 30 X 70 mm tungsten fins with five 0.5-mm-thick tungsten septa inside. All the fins are driven to swing back and forth synchronously with a span of 60°. Subsequently, the foci of the fins sweep across the whole field of view (21 cm ф). Linear sampling is selectable from 2.5 mm to 20 mm by combining ring rotations. The slice thickness is varied for 20, 25 and 30 mm FWHM by selecting a slice collimator with a slit of 16, 20 and 24 mm width respectively. SET study with 9 9 Tcm showed resolutions of 7.5 mm and 10 mm FWHM at the 10-cm radius and at the centre respectively, and sensitivities of 1 0 and 2 1 kcounts-s -1 juCi’ 1 -ml with a 2 0 cm ф uniform activity of 133Xe and 99 Tcm respectively. The fastest scan enables 2-s unit sampling with a resolution of 20 mm FWHM, which enables dynamic study with the 133Xe clearance method. A collimator of PET consists of two doughnut-shaped lead plates to collimate the slice plane. A gap between the doughnuts is selectable at 12, 20 and 28 mm. Linear sampling in PET is 20 mm at the steady ring, 10 mm employing the ring rotation, and 3 mm with the ring rotation and wobbling, the ring of which was designed with a 7.3-mm ф. PET study by 68Ga with 9 points wobbling showed a resolution of 12 and 10 mm FWHM at the 10 cm radius and at the centre respectively. The system sensitivity was 7, 20 and 30 kilo true events per jUCi/ml for the 12, 20 and 28 mm slice gaps respectively. At present the system is being used clinically for cerebral blood flow studies with 133Xe inhalation and 81Krm constant infusion, and for blood brain barrier studies with 99 Tcm 0 4 or 6 SGa-EDTA, and has given many positive results.
153
154
1.
KANNO et al.
INTRODUCTION
The most serious hindrance to the widespread distribution o f a positron emission tomograph (PET), in spite o f its excellent physical and physiological advantages over a single-photon emission tomograph (SET) [ 1], is that PET requires too many operators and is too expensive for an in-house cyclotron in a normal hospital. One solution is for an emission tomograph to combine PET and SET in one system [2].
2.
SYSTEM DESIGN
2.1. Nal ring detector The Headtome uses a circular ring array o f 64 Nal (Tl) detectors. The diameter o f the ring is 42 cm. The Nal crystal is 16 mm wide, 28 mm high and 70 mm long. A reflector and an aluminium can add an extra 1 mm on both sides. A photomultiplier tube (PMT) is photo-coupled directly to the Nal crystal. Between adjacent crystals is placed a lead wedge o f 2 mm at the innermost width. The detector ring is installed in the gantry with .the capability o f rotating and wobbling in a slice plane (Fig. 1). The ring is rotated in SET by a multiple of one-eighth of a crystal-crystal angle (Д 7 ) according to a sampling resolution, and in PET by one half o f Д 7 . The wobbling with a diameter of 7.3 mm is employed in PET to improve linear sampling [3]. On the inner face o f the crystal ring two collimators can slide axially for each PET and SET function (Fig. 2). 2.2. Collimator for SET A collimator for SET is one o f the most successful configurations in the present system. The collimator basically consists o f 64 sets o f tungsten fins located on each crystal. One set o f the fin collimator is a main fin consisting o f a 2-mm-thick 30 X 70-mm tungsten plate and a sub-fin o f a composite o f five 0.5-mm-thick tungstens, three parallel to the main fin and two perpendicular (Fig. 3). The main fin and sub-fin are synchronously driven by a big wheel gear. At any inclining angle o f the fin, according to the nature o f this structure, the adjacent main fins always maintain an angle o f 5.6°, one-64th o f 360°, which makes an intrinsic focus point on the circle, whereas the sub-fin was designed to have a focus with fixed length o f 6 cm from the collimator face. The swing angle is fully computer controlled. The collimator is inclinable up to 30° in each direction to see the whole field o f view (FO V) o f 21 cm. The axial thickness o f the slice is selectable with exchangeable three-slice collimators, which are made
155
IAEA-SM-247/13
detectors and a photomultiplier tube is located on a slice plane. Inside the ring face two collimator units slide rearwards (upper) for positron detection and frontwards (lower) for single-photon detection.
S in g le
P h o to n
D etectio n
P o sitro n
A n n ih ila tio n
D etectio n
FIG.2. T w o configurations of the hybrid detector of the Headtome. In single-photon detection (left) 64 moving fin collimators swing back and forth up to 30° and each collimated ray sweeps the full field of view. In positron annihilation detection (right) conventional coincidence is used. The view angle (ф) of the coincidence line is 30°.
156
KANNO et al.
FIG.3. Moving fin collimator. O n e collimator unit consists of a fin with a 2-mm-thick, 70-mm-long and 30-mm-high tungsten plate and a sub-fin with a composite of five 0.5-mm-thick tungsten plates, three parallel to and two perpendicular to the main fin (bottom right). The 64 units of these collimators are driven to swing back and forth synchronously by wheel gear.
o f 5-mm-thick lead cylinders with a slit o f 16, 20 and 24 mm width across the slice plane.
2.3. Collimator for PET A collimator for PET consists o f a slice mask and a scatter mask. The slice mask is made o f two doughnut-shaped lead plates o f 1-cm thickness and 8-cm radial length. The space between tw o doughnuts is selectable for 12, 20 and 28 mm. The scatter mask is made o f a 0.5-mm-thick lead belt just covering the inner face o f the crystal ring [4].
2.4. Electronics The outputs o f the detectors are fed, through pre-amplifiers, to timing single channel analysers (TSCA), in which the pulse heights are discriminated by upper and lower levels set equally in 64 channels, and timings are selected by a constant fraction technique. The time resolution is 5.3 ns FWHM. In the coincidence logic, 64 detector lines are organized into eight group pulses and checked for even
IAEA-SM-247/13
157
parity [5]. The coincidence time window is 20 ns. At every coincidence event two detector addresses are encoded to two 6 bits and fed to the first-in-first-out (FIFO) memory. The single photon logic is simply encoding every event to the 6 bits detector address and the 6 bits collimator angle and stored in FIFO. The data in FIFOs are subsequently transferred to two 4096 words buffer memories, then into a PDP-11/34 having 64 К words. 2.5. Data sampling The sampling format can be selectable from four scan modes in SET and three scan modes in PET according to the linear sampling and feasible scan time desired (Table I). 2.6. Data correction and reconstruction In SET ray sums are initially corrected for collimator efficiencies for each swing angle and are sorted into the parallel fashion projection. The algorithm mainly used is a filtered back projection technique and partly an interactive technique. Absorption correction in SET is carried out so as to normalize the measured ray sums o f the object by the ray sums o f the uniform activity pool and to correct the object geometry. In PET accidental events are initially removed by subtracting values estimated by single count rates and a time window, then the efficiencies o f coincidence pairs are corrected with ray sums assessed from the uniform activity. Assuming that the scattered events are distributed uniformly through the whole coincidence field, scattered events are taken o ff as subtracting mean value o f ray sums occurring outside the FOV. The absorption correction in PET is being developed.
3.
RESULTS AND DISCUSSIONS
The Headtome has been clinically used with SET since November 1979. The electronics adjustment in PET was concluded in March 1980. 3.1. SET performances The fin collimator revealed some favourable characteristics. Efficiency was highest at the neutral point o f the inclining angle, resulting in the lowest increase o f statistical noise induced by absorption correction, and high spatial resolving power with the minimum sacrifice o f sensitivity. The collimator also features a high speed capability for dynamic study in SET.
L/l С»
TABLE I. HEADTOME SCAN PARAMETERS
Single-photon scan
Rotation steps3
Collimator swing (degree)
Linear sampling (mm)
Angular sampling (degree)
High speed
None
21 X 2.8
2 0 .6
2 .8
2
Medium speed
2 X D/2
11 X 5.6
10.3
5.6
8
Medium resolution
4 X D/4
11 X 5.6
5.2
5.6
20
High resolution
8
X D/ 8
11 X 5.6
2 .6
5.6
40
Positron scan
Rotation steps
Wobbling points
Linear sampling (mm)
Angular sampling (degree)
Unit scan time (s)
Low resolution
None
None
2 0 .6
2 .8
Medium resolution
2 X D/2
None
10.3
2 .8
2
High resolution
2 X D/2
5 -1 0
3
2 .8
20
1
KANNO et al.
a D = crystal-crystal interval = 20.6 mm.
Unit scan time (s)
159
IAEA-SM-247/13 SINGLE-PHOTON TANGENTIAL RESOLUTION
Distance from Centre (cm)
FIG.4. Spatial resolution of SET. The high resolution m o d e revealed a tangential resolution of 10 m m and 7.5 m m F W H M at the centre and at the 1 О-cm radius respectively. Even at the lowest scan mode, the high speed scan, a resolution of 24 m m and 20 m m F W H M was obtained respectively.
TABLE II.
Single-photon
SLICE THICKNESS AND ITS SENSITIVITY
Slice setting (mm)
Image thickness (mm)
Sensitivity3 (kcounts-s -1 -jLtCi-1 -ml)
16 Slit
20 FWHM
16b
20
25
20
24
30
26
12 Gap Positron
6
FWHM
7C
20
10
20
28
13
30
а Measured by 99 Tcm and 68Ga for single-photon and positron respectively b These are measured at the energy window of 110 to 170 keV, thus including scattered events. c These are values after subtraction of the accidental and the scattered events.
KANNO et al.
160
POSITRON RESOLUTION BY 6BGa 30 Ring Rotation + 9 Points Wobbling
FWTM
♦---- i —
*-A
20 o o С Л
10
a
Tangential
• Radial
2 1Ш10 68Ga Line Source - in 20 cm0 Water Pool At Energy Window of (100-700) keV
0
0
2
Ц
6
8
10
Distance from Centre (cm)
FIG.5. Spatial resolution of PET. The tangential resolution with the ring rotation and 9 points wobbling showed 11 m m and 13 m m F W H M and 22 m m and 2 5 m m F W T M at the centre and at the 10-cm radius respectively, and a radial resolution of 12 m m and 13 m m F W H M and 21 m m and 22 m m F W T M respectively.
Spatial resolutions examined using a 2-mm ф line source in a 20-mm ф water pool are shown in Fig. 4. Higher resolutions were obtained at the periphery than at the centre. The high resolution mode gave 7.5 mm and 10 mm FWHM at the 10-cm radius and at the centre respectively. The lowest resolution scan, the high speed mode, gave 20 mm and 24 mm FWHM at each. An axial resolution of slice thickness measured by a 0.5-mm-thick " T c m-plane source in the 20-cm ф water pool was 20 mm, 25 mm and 30 mm FWHM for slice collimators o f 16 mm, 20 mm and 24 mm slit respectively (Table II). Only slight differences o f less than 10% o f FWHM were observed through various radii. Sensitivities were measured for each slice collimator using a 20-cm ф cylinder filled with " T c m or 133Xe solution (Table II). Sensitivities were almost proportional to the slice thickness. With the standard slice collimator with 20-mm slit, the sensitivities for " T c m and 133Xe were 21 and 10 kcounts s ' 1-/uCi-1 ml respectively .1
1 1 Ci = 3.70 X 10 10 Bq.
IAEA-SM -247/13
80-248
Y.M.
161
46years, M
Anterior Trunk Occlusion of Right MCA.
XCT
81Krm Infusion ECT (Regional Cerebral Perfusion of right MCA Distribution)
FIG. 6. Cerebral blood perfusion measured by ы K rm equilibrium imaging. The patient with occlusion o f the middle cerebral artery was exam ined by intracarotid continuous infusion o f glK rm . The defective area farrow) was clearly shown to be more widespread than that o f low density in X-ray CT.
Scatter fractions in the final reconstructed images were examined using a 20-cm ф water pool with uniform activity containing a 5-cm ф cylinder without activity. An enormous scatter fraction was found in a study with " T c m, 45% and 25% o f surrounding area at the centre and at the 10-cm radius respectively.
3.2. PET performances PET showed increased quality, particularly in uniformity in response through out the FOV, than with SET. Spatial resolutions measured with the same phantom used in SET revealed a higher resolution at the centre than at the periphery (Fig. 5). The high resolution mode with 9 points wobbling showed 11 mm and 13 mm FWHM at the centre and at the 10-cm radius respectively.
162
KANNO et al. Regional CBF Study by 133Xe 80-765
M.S,
In tra Venous In je c tio n and HEADTOME
L e ft Middle Cerebral A r te ria l Occlusion
FIG. 7. Quantitative cerebral blood flo w calculated from seq u entia l 133Xe clearance images. Following the intravenous slow injection o f 30 mCi 133Xe saline, sequential images o f 20 s were measured simultaneously with the end-tidal I33X e concentration curve. A cerebral blood flo w o f about 500 pixels each measuring 7 X 7 mm was calculated by the time integrating technique.
The sensitivities measured were 7, 20 and 30 kilo true events per цСi/ml with the same phantom filled with a 68Ga solution for a 12-mm, 20-mm and 28-mm slice gap respectively. These are the values after subtracting the accidental and scattered events mentioned in Section 2.6. The sensitivity o f the Headtome is insufficient, being one-fifth o f that o f Positome II employing BGO; however, in PET a dynamic study is inherently less important than in SET because most o f the PET studies are performed under a physiological equilibrium state o f an emitter in the tissue. Scatters observed in the final reconstructed images were far less than in SET, whereas on the ray sum projection there were 13%, 19% and 24% scatters with slice gaps o f 12 mm, 20 mm and 28 mm respectively, at an energy window of 100 to 700 keV. 3.3. Qinical applications The primary purpose o f the Headtome was to measure a local cerebral blood flow by SET. Figure 6 shows images o f the cerebral blood perfusion o f a patient, with occlusion o f the right middle cerebral artery, measured during an 81Krm-infusion via the right internal carotid [ 6], with high resolution scan, 40 s per slice, and with a 16-mm slit in the slice collimator.
163
IAEA-SM -247/13
:
*
.
(
* Ga-EDTA
BBB study
Case 69-0461 Occlusion of Left Posterior Cerebral Artery
FIG.8. Blood brain barrier study by 68Ga-EDTA. Posterior cerebral artery occlusion measured follow ing intravenous injection o f 2 mCi 6SGa-EDTA. A high resolution scan with 5 points wobbling was carried out. The scan time per slice was 2 min.
An inert gas cerebral clearance method is the only method that can give a quantitative local cerebral blood flow (CBF). Figure 7 shows the tomographic CBF o f a patient with occlusion o f the right middle cerebral artery measured by a 133Xe-intravenous injection technique. The scan was made at a level o f 5 cm above the OM-line using a 24-mm slice collimator with high speed scan o f 2 s. From the sequential images reconstructed from data integrated every 20 s, CBF values o f every pixel were calculated by ‘time integrating’ method [7]. Conventional blood brain barrier (BBB) studies are performed routinely using a " T c m-pertechnetate or a 68Ga-EDTA. Figure 8 shows 68Ga-EDTA study on a patient with left posterior cerebral artery occlusion examined following a 2 mCi intravenous injection with high resolution scan with 7 points wobbling and a slice gap o f 20 mm.
4.
CONCLUSION
A hybrid-type emission tomograph for imaging SET and PET was designed and constructed. High performances were obtained in experiments and clinical applications. The hybridity allows a high flexibility in physiological studies in the clinic without an in-house cyclotron.
164
KANNO et al.
REFERENCES [ 1]
PH ELPS, M.E., Emission com puted tom ography, Sem in. Nucl. Med. 7 (19 7 7 ) 337.
[2]
UEM URA, K., KANNO, I., M IURA, S., M IURA, Y., KAWATA, Y., TAKA HASHI, S., H IR O SE , K., KOGA, K., HEADTOME: A new hybrid emission tom ograph and its application for brain study, J. Nucl. Med. 21 (1 9 8 0 ) P15. ВОШ ., C., e t al., A c o m p u te r assisted ring d e te cto r po sitro n cam era system for reco n stru ctio n tom ography o f the brain, IEEE Trans. Nucl. Sei. NS-25 (19 7 8 ) 624. D E R E N Z O , S.E., et al., High resolution com puted tom ography of positron em itters, IEEE Trans. Nucl. Sei. N S -2 4 (1 9 7 7 ) 544.
[3 ] [4] [5] [6] [7]
CHO, Z .H ., e t al., Coincidence electronics for the circular ring transaxial p ositron cam era, Nucl. Instrum . M ethods 148 (1 9 7 8 ) 415. FA Z IO , F., e t al., Assessm ent o f regional cerebral blood flow by co n tin u o u s carotid infusion o f kry p to n -8 Im , J. Nucl. Med. 18 (1 9 7 7 ) 962. KANNO , I., LASSEN, N.A ., Two m ethods for calculating regional cerebral flow from em ission co m puted tom ography o f inert gas c oncentrations, J. C om put. Assist. T om ogr. 3 (1 9 7 9 ) 71.
IAEA-SM-247/54
R O T A T IO N A L P O S IT R O N C O M P U T E D TOM OGRAPHS E. TANAKA, N. NOHARA, M. YAMAMOTO, T. TOMITANI, H. MURAYAMA, Y. TATENO National Institute o f Radiological Sciences, Chiba-shi K. ISHIMATSU Hitachi Medical Corporation, Kashiwa-shi K. TAKAMI Central Research Laboratory, Hitachi Ltd, Kokubunji-shi, Japan
Abstract ROTA TIO N A L PO SIT R O N COM PUTED TOM OGRAPHS. A description and analysis are presented o f positron-em ission c o m p u ted tom ographs w hich utilize th e co n tin u o u s ro ta tio n scan o f a ring d e te cto r array. T he d e te cto rs are arranged in a particular irregularity d eterm ined by a com puter. T he aim o f th e ro ta tio n scan is to exceed the sam pling requirem ent to recover intrinsic d e te cto r reso lu tio n w ith high sam pling redundancy. It has already been show n th a t fine and unifo rm sam pling can be achieved w ith continuous m ultiple ro ta tio n . T o o b tain an angularly invariant sam pling w ith lim ited angle (< 3 6 0 ° ) ro ta tio n , the angle o f ro ta tio n m ust be 360°/n w here n is an odd integer ( > 3 ), and the d e te cto r array should be com posed of n iden tical subarrays arranged in every 360 / n. T he detecto rs m ust also be arranged o n a non-circular ring to avoid c o n ce n tra tio n o f sam pling points. An exam ple of a d e te c to r array for 120° ro ta tio n (n = 3 ) show s excellent sam pling characteristics. The effect o f linear and angular sam pling on spatial resolution is discussed taking in to account the sm oothing effect o f data binning. It is show n th a t th e sm oothing effect is n o t appreciable, if the N yquist frequencies for sam pling are pro p erly chosen. A ro tary p o sitro n tom ograph, Positologica, w hich has been developed fo r brain study, has a circular array of 64 BGO crystals. T he spatial reso lu tio n is 5.8 m m FWHM at th e centre and less th an 9 m m FWHM in the field of view. T he p o in t spread functions well reflect the results o f th eo re tic al analysis.
1.
INTRODUCTION
One aim in recent improvements in positron-emission computed tomography (PECT) is high detection sensitivity with high spatial resolution. A circular ring array o f BGO crystals is a suitable choice for this goal at present. In the ring
165
166
TANAKA et al.
system, the geometric sampling problem is o f particular importance, because the linear sampling interval in the stationary mode is not fine enough to recover the resolution capability o f detectors. The half detector angle rotation can halve the sampling interval [ 1 ], but is still insufficient for optimization. A wobbling m otion is therefore implemented in some recent devices [2, 3]. We have proposed the continuous rotation o f a ring array o f detectors arranged with a particular non-uniform spacing [4]. The advantage o f the rotational scan is that fine and uniform sampling is achieved with high sampling redundancy. The high sampling redundancy reduces the effect o f the instability o f the detector gain. The continuous rotation scan is divided into two types: multiple rotation and limited angle (< 3 6 0 °) rotation. The former can offer a rather arbitrary scan speed, but some design com plexity is inevitable because o f the electric power supply to the rotary part and data transmission from the rotary part to the stationary part. We have recently developed a system of this type for brain study using 64 BGO crystals [5, 6 ]. In this system, electric power is supplied through a slip ring, and data transmission is made through a parallel multi-bit rotary photocoupler. In the limited angle rotation type, a 360° rotation may be feasible, but a smaller angle rotation may be preferable mechanically if the required sampling property is achieved. The sampling problems in the rotational PECT are discussed together with a description o f the performance o f our head PECT system with emphasis on spatial resolution.
2.
DETECTOR ARRANGEMENT
A feature o f the rotational PECT is that the detectors are arranged on a ring so that the sampling density in projections is as uniform as possible in the range o f interest, where the sampling density is the number o f coincidence lines falling in a projection bin. The arrangement can be determined by trial and error or by a computer iteration search program, the latter method giving better results [7, 8]. In the iteration method, the position of each detector is moved along the ring separately in turn, and at each step the detector is fixed at a position where a certain figure o f merit on the sampling density uniformity is maximized. The cycle o f iteration is repeated until all detectors no longer move. An irregular arrangement requires a certain amount o f marginal space between detectors, and may result in reduction o f geometric efficiency. In the 360° rotation, however, it has been shown that a small total margin (in 360°) o f the order of one detector width is sufficient to obtain satisfactory sampling characteristics. Distributions o f both the sampling densities and detector sensitivities in projections are independent o f the view angle. This property simplifies the correction o f projection data for non-uniformity o f the distributions.
IAEA-SM-247/54
167
m E
e 10-
со
ÛT Ш CL > ► -
1
<л 5 z
Ш
Q
O Z ^
< W
F IG .L
0
r 11
0
y 1Y * 11u 2
S11 T 11 Гд l V l l f l l l | t r l >iu t I
U
6
e
10
DISTANCE FROM CENTRE (cm )
D e te c to r arrangem ent f o r a 1 2 0 ° ro ta tio n system and its sam pling density d is trib u tio n .
In the smaller angle (< 3 6 0 ° ) rotation, the sampling density distribution generally depends on the view angle, which in practice is not convenient. A possible way of realizing angularly invariant sampling is to divide the entire circle into n(=integer) identical segments, and to allocate identical sub-arrays into them. The angular range o f rotation is 360°/n. In such arrays, however, sampling points concentrate at distances ±Rd cos(mîr/n) from the centre, where R(j is the detector ring radius and m is a positive integer smaller than n. If n is even, a concentration occurs at the centre o f the projections, and hence n must be odd. For n = 3 (120° rotation), concentrations occur at R = ±Rd/2. One way to avoid the concentrations is to arrange the detectors on a non-circular ring so that the detectors in each segment have different distances from the centre. Figure 1 shows an example o f such arrays and its sampling density distribution. The arrangement is searched by computer iteration assuming that the minimum centreto-centre spacing o f detectors is 16 mm whereas the average spacing is 18.2 mm. The distance o f the detectors from the centre is changed linearly with the angle in a range 20—21.6 cm.
3.
EFFECT OF SAMPLING ON THE SPATIAL RESOLUTION
If we neglect the positron range and angular deviation o f the annihilation photons, the in-plane spatial resolution is limited by the spatial resolution of the detectors and the resolution imposed by linear and angular sampling. We shall
168
TANAKA et al. MT F
(a) DETECTOR [ TANGENTIAL)
(Ы BINNING LINEAR
ttR
(C)
ANGULAR BINNING
N
N / ttR
F IG .2.
D e te c to r reso lu tio n and sm oo th in g e ffe ct o f b in n in g in linear and angular sampling.
consider first the detector response for the coincidence detection assuming rectangular detectors (crystals). The spillar effect o f photons from detectors is neglected. The detector response at the ring centre is isotropic, and is triangular with a FWHM equal to half the detector width d. The response at an off-centre position is not isotropic, and its shape is represented by a radial and tangential response. (N ote that these responses refer to the detector and not to the reconstructed images.) The radial response is usually broadened with increasing distance from the centre, but the broadening is improved by the use o f shielding septa between the crystals. The tangential response becomes more rectangular as the distance and the FWHM increases accordingly. In Fig.2(a), the tangential response and its MTF (modulation transfer function) are shown for R = 0 and R = R(j/2. The two MTFs go to zero at = 2/d and u'd = 4/(3d) respectively, although these MTFs still have higher frequency components. In the rotational PECT, the original sampling points in projections distribute randomly, and binning o f projections with a sampling interval, a, is equivalent to applying rectangular smoothing to the original projections before the sampling. The smoothing function and its MTF are shown in Fig.2(b). The MTF goes to zero at 1/a, which is twice as high as the Nyquist frequency, t>a = l/(2 a ), o f the linear sampling. Angular binning o f projections imposes tangential smoothing on the detector response at the off-centre positions. The angle o f a coincidence line is obtained by summing the angle o f a rotating gantry and the angle o f the line with respect to co-ordinates fixed at the gantry. If both the angles are binned with Д0
IAEA-SM-247/S4
169
DISTANCE FROM CENTRE F IG .3 . 2
Responses o f the ro ta ry p o sitro n tom ograph Positologica f o r a line source I 6 *G a,
mm ф ) a t various points.
independently, the response function o f the tangential smoothing at distance R is triangular with a FWHM equal to RA0 = 7rR/N, where N is the number o f views (see Fig.2(c)). The MTF o f the angular smoothing goes to zero at N /( 7rR) which is twice as high as the Nyquist frequency, vg = N/(27rR), o f the angular sampling. (The response is somewhat improved by finer sampling o f two angles before summation.) From the remarks above on the MTFs and from the requirements imposed by the sampling theorem for accurate (i.e. artifact-free) image reconstruction, a reasonable choice o f the sampling interval will be va > and vg > v'A if we neglect the higher frequency components o f the detector response, or more preferably va = (3/2)i>d and Vg = vd taking into consideration the higher frequency. The latter choice is rewritten as a = d /6
and
N s 47rR/d
(1)
With this choice, image blurring due to the smoothing effects o f both the linear and angular binning are not appreciable.
4.
A PROTOTYPE ROTARY POSITRON TOMOGRAPH: POSITOLOGICA
A positron tomograph o f the continuous multiple rotation type has been developed for studying the human brain. The device has a circular array o f 64
170
TANAKA et al.
DISTANCE FROM CENTRE ( cm )
F IG .4. of
2
FW HM s o f the responses o f the Positologica. E xp e rim e n ta l data are f o r a line source
mm ф.
A : R a d ia l response o f reconstructed image. B:
Tangential response o f reconstructed image.
C:
Tangential response observed in p ro je ctio n .
D : Tangential geom etric response o f detectors.
13NH3 F IG .5.
Images o f
are a b o u t 2
13 N H 3
11CO
and n CO o b taine d w ith a n o rm a l subject (O M + 4 cm ). T o ta l counts
X 1 0 6 each. P h o to n abso rp tio n was corrected by transmission data.
IAEA-SM-247/54
171
12 X 20 X 26mm BGO crystals which is rotated at a speed o f 60 rev/min. The detector ring radius is 22 cm. Taking into consideration the sampling require ment given by E q .(l), we adopted the sampling interval a = 2 mm (=d/ 6) and the number o f views N = 128 ( = 47rR/d for R = 12 cm). The detector arrangement was determined by computer iteration. The sampling density per 2 mm bin is 11.9 average, with a minimum o f 11, in the field o f view. The sensitivity is about 17 к counts's -1 per /uCrml for a 20 cm ф water phantom with 2 cm slice .1 Figure 3 shows the response functions o f reconstructed images o f a line source ( 68Ga, 2mm Ф) placed in air perpendicularly to the detector plane. Figure 4 shows various FWHMs as functions o f the distance from the centre. The excellent resolution at the centre is because the position corresponds exactly to a sampling point for all views, and this fact decreases blurring from linear binning, source size, angular deviation o f photons, etc. The radial response at off-centre will be improved by the use o f a septal shield between the crystals. Figure 5 shows images o f 13NH 3 and 11CO obtained with a normal subject.
5.
CONCLUSION
It has been shown that the rotational PECT provides a sampling capability to recover the principal com ponent o f the detector resolution by a suitable choice o f linear and angular sampling intervals. In the limited angle ( 120°) rotation, a satisfactory sampling characteristic can be obtained with a non-circular array o f detectors. The rotary PECT system, Positologica, gives an excellent imaging performance, although the practical resolution capability may be limited by the necessity of software smoothing to improve counting statistics.
ACKNOWLEDGEMENTS The authors thank T.A. Iinuma, Y. Suda and M. Endo for their assistance in image reconstruction. This work was partly supported by a grant from the Ministry o f Health and Welfare, Japan.
REFERENCES [ 1] CHO, Z.H., CHAN, J.K., ERIKSSON, L., Circular ring transverse axial positron camera for 3-dimensional reconstruction of radionuclides distribution, IEEE Trans. Nucl. Sei. NS-23 (1976) 613. 1 1 Ci = 3.70 X 1010 Bq.
172
TANAKA et al.
[2] BOHM, C., ERIKSSON, L., BERGSTROM, M., et al., A computer assisted ring detector positron camera system for reconstruction tomography of the brain, IEEE Trans. Nucl. Sei. NS-25 (1978) 624. [3] TER-POGOSSIAN, M.M., MULLANI, N.A., HOOD, J.T., et al., Design considerations for a positron emission transverse tomograph (PETT V) for the imaging of the brain, J. Comput. Assist. Tomog. 2 (1978) 539. [4] TANAKA, E., NOHARA, N., YAMAMOTO, M., et al., “Positology” - The search for suitable detector arrangements for a positron ECT with continuous rotation, IEEE Trans. Nucl. Sei. NS-26 (1979) 2728. [5] TANAKA, E., NOHARA, N., TOMITANI, T., et al., A positron emission computed tomography: “POSITOLOGICA” , Radioisotopes (Tokyo) 29 (1980) 302. [6 ] NOHARA, N„ TANAKA, E., TOMITANI, T., et al., POSITOLOGICA: A positron ECT device with a continuous rotating detector ring, IEEE Trans. Nucl. Sei. NS-27 (1980) 1128. [7] YAMAMOTO, M., NOHARA, N., TANAKA, E., A new method for fine and uniform sampling in positron emission CT, Proc. 6 th Intern. Conf. Information Processing in Medical Imaging, Paris, 2 - 6 July 1979 (1980) 201. [ 8 ] YAMAMOTO, M., Detector arrangement and sampling characteristics in a rotary positron emission CT, Phys. Med. Biol, (accepted for publication).
IAEA-SM-24 7/202
I n v ite d R e v ie w P a p e r
S IN G L E P H O T O N IM A G IN G N e w in s tr u m e n ta tio n a n d te c h n iq u e s G. MUEHLLEHNER, J. COLSHER University o f Pennsylvania, Philadelphia, Pennsylvania, United States o f America
Abstract SINGLE PHOTON IMAGING: NEW INSTRUMENTATION AND TECHNIQUES. The performance of Anger scintillation cameras continues to be enhanced through a series of small improvements which result in significantly better imaging characteristics. The most recent changes in camera design consist of: ( 1 ) the introduction of photomultipliers with better photocathode and electron collection efficiencies, (2) the use of thinner (3/8 or 1/4 in) crystals giving slightly better intrinsic resolution for low gamma-ray energies, (3) inclusion of a spatially varying energy window to compensate for variations of light collection efficiency, (4) event-by-event, real-time distortion removal for uniformity correction, and (5) introduction of new methods to improve the count-rate capability. Whereas some of these improvements are due to better understanding of the fundamentals of camera design, others are the result of technological advances in electronic components such as analogue-to-digital converters, micro processors and high-density digital memories. This trend is likely to continue for many years to come, culminating in replacement of the former analogue processing system by a digital processor. Whereas intrinsic camera performance has improved significantly, only modest gains have been achieved in collimator imaging characteristics. Thus, the system performance is approaching a limit set by the collimator, at least for the radiopharmaceuticals in current use. The development of single photon tomography in nuclear medicine has developed along two parallel paths. Multipinhole and rotating slant-hole collimator attachments provide some degree of longitudinal tomography, and are currently being applied to cardiac imaging. At the same time rotating camera systems capable of transverse as well as longitudinal imaging are being refined technically and evaluated clinically. Longitudinal tomography is of limited use in quantitative studies and is likely to be an interim solution to three-dimensional imaging. Rotating camera systems, on the other hand, not only provide equal resolution in all three dimensions but are also capable of providing quantitative accuracy. This is the result of progress in attenuation correction and the design of special collimators. Single photon tomography provides a small but noticeable improvement in diagnostic accuracy which is likely to result in widespread use of rotating camera systems in the future.
1 . INTRODUCTION T h is p a p e r in stru m e n ta tio n
re v ie w s r e c e n t d ev elo p m en ts in s in g l e p h o to n and te c h n iq u e s . S in c e th e Anger s c i n t i l l a t i o n
173
MUEHLLEHNER and COLSHER
174
cam era c o n tin u e s to b e th e p r i n c i p a l im ag in g d e v ic e u se d in n u c l e a r m e d i c i n e , t h e d i s c u s s i o n o f new i n s t r u m e n t a t i o n w i l l
be
l i m i t e d p r i m a r i l y t o im p ro v e m en ts i n i t . The t h r e e m o st b a s i c and im p o r ta n t im a g in g p a ra m e te rs - u n i f o r m i t y , r e s o l u t i o n and c o u n t- r a te c a p a b ility - have r e c e n tly o r a re c u r r e n tly b e in g im p ro v e d s i g n i f i c a n t l y . Each o f th e s e w i l l be d is c u s s e d in s e c tio n 2. F o r a r e v i e w o f o t h e r gamma c a m e r a t e c h n o l o g y i n c l u d i n g m u l t i c r y s t a l cam era s and s e m ic o n d u c to r d e t e c t o r s , th e reader
is
referred
to
M cK eig h en
(1 ).
S y ste m p e r fo r m a n c e an d c o n f i g u r a t i o n s a r e d i s c u s s e d i n s e c tio n 3. As t h e b a s i c d e t e c t o r i s p e r f e c t e d t h e c o l l i m a t o r h a s becom e th e l i m i t i n g f a c t o r in sy s te m r e s o l u t i o n . A dvances i n u ltr a s o u n d and x -ra y -c o m p u te d to m ography h a v e c a u s e d th e p r a c t i c e o f n u c l e a r m e d ic in e to s h i f t from t r a d i t i o n a l p r o c e d u re s such as b r a in sc a n n in g to w ard s i n v e s ti g a ti o n o f p h y s io l o g i c o r d y n am ic p r o c e s s e s s u c h a s g a te d c a r d i a c im a g in g . T h is t r a n s i t i o n , c o m b in ed w i t h t h e r a p i d e v o l u t i o n o f c o m p u te rs , h a s r e s u l te d in a s i g n i f ic a n t change in th e c o n fig u ra tio n o f a 's ta n d a r d ' A nger s c i n t i l l a t i o n cam era. I t now t e n d s t o a c c u m u l a t e t h e im a g e i n a d i g i t a l m em ory w i t h t h e c a p a b i l i t y f o r r a p i d d y n a m ic im a g in g a n d w i t h im a g e d i s p l a y an d r e c o r d i n g p e r form ed v i a a v id e o te r m in a l. As t h e q u a l i t y o f d i g i t a l d i s p l a y s i s im p ro v e d to e q u a l th e fo rm e r a n a lo g d o t - b y - d o t d i s p l a y s , d i g i t a l i m a g e r e c o r d i n g a n d d i s p l a y m ay b e c o m e t h e s t a n d a r d im a g e p r e s e n t a t i o n m ode. S i n g l e > - p h o t o n e m i s s i o n c o m p u te d to m o g r a p h y (SPECT) w i l l b e d is c u s s e d in s e c tio n 4. E m i s s i o n c o m p u t e d t o m o g r a p h y (E C T) te c h n iq u e s have lo n g b een i n v e s tig a te d in n u c le a r m e d ic in e . W h ile t h e r e c e n t e m p h a sis h a s b e e n on p o s i t r o n to m o g ra p h y a s a r e s e a r c h t o o l , s i n g l e - p h o t o n c o m p u ted to m o g ra p h y h a s b e e n e x p lo re d f o r c l i n i c a l n u c le a r m e d ic in e . A stro n g case can be m a d e t h a t s i n g l e p h o t o n ECT i s c a p a b l e o f q u a n t i t a t i o n e q u a l l i n g p o s i t r o n ECT. H ow ever, t h i s can b e a c h ie v e d o n ly w ith s y s te m s w h ich c o l l e c t p r o j e c t i o n d a ta from a f u l l 360 d e g r e e s . L o n g itu d in a l to m o g rap h ic a tta c h m e n ts such as 7 -p in h o le o r s l a n t - h o l e c o l l i m a t o r s p r o v i d e som e t o m o g r a p h i c e f f e c t by v ie w in g th e d i s t r i b u t i o n th ro u g h a lim ite d a n g u la r ra n g e .
2.
DETECTOR IMPROVEMENTS
The p e rfo rm a n c e o f A nger s c i n t i l l a t i o n c am era s c o n tin u e s to b e e n h a n c e d th ro u g h a s e r i e s o f s m a ll im p ro v em en ts w h ich ta k e n t o g e t h e r h a v e r e s u l t e d in s i g n i f i c a n t l y b e t t e r im a g in g c h a ra c te ristic s. Some o f t h e s e i m p r o v e m e n t s r e s u l t f r o m a b e t t e r u n d e r s ta n d in g o f cam era d e s ig n and im p ro v em en ts i n i n d i v i d u a l c o m p o n en ts su c h a s p h o t o m u l t i p l i e r t u b e s . O th ers a re
IAEA-SM-247/202
175
l a r g e l y t h e r e s u l t o f t e c h n o l o g i c a l a d v a n c e s i n e l e c t r o n i c com p o n e n ts such a s a n a l o g - t o - d i g i t a l c o n v e r te r s , m ic ro p ro c e s s o rs and h i g h - d e n s i t y d i g i t a l m em o ries. I t i s t h is l a t t e r d e v e lo p m ent t h a t h a s p e r m itte d th e d ev elo p m en t and im p le m e n ta tio n o f th e u n i f o r m i t y c o r r e c t i o n m eth o d s t h a t a r e d i s c u s s e d in th is se c tio n . A lso d is c u s s e d in t h i s s e c ti o n a r e i n t r i n s i c r e s o l u t i o n im p ro v e m en ts. F in a lly , p ro v in g count r a te p erfo rm an ce a re
2 . 1 .U n ifo rm ity
C o rre c tio n
N o n u n ifo rm itie s
in
p o te n tia l so lu tio n s d isc u sse d .
fo r
im
M eth o d s
th e
fie ld
flo o d
of
a sc in tilla tio n
cam era a r e c a u s e d by v a r i a t i o n s i n p o i n t- s o u r c e s e n s i t i v i t y and s p a t i a l d i s t o r t i o n s (2 ). P r e v io u s m eth o d s t o c o r r e c t f o r f i e l d n o n u n if o r m itie s i n s c i n t i l l a t i o n cam eras d id n o t a d d re s s th e fu n d a m e n ta l c a u s e s . R a th e r th e y sim p ly in c r e a s e d o r d e c re a s e d th e t o t a l c o u n ts a c q u ire d in a p a r t i c u l a r a r e a u sin g in f o r m a tio n from p r e v io u s ly a c q u ir e d f i e l d flo o d s ( 3 ,4 ) . S in c e th e y d id n o t c o r r e c t l y c o m p e n sa te f o r th e p rim a ry c a u s e s o f n o n u n i f o r m i t ie s , a r t i f a c t s and v a r io u s r e l a t e d p ro b lem s ( 2 ,5 ,6 ,7 ) c o u ld r e s u lt.N o n u n if o r m itie s due to v a r i a t i o n s i n p o in t- s o u r c e s e n s i t i v i t y sh o u ld b e c o r r e c te d u s in g th e s l i d i n g e n e rg y w in dow t e c h n i q u e ( 2 , 8 ) i n w h i c h t h e e n e r g y w in d o w i s a d j u s t e d on an e v e n t-b y -e v e n t b a s is to c o r re c t fo r lo c a l v a r ia tio n s in p h o to p eak p u ls e h e ig h t. N o n u n ifo rm itie s due to s p a t i a l d i s t o r t i o n s s h o u ld b e c o r r e c te d by re m o v in g th e d i s t o r t i o n s and r e c o rd in g th e d e te c te d e v en t in i t s a p p r o p ria te s p a t i a l lo c a tio n (9 ,1 0 ).
2 .1 .1 .
C o rre c tio n
fo r
P o in t-S o u rc e
S e n s itiv ity
V a ria tio n
V a ria tio n s in p o in t-s o u rc e s e n s i t iv i t y (co u n t r a te ) as a f u n c tio n o f p o s i t i o n on th e cam era f a c e a r e c au sed by s h i f t s in th e a v e ra g e p h o to p ea k p u ls e h e ig h t r e l a t i v e to th e energy a c c e p t a n c e w in d o w . T h e s e v a r i a t i o n s c a n r e s u l t f r o m a n im p ro p e r g a in s e ttin g of a p h o to m u ltip lie r as a r e s u l t o f in c o r r e c t tu n in g o r e l e c t r o n i c com ponent changes w ith tim e . They c an a l s o b e d u e t o an o p t i c a l d e s i g n w h e re l i g h t c o l l e c tio n e ffic ie n c y v a rie s as a fu n c tio n of p o s itio n . U s in g a s i m p l e s p a t i a l l y i n v a r i a n t e n e r g y w in d o w h a s a nu m b er o f d i s a d v a n t a g e s : 1) t h e e n e r g y r e s o l u t i o n i s w o rs e th a n i t s h o u ld b e b e c a u s e a l l r e g io n s o f th e cam era a r e t r e a t e d i n common, 2) from r e g io n to r e g io n and 3) th e ia tio n s at
th e p o in t-s o u rc e as th e p h o to p eak
d e s ig n m ust be a s a c rific e in
s e n s i t iv i t y w ill v ary s h i f t s r e l a t i v e to th e
co m p ro m ised t o m in im iz e s p a tia i re so lu tio n .
energy
v ar
w in d o w
176
F IG . l.
MUEHLLEHNER and COLSHER
B lo c k diagram o f event-by -event processor f o r po sition-dependent energy d iscrim in a tio n .
S in c e s p a t i a l v a r i a t i o n s in p h o top eak a m p litu d e change s lo w ly w ith p o s i t i o n i t i s p o s s i b l e to r e c o r d photop eak am p li tu d e in a 64 X 64 m a tr ix c o r r e sp o n d in g t o a r e g u la r a r ra y o f r e g io n s on th e s c i n t i l l a t i o n c r y s t a l and u s e t h i s in fo r m a tio n to c o r r e c t f o r th e v a r i a t i o n s . F ig u r e 1 shows a b lo c k diagram o f an e v e n t - b y - e v e n t p r o c e s s o r w hich u s e s th e p o s i t i o n c o o r d in a te s t o lo o k up th e s p a t i a l l y v a r y in g p h o top eak in fo r m a tio n and th en e i t h e r moves th e en erg y window (8 ) or a d j u s t s th e en erg y s i g n a l a m p litu d e (1 1 ) to com pensate f o r s p a t i a l v a r i a t i o n s in p h o to peak a m p litu d e . I t has b een shown (8 ) t h a t t h i s c o r r e c t io n i s v a l i d o v er a w id e ran ge o f en erg y window w id t h s , gamma-ray e n e r g i e s , i n c i d e n t cou n t r a t e s and v a r y in g s c a t t e r c o n d it io n s . A f u r t h e r r e fin e m e n t o f t h i s te c h n iq u e in v o lv e s ch an gin g th e w id th o f th e en erg y window to com pensate f o r p o s i t i o n a l v a r ia t i o n s in e n e r g y r e s o l u t i o n ( 1 0 ) . 2 . 1 . 2 . C o r r e c tio n f o r S p a t i a l D i s t o r t i o n S p a t i a l d i s t o r t i o n s a r e s y s t e m a t ic e r r o r s in th e p o s i t i o n in g o f s c i n t i l l a t i o n e v e n t s . Such d i s t o r t i o n s a r e cau sed by n o n lin e a r ch an g es in th e l i g h t d i s t r i b u t i o n in th e s c i n t i l l a t o r a s a f u n c t io n o f l o c a t i o n . S in c e th e l i n e a r Anger camera a r i t h m e tic scheme i s n o t a d eq u a te t o com pensate f o r t h e s e e f f e c t s , e v e n t s a r e n o t r e c o r d e d in t h e i r tr u e l o c a t i o n . The r e s u l t i n g e r r o r s ( l e s s th an 1 .0 mm) a re s m a ll compared to th e e v e n t - b y e v e n t e r r o r r e s u l t i n g from s t a t i s t i c a l u n c e r t a i n t i e s in th e
IAEA-SM-247/202
F IG .2.
177
B lo c k diagram o f on -line d is to rtio n rem oval processor. X a , Ya are the coordinates as
determ ined by the analog e lectronics and X c , Y c are the corrected coordinates.
num ber o f p h o to n s r e c e iv e d by each p h o t o m u l t ip l i e r . H ow ever, th e s e sm a ll d i s t o r ti o n s cau se v i s i b l e a r t i f a c t s b eca u se th e d isp la c e m e n ts a re a p p lie d to a l l e v e n ts in a p a r t i c u l a r re g io n . N o n u n if o r m itie s r e s u l t i n g from s p a t i a l d i s t o r t i o n a r e c au sed by l o c a l c o u n t c o m p re ssio n o r e x p a n s io n (1 2 ). To b e v i s u a l l y n o t i c e a b l e i n a n im a g e o f a l i n e p a t t e r n p h a n to m o r a S m ith o r t h o g o n a l - h o l e p a t t e r n s u c h d i s t o r t i o n s m u st e x c e e d s e v e ra l m illim e te rs in s p a t i a l d isp la c e m e n t; to be s ig n if ic a n t in c l i n i c a l im a g e s , t h e d i s t o r t i o n s m u st b e e v e n m ore s e v e r e . H o w e v e r , d i s t o r t i o n s may c a u s e u n a c c e p t a b l e f i e l d f l o o d v a r i a t i o n s when t h e d is p la c e m e n t i s l e s s th a n a m i l l i m e t e r . As an e x a m p l e , i f a c i r c u l a r a r e a o f 2 0 mm d i a m e t e r i s c o m p r e s s e d to w a rd s i t s c e n t e r fro m a l l d i r e c t i o n s ( a s w o u ld b e t h e c a s e o v e r t h e c e n t e r o f a p h o t o m u l t i p l i e r ) b y 0 . 4 mm, t h e e f f e c t i v e a r e a i s r e d u c e d f r o m 1 0 0 t t mm2 t o 9 2 i t m m ^. T h is c a u s e s an i n c r e a s e i n c o u n t d e n s i t y o f 8% w i t h a c o r r e s p o n d i n g r e d u c t i o n in count d e n s ity in th e su rro u n d in g a re a . Thus s p a t i a l d i s to r tio n s cause n o tic e a b le f ie ld - f lo o d n o n u n ifo rm itie s w e ll b e fo re d isp la c e m e n ts a re v is u a lly a p p a re n t in lin e p a tte r n im a g e s, and in d e e d a r e th e p rim a ry s o u rc e o f f i e l d - f l o o d a b n o r m a litie s . S in c e d i s t o r ti o n s a re s y s te m a tic and change slo w ly w ith p o s i t io n , th e e r r o r s can be c o rre c te d th ro u g h e v e n t-b y -e v e n t p ro c e s s in g d u rin g a c c u m u la tio n . F i r s t th e d is p la c e m e n ts m ust be
178
MUEHLLEHNER and COLSHER DISTORTED CAMERA
UNCORRECTED
F IG .3 .
CORRECTED
H ig h ly d is to rte d camera. F lo o d , s lit p a ttern and o rth o g o n a l hole p a tte rn images
(5, 1.5 and 2.5 m illio n counts respectively) w ith and w ith o u t co rre ctio n . The isotope was 9 9 Tcm
and a 15% energy w in d o w was used.
m e a su re d a c c u r a t e l y , w h ich can b e a c h ie v e d by u s in g a l i n e p a t te rn . S in c e t h e t r u e l o c a t i o n i s known a n d t h e a c t u a l ( d i s t o r t e d ) l o c a t i o n i n t h e im a g e i s m e a s u r e d , a c o r r e c t i o n d i s p l a c e m e n t c an be c a lc u la te d . T h is c a l c u l a ti o n i s p e rfo rm e d f o r a l l s o u rc e lo c a tio n s in th e f ie ld . I f th e s e c o r re c tio n d isp la c e m e n ts a re s t o r e d i n a m em ory, t h e n f o r e a c h e v e n t t h e X an d Y c o o r d i n a t e s can be d i g i t i z e d , and th e c o r r e c ti o n d is p la c e m e n ts can b e added to th e o r i g i n a l c o o r d in a te s to o b ta in th e c o r r e c t c o o r d in a te s of th e ev en t. F i g u r e 2 show s a b l o c k d ia g r a m o f a r e a l - t i m e , e v e n t - b y e v e n t p r o c e s s o r w h i c h h a s a 64 X 64 m em ory f o r d i s t o r t i o n c o e f f i c i e n t s and a h a rd w ire d i n t e r p o l a t o r . T h is l a t t e r i s n e c e s s a r y t o a s s u r e sm o o th t r a n s i t i o n o f t h e d i s t o r t i o n d is p la c e m e n ts
IAEA-SM-247/202
179
w h ich a r e a d d ed to th e o r i g i n a l X and Y c o o r d i n a t e s . If a s c i n t i l l a t i o n cam era i s d e s ig n e d f o r good s p a t i a l r e s o l u t i o n , larg e d is to r tio n s r e s u lt. F i g u r e 3 sh o w s a n e x a m p le o f a 37 tu b e la r g e f i e l d - o f - v i e w cam era b o th b e f o r e and a f t e r d i s t o r tio n re m o v a l, d e m o n s tra tin g t h a t d i s t o r t i o n s can b e th e m ajo r fa c to r in f ie ld -flo o d n o n u n ifo rm itie s. S in c e th e s e d is t o r ti o n s a re in h e r e n t in th e d e s ig n , th e y g e n e r a lly a re in d e p e n d e n t of e n e r g y w indow w i d t h , g a m m a -ra y e n e r g i e s , i n c i d e n t c o u n t r a t e s and
sc a tte r
c o n d itio n s
2 .2 ,In trin s ic
(9 ).
R e so lu tio n
Im p ro v em en ts
A num ber o f f a c t o r s h a v e c o n t r i b u t e d to th e im p ro v em en t i n i n t r i n s i c r e s o l u t i o n : 1) t h e i n t r o d u c t i o n o f p h o t o m u l t i p l i e r s w i t h b e t t e r e l e c t r o n c o l l e c t i o n e f f i c i e n c i e s , 2) t h e u s e o f t h i n n e r s c i n t i l l a t i o n c r y s t a l s , 3) i n c r e a s e i n t h e n u m b er o f p h o t o m u l t i p l i e r s and 4) im p ro v e d d e s i g n fre e d o m th r o u g h s p a t i a l l y v a r y i n g e n e r g y w ind o w s a n d s p a t i a l d i s t o r t i o n re m o v a l fo r u n ifo rm ity c o rre c tio n . V a rio u s m a n u fa c tu re rs o f s c i n t i l l a t i o n c a m e ra s u s e som e c o m b i n a t i o n o f t h e s e im p ro v e m e n ts w ith th e r e s u l t th a t th e b e s t i n t r i n s i c r e s o lu tio n of p re s e n t d a y l a r g e f i e l d - o f - v i e w c a m e r a s i s 3 . 5 - 4 . 0 mm FWHM. Each o f th e s e im p ro v e m en ts w i l l b e d is c u s s e d b e lo w . The im p a c t o f t h e s e a n d f u r t h e r im p ro v e m e n ts on s y s te m p e r f o r m a n c e w i l l b e d is c u s s e d in s e c tio n 3. 2 .2 .1 .
P h o to m u ltip lie r
P h o to m u ltip lie rs
D e v elo p m en ts w h ich
are
p rim a rily
in te n d e d
for
use
in
A nger s c i n t i l l a t i o n cam eras h av e r e c e n tl y been d e v e lo p e d . In p a r t i c u l a r , th e p h o to c a th o d e to dynode g e o m e try (13) o r f i r s t dynode s h a p e (14) h a s b e e n r e d e s ig n e d t o im p ro v e c o l l e c t i o n e ffic ie n c y . T h is r e s u l t s i n b e t t e r e l e c t r o n s t a t i s t i c s w h ich h a s a d i r e c t in f lu e n c e on b o th e n e rg y r e s o l u t i o n and s p a t i a l re so lu tio n . An e n e r g y r e s o l u t i o n o f 8 - 9 % h a s b e e n a c h i e v e d w i t h b o t h 2" a n d 3 " d i a m e te r p h o t o m u l t i p l i e r s a t a g am m a-ray e n e r g y o f 1 2 2 KeV ( 1 5 ) . A c o m p a ra b le num ber f o r th e im p ro v e m ent in s p a t i a l r e s o l u t i o n i s d i f f i c u l t to q u o te . 2 .2 .2 .
C ry sta l
W ith
th e
T h ic k n e s s in c re a se d
R e d u ctio n use
of
T c-99m
ra d io p h a rm a c e u tic a ls
and
th e in tr o d u c tio n of T l-2 0 1 , th e p erfo rm an ce of s c i n t i l l a t i o n c a m e ra s a t lo w e n e r g i e s h a s b eco m e i n c r e a s i n g l y i m p o r t a n t . R ed u cin g th e s c i n t i l l a t i o n c r y s t a l t h ic k n e s s fro m ^ in c h to e i t h e r 3 /8 in c h o r \ in c h , im p ro v e s i n t r i n s i c s p a t i a l r e s o l u t i o n b y a p p r o x i m a t e l y 1 mm FWHM f o r e n e r g i e s o f 1 4 0 KeV a n d b e lo w . W h ile v a r i o u s p u b l i c a t i o n s on t h i s s u b j e c t ( 1 6 ,1 7 ,1 8 ) a r e n o t i n c o m p le te a g re e m e n t, T a b le I p r e s e n t s th e e x p e c te d r e s o l u t i o n g a i n a n d s e n s i t i v i t y l o s s a t lo w e n e r g i e s ( 1 9 ) .
MUEHLLEHNER and COLSHER
180
TABLE I . RESOLUTION AND SENSITIVITY CHANGE AS RESULT OF CRYSTAL REDUCTION FROM % TO k, INCH
Iso to p e
In trin s ic re so lu tio n im p ro v em en t*
L oss o f se n sitiv ity
T l-2 0 1
1 .3
n e g lig ib le
T c-99m
1 . 0 mm FWHM
*The r e s o of m illim fu n c tio n ; p lie d by
mm FWHM
15%
l u t i o n im p ro v em en t i s s t a t e d i n te rm s e t e r s FWHM c h a n g e s f o r a l i n e s p r e a d b a r - p a t te r n r e s o lu tio n m ust be m u lti 1 .8 to g iv e co m p a rab le num bers.
The r e a s o n f o r im p ro v e d i n t r i n s i c s p a t i a l r e s o l u t i o n w i t h a th in n e r c r y s t a l i s n o t as o b v io u s as i t a p p e a rs a t f i r s t g lan c e. I t i s a p o p u la r m is c o n c e p tio n (18) t h a t th e im p ro v e m en t r e s u l t s fro m m o v in g t h e p h o t o m u l t i p l i e r s c l o s e r t o t h e o r ig in of th e l ig h t in a h -in c h c r y s ta l th an in a ^ -in c h c ry s ta l. I f t h i s w e re t h e c a s e , t h e sam e r e s u l t c o u ld b e a c h i e v e d by re d u c in g th e th ic k n e s s o f th e g la s s c o v e rin g th e c r y s t a l by ^i-in ch . T h is r e d u c t io n i n g la s s t h i c k n e s s , h o w e v er, d o e s n o t a ch iev e th e d e s ire d r e s u l t . M ore p r o b a b l e r e a s o n s a r e t h a t t h e red u ced c r y s t a l th ic k n e s s changes th e l i g h t d i s t r i b u t i o n a t th e p h o to m u ltip lie r s b e ca u se o f th e in d ex o f r e f r a c t i o n change a t th e c r y s t a l - t o - g l a s s i n t e r f a c e and b e c a u se o f a d d it i o n a l d e sig n free d o m i n p o s i t i o n i n g o p t i c a l m asks b e tw e e n th e c r y s t a l a sse m b ly and th e l i g h t p i p e . 2 .2 .3 .
Num ber o f
P h o to m u ltip lie rs
A p p ro x im a te ly 6 y e a r s ago th e A nger s c i n t i l l a t i o n cam era e v o l v e d fro m a d e v i c e w i t h 19 p h o t o m u l t i p l i e r s a n d a 1 0 - i n c h f ie ld - o f - v ie w to e i t h e r a h ig h r e s o l u ti o n ( 3 . 5 - 6 mm FWHM) 1 0 - i n c h f i e l d - o f - v i e w c a m e ra w i t h 37 p h o t o m u l t i p l i e r s ( 2 - i n c h d i a m e t e r ) o r a m e d i u m r e s o l u t i o n ( 5 - 8 mm FWHM) , 1 5 - i n c h f i e l d o f - v i e w c a m e r a w i t h 37 p h o t o m u l t i p l i e r s ( 3 - i n c h d i a m e t e r ) . W h ile t h e r e w e re a num ber o f v a r i a n t s , t h i s p a t t e r n p re d o m i n a te d w ith i n t r i n s i c r e s o l u t i o n s lo w ly c h a n g in g from th e h ig h e n d o f t h e r a n g e t o t h e low en d o v e r t h e y e a r s . As a l t e r n a t i v e m eth o d s f o r r e s o l u t i o n im p ro v e m en ts a r e b e in g e x h a u s te d ,
IAEA-SM-247/202
181
a new fa m ily o f l a r g e - f i e l d - o f - v i e w cam eras w ith 6 1 , 75 or 91 p h o t o m u lt ip lie r s h a s b een d e v e lo p e d . T h is b r u te f o r c e te c h n iq u e im proves r e s o l u t i o n by a p p r o x im a te ly 1 mm so th a t t h e s e la r g e f i e l d - o f - v i e w cam eras now h ave a p p r o x im a te ly th e same i n t r i n s i c r e s o l u t i o n a s 1 0 -in c h d ia m e te r cam eras.
2 . 2 . 4 . D i g i t a l P r o c e s s in g One o f th e prim e r e q u ir e m e n ts o f th e a n a lo g Anger p o s i t i o n in g e l e c t r o n i c s i s th e 'proper' sh ap e o f th e l i g h t r e sp o n se f u n c t io n . The l i g h t r e s p o n s e f u n c t io n i s th e l i g h t i n t e n s i t y m easured by a p h o t o m u lt ip lie r a s a f u n c t io n o f d is t a n c e from th e c e n t e r o f th e p h o t o m u lt ip lie r . Much e f f o r t h as been s p e n t on a n a l y t i c a p p ro a ch es to t h i s prob lem ( 2 0 , 2 1 ) , on m ethods to a c h ie v e th e p r o p e r r e s p o n s e by m a n ip u la tio n o f th e l i g h t d i s t r i b u t i o n ( 2 2 ) , and th e u s e o f n o n lin e a r p r e a m p lif ie r s ( 2 3 ) , and on d e l a y - l i n e m ethods ( 2 4 ) . In ea c h c a s e th e d e s ig n e r t r i e s t o o p t im iz e s p a t i a l r e s o l u t i o n w ith t h e c o n s t r a in t o f u n iform (o r n e a r ly u n ifo rm ) l i g h t c o l l e c t i o n f o r good en ergy r e s o l u t i o n and o f good l i n e a r i t y f o r a c c e p ta b le f lo o d f i e l d u n ifo r m it y . Through th e i n t r o d u c t io n o f d i g i t a l te c h n iq u e s fo r s p a t i a l l y v a r y in g en erg y d is c r im in a t io n and s p a t i a l d i s t o r t i o n rem oval a s d is c u s s e d in S e c t io n 2 . 1 , a m ajor con s t r a i n t h a s b een removed from s c i n t i l l a t i o n camera d e s ig n . F ig u r e 3 shows a s an in t e r m e d ia t e s t e p th e h ig h ly d i s t o r t e d but h i g h - r e s o l u t i o n im age o f th e a n a lo g p r o c e s s o r b e f o r e th e d ig it a l p ro cesso r c o r r e c ts fo r n o n lin e a r it ie s . T h is in c r e a s e d d e s ig n freedom can le a d to im proved r e s o l u t i o n compared w ith an a n a lo g p r o c e s s o r . 2 . 3 . C ount-R ate P erform ance The c o u n t - r a t e c a p a b i l i t y o f th e Anger s c i n t i l l a t i o n camera i s l i m i t e d by: 1) th e d eca y tim e o f N a l ( T l ) , 2) by th e camera e l e c t r o n i c s (m a in ly ch a r g e i n t e g r a t i o n , en e rg y norm al i z a t i o n and u n ifo r m ity c o r r e c t io n ) and 3) by d is p la y and d i g i t a l p r o c e s s in g d e v i c e s . The fu n d a m en ta l l i m i t a t i o n o f l i g h t decay o f th e s c i n t i l l a t i o n can be red u ced s i g n i f i c a n t l y th rou gh p u ls e s h o r t e n in g and co u n t l o s s e s due to e l e c t r o n i c d ead tim e can be m in im ized by a n a lo g b u f f e r i n g . U sin g t h e s e t e c h n iq u e s , th e c o u n t - r a t e c a p a b i l i t y o f th e Anger camera can be e x te n d e d s i g n i f i c a n t l y . 2 . 3 . 1 . A nalog B u f f e r in g One o f t h e problem s o f t e n e n c o u n te r e d in gamma camera sy ste m s i s t h a t th e 'fr o n t-e n d ' a n a lo g e l e c t r o n i c s can p r o c e s s
182
MUEHLLEHNER and COLSHER TABLE I I . DATA LOSSES FOR VARIOUS CAMERA AND COMPUTER CONFIGURATIONS
C am era s y s te m
20% D a t a
L oss
O u tp u t R a te *
U nio n
C a rb id e
C am era
A) W i t h o u t u n i f o r m i t y B) W i t h
u n ifo rm ity
c o rre c tio n
c o rre c tio n
C) R e c o r d e d i n DEC Gamma 1 1 co m p u ter w ith o u t u n if o r m ity c o rre c tio n S e a rle A) H i g h
8 1 К CPS 1 7 К CPS 1 7 К CPS
LFOV count ra te
mode
8 3 К CPS
B) R e c o r d e d i n DEC Gamma 1 1 co m p u ter cam era in h ig h c o u n t r a t e m ode
2 6 К CPS
C) R e c o r d e d i n DEC Gamma 1 1 c o m p u te r. C am era i n 10 m i c r o s e c . b u f f e r m ode
5 0 К CPS
^ M easu red w i t h p o i n t s o u r c e o f T c-99m w i t h o u t s c a t t e r i n g m a t e r i a l p r e s e n t , 20% e n e r g y w i n d o w .
p o s itio n in fo rm a tio n f a s t e r th an su b seq u e n t d i g i t a l p ro c e s s o rs . T h is i s show n i n T a b l e I I w h ic h l i s t s t h e o u t p u t r a t e a t w h ic h 20% o f t h e i n c o m i n g p o s i t i o n i n f o r m a t i o n i s l o s t d u e t o d e a d tim e and p i l e - u p . I f a c a m e r a i s o p e r a t e d a b o v e t h e 20% d a t a l o s s p o i n t , a n i n c r e a s e o f 1 0 К CPS i n t h e i n p u t r a t e r e s u l t s i n o n l y a 6 К CPS i n c r e a s e i n t h e o u t p u t r a t e . T hus, w h ile th e cam era can o p e r a te a t h ig h e r r a t e s any a d d i t i o n a l r a d i a t i o n b u r d e n o n t h e p a t i e n t i s o n l y 60% e f f e c t i v e . An a n a l o g b u f f e r u s u a l l y c o n s i s t s a n d -h o ld c i r c u i t s c a p a b le o f a c c e p tin g ra n d o m ly i n tim e and r e l e a s i n g th em a t su c c e e d in g c i r c u i t s w ith lo n g e r d e a d tim
o f a num ber o f sa m p le f a s t p u lse s o c cu rrin g r e g u la r i n te r v a l s to es. The c o u n t - r a t e
183
IAEA-SM-247/202
F IG .4 .
F ra c tio n o f counts lo st as a fu n c tio n o f in p u t data rate (R 0j and deadtim e ( t ) o f slower
o u tp u t circu its f o r no buffers, 2, 4 and in fin ite ly m any buffers. R elative c o u n t rate is R 0' r .
lo sse s
are
p lo tte d
in
F ig u re
4 as
a
fu n c tio n
of
th e
in p u t
ra te
(2 5). T h is p l o t assu m e s t h a t t h e sa m p le tim e o f t h e s a m p le a n d - h o ld i s n e g l i g i b l e com pared to th e d e a d tim e o f th e s u c c e e d in g c i r c u i t . As c a n b e s e e n , r e l a t i v e l y f e w b u f f e r s d r a m a t i c a l l y i n c r e a s e t h e c o u n t r a t e a t w h i c h 20% o f t h e c o u n t s a r e lo st. M ore t h a n 3 o r 4 b u f f e r s r e s u l t i n l i t t l e g a i n . T ab le I I i n c l u d e s c o u n t r a t e d a t a f o r a s y s t e m ( S e a r l e LFOV a n d DEC Gamma 1 1 ) w h i c h s h o w t h a t w h e n t h e c a m e r a i s o p e r a t e d i n t h e h i g h c o u n t - r a t e m ode a l a r g e f r a c t i o n o f t h e i n f o r m a t i o n i s lo st. T h r o u g h b u f f e r i n g t h e 20% d a t a l o s s r a t e c a n b e a l m o s t d o u b led . 2 .3 .2 .
P u lse
S h o rte n in g
A f t e r a gamma r a y i n t e r a c t s i n N a l ( T l ) , t h e r e s u l t i n g l i g h t i s e m i t t e d w i t h a d e c a y tim e o f 240 n s e c and r e q u i r e s 1 0 0 0 n s e c f o r 98% l i g h t c o l l e c t i o n . W ith o u t s p e c i a l p u ls e h a n d lin g t e c h n i q u e s , t h i s slo w d e ca y w i l l r e s u l t i n p u l s e p i le - u p a t h ig h d a ta r a te s and s e r i o u s l y l i m i t th e co u n t r a t e c a p a b i l i t y a c h i e v a b l e w ith N a l ( T I ) . The e l e c t r i c a l s i g n a l c o r re s p o n d in g to th e l i g h t e m is s io n from N a l( T l) c an be sh o rte n e d by o n ly u sin g a f r a c t io n o f th e l i g h t e m itte d . T h is te c h n iq u e h a s b e e n p r e v io u s ly d e s c rib e d (2 6 ,2 7 ) and i s b e in g a p p lie d to A nger cam eras (2 8 ,2 9 ) . S in c e o n ly a f r a c t io n o f t h e l i g h t i s u s e d , t h e r e w i l l b e som e r e s o l u t i o n l o s s . T ab le I I I su m m arizes c a l c u l a t e d c o u n t - r a t e c a p a b i l i t y and r e s o l u t i o n lo s s as a f u n c tio n o f th e i n t e g r a t i o n tim e . I t i s assum ed
184
.
MUEHLLEHNER and COLSHER
TABLE I I I .
E F F E C T OF PU LSE SHORTENING ON CAMERA PERFORMANCE
O u tp u t co u n t r a t e 20% d a t a l o s s
In te g ra tio n tim e
L ig h t E m itte d
R e so lu tio n FWHM
89000
CPS
1000 n se c
98%
4 . 0 mm
223000
CPS
400 n s e c
81%
4 .4
372000
CPS
240 n s e c
63%
5 . 0 mm
890000
CPS
100 n se c
34%
6 . 9 mm
mm
t h a t th e p u ls e i s s h o rte n e d to c o rre sp o n d to th e l i s t e d i n t e g r a tio n tim e . By u s i n g a v a r i a b l e i n t e g r a t i o n t i m e ( 2 9 ) , t h e r e s o l u t i o n l o s s c a n l i k e l y b e re d u c e d f u r t h e r . N o te t h a t w i t h an i n t e g r a t i o n t i m e a s s h o r t a s 2 4 0 n s e c , 63% o f t h e l i g h t i s s t i l l c o lle c te d . T h u s a n i n t r i n s i c r e s o l u t i o n o f 5 mm FWHM c a n b e a c h i e v e d i n a s y s t e m w h i c h h a s 4 mm FWHM r e s o l u t i o n w ith a lo n g i n te g r a t i o n tim e . W h ile p u l s e s h o r t e n i n g h a s n o t y e t b e e n im p le m e n ted i n c o m m e rc ia lly a v a i l a b l e cam eras to th e a u t h o r s ' k n o w le d g e, i t o f f e r s c o n s i d e r a b le p ro m is e s i n c e th e c o u n t - r a t e c a p a b i l i t y o f t h e A nger c a m e ra s c an b e im p ro v e d d r a m a tic a lly w ith o n ly a sm a ll lo s s o f i n t r i n s i c r e s o l u ti o n . 3.
SYSTEM PERFORMANCE
The p r e v io u s s e c t i o n d i s c u s s e d im p ro v e m e n ts i n d e t e c t o r p e rfo rm a n c e and c o n c e n tr a te d on im p ro v ed i n t r i n s i c r e s o l u t i o n . I n t h e q u e s t f o r e v e r b e t t e r i n t r i n s i c r e s o l u t i o n , i t i s im p o r t a n t t o rem em ber t h a t i t i s b u t o n e o f t h e f a c t o r s i n f l u e n c i n g sy ste m r e s o l u t i o n . The f o u r m a jo r c o n t r i b u t i n g p h y s i c a l f a c t o r s t o s y s te m r e s o l u t i o n a r e : 1) i n t r i n s i c r e s o l u t i o n , 2) e n e r g y r e s o l u t i o n , 3) c o l l i m a t o r r e s o l u t i o n and 4) s t a t i s t i c a l a c c u r a c y . In a d d i t i o n , p a t i e n t m o tio n and d i s p l a y q u a l i t y a f f e c t c l i n i c a l re so lu tio n . E n e r g y r e s o l u t i o n i s no w t y p i c a l l y b e l o w 11% j u s t i f y i n g a 15% e n e r g y w i n d o w f o r i m p r o v e d s c a t t e C o llim a to r p e rfo rm a n c e h a s r e c e n t l y b e e n im p ro v e d s e r ie s o f changes r e s u ltin g in c o llim a to rs w ith th
f o r T c-99m , r re je c tio n . th ro u g h a in n e r w a lls
( 0 . 2 mm f o r h i g h r e s o l u t i o n c o l l i m a t o r s ) a n d s m a l l e r h o l e s ( a p p r o x i m a t e l y 1 m m ). I n a d d i t i o n , m any c l i n i c a l s t u d i e s a llo w a c c u m u la tio n o f s u f f i c i e n t c o u n ts to ta k e a d v an tag e of th e im p ro v e d p e rfo rm a n c e w i t h o u t u n d u e r a d i a t i o n e x p o s u r e to
IAEA-SM-247/202
F IG .5.
185
The R o llo p h a nto m as im aged w ith a state o f the a rt system (A ) in 1 9 7 7 (fro m :
N uclear M edicine Physics, In stru m e n ta tio n , and Agents (F.D. R O L L O , E d.), M osby, St. L o u is (1977) 39 7 ; and (B) in 1 980 (courtesy o f D.E. Persyk, A p p lie d Research D e p t., Siemens Gammasonics).
th e p a tie n t. F ig u r e 5 su m m ariz es t h e s e im p ro v e m en ts p i c t o r i a l l y : t h e R o l lo p h a n to m i s im ag ed w i t h a s t a t e o f t h e a r t s y s te m i n 197 7 a n d a g a i n i n 1 9 8 0 . T h is p h a n to m w as c h o s e n s i n c e c o l d s p o t im a g in g w i t h low c o n t r a s t i s t h e m o st d e m a n d in g im a g in g s i t u a t i o n in n u c le a r m e d ic in e . T h i s im a g e ( F i g . 5B) r e p r e s e n t s a s y s t e m r e s o l u t i o n o f l e s s t h a n 7 mm. A t t h i s p o i n t , i n t r i n s i c r e s o l u t i o n i s n o t a d o m in a n t c o n t r i b u t i n g f a c t o r to sy ste m p e rfo rm a n c e . In d eed , even i f i n t r i n s i c r e s o l u t i o n w ere to be re d u c e d to z e r o , sy s te m r e s o l u t i o n w o u l d i m p r o v e b y l e s s t h a n 20% ( f r o m 7 . 6 mm t o 6 . 4 mm FWHM) f o r a h i g h r e s o l u t i o n c o l l i m a t o r a t 1 0 cm f r o m t h e f a c e o f th e cam era. Thus sy ste m p e rfo rm a n c e i s a p p ro a c h in g a l i m i t s e t by th e c o llim a to r . S in c e th e p ro c e ss o f c o llim a tio n is r e a s o n a b ly w e l l u n d e r s to o d , im p ro v e m en ts i n c o l l i m a t i o n a r e a c h ie v e d o n ly th ro u g h 'f in e tu n in g E v e n a 20% i m p r o v e m e n t in r e s o l u ti o n w ith no s a c r i f i c e in s e n s i t i v i t y i s d i f f i c u l t to ach iev e . O rgan m o tio n d ue to r e s p i r a t i o n i n p r o c e d u r e s su c h as l i v e r s t u d i e s i s t y p i c a l l y 1 cm, s o m o re c a n b e g a i n e d by em p lo y in g m o tio n c o r r e c t i o n d e v ic e s (3 0 ,3 1 ) th a n by f u r t h e r im p ro v e m en ts i n i n t r i n s i c r e s o l u t i o n . A lso th e p r o p e r d i s p l a y o f i m a g e s b e c o m e s i n c r e a s i n g l y i m p o r t a n t a s im a g e q u a l i t y im pro v es. T h is i s p a r t i c u l a r l y t r u e o f d i g i t a l sy s te m s . The d i s p la y sy s te m s h o u ld n o t r e s u l t in any p e r c e p ti b l e l o s s o f r e s o lu tio n . F o r m odern l a r g e - f i e l d - o f - v i e w c am eras th e sy ste m s h o u l d b e c a p a b le o f a c q u i r i n g 256 X 256 im a g e s . F u rth e r d is c u ssio n o f t h is p o in t is beyond th e scope o f t h is p a p e r.
186
MUEHLLEHNER and COLSHER
4. S I N G L E P H O T O N T O M O G R A P H Y E m is s io n c o m p u ted to m o g ra p h y h a s b e e n a tt e m p t e d w i t h v a r y in g s u c c e s s i n n u c l e a r m e d ic in e an d w as e v a l u a t e d c l i n i c a l l y a s e a r l y a s 1973 ( 3 2 ) . H ow ever, i t w as o n ly a f t e r th e d e v e lo p m en t o f x - r a y com p u ted to m o g ra p h y t h a t i n t e r e s t i n t h i s m o d a l i t y in c re a se d in n u c le a r m e d ic in e . A t t h i s p o i n t tw o a p p r o a c h e s a r e b e in g e v a l u a t e d : 1) l o n g i t u d i n a l to m o g ra p h y i n t h e fo rm o f 7 - p in h o le c o ll i m a t o r s ( 3 3 ) , s l a n t - h o l e c o ll i m a t o r s (3 4 ,3 5 ) and s c a n n i n g c a m e ra s y s t e m s (3 6 ) an d 2) t r a n s v e r s e to m o g ra p h y u s i n g e ith e r ro ta tin g sy ste m s (4 0 ).
cam eras
(3 7 ,3 8 ,3 9 )
or m u ltic ry s ta l
sc a n n in g
I n r e c e n t y e a r s l o n g i t u d i n a l com p u ted to m o g ra p h y w it h t h e 7 -p in h o le c o llim a to r h as found w id e sp re a d a p p li c a t io n in c a r d ia c im a g in g (3 3 ). I n a n a t t e m p t t o o v e rc o m e som e o f i t s p ro b le m s, su ch as v a r ia b le s l i c e th ic k n e s s and d e p th r e s o l u ti o n , v a rio u s c o n fig u ra tio n s o f s la n t- h o le c o llim a to rs a re b e in g d e v elo p ed and e v a lu a te d ( 3 4 ,3 5 ). U n fo rtu n a te ly , c li n i c a l e v a l u a t i o n s a r e sh o w in g l i t t l e im p ro v e m e n t i n d i a g n o s t i c a c c u r a c y (4 1 ,4 2 ). M any o f t h e p r o b l e m s w i t h t h e s e t e c h n i q u e s a r e t h e r e s u l t o f i n c o m p l e t e a n g u l a r s a m p li n g w h ic h c a n l e a d t o im a g e a r t i f a c t s (43) and p o o r d e p th r e s o l u t i o n . U ltim a te ly , lo n g i t u d i n a l tom ography i s l i k e l y to f i n d a p p l i c a t i o n s in v e ry s p e c i f i c c l i n i c a l s i t u a t i o n s w here im p ro v ed l o c a l i z a t i o n and l e s i o n c h a r a c t e r i z a t i o n w i l l y i e l d som e a d d i t i o n a l u s e f u l in fo rm a tio n . E a rly t r a n s v e r s e s e c ti o n cam eras w ere u s u a lly s i n g l e s l i c e d e v ic e s ( 4 4 ,4 5 ) ; m ore r e c e n t a p p r o a c h e s h a v e m u l t i s lic e c a p a b ility . R o ta tin g cam era sy ste m s ( 3 6 ,3 7 ,3 8 ) p ro v id e t h e c a p a b i l i t y t o im a g e a n o r g a n w i t h a p p r o x i m a t e l y e q u a l r e s o l u t i o n i n a l l t h r e e d i r e c t i o n s t h r o u g h o u t t h e im ag ed v o lu m e w ith good q u a n ti t a t i v e a c c u ra c y (4 6 ). T h is i s p a r t i c u l a r l y tru e in th e b ra in . The m u l t i c r y s t a l s c a n n e r d e v e lo p e d by S to k e ly e t a l . (40) e m p h a s iz e s h ig h c o u n t - r a t e c a p a b i l i t y f o r q u a n t i t a t i o n o f f a s t d y n am ic f u n c t i o n s t u d i e s . The a d v a n t a g e s and d is a d v a n ta g e s a lo n g w ith th e te c h n iq u e s and p ro b lem s h av e b e en d is c u s s e d in r e c e n t re v ie w p a p e rs on th e s u b j e c t (4 3 ,4 7 ) and,
th e re fo re ,
o n ly
4 .1 .
B e n efit
C o m p u te d T o m o g ra p h y
of
sp e c ific
a sp e cts
w ill
be
d isc u sse d
h ere.
A q u e s t i o n w h ic h n e e d s t o b e a d d r e s s e d i s : W hat c a n b e a c h i e v e d by a p p ly i n g c o m p u te d to m o g r a p h ic t e c h n i q u e s t o n u c l e a r m e d ic in e and w hat i s th e chance o f a c h ie v in g t h i s g o a l? As a p a r a l l e l i n r a d i o l o g y , c o m p u ted to m o g ra p h y a llo w e d t h e v i s u a l i z a t i o n o f v e r y lo w c o n t r a s t s t r u c t u r e s n o r m a l l y n o t s e e n on x - r a y im ag es. In a d d itio n , th e p re s e n ta tio n of in fo rm a tio n in
187
IAEA-SM-247/202
t r a n s v e r s e s e c ti o n a llo w s b e t t e r v i s u a l i z a t i o n o f a n a to m ic s tr u c tu r e s and a v o id s c o n fu sio n due to o v e rly in g and u n d e rly in g s tru c tu re s. I t i s a p p a re n t th a t th e l a t t e r b e n e f it can r e a d ily be tr a n s f e r r e d t o n u c l e a r m e d i c i n e : m uch o f t h e d i a g n o s t i c i n f o r m a t i o n i s d e r iv e d from p a t t e r n r e c o g n i t i o n o f n o rm a l v e r s u s ab n o rm a l s t r u c t u r e s and d epends on v i s u a l r e c o g n itio n o f s l i g h t d i f f e r ences in e ith e r shape o r in te n s ity . A tra n sv e rse p re se n ta tio n a d d s new i n f o r m a t i o n b u t m ak e s i t m o re d i f f i c u l t t o r e c o g n i z e s l i c e - t o - s l i c e v a r ia tio n s in in te n s ity and sh ap e. T h is i s sh o w n i n F i g u r e 6 w h ic h d e p i c t s l o n g i t u d i n a l a n d t r a n s v e r s e c u ts th ro u g h a l i v e r . W h ile th e t r a n s v e r s e s e c t i o n s a llo w good v is u a liz a tio n of in te r n a l su rfa c e s , th e lo n g itu d in a l se c tio n s a llo w b e t t e r a p p r e c ia tio n o f th e sh ap e in th e a x ia l d i r e c ti o n . Thus th e g o a l m u st b e a th r e e - d im e n s io n a l r e p r e s e n t a t i o n o f th e o b j e c t w ith th e a b i l i t y to v iew s l i c e s i n any d i r e c t i o n , h o p e f u lly w ith eq u al s p a t ia l r e s o lu tio n in a l l 3 ax es. The p o t e n t i a l b e n e f i t o f c o n t r a s t im p ro v e m en t m u st b e e x a m in e d m ore c l o s e l y , e s p e c i a l l y s i n c e im a g e r e f o c u s i n g t e c h n iq u e s w h ich h a v e b e e n v e ry v a lu a b l e i n o t h e r f i e l d s h a v e n o t p r o v e n t o b e o f m uch b e n e f i t i n n u c l e a r m e d i c in e d u e t o t h e l im i te d s t a t i s t i c a l a c c u ra c y o f th e d a ta (48) . B ecause o f th e f i l t e r e d b a c k p ro je c tio n p ro c e s s , th e e x p e c te d u n c e r t a i n t y o f th e d a ta i n a p i x e l i s n o t sim p ly r e l a t e d to th e num ber o f c o u n ts a s one e x p e c ts from P o is s o n sta tistic s , but b e e n show n (4 9 ) g iv en
a lso th a t
to th e num ber o f d a ta p o i n ts . th e u n c e r ta in ty fo r a u n ifo rm
I t has d isc is
by 120(no. o f r e s o l u ti o n c e l l s ) rm s% = ------------------------------------------------------- ---------------------------(to ta l no. e v e n ts )2
Thus a s l i c e c o n ta in in g 10^ t o t a l e v e n ts and h a v in g 400 reso l u t i o n c e l l s ( 2 0 cm c i r c u l a r o b j e c t w i t h 1 . 0 c m d i a m e t e r r e s o l u t i o n c e l l s ) h a s a n u n c e r t a i n t y o f 10% p e r p i x e l . A c o m p le te s t u d y o f a n o r g a n w i t h 10 s l i c e s r e q u i r e s t h e a c c u m u l a ti o n o f 10^ c o u n t s , s i g im a g e . On t h e v e rse se c tio n s f a c to r of 2 .8 . t io n a l c o u n ts . s t a ti s t i c s are p a tie n t and by
n i f i c a n t l y m ore t h a n i s n e e d e d f o r a p r o j e c t i o n o t h e r h a n d , i t h a s b e e n show n (5 0 ) t h a t t r a n s h a v e h i g h e r c o n t r a s t th a n p r o j e c t i o n im ages by a To o b t a i n th is g a in in c o n tr a s t re q u ir e s a d d i H o w e v e r, i n m any c l i n i c a l s t u d i e s t h e a v a i l a b l e l i m i t e d b y t h e m axim um r a d i a t i o n d o s e t o t h e t h e m axim um p r a c t i c a l i m a g i n g t i m e , s u g g e s t i n g
t h a t ECT s y s t e m s e ffe c tiv e .
sh o u ld
have
in c re a sed
se n sitiv ity
to
be
MUEHLLEHNER and COLSHER
189
IAEA-SM-247/202
ч
F IG . 6.
Щ
i
(A ) Transverse liv e r scan using 99Tcm -labelled su lp h u r c o llo id (4 m C i). T ota l
counts = 9 . 6 - 106, 16 m inutes scan tim e, slice thickness 1.3 cm ; (B) same data as A b u t reorganized to generate lo n g itu d in a l section images (fro m R e f. [3 7 ]).
4 .2 .Q u a n tita tiv e
A ccuracy
W h ile m uch h a s b e e n s a i d a b o u t t h e p r o m is e o f q u a n t i t a t i o n w i t h p o s i t r o n e m i s s io n c o m p u te d to m o g ra p h y ( 5 1 ) , i t i s o n ly r e c e n t l y t h a t som e o f t h e p r o b l e m s a r e b e i n g d i s c u s s e d i n t h e l i t e r a t u r e (5 2 ,5 3 ). Q u a n tita tiv e sin g le g rap h y i s a to u g h e r p ro b lem a n a l y t i c a l l y m atio n s and i t e r a t i v e a p p ro a c h e s. T hus, g raphy lo s e s a c c u ra c y in th e p r a c t i c a l a
p h o t o n c o m p u te d to m o and depends on a p p ro x i w h i l e p o s i t r o n tom o p p lic a tio n of a
190
MUEHLLEHNER and COLSHER
T A B L E IV. R E S O L U T I O N I N R E C O N S T R U C T E D I M A G E
D ista n ce of
from
c en ter
H ig h
ro ta tio n
re so lu tio n
p a ra lle l
h o le
c o llim a to r (fro m 37)
1 6 mm
0 cm
S p e c ia l p a ra lle l
ECT h o le
c o llim a to r (fro m 57)
Fan-beam c o llim a to r (from
37)
1 7 . 1 ram
1 2 mm
2 .5
cm
1 5 . 4 mm
1 6 .6
mm
1 2 mm
5 .0
cm
1 4 .7
mm
1 5 .9
mm
1 2 mm
7 .5
cm
1 4 . 1 mm
1 5 .1
mm
1 2 mm
A l l v a l u e s a r e m e a s u r e d FWHM v a l u e s i n p r e s e n c e o f s c a t t e r m a t e r i a l w i t h r a d i u s o f r o t a t i o n 1 5 - 2 0 cm. In trin sic r e s o l u t i o n o f c a m e r a s u s e d w a s 7 - 8 mm FWHM.
's o l v a b l e ' p ro b le m , s i n g l e p h o to n to m o g ra p h y g a in s a c c u r a c y i n t h e p r o c e s s o f a p p r o x im a tin g an 'u n s o l v a b le ' p ro b le m . The m a jo r s o u r c e s o f i n a c c u r a c i e s a r e : 1) ra n g e o f a n g u l a r s a m p li n g , 2) u n i f o r m i t y o f s p a t i a l r e s o l u t i o n , 3) u n i f o r m i t y o f s e n s i t i v i t y , 4) d e t e c t o r i m p e r f e c t i o n s an d 5) a t t e n u a t i o n c o r r e c t i o n . Each o f th e s e w i l l b e d is c u s s e d in tu r n . In a d d itio n , s c a tte re d r a d i a t i o n from th e p a t i e n t i n f lu e n c e s q u a n t i t a t i v e a c c u r a c y ; w h ile i t s e f f e c t s a r e w e ll u n d e rs to o d in p la n a r im a g in g , i t s in flu e n c e in tra n s v e rs e re c o n s tru c tio n s has n o t been stu d ie d in d e ta il. 4 .2 .1 .
A n g u la r
S a m p lin g
L o n g itu d in a l to m o g ra p h ic d e v ic e s su ch as th e 7 - p in h o le o r v a r io u s c o n f ig u r a ti o n s o f s l a n t - h o l e c o ll i m a t o r s v iew th e o b j e c t from a l im i te d a n g u la r r a n g e . T h is does n o t a llo w a c c u ra te q u a n tita tiv e d e te rm in a tio n o f tr a c e r c o n c e n tra tio n s in an unknown d i s t r i b u t i o n . L im ite d q u a n t i t a t i o n i s p o s s ib le i f t h e o b j e c t i s s m a ll (54) o r i f th e o b j e c t i v e s im p ly i s th e d e t e r m i n a t i o n o f v o lu m e o f i n f a r c t e d t i s s u e . In th e subsequent s e c ti o n s o n ly t r a n s v e r s e to m o g ra p h ic sy ste m s su ch as r o t a t i n g cam era sy ste m s (3 7 ,3 8 ,3 9 ) o r m u l t i d e t e c t o r r o t a t i n g sy ste m s (40) a re c o n sid ere d fo r q u a n tita tio n . W ith t h e s e l a t t e r , th e num ber o f a n g u la r sa m p le s i s t y p i c a l l y c h o se n i n a c c o rd a n c e w ith th e sa m p lin g th e o ry to a v o id undue lo s s o f r e s o l u ti o n o r a l i a s i n g a r tif a c ts (5 5 ). T h is i m p l i e s t h a t f o r a 20-cm o b j e c t t o b e i m a g e d w i t h a s p a t i a l r e s o l u t i o n o f 1 0 mm a t o t a l o f 1 0 0 p r o je c tio n s are needed.
IAEA-SM-247/202 4 .2 .2 .
U n ifo rm ity
of
S p a tia l
191
R e so lu tio n
D e te c to r sy ste m s u s in g p a r a l l e l o r n e a r l y p a r a l l e l h o le c o lli m a t io n s u f f e r from s i g n i f i c a n t lo s s o f r e s o l u ti o n a s a f u n c tio n o f d i s t a n c e from th e c o l l i m a t o r . U n fo rtu n a te ly , a u th o r s p r a i s i n g th e a d v a n ta g e s o f p o s i t r o n im a g in g o f t e n f a i l to s e p a r a t e a t t e n u a t i o n p ro b le m s fro m r e s o l u t i o n p ro b le m s (56) and th u s g iv e an e rro n e o u s im p re s s io n . T a b l e IV sh o w s t h e v a r i a t i o n o f s y s te m r e s o l u t i o n i n t h e r e c o n s t r u c t e d im a g e f o r a n A n g e r s c i n t i l l a t i o n c a m e ra w i t h 1) a p a r a l l e l h o l e c o l l i m a t o r , 2) a c o n v e r g i n g c o l l i m a t o r , 3) a p a r a l l e l h o l e c o l l i m a t o r e s p e c i a l l y d e s i g n e d f o r ECT ( 5 7 ) . U n ifo rm ity o f r e s o l u ti o n is im p ro v e d s i g n i f i c a n t l y by t a k i n g t h e g e o m e tr i c m ean o f o p p o s in g ( 1 8 0 ° ) v i e w s a n d i n d e e d v a r i e s b y o n l y 20% i n t h e r e c o n s t r u c t e d im ag es ( 5 0 ) . T h is v a r i a t i o n i s co m p a rab le to t h a t e n c o u n te re d in
p o sitro n
4 .2 .3 .
to m o g ra p h ic
U n ifo rm ity
of
sy ste m s
(5 8 ,5 9 ).
S e n sitiv ity
M o s t ECT s y s t e m s ( s i n g l e p h o t o n t y p i c a l l y em p lo y n o r m a l i z a t i o n o r u n c e d u re s to c o m p e n sa te f o r v a r i a t i o n s slic e . T h is i s e s s e n t i a l s i n c e e v en
as w e ll as p o s itro n ) ifo rm ity c o rre c tio n p ro of s e n s itiv ity w ith in a sm a ll v a r ia ti o n s in s e n s i
t i v i t y g e n e ra te c ir c u la r a r t i f a c t s (6 0 ). A p ro b le m w h ich i s f r e q u e n tly o v e rlo o k e d i s s l i c e - t o - s l i c e v a r i a t i o n s . S in g le s l i c e sy s te m s h a v e a c o u n tin g p r o f i l e w h ich te n d s to be G a u s s ia n in shape a c ro ss th e th ic k n e ss o f th e s l i c e ; 's lic e th ic k n e ss' is th e n th e fu ll-w id th -a t-h a lf-m a x im u m o f th e c o u n tin g p r o f i l e . M u l t is li c e sy ste m s c o n s i s t in g o f s e v e r a l in d iv i d u a l s l i c e s w i l l th e n have a c o u n tin g p r o f i l e o f o v e rla p p in g G a u ssia n -sh a p e d c u rv e s o f v a ry in g a m p litu d e (6 1 ). Such sy ste m s can o n ly g iv e q u a n t i t a t i v e in fo rm a tio n on th e a ssu m p tio n t h a t th e a c t i v i t y d i s t r i b u t i o n i s e s s e n t i a l l y u n ifo rm a c ro s s a s l i c e . I f , for e x a m p le , a p o i n t s o u r c e w h ich i s s m a ll com pared t o t h e s l i c e t h i c k n e s s i s p l a c e d a n y w h e re i n t h e s e n s i t i v e v o lu m e , i t w ou ld be d i f f i c u l t i f n o t im p o s sib le to d e te rm in e i t s a c t i v i t y . T h is p r o b l e m i s a v o i d e d i n r o t a t i n g c a m e r a ECT s y s t e m s , w h e r e t h e d e te c to r s e n s i t i v i t y i s u n ifo rm a t r i g h t a n g le s to th e s l i c e o rie n ta tio n . Thus th e c o u n ts from a p o i n t s o u rc e w i l l b e r e c o r d e d i n o n e s l i c e o r w i l l b e d i s t r i b u t e d am ong tw o s l i c e s s u c h t h a t t h e sum o f tw o s l i c e s w i l l a c c u r a t e l y r e p r e s e n t t h e a c t i v ity in th e p o in t so u rc e . H o w ev er, p a r t i a l v o lu m e e f f e c t s (6 2 ) w ill g e n e ra te v is u a l a r t i f a c t s u n le ss th in s lic e s a re reco n stru c te d . 4 .2 .4 .
th e
D e te c to r
Im p erfe ctio n s
S y ste m s c o n s i s t i n g o f m u l t i p l e s m a ll d e t e c t o r s p ro b lem t h a t one d e te c to r i s n e c e s s a ry f o r each
s u f f e r from sim u lta n e o u s
192
MUEHLLEHNER and COLSHER
r e s o lu tio n e le m e n t. The A nger s c i n t i l l a t i o n cam era h a s s e v e r a l th o u sa n d r e s o l u t i o n e le m e n ts w ith o n ly a s in g le la r g e c r y s t a l a n d t y p i c a l l y 37 p h o t o m u l t i p l i e r s . H ow ever, s l i g h t i m p e r f e c t io n s in i t s l i n e a r i t y and c o o rd in a te a d ju stm e n t can cau se n o tic e a b le p r o b l e m s i n ECT ( 4 7 , 6 3 ) . M any o f t h e s e p r o b l e m s c a n b e a v o i d e d th ro u g h s p a t i a l d i s t o r t i o n rem o v al as d e s c rib e d above and c a r e f u l a d ju s tm e n t and m o n ito rin g o f th e e le c t r o n i c c e n te r o r r o t a tio n (63). 4 .2 .5 .
A tte n u a tio n
C o rre c tio n
I n i t i a l l y r e c o n s t r u c ti o n s w ere done w ith s im p le a t t e n u a t i o n c o r r e c ti o n schem es t h a t in v o lv e d u s in g th e a r it h m e t ic o r g eo m e t r i c m ean o f o p p o s in g v ie w s ( 4 3 ) . F o r m ore a c c u r a t e a t t e n u a t i o n c o r r e c t i o n , tw o b a s i c a p p r o a c h e s m ay b e u s e d . T ra n sm issio n d a t a m ay b e t a k e n a n d u s e d th e shape of th e o b je c t o r t iv e ly , th e shape can be o a tte n u a tio n c o r re c tio n can a tio n fa c to rs .
w ith o u t any a tte n u a tio n b ta in e d from be perfo rm ed
a ssu m p tio n s re g a rd in g c o e ffic ie n t. A ltern a a n a to m ic a l la n d m a rk s and w ith e stim a te d a tte n u
T ra n sm issio n te c h n iq u e s f o r a tt e n u a t io n c o r r e c tio n s u f f e r f ro m tw o m a j o r p r o b l e m s . The tr a n s m is s io n d a ta m ust u s u a l ly be ta k e n b e f o re th e e m iss io n sc a n to a v o id in te r f e r e n c e . O fte n t h e e m i s s i o n s c a n c a n n o t b e p e r f o r m e d u n t i l a b o u t 30 m in u te s a f t e r in je c ti o n in o rd e r to o b ta in good ta rg e t-to -b a c k g ro u n d ra tio s. The se q u e n c e o f t r a n s m i s s io n s c a n , i n j e c t i o n , tim e d e la y , and e m is s io n sc a n im p lie s t h a t th e p a t i e n t m ust re m a in m o tio n le s s f o r p e rio d s o f a p p ro x im a te ly one h o u r. T h is i s v e ry d i f f i c u l t even fo r a c o o p e ra tiv e in d iv id u a l. The re g is tra tio n b e tw e e n t h e t r a n s m i s s io n and e m is s io n d a t a can e a s i l y b e i n e r r o r by 5 m illim e te r s o r m ore. T h e s e p r o b l e m s m ay b e a v o i d e d by p e rfo rm in g th e tr a n s m is s io n sc a n su b se q u e n t to th e e m issio n stu d y . T h is i s p o s s ib le w h ere th e r a d io is o to p e i s q u ic k ly c le a r e d from th e o rg a n o f i n t e r e s t . I t can a ls o be done by u sin g a h ig h e r energy tra n s m is s io n so u rc e . In a d d itio n , th e tra n s m iss io n d a ta a re s t a t i s t i c a l l y n o isy . S m o o th in g t h e d a t a re d u c e s t h i s p ro b lem b u t adds u n c e r ta in ty a b o u t th e lo c a tio n o f th e b o u n d a rie s.
of
The a l t e r n a t i v e m eth o d o f a t t e n u a t i o n c o r r e c t i o n u s e s one s e v e r a l te c h n iq u e s to f in d th e body c o n to u r and th e n u se s
e i t h e r th e a ssu m p tio n o f c o n s ta n t a t t e n u a t i o n c o e f f i c i e n t w ith in t h e body o r assu m es known a t t e n u a t i o n c o e f f i c i e n t s w i t h i n v a rio u s re g io n s . The body c o n to u r c an b e o b ta in e d by e i t h e r f i t t in g an e l l i p s e a ro u n d th e e m is s io n sc a n (53) , u s in g a tr a n s m is s io n sc a n f o r o u t li n e d e te r m in a tio n o n ly ( 3 7 ,6 4 ) , o r u s in g a s e c o n d e n e r g y w in d o w t o l o o k a t s c a t t e r e d r a d i a t i o n ( 3 7 ) . T hese m e th o d s a l l h a v e o b v io u s e r r o r s a s s o c i a t e d w i t h th em w h ic h a r e
193
IAEA-SM-247/202
now b e i n g s t u d i e d ( 4 6 ) a n d a r e t o l e r a b l e i f t h e o b j e c t i v e i s q u a n t i t a t i o n w i t h 10-20% o v e r a l l a c c u r a c y . The a b o v e -m e n tio n e d p r o b l e m s a r e g e n e r a l l y c om m on t o b o t h p o s i t r o n a n d s i n g l e p h o t o n ECT. Once t h e a t t e n u a t i o n c o e f f i c i e n t i s assu m ed o r m e a su re d and body c o n to u r i s known, th e e f f e c t s o f a tt e n u a t io n can be c o rre c te d . A lth o u g h th e m eth o d s a r e n o t a s s im p le a s in p o s i t r o n im a g in g (56) , v a r io u s te c h n iq u e s h a v e b e e n d e v e lo p e d . F or c o n s ta n t a tt e n u a t io n , th e m a th e m a tic a l i n t r a c t a b i l i t y h as been o vercom e, and d i r e c t c o n v o lu tio n a lg o rith m s h av e been d e riv e d (6 5 ,6 6 ,6 7 ). F o r t h e m ore g e n e r a l c a s e o f v a r y in g a t t e n u a tio n , i t e r a t i v e p ro c e d u re s have been d ev elo p ed (3 8 ,6 8 ). 5.
SUMMARY
In th e l a s t s e v e r a l y e a r s th e p e rfo rm a n ce o f th e Anger s c i n t i l l a t i o n c a m e r a h a s im p ro v e d s i g n i f i c a n t l y a n d i s co m in g c lo s e r to th e 'i d e a l d e te c to r ' p a r t i c u l a r l y in th e a re a s o f in tr in s ic s p a tia l re s o lu tio n , fie ld -flo o d u n ifo rm ity , s e n s iti v ity and c o u n t-ra te c a p a b ility . L i m i t a t i o n s im p o sed by c o l l i m a t o r c h a r a c t e r i s t i c s a r e b e c o m i n g d o m i n a n t w i t h o n l y s m a l l (20% ) im p ro v e m en ts i n s i g h t . P ro g ress
in
n u c le a r m e d ic in e
sin g le
p h o to n
im a g in g
in
th e
f u t u r e i s l i k e l y t o come i n d a t a a n a l y s i s , p a r t i c u l a r l y f o r fu n c tio n s tu d ie s d is c u s s e d e lse w h e re in th e s e p ro c e e d in g s as w e l l a s th r o u g h i n c r e a s e d u s e o f e m i s s io n c o m p u ted to m o g ra p h y u s in g r o t a t i n g cam era sy s te m s . T hese have th e ad v an tag e o f p e r m ittin g c u r r e n t p ro c e d u re s and - a t m odest in c re m e n ta l c o s t a llo w a c q u i s i t io n o f th r e e - d im e n s io n a l in fo rm a tio n f o r p re s e n t a t i o n o f t r a n s v e r s e and l o n g it u d i n a l s e c ti o n s . T h is im ag in g mode g i v e s a d d i t i o n a l c l i n i c a l i n f o r m a t i o n i n s e l e c t d i a g n o s t i c pro ced u res. In a d d itio n , th e s e in s tru m e n ts a re c a p a b le of p ro v id in g q u a n t i t a t i v e in fo rm a tio n and a re th u s u s e f u l r e s e a rc h to o ls.
REFERENCES [1] MCKEIGHEN, R.E., “A review of gamma camera technology for medical imaging” , Imaging for Medicine 1 (NUDELMAN, S., Ed.), Plenum Press, New York (1980). [2] TODD-POKROPEK, A.E., ERBSMANN, F., SOUSSALINE, F., “The non-uniformity of imaging devices and its impact in quantitative studies”, Medical Radionuclide Imaging (Proc. Symp. Los Angeles, 1976) 1, IAEA, Vienna (1977) 67. [3] GRAHAM, L.S., OLCH, A., USZLER, J.M., et al., Automatic field uniformity corrections in scintillation cameras, SPIE Semin. Proc. 152 (1978) 127. [4] MORRISON, L.M., BRUNO, F.P., MAUDERLI, W., Sources of gamma camera image inequalities, J. Nucl. Med. 12 (1971) 785.
194
MUEHLLEHNER and COLSHER
[5] PADIKAL, T.N., ASHARE, A.B., KEREIAKES, J.G., Field flood uniformity correction: benefits or pitfalls? J. Nucl. Med. 17 (1976) 653. [6 ] JANSSON, L.G., PARKER, R.P., Pittfalls in gamma camera field uniformity correction, Br. J. Radiol. 48(1975) 408. [7] WICKS, R., BLAU, M., Effect of spatial distortion on Anger camera field-uniformity corrections (concise communication), J. Nucl. Med. 20 (1979) 252. [8 ] STEIDLEY, J.W., KEARNS, D.S., HOFFER, P.B., Uniformity correction with the Micro-Z processor, J. Nucl. Med. 19 (1978) 712. [9] MUEHLLEHNER, G., COLSHER, J.G., STOUB, E.W., Correction for nonuniformity in scintillation cameras through removal of spatial distortion, J. Nucl. Med. 21 (1980) 771. [10] KNOLL, G.F., BENNETT, M.C., KORAL, K.F., et al., Removal of gamma camera nonlinearity and nonuniformities through realtime signal processing, Vlth International Conference on Information Processing in Medical Imaging, Paris, 1979. [11] ARSENEAU, R., private communication. [12] WOLFF, J.R., “Calibration methods for scintillation camera systems” , Quantitative Organ Visualization in Nuclear Medicine, University of Miami Press, Coral Gables, Florida (1971) 229. [13] RCA “teacup” design used in C31061 andC31071. [14] SRC LABORATORIES “gamma target” design used in SRC GT52B01 and GT75B01 photomultipliers. [15] PERSYK, D.E., MOI, T.E., State of the art photomultipliers for Anger cameras, IEEE Trans. Nucl. Sei. NS-26 (1979) 615. [16] SANO, R.M., TINKEL, J.B., LA VALLEE, C.A., et al., Consequences of crystal thickness reduction on gamma camera resolution and sensitivity, J. Nucl. Med. 19 (1978) 712. [17] CHAPMAN, D., NEWCOMER, K., BERMAN, D., et al., Half-inch vs. quarter-inch Anger camera technology: resolution and sensitivity differences at low photopeak energies, J. Nucl. Med. 20(1979) 610. [18] ROYAL, H.D., BROWN, P.H., CLAUNCH, B.C., Effects of a reduction in crystal thickness on Anger camera performance, J. Nucl. Med. 20 (1979) 977. [19] MUEHLLEHNER, G., Effect of crystal thickness on scintillation camera performance, J. Nucl. Med. 20(1979) 992. [20] BAKER, R.G., SCRINGER, J.W., An investigation of the parameters in scintillation camera design, Phys. Med. Biol. 12 (1967) 51. [21 ] SVEDBERG, J.B., Image quality of a gamma camera system, Phys. Med. Biol. 13 (1968) 597. [22] MARTONE, R.J., GOLDMAN, S.C., HEATON, C.C., Scintillation camera with light diffusion system, U.S. Patent No. 3784819 (1974). [23] KULBERG, G.H., VAN DIJK, N., MUEHLLEHNER, G., Improved resolution of the Anger scintillation camera through the use of threshold preamplifiers, J. Nucl. Med. 13 (1972) 169. [24] HIRAMOTO, T., TANAKA, E., NOHARA, N., A scintillation camera based on delayline time conversion, J. Nucl. Med. 12 (1971) 160. [25] MOLL, G., Nuclear electronics: scaler or buffer storage for higher resolution in pulse counting equipments in radiation measurement, Kerntechnik 8 (1966) 404. [26] AMSEL, G., BOSSHARD, R., ZAJDE, C., Shortening of detector signals with passive filters for pile-up reduction, Nucl. Instrum. Methods 71 (1969) 1. [27] BRASSARD, C., Fast counting with Nal spectrometers, Nucl. Instrum. Methods 94 (1971) 301. [28] MUEHLLEHNER, G., BUCHIN, M.P., DUDEK, J.H., Performance parameters of a positron imaging camera, IEEE Trans. Nucl. Sei. ÑS-23 (1976) 528.
IAEA-SM-247/202
195
TANAKA, E., NOHARA, N., MURAYAMA, H., Variable sampling-tirae technique for improving countrate performance of scintillation detectors, Nucl. Instrum. Methods 158(1979) 459. TURNER, D.A., FORDHAM, E.W., ALI, A., et al., Motion corrected hepatic scintigraphy: an objective clinical evaluation, J. Nucl. Med. 19 (1978) 142. MCKEIGHEN, R.E., Improved means of correcting motion blurring in scintigraphic images, Phys. Med. Biol. 24 (1979) 353. KUHL, D.E., SANDERS, T.P., Characterizing brain lesions with use of transverse section scanning, Radiology 98 (1971) 317. VOGEL, R.A., KIRCH, D., LEFREE, M., et al., A new method of multiplanar emission tomography using a seven pinhole collimator and an Anger scintillation camera, J. Nucl. Med. 19 (1978) 648. NALCIOGLU, O., MORTON, M.E., MILNE, N., Computerized longitudinal tomography (CLT) with a bilateral collimator, IEEE Trans. Nucl. Sei. NS-27 (1980) 430. CHANG, W., LIN, S.L., HENKIN, R.E., SALO, B.C., A multisegmental slant hole tomo graphic collimator (MUST): A new tomographic gamma camera system, J. Nucl. Med. 21 (1980) P28. ANGER, H.O., “Tomography and other depth discrimination techniques”, Instrumentation in Nuclear Medicine 2 (HIÑE, G.J., SORENSON, J.A., Eds), Academic Press, New York, (1974) 61. JASZCZAK, R.J., CHANG, L., STEIN, N.A., et al., Whole-body single-photon emission computed tomography using dual, large-field-of-view scintillation cameras, Phys. Med. Biol. 24(1979) 1123. BUDINGER, T.F., GULLBERG, G.T., Three-dimensional reconstruction in nuclear medicine imaging, IEEE Trans. Nucl. Sei. NS-21 (1974) 2. KEYES, J.W., ORLANDEA, N., HEETDERKS, W.J., et al., The humangotron - a scintillation camera transaxial tomograph, J. Nucl. Med. 18 (1977) 381. STOKELY, E.M., SVEINSDOTTIR, E., LASSEN, N.A., et al., A single photon dynamic computer assisted tomograph (DCAT) for imaging brain function in multiple cross sections, J. Comput. Assist. Tomogr. 4 (1980) 230. GREEN, A., ALDERSON, P., BERMAN, D., et al., A multicenter comparison of standard and 7 pinhole tomographic myocardial perfusion imaging: ROC analysis of qualitative visual interpretation, J. Nucl. Med. 21 (1980) P70. HENKIN, R.E., HALE, D.J., SALO, B.C., et al., 7 pin-hole tomography. Is it worth it? J. Nucl. Med. 21 (1980) P69. BUDINGER, T.F., Physical attributes of single-photon tomography, J. Nucl. Med. 21 (1980) 579. KUHL, D.E., EDWARDS, R.Q., RICCI, A.R., et al., The Mark IV system for radionuclide computed tomography of the brain, Radiology 121 (1976) 405. STODDART, H.F., STODDART, H.A., A new development in single gamma transaxial tomography: Union Carbide focused collimator scanner, IEEE Trans. Nucl. Sei. NS-26 (1979) 2710. JASZCZAK, R.J., COLEMAN, R.E., WHITEHEAD, F.R., Physical factors affecting quantitative measurements using camera-based single photon emission computed tomo graphy (SPECT), IEEE Trans. Nucl. Sei. (to be published 1981). JASZCZAK, R.J., COLEMAN, R.E., LIM, C.B., Single photon emission computed tomography, IEEE Trans. Nucl. Sei. NS-27 (1980) 1137. INTERNATIONAL ATOMIC ENERGY AGENCY, “IAEA Co-ordinated Research Programme on the Intercomparison of Computer-assisted Scintigraphic Techniques:
MUEHLLEHNER and COLSHER
Third Progress Report” , Medical Radionuclide Imaging (Proc. Symp. Los Angeles, 1976) 1, IAEA, Vienna (1977) 585. BUDINGER, T.F., DERENZO, S.E., GREENBERG, W.L., et al., Quantitative potentials of dynamic emission computed tomography, J. Nucl. Med. 19 (1978) 309. JASZCZAK, R.J., MURPHY, P.H., HUARD, D., et al., Radionuclide emission computed tomography of the head with 99 Tcm and a scintillation camera, J. Nucl. Med. 18 (1977) 373. PHELPS, M.E., HOFFMAN, E.J., MULLANI, N.A., et al., Application of coincidence detection to transaxial reconstruction tomography, J. Nucl. Med. 16 (1975) 210. HOFFMAN, E.J., HUANG, S.C., PHELPS, M.E., Quantitation in positron emission computed tomography: 1. Effect of object size, J. Comput. Assist. Tomogr. 3 (1979) 299. HUANG, S.C., HOFFMAN, E.J., PHELPS, M.E., et al., Quantitation in positron emission computed tomography: 2. Effects of inaccurate attenuation correction, J. Comput. Assist. Tomogr. 3(1979)804. MUEHLLEHNER, G., ATKINS, F., HARPER, P.V., “Positron camera with longitudinal and transverse tomographic capabilities” , Medical Radionuclide Imaging (Proc. Symp. Los Angeles, 1976) 1, IAEA, Vienna (1977) 291. HUESMAN, R.H., The Effects of a Finite Number of Projection Angles and Finite Lateral Sampling of Projections on the Propagation of Statistical Errors in Transverse Section Reconstruction, Lawrence Berkeley Laboratory Report LBL-4773 (1976). PHELPS, M.E., HOFFMAN, E.J., MULLANI, N.A., et al., Application of annihilation coincidence detections to transaxial reconstruction tomography, J. Nucl. Med. 16 (1975) 210. KIRCOS, L.T., LEONARD, P.F., KEYES, J.W., An optimized collimator for single-photon computed tomography with a scintillation camera, J. Nucl. Med. 19 (1978) 322. DERENZO, S.E., BUDINGER, T.F., CAHOON, J.L., et al., The Donner 280-crystal high resolution positron tomograph, IEEE Trans. Nucl. Sei. NS-26 (1979) 2790. ERIKSSON, L., BOHM, C., BERGSTROM, M., One year experience with a high resolution ring detector positron camera system: present status and future plans, IEEE Trans. Nucl. Sei. NS-27 (1980) 435. TODD-POKROPEK, A., ZUROWSKI, S., SOUSSALINE, F., Non-uniformity and artifact creation in emission tomography, J. Nucl. Med. 21 (1980) P38. BUDINGER, T.F., DERENZO, S., GULLBERG, G.T., et al., Trends and prospects for circular ring positron cameras, IEEE Trans. Nucl. Sei. NS-26 (1979) 2742. GLOVER, G.H., PELC, N.J., Nonlinear partial volume artifacts in X-ray computed tomography, Med. Phys. 7 (1980) 238. JASZCZAK, R.J., COLEMAN, R.E., “Selected processing techniques for scintillation camera based on SPECT systems” , Proc. Tenth Symposium on the Sharing of Computer Programs and Technology in Nuclear Medicine, Society of Nuclear Medicine, New York (1980). HUANG, S.C., PHELPS, M.E., CARSON, R.E., et al., An improved attentuation correction method for positron emission computed tomography, J. Nucl. Med. 21 (1980) P109. TRETIAK, O.J., METZ, C., The exponential radon transform, Soc. Indus. Appl. Math. (in press). BELLINI, S., PIACENTINI,M., CAFFORIO, C., et al., Compensation for tissue absorption in emission tomography, IEEE Trans. Acoust. Speech Signal Process ASSP-27 (1979) 213. GULLBERG, G.T., The attenuated radon transform: Theory and application in medicine and biology, PhD Thesis, Berkeley, University of California, 1979. CHANG, L.T., A method for attentuation correction in radionuclide computed tomography, IEEE Trans. Nucl. Sei. NS-2S (1978) 638.
IAEA-SM-247/202
197
DISCUSSION T.D. CRADDUCK: In your presentation you have emphasized that the so-called non-uniformity o f scintillation camera response is due primarily to the distortion in the image or, more accurately, the malpositioning o f events. You have described a system where uniformity correction is applied first by energy correction, then by linearity correction. Would you care to comment on systems that correct for non-uniformity on the basis o f the idea that the non-uniformity is caused by variations o f sensitivity, i.e. those systems that add or subtract counts to produce uniformity? G. MUEHLLEHNER: Uniformity correction devices that add or subtract counts do not properly compensate for the basic cause o f non-uniformity but provide only a cosmetic fix which gives visually acceptable images. These systems hamper rather than help when quantitative information is required (for instance, see Ref. [2] o f the paper). A.E. TODD-POKROPEK: I should like to thank you for re-stating the conclusions o f our paper on spatial distortion and uniformity corrections presented at the IAEA Symposium in Los Angeles in 1976 (based, o f course, on your own pioneering work ).1 Could you comment on the problem o f the dependence of attenuation correction on the activity distribution? The attenuation suffered by a uniform source is different from the attenuation o f a series o f point sources in the same attenuating medium. Therefore, a transmission scan does not, presumably, contain enough information to correct both these types o f distribution simul taneously. What is the ‘effective’ value for p if it is dependent on the amount of scatter? Could you comment on this problem and the current development of attenuation correction algorithms? G. MUEHLLEHNER: To the best o f my knowledge, attenuation (and, therefore, attenuation correction) is not a function of the activity distribution but only o f the size o f the object and the attenuation coefficient within the object. We have found that the iterative approach described by L. Chang (Ref. [68 ] in the paper) works well for various object sizes and shapes. I would expect that the variations as a function o f activity distribution which you observe are due to the influence o f scattered radiation. Subtraction techniques for scattered radiation are not yet fully explored. R.M. ADAMS: I should like to congratulate you on an excellent presentation. My question is, when using a system with ultra-short pulse clipping, is it still possible to use your scheme for distortion correction without protracting the dead time significantly at high count rates?
1 TODD-POKROPEK, A.E., ERBSMANN, F., SOUSSALINE, F., “The non-uniformity of imaging devices and its impact in quantitative studies” , Medical Radionuclide Imaging (Proc. Symp. Los Angeles, 1976) 1, IAEA, Vienna (1977) 67.
198
MUEHLLEHNER and COLSHER
G. MUEHLLEHNER: If the analogue section o f a scintillation camera using pulse shortening is operated at a total output rate of, for example, 500 000 counts per second, then less than 250 000 counts per second will pass the energy selection window in a clinical study. If these pulses are de-randomized in a buffer circuit, then the time available is approximately 4 microseconds between events, which is more than enough to perform digital distortion removal on an event-by-event basis. J.-C. ROUCAYROL: Apart from improving the uniformity o f response, is there any reason for using two cameras instead o f one in the equipment that you have shown? In our own system we use only one camera, rotate it through 360° and add the opposite projections. G. MUEHLLEHNER: Transverse section imaging devices need very high sensitivity in order to accumulate sufficient counts, i.e. approximately 106 per slice, for acceptable images. There will probably be some clinical studies where, say, 20 minutes imaging time obtained with two cameras is acceptable but 40 minutes imaging time with a single camera is excessive. The decreased imaging time is thus the main reason for using two cameras in transverse section imaging. P. SHARP: In practice, is there any real advantage in using a crystal which is thinner than the standard £-in thick crystal? G. MUEHLLEHNER: This question is discussed in more detail in the written version o f the paper and in the references. Briefly, there is some small gain for " T c m and a very real gain for 201T1, since for 201T1 a thinner crystal results in negligible loss o f sensitivity and a clinically significant gain in resolution. S. KESZTHELYI-LANDORI: What is your opinion concerning an Anger camera system with a xenon gas scintillation detector (Policarpo detector)? The energy resolution o f this detector is known to be better than that o f Nal. G. MUEHLLEHNER: Gas-filled detectors promise both better energy and better position resolution. Unfortunately, they generally have significantly lower sensitivity than an Anger scintillation camera at 140 keV and above. At present, system resolution is largely determined by the resolution o f the collimator and not by the intrinsic camera resolution. If you are willing to sacrifice sensitivity, it is better to trade collimator sensitivity for better collimator resolution and not detector sensitivity for better intrinsic camera resolution.
IAEA-SM-247/203
I n v ite d R e v ie w P a p er P O S IT R O N C O M P U T E D T O M O G R A P H Y P r e s e n t a n d f u t u r e d e s ig n a lte r n a tiv e s * M.E. PHELPS, E.J. HOFFMAN, SUNG-CHENG HUANG, D.E. KUHL Division o f Nuclear Medicine and the Laboratory o f Nuclear Medicine and Radiation Biology, UCLA School o f Medicine, Los Angeles, California, United States o f America
Abstract POSITRON COMPUTED TOMOGRAPHY: PRESENT AND FUTURE DESIGN ALTERNATIVES. Positron computed tomography (PCT) has experienced a significant growth in the past seven years. At present, the number of PCT programs throughout the world are 18 in North America, 11 in Europe and 4 in Asia. The status of these programs ranges from active to building, and the primary area of investigation is the brain and heart, although programs focused on cancer and functional studies of the whole body also exist. There are also six commercial companies in the United States of America, Canada, Europe and Asia that have or are developing PCT programs. Although PCT system designs are still evolving and all the complex issues of system optimization are still not completely defined, some design concepts have achieved general acceptance. Circumferential designs employing hexagonal, octagonal or circular geometries are preferred since they maximize the tomographic plane efficiency. Bismuth germanate (BGO) has become the detector of choice because of its high efficiency (intrinsic photopeak and geometric through high packing densities) when the small detectors required for high spatial resolution are used. Caesium fluoride detectors are also being investigated because of their short coincidence time resolution and potential for time of flight measurements. Multiplane systems, consisting of stacked single plane geometries, have been developed to provide higher overall organ efficiency, typically with some compromise in image quality compared with a single plane system. The most poorly characterized aspect of multiplane systems is the slit and septa shielding design that has a large impact on random and scatter coincidences, count-rate capability, efficiency and uniformity of spatial resolution. The present PCT systems have spatial resolutions in the tomo graphic plane of about 8 to 17 mm and axial resolution of about 12 to 18 mm. PCT systems that have been or are now being developed are providing the means to demonstrate the unique capability of PCT to provide quantitative measurements of local biochemical and functional processes in man.
* This work was partially supported by the Department of Energy Contract DE-AM03-76-SF00012 and NIH Grants USPHS7-R0I-GM24839, P01-NS1 5654 and NS-02808.
199
200
PHELPS et al.
INTRODUCTION The interest in positron computed tomography (PCT) has dramatically increased over the past five to seven years (Table I). Whereas some o f this growth is due to the acceptance and development o f computed tomography techniques in general, it has been primarily generated by advancements in instrumentation, labelled compounds, tracer kinetic models and successful applications o f PCT primarily to studies o f the brain, heart and pancreas in man. The aim o f these studies has been extended beyond tomographic delineation o f organs from surrounding structures or examination o f local regions within an organ to provide quantitative measurements o f the local tissue radioactivity concentrations. This quantitative measurement introduces a potential in PCT for the implementation o f tracer kinetic models which, when applied with appropriately labelled compounds, can be used for the measurement o f physiological processes such as the rate of metabolism, blood flow, membrane transport, synthesis, receptor binding and other processes in both normal and diseased states in man. The approach is analogous to invasive animal studies in which tissue tracer concentrations o f compounds labelled primarily with 14C and 3H are measured by external counting o f excised samples or with autoradiography [1, 2]. However, the tissue tracer concentrations are measured non-invasively with PCT and compounds are labelled with positron-emitting radionuclides of n C, 13N, 150 , 18F, 68Ga, etc., or used directly such as in the case o f 77Kr, 82Rb, 38K and others. This analytical imaging capability places demanding requirements on PCT system design. Added to this are the different requirements for a system designed for the brain only as compared with a whole-body capability, static versus dynamic studies and the different variables o f design such as geometric configuration, detector materials, spatial resolution, system sensitivity or efficiency, count-rate capability, random and scatter coincidence rate, shielding design, linear and angular sampling, photon attenuation correction, image reconstruction algorithm, single versus multiplane designs and data display. An extensive presentation o f all these factors is beyond the scope o f this survey. However, certain o f the above factors will be discussed in the light o f different systems which have been or are now being developed.
HISTORICAL BACKGROUND 1 Imaging o f compounds labelled with positron-emitting radionuclides was first suggested in 1951 by Wrenn and co-workers [3] and Brownell as reported 1 Most of the historical background has been taken from DERENZO, S.E., Optimization of Side Shielding for Circular Positron Emission Tomographs, University of California, Lawrence Berkeley Lab. Rep. 8622 (Jan. 1979).
IAEA-SM-247/203
201
by Sweet [4], and the first successful positron scanner was described by Brownell and Sweet in 1953 [5]. In 1959 Anger presented his positron camera at the IAEA Symposium in Vienna [6 ]. These systems were designed and used for conventional 2-dimensional imaging o f positron radionuclide distributions in the body. The concept o f transverse section imaging with positrons was developed in 1962 by Rankowitz, Robertson and co-workers [7, 8 ] at Brookhaven National Laboratory. This system consisted o f a small diameter circular array o f 32 Nal(Tl) detectors for studies o f the brain, and employed linear superposition o f back projections for formation o f transverse section images (i.e. blurring tomography). Concurrently Kuhl, Edwards and co-workers [9, 10] were developing the concept o f transverse section scanning for single-photon-emitting radionuclides. Subsequent to these developments, Anger [11], Muehllehner and Wetzel [12], Todd-Pokropek [13], Myers and co-workers [14], Bowley and co-workers [15], Tanaka [16] and Anger [17] investigated and developed techniques for performing emission tom o graphy. All these approaches were limited by problems encountered from inade quate mathematical image reconstruction algorithms, computational hardware and number o f angular views, and image distortions due to photon attenuation or statistical limitations. In 1973 Kuhl, Edwards, Ricci and co-workers [10] presented a paper on quantitative section scanning in which they used a correction technique referred to as the orthogonal tangent correction. This approach employed orthogo nal scan profiles with multiplicative correction factors to compensate partially for the blurring present in their original approach that employed superimposition o f back projections. Although this algorithm was not a true reconstruction algorithm by present standards o f computed tomography2, it did provide an improvement in the tomographic resolution (i.e. contrast enhancement), and together with attenuation corrections and a normalization from assuming the head was similar to a uniform cylinder of activity, allowed the first use o f section scanning for measurement o f tissue activity concentrations and the measurement o f cerebral blood volume with " T c m-labelled red blood cells [19]. In 1976 Kuhl and co-workers published an article on the MARK IV [20] scanner which consisted o f a square array o f 32 Nal(Tl) detectors with a significant increase in efficiency over their previous scanners (MARK II and III), and which employed an exact reconstruction algorithm (orthogonal version o f algebraic Reconstruction technique, ART [21 ]. In 1974, Budinger and Gullberg [22] also published an article on single photon tomographic reconstructions with a rotating camera and an iterative least squares reconstruction algorithm. 2 The term computed tomography or computed tomograph refers to image reconstruction techniques or tomographs which employ exact or near exact reconstruction algorithms [18]. It should be noted that this is a necessary but not sufficient condition for quantitative emission computed tomography. The tomographic technique must provide appropriate correction for photon attenuation, linear and angular sampling, spatial resolution uniformity, statistical accuracy, etc. [18].
TABLE I. POSITRON COMPUTED TOMOGRAPHY PROGRAMS
Location
Area of investigation
Status
Type of tomograph 3
NORTH AMERICA UGLA, Los Angeles, CA
Brain & heart
Active
ECAT II & Neuro ECAT
Washington Univ., St. Louis, MO
Brain & heart
Active
PETT VI
Harvard Univ., Boston, MA
Brain, heart & lung
Active
PC I & II
Univ. of California, Berkeley, CA
Brain & heart
Active
Donner Circle
Univ. of Pennsylvania, Philadelphia, PA
Brain
Brookhaven National Lab. and New York Univ., Upton, NY
Brain & heart
Active Active Active
PETT V Pett III (PETT VI) Dual Headed Camera*5
Univ. of Chicago, Chicago, IL
Brain & heart
Oakridge Assoc. Univ., Oak Ridge, TN
Cancer
Active
ECAT II
Sloan-Kettering/Cornell, New York, NY
Cancer & brain
Active
PCT 4200c (PCT 4600)
National Inst, of Health, Bethesda, MD
Brain
Active
ECAT II (Also building a tomograph)
Univ. of Wisconsin, Madison, WI
Whole body
Active
ECAT II
Univ. of Miami, Miami, FLA
Brain
Building
(PETT V)
Univ. of Michigan, Ann Arbor, MI
Brain
Building
(PCT 4600)
Univ. of Texas, Houston, TX
Heart & brain
Univ. of Washington, Seattle, WA
Cancer
Building
Columbia Univ., New York, NY
Brain
Building
Cho Circle
McGill Univ., Montreal, Canada
Brain
Active
Positome III
Univ. British Columbia, Vancouver, BC
Brain
Building
Building
EUROPE
Frédéric Joliot Hosp., Orsay, France
Brain
Active
ECAT II
Nucl. Res. Center, Julich, FRG
Heart & brain
Active
ECAT II
Hanover Med. Sch., Hanover, FRG
Heart & brain
Active
PCT 4200
Hammersmith Hosp., London, UK
Heart & brain
Active
ECAT II
Whole body
Active
ECAT II
Brain
Active
ECAT II & (Neuro ECAT)
Max Planck Inst., Cologne, FRG
Brain
Building
Univ. of Uppsala, Uppsala, Sweden
Brain
Building
Univ. of Milan, Milan, Italy
Whole body
Building
Univ. of Naples, Naples, Italy
Brain
Building
Univ. of Pisa, Pisa, Italy
Heart
Building
PC-95 (Neuro ECAT)
Building
Turku, Finland
IAEA-SM-247/203
Univ. of Liège, Liège, Belgium Univ. of Ghent, Belgium
ASIA Nakano Chest Diseases Hospital, Tokyo, Japan
Lung
Building
National Radiological Inst., Chiba, Japan
Brain
Active
Positologica
Tohoku Univ., Sendai, Japan
Brain
Active
ECAT II
Research Inst. Brain & Blood Vessels, Akita, Japan
Brain
Active
Headtome
203
a Parentheses refer to systems on order or that are being constructed. b Searle Radiographies. c Cyclotron Corporation version of Brownell’s camera (PCT 4200 on loan).
TABLE II. COMMERCIAL SUPPLIERS OF PCT SYSTEMS 204
Company
Location
Detector11
Design
Status
systems sold
Whole-body units ORTEC, Inc.
Cyclotron Corp.
Oak Ridge, TN
Berkeley, CA
11
Nal(Tl) 38 dia. X 76 mm BGO
100.5 cm dia. hexagon single plane 99.8 cm dia. circle 1-3-image planes
Completed Under development
0
BGO 19.8 dia. X 30 mm
90 cm dia. circle 7- or 9-image planes
Under development
0
ORTEC, Inc.
Oak Ridge, TN
BGO 17 X 27 X 30 mm
65.8 cm dia. circle 1—5-image planes
Completed
2
Cyclotron Corp.
Berkeley, CA
BGO 19.8 dia. X 30 mm
60 cm dia. circle 7- or 9-image planes
Under development
2
Scanditronix
Uppsala, Sweden
BGO 10 X 20 X 30 mm
48 cm dia. circle 7-image planes
Under development
1
AECL
Ottawa, Canada
BGO 1 8 -2 1 X 30 X 30 mm
42 cm dia. circle 3-image planes
Under development
0
Hitachi Ltd.
Kashiwa, Japan
BGO 12 X 20 X 26 mm
44 cm dia. circle 1-image plane
Under development
0
Shimaelzu Siesakusho
Akita, Japan
Nal 16 X 28 X 70 mm
42 cm dia. circle 1-image plane
Under development
0
Detector dimensions are given for cylindrical detectors as dia. X thickness or for rectangular detectors as width X height X thickness. The AECL device has tapered crystals that are 18 mm at the front face and 21 mm at the back.
PHELPS et al.
Neurological units
IAEA-SM-247/203
205
In 1973, Chesler and co-workers [23] scanned a symmetrical phantom of positron activity with a pair o f detectors connected in coincidence and, because the phantom was symmetrical, this single scan profile was used for different angular projections to reconstruct the cross sectional activity distribution with a true tomographic reconstruction algorithm. The first truly quantitative and practical positron computed tomograph was the PETT II (PETT is an acronym proposed by Phelps in 1973 and stands for positron emission transaxial tomograph) developed by Phelps, Hoffman, Ter-Pogossian and co-workers [24, 25]. This system consisted o f a hexagonal array o f Nal(Tl) detector connected in coincidence between bank pairs and used a Fourier-based reconstruction algorithm [24, 25]. The PETT II system was used to develop the principles o f PCT, and for phantom and whole-body animal studies with 11CO-haemoglobin, I3NH 3 and 18F. In 1975, these same investigators published results from the whole-body PETT III that was developed for human studies [2 6 -2 8 ]. This system was composed o f a hexagonal array of 48 Nal(Tl) detectors. In 1976, Cho and co-workers [29, 30] presented a circular ring design comprised of 64 Nal(Tl) crystals. In 1976, the first commercial PCT system (ECAT) was constructed by ORTEC, Inc., and delivered to UCLA. This system was designed by Phelps and Hoffman [31 ]. Circular ring PCT systems have also been developed by Bohm, Eriksson and co-workers [32] (first systems to incorporate the wobble motion to improve sampling in circular tomographs); Derenzo, Budinger and co-workers completed the construction o f the Donner circular PCT in 1978 [33], and Yamamoto, Thompson and co-workers improved the original Brookhaven tomograph [34] and converted the design to the first system to employ bismuth germanate detectors [35] as originally suggested by Cho and co-workers [36]. The first PCT systems with simultaneous multiple plane capability were the dual-headed positron camera (PC) o f Brownell, Durham, Chesler and co-workers [37] and Muehllehner and co-workers’ [38, 39] use o f two large field o f view cameras which rotated about the patient. Multiple layer circumferential PCT systems have been developed by TerPogossian, Mullani, Hood and co-workers in the form o f the hexagonal, 7-image plane, Nal(Tl) detector PETT IV [40] for whole-body studies and the circular 7-image plane Nal(Tl) detector PETT V [41 ] for the brain. These systems employed a one-dimensional Anger position logic for identifying detector planes. Thompson, Yamamoto and co-workers also expanded their single ring PCT to a two-detector ring system (private communication). Ter-Pogossian, Mullani, Hood and co-workers are now developing a multiplane circular array with individual detectors of CsF. A small-diameter 5-detector plane (9-image planes) circular PCT with BGO detectors is also being developed by Brooks, Sank and DiChiro [42]. A circular ring system with 64 BGO detectors has been developed by Tanaka and co-workers [43] as well as a hybrid (PCT and single-photon CT) circular tomograph with 64 Nal(Tl) detector by Uemura and co-workers [44].
206
PHELPS et al.
WHOLE-BODY DEVICES
ANGER POSITRON CAMERA Rotote(180°)
MULTICRYSTAL OR WIRE CHAMBER POSITRON CAMERA Translate (few cm) Rotate 180°
CIRCULAR HEXAGONAL RING RING CAMERA CAMERA Translate (few cm) Rotate (few degrees)а/or Wobble 360° Rotate 60°
N E U R O L O G I C A L DEVICES
В
OCTAGONAL RING
CIRCULAR RING
CAMERA
CAMERA
Translate (few cm)
R otate (few d eg rees)
R otate 4 5 °
and/or Wobble 3 6 0 °
F IG . l. A . G eom etric designs o f w ho le -b o d y p o sitro n com puted tom ography devices illu s tra tin g the g eom etric co n figurations and req u ire d m otions. B. neurological p o s itro n com puted tomographs.
G eom etric con fig ura tio n s o f m ultiplan e
207
IAEA-SM-247/203
A number of commercial systems have been or are being developed at present (Table II). The principles o f PCT and systems that were developed up to 1977 have been reviewed in detail elsewhere [18].
GEOMETRIC CONFIGURATION It is obvious from the above that PCT system design has involved a variety of different strategies, which in fact are still in a state o f evolution. This situation defies a simple description. However, some concepts o f PCT system design have achieved general acceptance. Circumferential designs employing hexagonal, octagonal or circular geometries (F ig .l) are preferred because they maximize the single-plane efficiency. This is a key factor since whether a system has single or multiple planes, the minimum imaging time is determined by the time required to gather sufficient data for a single plane [45 ]. Single-plane PCT systems The design o f a single-plane system forms the basis o f PCT systems, although some modification is required when expanded to multiple planes. The parameters o f primary importance in PCT system design are the inter-detector diameter, detector packing, shielding design, detector size, intrinsic efficiency (total and photopeak), coincidence time resolution and sampling capabilities. These para meters will be examined individually but they are, o f course, all interrelated. The diameter (D) o f the ring (Fig. 2) will determine the geometric or solid angle detection efficiency (Í 2) for any given pair o f detectors since fia l/D 2
(1)
This would suggest that the ring diameter should be small to maximize the system geometric efficiency. However, the efficiency for random and scatter coincidence is also a function o f Í2, unfortunately on £22. For example, the amount o f scattered coincidence (SC) for a coincident pair o f detectors is given by [46] SC = £
S21S22e ,( E i)e 2(E¡)e
e
Aj
(2)
FOV
where £2, and Í 22 are the solid angles o f detectors 1 and 2, respectively, e^ E j) and 62(Ej) are the intrinsic detector efficiencies for the detection o f gamma rays from i with energy E¡ and for gamma rays from j with energy E¡, respectively; e ‘ ‘ a n d e j j are the absorptions o f gamma rays from i and j, respectively, Aj is the activity at i, and FOV is the detector field o f view.
208
PHELPS et al. SCATTER AND RANDOM COINCIDENCES
Д
Б
С
Axial Direction ---- ►
F IG .2 .
Schem atic illu s tra tio n o f single (A ) and m u ltip le plane (В, C) p o sitro n com puted to m o
g raphy devices illu s tra tin g shielding and septa designs. The shields and septa determ ine the geom etric angle o f acceptance f o r bo th co in cid e n t and n o n co in cid e n t rad ia tio n . The geom etric angle o f n o n co in cid e n t rad ia tio n is d eterm ined by the rin g diam eter (D j, the s lit shield and septa length (L ) and s lit opening (S). To m axim ize the ra tio o f true coincidences to those fro m scatter and random events, the g eom etric angle o f acceptance f o r n o n co in cid e n t rad ia tio n should be as close as possible to th a t f o r coincidence. Septa are em ployed in m ultiplan e devices to lim it the increasing angle o f acceptance as the num ber o f planes is increased ( A - C j. L im ita tio n s in septa resu lt fro m the need f o r them to be th in and sh o rt to m in im ize the angulation produced in the fo rm a tio n o f midplanes (see F ig .4 ) th a t c o n flicts the shielding requirem ents o f long th ic k septa. F O V is the tom ographic fie ld -o f-vie w .
From E q.(2) it is apparent that small values o f the solid angle will cause reductions o f scatter coincidences relative to true coincidences. Random coincidence, RC, is also proportional to П 2 but has an additional dependence on the system time resolution (2 r) as given by
RC = 2
£2iíÜ 2e¡ (Ej) e (E j)e- ^'X‘ e~^jXj AjA j2r
(3)
FOV
Because of this quadratic dependence o f scatter and random coincidence on the geometric efficiency Í2, there are system diameters below which the scatter and random coincidences will exceed the true coincidence rate. The primary design goal is to pick a system diameter which is large enough to give acceptable scatter and random rates while maintaining an adequate true coincidence sensitivity. A key factor in achieving this results from careful shielding designs (Figs 2 and 3). These shields are lead plates placed above and below the plane o f detectors and extend as close to the patient as possible. This is done to restrict the activity in the field o f view o f any detector to the cross-sectional slice from which true coincidences can be recorded, as discussed by Phelps and co-workers [24, 47],
209
IAEA-SM-247/203
Pb Shielding
F IG .3 .
Diagram o f s lit shield and septa design proposed by H o ffm a n and co-w orkers [4 8 ].
S is the s lit o r septa shield gap, T is the s lit o r septa gap a t the edge o f the fie ld -o f-v ie w and L is the length o f the s lit shield o r septa. Cross hatched area. As is the plane thickness in which a ll true coincidence data are assumed to originate. A ) is f o r a single plane design. B) diagram o f a m ultiplan e s lit shield design, w hich a llow s fle x ib ilit y f o r both lo w background tom ography em p lo yin g o n ly the straight-across image planes (i.e. three image planes in the exam ple) and high e fficie n cy when the interplane septa are rem oved and a d d itio n a l m idplane images are fo rm e d in the m idplane fro m the sum o f adjacent detectors. When in p o sitio n , the septa p rovide a very high scatter and random s rejectio n w ith im proved a xia l reso lu tio n and o n ly a sm all decrease in true coincidence e ffic ie n c y (see te xt). With the septa rem oved the system provides high e ffic ie n c y f o r the m axim um n u m ber o f image planes (i.e. fiv e in the exam ple) w ith a m in im u m d is to rtio n due to angulation in the m idplane (see F ig.5). D o tte d lines indicate the d ire c t plane and m idplane data co lle ctio n concept.
Hoffman and co-workers [26, 46, 48] and Derenzo and co-workers [49, 50]. Any activity viewed by the detector outside the slice is only a potential random or scatter coincidence, never a true coincidence event. Thus it is apparent that the length, L, o f the shields and the size o f the opening in the shield, S, will also be important in limiting random and scatter coincidences. Derenzo, Budinger and co-workers [49—51 ] have shown that T C aS2
(4)
R C a S 3/L
(5)
S C a S 4/L 2
( 6)
A shielding design proposed by Hoffman and co-workers [48 ] replaces the standard parallel slit shields with ones that are tapered such that the opening to
210
PHELPS et al.
the field o f view (FOV) is slightly less than at the detector as shown in Fig.3A. This dramatically reduces the scatter and random coincidence rate while only producing a small decrease in the true coincidence rate. For instance, with a 100-cm ring dia., 38.1 mm circular Nal(Tl) detectors, L = 25 cm, S = 38.1 mm and T = 25 mm (Fig.3), the scatter was reduced by a factor o f 2 whereas the true coincidence rate was decreased by only 15%. In fact, the true coincidence that is lost is predominately from the wings o f the line spread function (LSF). Thus, these shields not only reduce scatter and randoms but also improve the axial resolution [47]. Phelps and co-workers [23, 27, 47] and Hoffman and co-workers [26, 46, 48] have shown that the detector separation must be reasonably large to minimize the resolution variation within the field o f view and to produce the best absolute resolution for a given size detector (both in the image plane and axial direction). For example, with a ring diameter o f 45 cm, the resolution varies by about 30 to 40% across a 25-cm field o f view (FOV) with 17 mm wide by 28 mm high BGO detectors. This is contrasted with a ring dia. o f 75 cm where the resolution vari ation is 10 to 20%. This variation in spatial resolution is modulated in the image reconstruction which tends to make the variations somewhat less and the absolute values somewhat worse. Derenzo [52] has developed a figure o f merit that takes into account the detector ring diameter, shield lengths, the shield gap S, detector coincidence time resolution and activity levels. This formulation, although only for single-plane systems and not inclusive o f all important system design factors, is an attempt to develop a unified formulation for design optimization similar to the figure o f merit for collimated detectors developed by Beck and co-workers [53]. From Derenzo’s [52] formulation it would appear that the ring diameter should be somewhat less than or equal to a factor o f two larger than the FOV. This formu lation, however, neglects the resolution variation with distance across the FOV that would favour an increase in the ring diameter. This is even more o f an issue in multislice systems in terms o f the cross planes (Fig.4 and below). Thus it would appear that from all these conditions the ring diameter should in fact be twice the diameter o f the FOV or somewhat larger, as was originally proposed by Phelps and Hoffman and co-workers [23, 25, 26, 46, 47]. Multiplane PCT systems Circumferential multiplane PCT systems are being developed to increase the overall organ efficiency. These systems consist o f stacked single-plane geometries in which not only are the directly opposing detectors used to form an image plane but coincidences are recorded between the two neighbouring detector planes. These data are summed and assumed to be a good approximation o f the plane midway between adjacent planes, as shown in Fig.4. As discussed above, the
211
IAEA-SM-247/203
Line Spread Functions
Л
mm.
M
Line Spread Functions
J
L
A
A
A
7ГТГ
20cm 3Öcm
К
0
№
Directly Opposing Plane
Cross Plane
DIRECTLY O PPO SIN G and C RO SS P L A N E A X IA L R ES O LU T IO N Axial Direction
F IG .4.
/1/
70cm
ЕЕЕЕГ
Schem atic illu s tra tio n o f the variation in spatial resolution o f the m idplane axial
reso lu tio n as a fu n c tio n o f the d e te cto r rin g diam eter. With a sm all rin g diam eter, as shown to the le ft, the a x ia l reso lu tio n d ra m a tica lly varies fro m centre to edge o f the fie ld o f view IF O V j. A t the fa r extrem e o f the (F O V ) the line spread fu n c tio n s w ill a ctu a lly separate and regions in the m idplane w ill be p o o rly sampled. Increasing the d e tecto r ring diam eter, as shown to the rig h t, im proves the u n ifo rm ity o f the m idplane a xia l reso lu tio n . In a d d itio n this variation is m in im iz e d by close packing the detectors in the a xia l d ire ctio n and e m p lo yin g septa (n o t shown) w ith a m in im u m interference o f the g eom etric F O V o f the m idplane. This requires short, tapered septa w hich u n fo rtu n a te ly increases septa p e n etra tio n and increases random and scatter events.
variation in axial resolution with distance (i.e. from the centre to edge o f the FOV) can be significant if a small ring diameter is chosen. To minimize this variation, the detectors should be close packed in the axial direction (this also favours rectangular over cylindrical detectors), and the ring diameter should be somewhat larger than the single-plane requirements discussed above. To reduce the scatter and random coincidence rates, interplane septa are em ployed in multiplane designs (Fig.2). These septa, however, increase the axial resolution from the centre to the outer edge o f the field o f view. Thus, there is a conflict between minimizing the axial resolution variation with distance and the shielding requirement to minimize randoms and scatter. The former requires short septa tapered to a thin edge at the FOV (Fig.4) and the latter requires long septa with straight sides (Fig.2) or tapered in the opposite direction (Fig.3). The present systems employing fixed septa can have a resolution variation throughout the FOV o f 40 to 200% depending on the ring diameter, septa length and thickness, detector size and shape, and
212
PHELPS et al.
detector packing density [45]. This spatially variant system response has serious implications in quantification o f tissue activity concentrations and partial volume effects [54]. It should also be appreciated that the interplane data are a large fraction o f the total system efficiency in a multiplane geometry. For example, with 2, 3, and 5 detector plane systems, the cross plane has 1.0, 1.3 and 1.6 times as much efficiency as the total from all the straight-across planes [45 ]. Multiplane PCT system designs generally compromise the quality o f the straight-across plane to some degree compared with a single-plane device. However, this appears warranted to provide high overall organ efficiency. An alternative approach is employed in the Neuro ECAT (ORTEC, Inc.) in which the system provides an electronically controlled selection of the inverse tapered shields shown in Fig.3 (low randoms and scatter, improved axial resolution and high count-rate capabilities) with only the straight-across planes or alternatively retraction o f the septa for maximum sensitivity (e.g. for tracers with low tissue concentrations) and minimum distortion o f the interplane data. The latter mode has a significant increase in scatter and randoms (when used for low tissue concentration, i.e. receptor studies, randoms are reduced because o f reduced activity levels). Although shielding designs of single-plane systems are rather straightforward, there are serious pitfalls in the design o f interplane shielding in multiplane devices that are not yet well characterized, and more definitive studies are required.
Hexagonal and octagonal array Generally, the advantages o f the hexagonal and octagonal geometries are as follows. They provide a straight-sided design with the following advantages: ( 1 ) allows maximum and uniform sampling resolution in both linear and angular directions (i.e. maximizes recovery o f inherent detector resolution); ( 2 ) sampling resolution is not limited or determined by detector size; (3) high packing density for high geometric efficiency; (4) redundant sampling which provides more uniform error distribution, general noise reduction, protection against errors from detector instability, miscalibration and motion artifacts (movement o f activity, organ or patient); (5) small linear motion (detector width) and angular rotation (60° or 45°); (6) high intrinsic detector efficiency since small detectors are not required to improve sampling; (7) rapid scan capability; (8) shadow shields can be used for high resolution options with increased sampling resolution to assure maximum resolution recovery. The disadvantages o f the hexagon and octagon are: (1) they do not use all the geometric efficiency in the plane; (2 ) motion o f the gantry is required; (3) short scan times require significant data collection during movement o f detectors.
IAEA-SM-247/203
213
Circular array Advantages o f the circle are: (1) maximum geometric efficiency; (2) images, although with a loss in resolution, can be reconstructed without moving the gantry; resolution can be improved with a small rotation (i.e. 1/2 detector spacing) and/or 360° wobble; (3) some degree o f redundant sampling by over-rotation o f the gantry; (4) high detector packing geometry; (5) rapid scan capability. Disadvantages o f the circle are: (1) Intrinsic detector efficiency is typically reduced because small detectors are required to improve sampling; this is par ticularly true if small Nal(Tl) o f CsF detectors are employed. Bismuth germanate (BGO) detectors reduce this problem significantly [33, 35, 36, 55]. (2) The majority o f geometric lines o f response (LOR) in the circle that are not used in the hexagon or octagon have lower intrinsic efficiency and significant angulation error. For example, the large angle lines o f the fan beam in the circle have lower efficiency than straight across or small angle lines because o f the oblique angle o f incidence and small effective detector thickness (this also causes the line spread function to become assymetric [32, 51 ]. This drop in efficiency occurs for true coincidence but not for scatter and randoms, so the S/N is reduced for the wide angle lines. This drop in coincidence efficiency can be as high as a factor o f 2 at the largest angles, even for BGO o f the sizes required for high resolution circular systems. This factor is even higher for small Nal(Tl) or CsF detectors and increases as circle diameter is reduced. These LORs are a small number compared with the total, but are a majority o f the lines not used in the hexagon or octagon. (3) The efficiency across the FOV in the circle has marked non-uniformity when the wobble m otion is used. (4) Although images can be reconstructed without detector movement, resolution is lower and/or non-uniform due to limited sampling. (5) The requirement o f 360° wobble and half detector rotation remove the advantage o f stationarity in the circle over the linear scan and 60° or 45° rotation o f the hexagon or octagon. There are many aspects o f the different geometric designs which still need to be analytically characterized before absolute conclusions can be drawn. Because the apparent high geometric efficiency o f the circle is not as large as one would initially expect and because o f a somewhat lower intrinsic detector efficiency, it is hard to see how the circle geometry will provide significantly higher efficiency than a hexagon or octagon o f the same diameter. In fact, if small Nal(Tl) or CsF detectors are employed in the circle, the hexagon can have higher efficiency. The sampling question with the circle needs to be better characterized in terms of resolution recovery and noise from the non-uniform efficiency. The minimum scan time and distortion due to the movement during rapid scans in the particular data collection scheme o f all systems needs to be examined in more detail, although statistical considerations are probably the major factor in minimum scan times.
214
F IG .5.
PHELPS et al.
Images o f a m u ltip le line source phantom (Derenzo p h a nto m [ 3 3 ] ) show ing the
e ffe c t o f sampling. The distance between adjacent line sources is 10, 12, 14, 16, 18 and 20 mm in the six sectors o f the phantom . The diam eter o f the line sources were equal to one fo u rth o f the spacing in each group. The image on the le ft was reconstructed using a sam pling distance th a t is one fo u rth the image resolution whereas the image on the rig h t em ployed a sam pling distance one h a lf the image reso lu tio n . The sam pling a rtifa cts (aliasing) are rea d ily apparent in the image to the rig h t. This is clear fro m the n o n -u n ifo rm shape and in te n sity o f the line sources com pared w ith the image a t the le ft w hich is sampled a p p ro p ria te ly (see te xt). N o te th a t the line sources w ith spacings equal to
10
and
12
mm appear to be b e tter resolved in the image to the rig ht.
However, this is o n ly an oscilla tin g a rtifa c t caused by under sam pling (i.e. the apparent p a ttern is n o t th a t o f the actual lines sources). F ro m R ef.
[50].
RANDOM AND SCATTER COINCIDENCE AND ATTENUATION CORRECTION Detailed studies on the magnitudes, image distortion and correction schemes for random and scatter coincidences are provided elsewhere [49, 56]. However, it should be pointed out that randoms and scatter not only lower image contrast in PCT but they also can produce low and high frequency distortions in the image and can effectively lower the true efficiency o f a system [49, 56]. For example, increases in efficiency or number o f counts detected are done to reduce noise and improve accuracy. However, if random and scatter coincidences are a signifi cant fraction of the data then they effectively reduce the system efficiency by increasing image noise whether they are left in the data or subtracted. Subtraction of randoms and scatter increases noise caused by propagation o f statistical error (and exactness o f the approach) from the arithmetical operations o f subtraction. Correction schemes and minimization o f errors in the correction for photon attenuation, including automated edge finding routines, have been discussed in detail by Huang and co-workers [57].
IAEA-SM-247/203
F IG . 6 .
215
Images o f a p a rallel bar phantom w ith bar w id th s and spacing equal to 9 m m and an
in trin s ic d e tecto r reso lu tio n o f 9 m m. The upper le ft hand image was reconstructed fro m data collected w ith a sam pling distance o f 2.85 m m. The upper rig h t hand image was collected w ith a sam pling o f 5 .7 m m. The tw o lo w e r images were reconstructed using a sam pling o f 11.4 mm and the image a t the lo w e r rig h t hand corner was taken w ith a bar phantom displaced by 5 mm com pared w ith the image a t the lo w e r le ft. The reco n stru ction f ilt e r fu n c tio n s were adjusted to p rovide equal reso lu tio n recovery responses to illu stra te solely the e ffe c t o f sampling. The images in the upper le ft, upper rig h t and lo w e r le ft illu stra te the e ffe c t o f sam pling when the sam pling distance is a b o u t one th ird , one h a lf and equal to the spatial reso lu tio n in the image. The resu lta n t aliasing produces d is to rtio n s in shape, in te n sity and p o sitio n . The images a t the b o tto m illu s tra te th a t the sam pling a rtifa c t also becomes dependent upon the source position.
SAMPLING Linear and angular sampling can be both determining and limiting factors in a PCT system. The basic linear sampling requirement is set by the sampling theorem, that states that the distance between samples should be half the maximum resolution to be recovered in the image (i.e the cutoff frequency o f the recon struction filter function). Budinger and co-workers [51 ] and Huang and co workers [58] have demonstrated the aliasing artifacts that occur in PCT when larger sampling distances are employed. Huang and co-workers [58] have shown that under realistic conditions (i.e. sampling theorem applies to ideal conditions o f
T A B L E III. C O M P A R IS O N O F SOM E O F T H E P R O P E R T IE S O F P O T E N T IA L PCT D E T E C T O R S
BGO
CsF
Plastic
Ge
12.53
8 3 ,3 2 ,8
55.9
6 .1
32
1.03
5.38
7.13
4.61
Mean decay time (ns)
230
300
5
3
Light yield relative to Nal (%)
100
10
5
18
0.5
< 0 .2
<
3 -4
5 -1 0
Density (g/cm3)
Best measured coinc. time resolution (ns; FWHM)
3.67
1.3
5.0
1 .0
6 -1 2
12-24
3 -4
Energy resolution (FWM at 511 keV, %)
7 -9
15-20
27
Thickness to stop 90% of 511 keV (m m )
68
25
59
190
52
Percentage of interactions that are photoelectric at 511 keV
19
38
19
0
5
37
90 79
46 25
7.8
40
7.3
19
88
49 29
0
26
0
10
Intrinsic coinc. efficiency*5 20 X 30 X 50 mm (%) 4 X 30 X 50 mm (%) Photofraction (%) 20 X 30 X 50 mm 4 X 30 X 50 mm Requires p ro te c tiv e can
21
41 25 Yes Hydroscopic
71 No
Yes Very hydroscopic
0
_
No
.5 -1 .0
Yes, cooled to liquid N 2
40 8 0 -2 0 0
None
< 1%
Yes, only to contain gas
a Limit set by finite propagation in time of flight and total system timing considerations (see text). Energy threshold of 100 keV for Nal, CsF, plastic and Ge. Threshold of 300 k e V for BGO. Data from Derenzo, S.E., M o n te Carlo calculations of the detection efficiencies of Nal(Tl), BGO, CsF, Ge and plastic detectors for 511 keV photons (unpublished data).
b
PHELPS et al.
Realistic system coinc. time resolution (ns; FWO. lM)a
Multiwire chamber
216
Atomic number
Nal
IAEA-SM-247/203
217
band limited signals, sine functions, exact positions o f the samples, etc.), to reduce image distortions due to sampling errors, the sampling distance should be one third to one quarter the maximum resolution to be recovered. Figures 5 and 6 illustrate distortions due to sampling errors. A point seldom made is that the choice o f sampling distance is also related to the intrinsic detector resolution. Obviously, if the sampling distance is large compared with the detector resolution (i.e. unsampled gaps in the scan profiles), then the high resolution content intrinsic to the detector will not be covered and sampling artifacts will also result. Thus the sampling distance should also be < 1/3 to 1/4 the intrinsic detector resolution. The angular sampling requirement is given by the equation for the required number o f angular projections in CT [59]. Number o f angles (over a 180° arc) = (7rD)/2 a
(7)
where D is the diameter o f the FOV and a is the linear sampling distance. Equation (7) basically states that the linear sampling distance and arc length of the angular sampling should be about equal.
DETECTOR MATERIAL Some properties o f present detectors for PCT are listed in Table III. Plastic detectors have the advantage o f a rapid decay time and therefore excellent coincidence time resolution, but their low overall and photopeak efficiency and poor energy resolution for 5 1 1-keV radiation limit their usefulness, as evidenced by the fact that no PCT system has been built using plastic scintillators. Multiwire chambers have excellent intrinsic spatial resolution but low intrinsic efficiency and poor energy discrimination and moderate to poor coincidence time resolution [60, 61 ]. Jeavons and co-workers [61 ] have, however, recently shown significant improvements in the time resolution o f their gas proportional chamber PCT system (i.e. 2 r = 40 ns). Germanium semiconductor detectors can provide excellent energy resolution, replacement o f light collection using bulky photo tubes with direct charge collection and compact electronics. Position-sensitive charge collection can also be used for identification o f the location o f events within relatively large detector arrays [62]. However, whereas the total interaction efficiency for 511 keV interactions is high in Ge, the photofraction is low (i.e. only about 5% o f the interactions occur by the photoelectric process). Selection of photopeak events to take advantage o f the high-energy resolution yields a low intrinsic efficiency. The material cost and liquid N 2 cooling is also a detriment. Although the detectors mentioned above should not be com pletely ruled out, it would appear that the best detector choices today are BGO, CsF and
218
PHELPS et al.
Nal(Tl). Each o f these materials has advantages and disadvantages. If one makes the assumption that for a PCT system to be competitive today, the detectors must have a width o f < 18 mm to provide a final image resolution o f < 10 mm, then the following comparison can be made. The lower intrinsic and photopeak efficiency o f Nal(Tl) and CsF, the fact that Nal(Tl) and CsF are both hygroscopic (CsF even more than Nal(Tl)) which may limit their lifetime, and reduction in packing density due to their protective can, and interdetector scatter o f radiation (this can be reduced by interdetector shielding at the expense o f geometric efficiency) are all disadvantages relative to BGO. The advantage of Nal(Tl) and CsF is their better coincidence time resolution [63, 64] compared with BGO(Table III). The advantage o f coincidence time resolution for Nal(Tl) and CsF may be somewhat artificial. If one considers that the important factor is the ratio of true coincidence to random coincidences, and that the magnitude o f random coincidences is determined by the detector singles count rate (Eq.(3)) and the detector time resolution (2 r), then one can determine an ‘effective time resolution’ for different detectors. The effective time resolution (ETR) will be 2 г multiplied by relative factors which affect the ratio o f true to random coincidences. We shall assume that the energy threshold for BGO is set at about 350 keV and CsF is at 100 keV, and that sufficient time integration or a fast-slow coincidence circuit is used such that the energy resolution of BGO is about 20 to 30% at 511 keV. We shall also use a value o f the system 2 r for CsF o f 3.5 ns (this value is chosen because o f realistic considerations for the system time resolution in large detector arrays o f PCT systems, and the fact that if shorter times are used the efficiency at the edges o f the FOV will decrease because o f the finite propagation time in the time o f flight o f the two 511 keV photons) and 20 ns for BGO (time resolution about 10 to 15 ns is clearly achievable with optimum crystal-phototube coupling). From simple 2 r time resolution, it would appear that CsF is 5.7 times better than BGO in terms o f random coincidences. However, for the small detectors given above, BGO has an increased true coincidence efficiency (including the higher packing density) o f 1.8 to 3.0 over CsF ([52], assuming CsF has about the same efficiency as Nal(Tl)). Thus the ETR o f BGO relative to CsF is now 11 to 6.7 ns instead o f 20. The higher energy threshold o f BGO also allows a reduction in singles count rate (the energy threshold could also be raised for CsF but its low photofraction would cause a significant reduction in efficiency). The magni tude o f this factor is hard to estimate because it depends on the particular tomographic design, but where it might be relatively small in a well-designed single-plane system, it can become significant in a multiplane system. A reasonable estimate o f this factor might be 1.5 in favour o f BGO. This would further reduce the ETR o f BGO to 7.3 to 4.4 ns. Thus it is possible that in a realistic system the ETR o f BGO could be similar to CsF (and with improved detector-phototube coupling even equal to CsF). O f course, if larger detectors are compared then the magnitude o f the above comparison diminishes.
IAEA-SM-247/203
219
Using the example above, BGO would be favoured because at a similar randoms fraction it would have (1) increased efficiency (1.8 to 3.0), (2) somewhat lower scattered coincidences due to the higher energy threshold, (3) lower cross talk between neighbouring detectors which causes mispositioning if both events are accepted or lower efficiency if rejected or absorbers placed between detectors (i.e. lower packing fraction), (4) and a more durable detector material. This same conclusion has been reached by Derenzo [52] in his figure o f merit, even though he used a time resolution o f 3.3 ns for CsF and 30 for BGO. A possible unique application of CsF is for combining time-of-flight (TOF) positioning with the mathematical image reconstruction technique. In this approach the TOF difference in each coincidence event would be measured and the back projection process in the image reconstruction weighted by the TOF information. Allemand and co-workers [63] have reported a time resolution for small CsF detectors o f 0.5 ns which is equivalent to about a 10-cm resolution. These investi gators estimate that this added low frequency information would improve the S/N ratio in the reconstructed image o f a 32-cm dia. cylinder o f uniform activity by a factor o f 2.9. Whereas this TOF spatial resolution would be o f little value for the brain, it may be significant in the torso. However, what realistic time resolution can be achieved in the large multiple coincidence arrays o f PCT systems is yet to be determined.
SPATIAL RESOLUTION If scatter and random coincidences have been minimized, and appropriate sampling and attenuation corrections are provided, then the determining physical parameters o f spatial resolution in the final images are detector size and statistical noise. Because o f the particular type o f error propagation in CT, large numbers o f counts are required to attain reasonable noise levels in the tomographic images, as shown in Table IV. It has also been shown that spatial resolution (R) in CT is inversely proportional to the number o f counts (N) as given by [65 ] R c t l/N 3
(8)
Equation (8) assumes that the noise level per resolution element is maintained at a constant level as R is improved. This relationship has limited resolution improve ments in both X-ray and emission CT since large increases in either efficiency or dose are required to make even small improvements in spatial resolution. There are two approaches to resolution improvement in PCT. The first is the conventional approach o f decreasing the size o f the detector and increasing the cutoff frequency (i.e. to recover high resolution) in the filter function o f the reconstruction algorithm. However, this increases the noise at a third power
220
PHELPS et al.
TABLE IV. STATISTICAL ERROR IN PCT [45]
Statistical error3
Number of counts required at different image resolutions*5 (cm)
(± 1 SD in %)
0.5
1
±5
1.7 X 108
2 .1
±10
±15
2
3
107
2.6 X 106
7.7 X 10s
4.5 X 107
5.6 ± 106
7.0 X 10*
2.1 X 10s
1.9 X 107
2.4 X 106
3.0 X 105
8.9 X 104
±
a Upper bound of statistical error. b For 20 cm dia. object and with linear sampling distance Дх = 1 /2 image resolution. If object diameter is increased to 30 cm then number of counts required must be increased by 3.4. Error analysis is an upper bound error limit and using a ramp filter function.
FR EQUENCY ( m m '1)
FIG .7.
The m o d u la tio n transfer fu n c tio n s (M T F ) fro m detectors w ith 3.5 and 12 mm F W H M
resolutions. N o te n o t o n ly the extension to higher sp a tia l frequencies w ith the 3.5 mm reso lu tio n d e te c to r b u t also the increased a m p litu d e a t the lo w e r frequencies. This a m p lific a tio n o f the lo w e r frequencies produces higher reso lu tio n in the reconstructed image even i f the c u t o f f freq u e n cy o f the reco n stru ction f ilt e r fu n c tio n f o r the
1 2
-mm d e tecto r is em ployed
(see te x t). Thus, the higher in trin s ic reso lu tio n allow s reso lu tio n im p ro ve m en t w ith o u t noise a m p lific a tio n caused by the increase in the c u to ff freq u e n cy in the reco n stru ction (i.e. noise a m p litu d e is dependent upon the re co n stru ction filte r) .
IAEA-SM-247/203
F IG . 8 .
221
R econstructed images fro m co m p u te r sim ulations o f a bar p hantom w ith 11-mm bar
w id th s and spacings. The re co n stru ctio n f ilt e r fu n c tio n is the same f o r each image and had a c u to ff freq u e n cy equivalent to the recovery o f 1.2-cm resolution. A lth o u g h , images were reconstructed w ith o u t noise added, the noise w o u ld be the same in each image since the same reco n stru ction f ilt e r fu n c tio n was em ployed. Images illu stra te the d ra m a tic im p ro ve m en t in spatial reso lu tio n by s im p ly increasing the in trin s ic d e te cto r reso lu tio n (value shown in upper le ft hand corner o f each image). This approach, using high in trin s ic d e te cto r reso lu tio n , allows sig n ific a n t resolution im p ro ve m en t w ith o u t increasing noise a t the th ird p o w e r o f the resolution im p ro ve m en t (see te xt).
rate (Eq.(8)) and therefore dramatically increases the number o f counts required to obtain statistically valid images at the higher resolution (Table IV). Another approach suggested by Phelps and co-workers (unpublished data) is to decrease the size o f the individual detectors or detection elements to limits determined by efficiency and cost, and not to increase the cutoff frequency in the reconstruction algorithm. The effect o f the small individual detector elements is not only to extend the frequency content of the detector signal to higher values but also to increase the amplitude o f the lower frequencies, as shown in Fig. 7. These data are then used to reconstruct the image without raising the cutoff frequency o f the reconstruction filter function (i.e. even though the detector may have an intrinsic resolution o f a few mm, the cutoff frequency would be set to recover up to say > 1 0 mm; this approach will yield a significant image resolution improve ment without increasing image noise; this allows resolution improvement when the image is statistically limited (i.e. the typical limitation). An example o f this approach is shown in Fig.8.
222
PHELPS et al.
The implementation o f this technique requires a detector material that can maintain a high efficiency when the size is very small. It would appear that BGO meets this requirement. This concept also increases the value of the high intrinsic resolution o f the multiwire gas chambers, but it would appear that until their intrinsic efficiency is increased they are still behind the other detector choices. The tomographic system would also have to be able to provide linear sampling distances < 1/2 the intrinsic detector resolution even though the reconstructed image resolution would be considerably worse than the intrinsic detector resolution. Spatial resolution in the final image with the present operational tomographs is typically about 15 mm (full width half maximum (FWHM) o f a line source) in the image plane and from 16 to 30 mm in the axial resolution when employed in actual in vivo human studies. In special cases, spatial resolution o f about 10 mm has been demonstrated [30, 66, 67]. As discussed above, a major limitation has been from the high statistical requirements o f PCT. Specially designed brain units, and the Donner unit [32, 51 ] with the change over from Nal(Tl) to BGO should have the capability o f providing final image resolutions o f about 10 mm or slightly better on a routine basis. Dramatic improvement in resolution beyond 8 mm seems unrealistic unless methods can be employed to circumvent the inverse third power relationship in Eq.(8) (e.g. by the above method or TOF). To make things even worse, improvements in the axial resolution are at the cost o f a reduction in efficiency by 1/(axial resolution)2. Thus the total tomographic resolution is inversely related to the fifth power o f efficiency (remembering that efficiency increases with concomitant increases in noise from scatter and random coincidences, dead time effects and other sources may not represent a significant increase in true efficiency).
CONCLUSION PCT system design is com plex, and involves optimization and tradeoffs in numerous design parameters. A simple characterization o f a system in terms of spatial resolution by the FWHM o f a line source and gross efficiency can be extremely misleading. Spatial resolution (i.e. degree o f spatial frequency recovery, not simply a FWHM estimate) is a variable used to increase signal (S) and to extend spatial frequency recovery. Efficiency is a variable to decrease noise (N). Tradeoffs between these two variables are usually required to derive an optimum signal-tonoise (S/N ) in the final image. There are, however, many other factors which affect S/N and must be taken into account in PCT system design, as discussed above and elsewhere [45] (i.e. random and scatter coincidences, spatially variant resolution and efficiency, sampling, attenuation correction, non-uniform detector response, patient, organ and activity movement, type o f study to be performed, recon struction filter function and others). When these factors produce noise comparable
IAEA-SM-247/203
223
with that of limitations from intrinsic spatial resolution and statistical noise, then their noise contributions must be reduced to allow significant improvements in overall system S/N (i.e. further improvements in efficiency and spatial resolution will not significantly improve the image S/N). This must also be done to provide analytical measurement capabilities as opposed to simply providing a qualitative imaging device. Thus system evaluation and characterization must be carried out under realistic conditions and by multiparameter analysis. Many o f the relationships between the variables in PCT system design are now being characterized in detail [49, 50, 52, 54, 5 6—58], and initial efforts to develop a figure of merit in which some o f these factors are consolidated into a single equation are being made by Derenzo [52]. However, more fundamental work is needed to understand better each o f the variables and how to judge the necessary tradeoffs required in a final system design. Many o f these tradeoffs require definition o f priorities in the types o f studies for which the PCT system will be used (i.e. highest possible spatial resolution with a static radioactivity distribution versus poorer spatial resolution with the highest possible dynamic scan capability). Thus system design efforts must go hand in hand with the development of the application areas (i.e. primary organs of interest, types of labelled compounds, types o f tracer kinetic models, etc.). At present, circumferential PCT systems (hexagon, octagon and circle) with BGO detectors are the most common and successful. However, the choice of detector size, ring diameter and shielding design (particularly in multiplane systems) is still quite variable in different systems (i.e. Hoffman and co-workers3). The actual image resolution and true system efficiencies have not been well characterized for many of the systems that have been or are being developed. Therefore these were not listed for comparison in this survey. However, when these comparisons are made they should be from a signal-to-noise point o f the view using tests conducted under realistic conditions in w hich they will be used, rather than from a simple measurement o f spatial resolution at the centre o f the image plane and gross efficiency as discussed above. These tests can be performed with phantoms to assess the positional and frequency dependence and magnitude o f spatial resolution in the reconstructed images, the image signal-to-noise, scatter and random fractions, count-rate capabilities and efficiency. Phantoms for this purpose have been discussed by Phelps and co-workers [31, 45], Derenzo and co-workers [33], Hoffman and co-workers [54, 56], and Huang and co-workers [57, 58]. With multiplane systems, phantoms need to be designed that will also allow the determination o f the variation in axial resolution within the field of view o f the midplane. 3 HOFFMAN, E.J., SUNG-CHENG HUANG, PLUMMER, D.L., PHELPS, M.E., KUHL, D.E., “Influence of non-uniform resolution on image quality and quantitation in positron-emission-computed tomography (PCT)” , these Proceedings, IAEA-SM-247/92 (Poster Presentation).
224
F IG .9 .
PHELPS et al.
M eta b o lic stu d y w ith 18F -2-fluoro-2-deoxyglucose show ing the response o f the p rim a ry
(PVC ) and associative (A VC) visual co rte x fro m eyes closed c o n tro l to w h ite lig h t stim u la tio n (570 Ix ) and to a com plex visual scene o f a park. The c o n tro l and w h ite lig h t stim u la tio n are the same subject whereas the com plex stim u la tio n was a d iffe re n t subject (i.e. levels fro m th ird subject do n o t exa ctly correspond to the images in colum ns 2 and 3 and the sketches. N o te the increase in m etabolic o r fu n c tio n a l a c tiv ity o f the visual co rte x fro m the quiescent state o f eyes closed to increasing c o m p le x ity o f the visual s tim u la tio n . A lso note th a t in the eyes closed stu d y one can clearly see the de lin e a tion o f the p o sterio r-p a rie tal co rte x fro m the Brodm an area 19 o f the associative visual cortex. F ro m R ef. [65].
Careful system design and evaluation is necessary not only to determine the true tomographic resolving power o f a particular PCT but also to establish its measurement accuracy. This latter factor is essential since one o f the greatest facilities of PCT is the capability to perform a measurement o f local organ function that has not been possible in the past. Thus a PCT must be both an imaging and analytical measurement device.
IAEA-SM-247/203
225
The role o f CsF as a fast scintillator and the practical importance o f its use in time-of-flight remains to be clearly determined. The successful applications o f PCT, primarily in the brain and heart, have provided the necessary incentive for the further development o f PCT systems (see reviews by Phelps [68], Schelbert and co-workers [69], Ter-Pogossian [70], Wolf [71], and Table I and Fig.9); Alavi and co-workers [72], Brownell [73], Kirchner [7 4 ].4 It is not easy to predict the changes that will occur in this field during the next decade but it is clear that it has now advanced to the point that it will be given a chance to see whether it has the character and value to succeed.
REFERENCES [1] PHELPS, M.E., HOFFMAN, E.J., HUANG, S.C., et al., “Positron tomography: In vivo autoradiographic approach to measurement of cerebral hemodynamics and metabolism” , Cerebral Function, Metabolism and Circulation (INGVAR, D.H., LASSEN, N.A., Eds), Munksgaard, Copenhagen ( 1977) 192. [2] PHELPS, M.E., HOFFMAN, E.J., KUHL, D.E., “Physiologic tomography (PT): A new approach to in vivo measure of metabolism and physiologic function” , Medical Radio nuclide Imaging (Proc. Symp. Los Angeles, 1976) I, IAEA, Vienna (1977) 233. [3] WRENN, E.R., GOOD, M.L., HANDLER, P., The use of positron emitting radioisotopes for the localization of brain tumors, Science 113 (1951) 525. [4] SWEET, W.H., Uses of nuclear disintegrations in the diagnosis and treatment of brain tumors, New Engl. J. Med. 24 (1951) 875. [5] BROWNELL, G.L., SWEET, W.H., Localization of brain tumors with positron emitters, Nucleonics 11 (1953) 40. [6 ] ANGER, H.O., ROSENTHAL, D.J., “Scintillation camera and positron camera”, Medical Radioisotope Scanning (Proc. Seminar Vienna, 1959), IAEA, Vienna (1959) 59. [7] RANKOWITZ, S., ROBERTSON, J.S., HIGINBOTHAM, W.A., et al., Positron scanner for locating brain tumors, I.R.E. Int. Conv. Rec. 10 9 (1962) 49.
4 See also: KUHL, D.E., PHELPS, M.E., ENGEL, J., Jr., “Emission computed tomo graphy of fluorine-18-fluorodeoxyglucose and nitrogen-13-ammonia in stroke and epilepsy”, these Proceedings, IAEA-SM-247/88; JONES, T., RHODES, C.G., HEATHER, J.D., FORSE, G.R., LAMMERTSMA, A.A., FRACKOWIAK, R.S.F., LENZI, G.L., SELWYN, A.P., ALLAN, R.M., BUCKINGHAM, P.D., “The ability to measure quantitatively in three dimensions regional tissue metabolism and blood flow using oxygen-15 and a positron emission tomograph” , these Proceedings, IAEA-SM-247/23; PARTAIN, C.L., PRICE, R.R., JAMES, A.E., Jr., HÜBNER, K.F., MAHALEY, M.S., SCATLIFF, J.H., ROBERTSON, J.T., “Positron emission tomography: Brain tumour imaging using carbon-11 amino acids”, these Proceedings, IAEASM-247/79 (Poster Presentation); MYERS, W.G., BENUA, R.S., YEH, S.D.J., BIGLER, R.E., GRAHAM, M.C., DAHL, J.R., LEE, R„ REIMAN, R.E., BADING, J.R., LAUGHLIN, J.S., “Kinetics of potassium-38 in heart intercompared by means of a positron-electron transmutation [PET] tomograph and a rectilinear high-energy gamma (HEG) scanner”, these Proceedings, 1AEA-SM-247/82; WAGNER, H.N., Jr., “New perspectives in nuclear imaging” , these Proceedings, IAEA-SM-247/200.
2 26
PHELPS et al.
[8 ] ROBERTSON, J.S., MARR, R.B., ROSENBLUM, B., et al., “Thirty-two crystal positron transverse section detector”, Tomographic Imaging in Nuclear Medicine (FREEDMAN, G.S., Ed.), New York Society of Nuclear Medicine (1973) 142. [9] KUHL, D.E., EDWARDS, R.Q., Image separation radioisotope scanning, Radiology 80 (1963) 653. [ 1 0 ] KUHL, D.E., EDWARDS, R.Q., RICCI, A.R., et al., Quantitative section scanning using orthogonal tangent correction, J. Nucl. Med. 14 (1973) 196. [ 1 1 ] ANGER, H.O., “The scintillation camera for radioisotope localization”, Radioisotopen in der Lokalizations-Diagnostik (HOFFMAN, G., SCHEER, K.E., Eds), Stuttgart, E.K. Schattauer-Verlag (1967) 18. [ 1 2 ] MUEHLLEHNER, G., WETZEL, R.A., Section imaging by computer calculation, J. Nucl. Med. 12(1971) 79. [13] TODD-POKROPEK, A.E., “The formation and display of section scans” , Proc. Symp. American Congress of Radiology, 1971, Excerpta Medica, Amsterdam (1972) 545. [14] MYERS, M.J., KEYES, W.I., MALLARD, J.R., “An analysis of tomographic scanning systems”, Medical Radioisotope Scintigraphy (Proc. Symp. Monte Carlo, 1972) 1, IAEA, Vienna (1973) 331. [15] BOWLEY, A.R., TAYLOR, C.G., CAUSER, D.A., et al., A radioisotope scanner for rectilinear arc, transverse sections and longitudinal section scanning (ASS—The Aberdeen Section Scanner), Br. J. Radiol. 46 (1973) 262. [16] TANAKA, E., “Multi-crystal section imaging device and its data processing”, Proc. Thirteenth Congress of Radiology, Madrid, 1973, Excerpta Medica, Amsterdam (1973) 81. [17] ANGER, H.O., “Multiple plane tomographic scanner”, Tomographic Imaging in Nuclear Medicine (FREEDMAN, G.S., Ed.), Society of Nuclear Medicine, New York (1973) 2. [18] PHELPS, M.E., Emission computed tomography, Semin. Nucl. Med. 7 (1977) 337. [19] KUHL, D.E., REIVICH, M., ALAVA, A., et al., Local cerebral blood volume determined by three-dimensional reconstruction of radionuclide scan data, Circ. Res. 36 (1975) 610. [ 2 0 ] KUHL, D.E., EDWARDS, R.Q., RICCI, A.R., et al., The MARK IV system for radionuclide computed tomography of the brain, Radiology 21 (1976) 405. [ 2 1 ] GORDON, R., HERMAN, G.T., Three-dimensional reconstruction from projections: A review of algorithms, Int. Rev. Cytol. 38 (1974) 11. [2 2 ] BUDINGER, T.F., GULLBERG, G.T., Three-dimensional reconstruction in nuclear medicine emission imaging, IEEE Trans. Nucl. Sei. NS-21 3 (1974) 2. [23] CHESLER, D.A., “Positron tomography and three dimensional reconstruction techniques” , Tomographic Imaging in Nuclear Medicine (FREEDMAN, G.S., Ed.), Society of Nucl. Med., New York (1973) 176. [24] PHELPS, M.E., HOFFMAN, E.J., MULLANI, N.A., et al., Applications of annihilation coincidence detection to transaxial reconstruction tomography, J. Nucl. Med. 16 (1975) 210. [25] TER-POGOSSIAN, M.M., PHELPS, M.E., HOFFMAN, E.J., et al., A positron emission transaxial tomograph for nuclear medicine imaging (PETT), Radiology 114 (1975) 89. [26] PHELPS, M.E., HOFFMAN, E.J., MULLANI, N.A., et al., “Transaxial emission recon struction tomography: Coincidence detection of positron-emitting radionuclides”, Non-Invasive Brain Imaging, Radionuclides and Computed Tomography (DeBLANC, H., SORENSON, J.A., Eds), Society of Nucl. Med., New York (1975) 87. [27] HOFFMAN, E.J., PHELPS, M.E., MULLANI, N.A., et al., Design and performance characteristics of a whole body positron transaxial tomograph, J. Nucl. Med. 17 (1976) 493. [28] PHELPS, M.E., HOFFMAN, E.J., MULLANI, N.A., et al., Design considerations for a positron emission transaxial tomograph (PETT III), IEEE Trans. Nucl. Sei. NS-23 (1976) 516.
IAEA-SM-247/203
227
CHO, Z.H., CHAN, J.K., ERIKSSON, L., Circular ring transverse axial positron camera for 3-dimensional reconstruction of radionuclides distribution, IEEE Trans. Nucl Sei. NS-23 1 (1976)613. CHO, Z.H., COHEN, M.B., SINGH, M., et al., Performance and evaluation of the circular ring transverse axial positron camera (CRTAPC), IEEE Trans. Nucl. Sei. NS-24 1 (1977) .530. PHELPS, M.E., HOFFMAN, E.J., HUANG, S.C., et al., ECAT: A new computerized tomographic imaging system for positron emitting radiopharmaceuticals, J. Nucl. Med. 19 (1978)635. BOHM, Chr., ERIKSSON, L., BERGSTROM, M., et al., A computer assisted ring detector positron camera system for reconstruction tomography of the brain, IEEE Trans. Nucl. Sei. NS-25 1 (1978) 624. DERENZO, S.E., BUDINGER, T.F., CAHOON, J.L., et al., The Donner 280 crystal high resolution positron tomograph, IEEE Trans. Nucl. Sei. NS-26 (1979) 2790. YAMAMOTO, Y., THOMPSON, C.J. MEYER, E., et al., Dynamic positron emission tomography for study of cerebral hemodynamics in a cross-section of the head using positron-emitting n Ga-EDTA and 77 Kr, J. Comput. Assist. Tomogr. 1 (1977) 43. THOMPSON, C.J., YAMAMOTO, Y.L., MEYER, E., Positome II: A high efficiency positron imaging system for dynamic brain studies, IEEE Trans. Nucl. Sei. NS-26 (1979) 582. CHO, Z.H., FARUKHI, S., New bismuth germanate scintillation crystal —a potential detector for the positron camera application, J. Nucl. Med. 18 (1977) 840. BROWNELL, G.L., BURNHAM, C.A., CHESLER, D., et al., “Transverse section imaging of radionuclide distributions in heart, lung and brain” , Reconstruction Tomography in Diagnostic Radiology and Nuclear Medicine (.TER-POGOSSIAN, M.M., PHELPS, M.E., BROWNELL, G.L., Eds), University Park Press, Baltimore (1977) 293. MUEHLLEHNER, G., BUSCHIN, M.P., DUDEK, J.H., Performance parameters of a positron imaging camera, IEEE Trans. Nucl. Sei. NS-23 1 (1976) 528. MUEHLLEHNER, G„ ATKINS, F., HARPER, P.V., Positron camera with longitudinal and transverse tomographic capabilities (Proc. Symp. Los Angeles, 1976) 1, Medical Radionuclide Imaging, IAEA, Vienna (1977) 291. TER-POGOSSIAN, M.M., MULLANI, N.A., HOOD, J., et al., A multislice positron emission computed tomograph (PETT IV) yielding transverse and longitudinal images, Radiology 128(1978) 477. TER-POGOSSIAN, M.M., MULLANI, N.A., HOOD, J.T., et al., Design considerations for a positron emission transverse tomograph (PETT V) for imaging of the brain, J. Comput. Assist. Tomogr. 2 15 (1978) 539. BROOKS, R.A., SANK, V.M., DiCHIRO, G., et al., Design of a high resolution positron emission tomograph: The Neuro-Pet., J. Comput. Assist. Tomogr. 4 (1980) 5. TANAKA, E., NOHARA, N., TOMITANI, T., et al., A rotary positron emission tomo graph: “Positologica”, J. Nucl. Med. 21 (1980) 15 (abstract). UEMURA, K., KANN, I., MIURA, S., et al., Headtome: A new hybrid emission tomo graphy and its applications for brain study, J. Nucl. Med. 21 (1980) 15 (abstract). PHELPS, M.E., HOFFMAN, E.J., HUANG, S.C., et al., Design considerations in positron computed tomography (PCT), IEEE Trans. Nucl. Sei. NS-26 (1979) 2746. HOFFMAN, E.J., PHELPS, M.E., An analysis of some of the physical aspects of positron transaxial tomography, Comput. Biol. Med. 6 (1976) 345. PHELPS, M.E., HOFFMAN, E.J., MULLANI, N.A., et al., “Some performance and design characteristics of PETT III”, Reconstruction Tomography in Diagnostic Radiology and
228
[48]
[49]
[50]
[51 ] [52] [53]
[54] [55]
[56]
[57]
[58]
[59]
[60] [61]
PHELPS et al.
Nuclear Medicine (TER-POGOSSIAN, M.M., PHELPS, M.E., BROWNELL, G.L., Eds), University Park Press, Baltimore (1977) 371. HOFFMAN, E.I., PHELPS, M.E., RICCI, A.R., et al., “Optimization of system design parameters for emission computed tomography”, Frontiers of Engineering in Health Care, IEEE (79CH1440-7), New York (1979) 363. DERENZO, S.E., ZAKLED, H., BUDINGER, T.F., Analytical study of high-resolution positron ring detector system for transaxial reconstruction tomography, J. Nucl. Med. 7 (1977) 337. DERENZO, S.E., Optimization of Side Shielding for Circular Positron Emission Tomographs, University of California, Lawrence Berkeley Laboratory, Report 8622 (January 1979). BUDINGER, T.F., DERENZO, S.E., GULLBERG, G.T., et al., Trends and prospects for circular ring positron cameras, IEEE Trans. Nucl. Sei. NS-26 (1979) 2742. DERENZO, S.E., Method for optimizing side shielding in positron emission tomographs and comparing detector materials, J. Nucl. Med. (in press). BECK, R.N., ZIMMER, L.T., CHARLESTON, D.B., et al., “Advances in fundamental aspects of imaging systems and techniques” , Medical Radioisotope Scintography (Proc. Symp. Monte Carlo, 1972) 1, IAEA, Vienna (1973) 3. HOFFMAN, E.J., HUANG, S.C., PHELPS, M.E., Quantification in positron emission computed tomography: 1. Effect of object size, J. Comput. Assist. Tomogr. 3 (1979) 299. DERENZO, S.E., VULETICH, T., Scintillation Performance of Bismuth Germanate Crystals at Low Temperatures, University of California, Lawrence Berkeley Laboratory, Report 47466 (June 1978). HOFFMAN, E.J., HUANG, S.C., PHELPS, M.E., et al., Quantification in positron emission computed tomography: 4. Effect of accidental coincidences, J. Comput. Assist. Tomogr. (in press). HUANG, S.C., HOFFMAN, E.J., PHELPS, M.E., et al., Quantification in positron computed tomography. 2. Effects of inaccurate attenuation correction, J. Comput. Tomogr. 3 (1979) 804; HUANG, S.C., PHELPS, M.E., CARSON, R., et al., An improved method for the attenuation correction, J. Nucl. Med. (in press). HUANG, S.C., HOFFMAN, E.J., PHELPS, M.E., et al., Quantification in positron emission computed tomography: 3. Effect of sampling, J. Comput. Assist. Tomogr. (in press). SNYDER, D.L., COX, J.R., Jr., “An overview of reconstructive tomography and limitations imposed by finite number of projections” , Reconstruction Tomography in Diagnostic Radiology and Nuclear Medicine, (TER-POGOSSIAN, M.M., PHELPS, M.E., BROWNELL, G.L., Eds), University Park Press, Baltimore (1977) 3. LIM, C.B., CHU, D., KAUFMAN, L., et al., Initial characterization of a multiwire proportional chamber positron camera, IEEE Trans. Nucl. Sei. NS-22 1 (1975) 388. JEAVONS, A., TOWNSEND, D., FREY, P., DONATH, A., “A large-area stationary positron camera using wire chambers” , these Proceedings, IAEA-SM-247/65.
[62] KAUFMAN, L., WILLIAMS, S.H., HOSIER, K., et al., An evaluation of semiconductors for positron tomography, IEEE Trans. Nucl. Sei. NS-26 (1979) 648. [63] ALLEMAND, R., GRESSET, C., VOCHER, J., Potential advantages of a cesium fluoride scintillator for a time of flight positron camera, J. Nucl. Med. 21 (1980) 153. [64] MULLANI, N.A., FICHE, D.C., TER-POGOSSIAN, M.M., Cesium fluoride: a new detector for positron emission tomography, IEEE Trans. Nucl. Sei. NS-27 (1980) 572. [65] PHELPS, M.E., HOFFMAN, E.J., GADO, M., et al., “Computerized transaxial transmission reconstruction tomography”, Non-Invasive Brain Imaging, Computer Tomography and
IAEA-SM-247/203
[6 6 ] [67] [6 8 ] [69] [70] [71] [72] [73] [74]
229
Radionuclides (DE BLANC, H., SORENSON, J., Eds), Society of Nuclear Medicine, New York (1975) 111. BROWNELL, G.L., ACKERMAN, R.H., STRAUSS, H.H., et al., Preliminary imaging results of 18 F-2-fluoro-2-deoxy-D-glucose, J. Comput. Assist. Tomogr. 4 (1980) 437. BUDINGER, T.F. (private communication). PHELPS, M.E., Positron computer tomography studies of cerebral glucose metabolism in man: Theory and application in nuclear medicine, Semin. Nucl. Med. (in press). SCHELBERT, H.R., HENZE, H„ PHELPS, M.E., Emission tomography of the heart, Semin. Nucl. Med. (in press). TER-POGOSSIAN, M.M., Special characteristics and potential for dynamic function studies with PETT, Semin. Nucl. Med. (in press). WOLF, A., Special characteristics and potential of radiopharmaceuticals for emission tomography, Semin. Nucl. Med. (in press). ALAVI, A., REIVICH, M., GREENBERG, J., et al., Mapping of functional activity in the brain with F-18-fluoro-deoxyglucose, Semin. Nucl. Med. (in press). BROWNELL, G.L., Special characteristics and potential of transverse section imaging and peripheral vascular disease, Semin. Nucl. Med. (in press). KIRCHNER, P., Special characteristics and potential of the role of positron emission computerized tomography in retroperitoneal disease, Semin. Nucl. Med. (in press).
DISCUSSION A.E. TODD-POKROPEK: Would you like to comment on the use o f ‘timeof-flight’ techniques such as described by Allemand and co-workers? 1 M.E. PHELPS: These researchers have shown a time resolution o f 0.5 ns for CsF, which, in view o f Allemand’s expertise, probably represents the best possible value. This corresponds to a spatial resolution o f about 7 to 10 cm. Whereas this value appears to be high compared with the desired < 1 cm resolution in PCT, it does provide additional information to constrain the mathematical image reconstruction. At present, the reconstruction algorithms are typically unbounded and unconstrained with no information concerning the distribution o f data within each line integral. To add even the low frequency information o f time o f flight (TOF) to the line integrals would provide a reduction in image noise. Obviously, the better the TOF spatial resolution the more noise reduction will occur. The actual time resolution, spatial resolution and noise reduction with PCT system designs for different image resolution and object frequency distributions need to be further investigated. T. JONES: You mentioned that in multislice positron ECATs, the design o f the inter-plane septa still needs to be improved in order to achieve quantitation in the between-slice planes. Since there are relatively few options open for this, could you comment on what could be achieved? 1 ALLEMAND, R., GRESSET, C., VACKER, J., Potential advantages of a cesium fluoride scintillator for a time of flight positron camera, J. Nucl. Med. 21 (1980) 153.
230
PHELPS et al.
M.E. PHELPS: My point was really that system optimization for multiplane systems is different from that for single plane systems, and a focal point o f this optimization is the shielding and septa design. At present, most systems represent stacks o f single planes. The conflict between thick long septa to reduce randoms and scatter, and short knife-edge septa to reduce axial resolution variation, needs to be better appreciated, for instance by examining the compromises resulting from changes in ring diameter, crystal packing, septa geometry, detector materials and the number o f planes. Many of these factors are discussed in the text and in the poster presentation by Hoffman and co-workers.2 I think the true number of options and their relative merits have yet to be determined.
2 HOFFMAN, E.J., SUNG-CHENG HUANG, PLUMMER, D.L., PHELPS, M.E., KUHL, D.E., “Influence of non-uniform resolution of image quality and quantitation in positron-emissioncomputed tomography (PCT)”, these Proceedings, IAEA-SM-247/92 (Poster Presentation).
IAEA-SM-247/139
P O T E N T I A L A N D LIMITS O F Q U A N T I T A T I V E S T U D I E S IN EMISSION T O M O G R A P H Y F. SOUSSALINE, A.E. TODD-POKROPEK* D. COMAR, C. RAYNAUD, C. KELLERSHOHN S.H.F J . Département de Biologie, C.E.A., Hôpital d’Orsay, • Orsay, France
Abstract POTENTIAL AND LIMITS OF QUANTITATIVE STUDIES IN EMISSION TOMOGRAPHY. Quantitative extraction of scintigraphic data such that absolute measurements of regional activity and functional or metabolic parameters of regional activity can be greatly improved by use of emission computer tomography. However, the principal factors influencing the precision of the results must be analysed. Thus, the (positron emission computer tomograph) ORTECECAT system and the (single-photon computer tomograph) GE 400T rotating gamma camera system interfaced to an Informatek-Simis 3 computer system were tested. The overall sensitivity and, correspondingly, the signal-to-noise ratio, the non-stationarities of the system transfer function (STF), the attenuation correction, the tail effects (shape and amplitude of the STF) and the contrast variation as a function of the relative dimension of object structures, have been successively studied using specially designed phantoms. Such an analysis is needed to establish the corrections required to improve the precision of quantitative data and to design system and clinical protocols where the sources of such errors are minimized.
1.
INTRODUCTION
For many years, the quantitative use o f scintigraphic data has been sought. Major problems have been the elimination o f over- and underlying activity, and attenuation correction. It appears that isotope tomography can provide data for which these problems are considerably reduced. Two general classes o f tomographic systems will be considered here. They are called: SPECT - Single-Photon Emission Computer Tomography, and PECT — Positron Emission Computer Tomography. Quantitation, in this context, is used to imply a measurement either o f activity in a region which may be expressed in /uC icm -1 (MBqÉI -1) or a para meter such as volume, flow rate, metabolic extraction, which can be expressed in absolute terms.
* Also at Faculty of Clinical Sciences, University College, London, United Kingdom.
231
232
SOUSSALINE et al.
Whereas considerable progress in the use o f tomographic systems has been achieved [1—5], and good clinical results obtained, a number o f factors influencing precision need to be studied. These can be grouped as follows: (1 ) The overall sensitivity o f the system, and thereby the statistical noise. (2) The spatial variation o f the system transfer function (STF) and other such non-stationarity. (3) Attenuation. (4) The tail effects o f the STF (shape and amplitude at distance o f the STF resulting from Compton scatter in the object, and from the actual reconstruction algorithm used). (5) Contrast variation related to the relative dimension o f object structures and the spatial resolution, the signal-to-noise ratio and the tail effects. Although inter- and intra-experimental precision ( l /ст2) can both be defined, it is the former, the ability to obtain good quantitative estimates from a series of different tomograms, that seems most important. Intra-experimental precision is really only an estimate o f linearity [ 1 ]. The influence o f each o f these classes o f factors in limiting the potential of quantitation will be treated successively.
2.
MATERIAL
Practical results were obtained using two systems. A positron PECT system, the ORTEC-ECAT II, which has been described elsewhere [6,7 ], was used for obtaining phantom and patient data [8]. Standardized phantoms, such as proposed by Phelps and co-workers [7 ,9 ], as well as certain others specially designed, were employed. A single-photon SPECT system, the GE 400T rotating gamma camera connected to an Informatek-Simis 3 computer, was also tested, patient and phantom results also being obtained [ 10 , 1 1 ]. The former system has a transverse FWHM o f s= 14 mm, a slice thickness of = 19 mm, and a sensitivity o f = 13 000 counts/s juCi-1 • cm -1 with a 19-cm cylinder.1 The latter system has a transverse FWHM o f 16 mm, a slice thickness FWHM o f = 19 mm, sensitivity o f = 5000 counts/s /xCi-1 • cm-1, up to 64 slices being obtained simultaneously. The actual values o f these characteristics are strongly dependent on the protocol used for the measurement. Standard clinical values have been quoted. 3.
SENSITIVITY: SIGNAL-TO-NOISE RATIO
It is very important to realize that in tomography the total number o f events recorded per second does not necessarily define the signal-to-noise ratio. Not all 1 1 Ci = 3.70 X 1010 Bq.
IAEA-SM-247/139
F IG .l.
233
T om ographic s ta tistica l precision. The fra c tio n a l standard d e viation (F S D j is p lo tte d
as a fu n c tio n o f to ta l counts N j f o r various values o f the num ber o f pixe ls n p and also fo r co n ventional imaging.
counts are useful. Random (and Compton) events in PECT and Compton events in SPECT need to be eliminated. In PECT, system performance is determined by the true coincidence rate much more than by the total coincidence rate. However, provided the count rates are not too high, counting delayed coincidences to estimate the random rate, and subtracting from the total coincidence rate, gives a good approximation to the true coincidence rate. Time o f flight techniques [ 1 2] can significantly improve the way in which the system firstly selects true coincidences, and secondly establishes their origin, both o f which should enhance image quality. The expression [13]:
234
SOUSSALINE et al.
where FSD is the fractional standard deviation per pixel, К a constant depending on the filter and the pixel size used, nP the number o f pixels in the reconstructed object, and NT the total number o f counts detected, seems to describe well the signal-to-noise ratio in SPECT and PECT systems. The sensitivity defines NT, and therefore as image quality is proportional to 1 /v Ñ j f o r conventional images. The fractional standard deviation is, however, generally worse in tomograms, necessitating a larger number o f counts to be collected to obtain the same signal-to-noise ratio, as illustrated in Fig. 1. Statistical quality can be improved by the time o f flight techniques for PECT; correct design o f filters, and specifically in the choice o f window function; matching o f pixel size to the object investigated; and by increasing the solid angle, e.g. two-headed SPECT systems [14]. Compton rejection, with SPECT, may be improved by using a collimator with longer denser septa (e.g. Tantalum tubes) and by offsetting slightly the energy window towards higher energies. This improves significantly the contrast and spatial resolution, but should be used carefully as it can produce important ring artefacts due to the non-uniformity o f the camera. With PECT, Compton events should if possible be minimized rather than just subtracted. This is achieved by a good detection system design in terms o f geometric discrimination o f scatter.
4.
NON-STATIONARITIES OF THE SYSTEM RESPONSE FUNCTION
In PECT, the hexagonal sampling geometry and the RT (rotate translate) motion o f the ECAT allows the system response function to be reasonably invariant and isotropic (less than 10% FWHM variation in the 40-cm diameter field). This is not the case in PECT systems with ring geometry, because of irregular spatial sampling, unless improved by a wobbling motion [15]. Longi tudinal (Z) variation o f the system response function is also minimized as a result o f the geometry and the ‘shadow shields’ (collimator or gap) used with the ECAT (less than a 10% change from the centre to the edge o f the field). In SPECT, specifically when considering the use o f a rotating gamma camera such as the GE 400T, the non-uniformity o f the system response in the tom o graphic field, the non-linearities (differential and integral) can introduce important artefacts in the tomographic quality and the ability to extract quantitative data. This has been studied in detail, in terms o f noise-power spectra amplification, and has been presented elsewhere [16]. Figure 2 illustrates the effect o f the nonuniform sensitivity o f the camera, for a uniform source when reconstructing a tomographic image. It can be noted that important ring artefacts were created when uniformity correction was not performed, centred on the rotation axis, with amplitude increasing ~ as the inverse o f the artefact distance from the axis.
IAEA-SM-2 47/139
F IG .2.
235
The e ffe c t o f n o n -u n ifo rm ity when reco n stru cting a u n ifo rm source w ith a SPECT
system. On the le ft the results obtained w ith o u t u n ifo rm ity co rre ctio n are com pared w ith (on the rig h t) the results a fte r u n ifo rm ity correction.
F IG .3.
The e ffe c t o f atte nu a tio n . A u n ifo rm source is shown a fte r reco n stru ction , above as
an image, below as a p ro file through the image, on the le ft f o r a P E C T system, and on the rig h t f o r a SPECT system.
236
F IG .4.
SOUSSALINE et al.
The P E C T image and p ro file as shown in Fig. 3, is shown on the le ft before correction
f o r a tte nu a tio n , and on the rig ht, a fte r correction f o r attenuation.
F IG .5.
A p ro file through a series o f fiv e p o in t sources in scattering m edium is shown (lo w e r
curve J a fte r filte re d back-projection reconstruction, together w ith the p ro file o b taine d a fte r use o f the a tte nu a tio n -co rrectin g R IM a lg o rith m (upper curve).
IAEA-SM-247/139
5.
237
ATTENUATION
The amplitude o f the attenuation effect, for comparable geometry o f the radioactive distribution, and the ability to correct accurately for this effect, introduces a definitive distinction between SPECT and PECT. As illustrated in Fig. 3, the apparent attenuation o f 140 keV single photons ( 99Tcm) in a 20-cm diameter uniformly filled cylinder is significantly less important than for 511 keV coincident photon emission (68Ga) with the same geometry f 17] (contrary to the ratio o f their attenuation coefficients). In PECT, attenuation correction can, in principle, be performed exactly. The problems here are essentially statistical. The line integrals o f attenuation correction factors need to be defined precisely, either analytically (finding the edge o f the object and assuming values for p), or by measurement using an external source [6]. Considerable amplification o f the noise can occur, and care must be taken. For quantitative studies, correction is performed in standard conditions using transmission data with and without the object in the field. With a ‘measured correction’ for a 1 M count image o f an extended source after a transmission scan containing 4M counts (as shown in Fig. 4), the FSD in the reconstructed image was found to be 12%. With ‘analytical’ correction, errors in the integral over the source distribution, resulting from mismatching o f either the object and the outline, or the value o f /л ,can be as large as 20%. With SPECT systems, two methods have been generally employed. The first method tested was an approximate compensation using the geometric average of two opposed projections weighted according to a function o f ray length [18], followed by the conventional filtered back projection (FBP). The second was a ‘regularizing iterative m ethod’ (RIM), where a converging analytical solution o f the attenuated projection operator exists, modified by a relaxation constant. Figure 5 shows the profiles for a phantom consisting o f five linear sources embedded in 20 cm o f lucite when reconstructed using the two methods. No generally applicable method o f attenuation correction in SPECT has as yet been found [19].
6.
TAIL EFFECTS OF THE SYSTEM TRANSFER FUNCTION (STF)
It has been observed that tails o f the STF exist with considerable power even at large distances from their centres. These tails originate partly from the reconstruction algorithm, and partly from the physical response o f the system. In PECT systems, the effect appears to be greater than in SPECT systems, in particular for large plane detectors [4, 20]. One effect o f these tails is to create significant bakground activity throughout the field o f view causing quantitative errors, and loss o f contrast. The tail effect which seems, in this case, to be due
238
SOUSSALINE et al.
T A B L E I.
C O N T R A S T A G A IN S T L E SIO N SIZ E
Contrast
0.85
Cold lesion size in diameter (ram)
30
TOTAL
F IG . 6 .
COUNTS
0.61 20
0.25 10
—►
E xp e rim e n ta l p o in ts ob taine d f o r the m in im u m detectable contrast a t various to ta l
counts f o r a SPECT system are p lo tte d together w ith the th e ore tica l curve predicted. Two comparable p o in ts (a p o in t surrounded by a square) are p lo tte d f o r co n ventional imaging.
essentially to the Compton events, has been measured on the ECAT PECT system using a ‘cold lesion’ phantom 20 cm in diameter. The contrast, defined as max - min/max + min (for the peak and background values), is shown in Table I for three sizes o f cylindrical lesions. In a more general tomographic situation, the minimal detectable contrast has been studied [16] for a 3.6-cm hot spot in a 24-cm uniform field as a function o f the total number o f counts, as plotted in Fig. 6. As predicted theoretically, the minimum contrast seems to be about 2.5 times worse in tomographic scanning than in conventional 2-D imaging.
IAEA-SM-247/139
F IG .
239
7. The recovery c o e ffic ie n t p lo tte d as a fu n c tio n o f the diam eter o f the region o f interest
(o r ob je ct) f o r the E C A T P E C T system.
1.
CONTRAST VARIATION AS A FUNCTION OF THE RELATIVE OBJECT DIMENSION
Another important effect related to the tails o f the STF and the system spatial resolution (in X and Z) is the contrast with respect to the relative dimension o f object structures. This is critical when measuring metabolic processes, perfusion, or volume in organs such as brain and heart. Given the system resolution in standard clinical conditions (FWHM = 15 mm), many structures o f interest are smaller than 2 X FWHM. As suggested by Hoffman and co-workers [ 2 1 ], the non-linear relationship between the measured counts on the image obtained and the actual activity concentration, as a function o f object size, must be evaluated. Using cylindrical and spherical ‘h ot’ lesions o f different diameters in an active background, the so-called ‘Recovery Coefficient’ (RC) parameter [21], being the ratio o f apparent-to-true radioactive concentration, has been measured as shown on Fig. 7. In standard clinical conditions with the ECAT PECT system, the loss o f contrast and quantitative accuracy for objects smaller than 2 X FWHM is extremely important, and must be determined before extracting meaningful quantitative information. Two approaches are then possible: correcting the data extracted by using the known RC, or avoiding the measurement o f structures smaller than s 2 X FWHM.
SOUSSALINE et al.
2 40 8.
C O N C L U SIO N
Despite the very considerable clinical success that has been achieved in ECT, great care must still be taken when extracting quantitative data. The sources o f error must be analysed in detail and, in general, corrections obtained to improve the precision o f such results. Such an analysis also permits the design and construction o f systems where, it is hoped, the need for such corrections is reduced.
ACKNOWLEDGEMENTS The authors would like to thank their many colleagues and, in particular, D. Plummer, Ph. Collard, A. Cao, S. Ricard, M. Crouzel and B. Loch.
REFERENCES [1] KUHL, D.E., EDWARDS, R.Q., RICCI, A.R., et al., Quantitative section scanning using orthogonal tangent correction, J. Nucl. Med. 14 (1973) 196. [2] PHELPS, M.E., HOFFMAN, E.J., HUANG, S.C., KUHL, D.E., ECAT: a new computerised tomographic imaging system for positron-emitting radiopharmaceuticals, J. Nucl. Med. 19 (1978) 635. [3] TER-POGOSSIAN, M.M., MULLANI, N.A., HOOD, J., et al., A multi-slice positron emission computed tomograph (PETT IV) yielding transverse and longitudinal images, Radiology 128(1978) 477. [4] BROWNELL, G.L., BURNHAM, C.A., WILENSKY, S., ARONOW, S., KAZEMI, S., STRIEDER, D., “New developments in positron scintigraphy and the application of cyclotron-produced positron emitters”, Medical Radioisotope Scintigraphy (Proc. Symp. Salzburg, 1968) 1, IAEA, Vienna (1969) 163. [5] YAMAMOTO, Y., THOMPSON, C.J., MEYER, et al., Dynamic positron emission tomography for study of cerebral hemodynamics in a cross-section of the head using positron emitting 6SGa EDTA and 77Kr, J. Comput. Assist. Tomogr. 1 (1977) 43. [ 6 ] PHELPS, M.M., HOFFMAN, E.J., MULLANI, N., et al., Application of annihilation coincidence detection to transaxial reconstruction tomography, J. Nucl. Med. 16 (1975) 210. [7] HOFFMAN, E.J., PHELPS, M.M., MULLANI, N., HIGGINS, C.S., TER-POGOSSIAN, M.M., Design and performance characteristics of a whole body transaxial tomograph, J. Nucl. Med. 17 (1976) 493. [ 8 ] SOUSSALINE, F., TODD-POKROPEK, A.E., PLUMMER, D., COMAR, D., LOCH, C., HOULE, S., KELLERSHOHN, C., The physical performance of a single slice positron tomographic system and preliminary results in a clinical environment, Eur. J. Nucl. Med. 4(1979) 237. [9] SOUSSALINE, F„ HOULE, S., PLUMMER, D., TODD-POKROPEK, A.E., “Information processing in medical imaging”, Potentials of Quantitative Studies in Positron Emission Tomography (Proc. 7th Int. Conf.) 8 8 , Inserm, Paris (1980) 215.
IAEA-SM-247/139
241
[10] RAYNAUD, C., TODD-POKROPEK, A.E., SOUSSALINE, F., ZUROWSKI, S., RICARD, S., KELLERSHOHN, C., Preliminary results obtained in emission computer tomography with a rotating conventional gamma-camera, J. Nucl. Med. 21 (1980) 16. [11] SOUSSALINE, F., TODD-POKROPEK, A.E., ZUROWSKI, S., HUFFER, E., RAYNAUD, C., “Physical characteristics of a single photon tomographic system using a conventional rotating camera” (Proc. 7th Int. Conf. ACNP, 1980), Clinical Utility of Single Photon Emission Tomography, Washington (in press). [12] ALLEMAND, R., GRESSET, C., VACKER, J., Potential advantages of a cesium fluoride scintillator for a time of flight positron camera, J. Nucl. Med. 21 (1980) 153. [13] BUDINGER, T.F., DERENZO, S.E., GULLBERG, G.T., GREENBERG, W.L., HUESMAN, R.H., “Emission computer axial tomography”, Medical Radionuclide Imaging (Proc. Symp. Los Angeles, 1976) 1, IAEA, Vienna (1977) 321 (Abstract only). [14] JASZCAK, R.J., SPECT: Single photon emission computed tomography, IEEE Trans. Nucl. Sei. NS-27 3 (1980). [15] BUDINGER, T.F., DERENZO, S.E., GULLBERG, G.T., HUESMAN, R.H., Trends and prospects for circular ring positron cameras, IEEE Trans. Nucl. Sei. NS-26 (1979). [16] TODD-POKROPEK, A.E., ZUROWSKI, S., SOUSSALINE, F., Artefact creation and non-uniformity in tomography, J. Nucl. Med. 21 (1980) (Abstract only). [17] PANG, S.C., GENNA, S., “Clinical utility of single photon emission tomography, effects of attenuation on emission tomography, its reconstruction non-isotropicity and statistics” (Proc. 7th Int. Conf. ACNP, 1980), Clinical Utility of Single Photon Emission Tomography, Washington (in press). [18] SORENSON, J.A., “Instrumentation in nuclear medicine”, Methods for Quantitative Measurement of Radioactivity in Vivo by Whole Body Counting 2, Academic Press, New York (1974) 311. [19] BUDINGER, T.F., GULLBERG, G.T., HUESMAN, R.H., “Emission computed tomo graphy”, Topics in Applied Physics 32, Springer Verlag, Berlin (1979) 147. [20] MUEHLLEHNER, G., BUSCHIN, M.P., DUDEK, J.H., Performance parameters of a positron imaging camera, IEEE Trans. Nucl. Sei. NS-23 (1976) 528. [21] HOFFMAN, E.J., HUANG, S.C., PHELPS, M.E., Quantitation in positron emission computed tomography: effect of object size, J. Comput. Assist. Tomogr. 3 (1979) 299.
IAEA-SM-247/31
SPET I
Figures o f merit fo r two multiple detector (single slice) and one area detector (multiple slice) single photon emission tomographic instruments* P.H. JARRITT, I.D. CULLUM, P.J. ELL Institute o f Nuclear Medicine, The Middlesex Hospital Medical School, London, United Kingdom
Abstract SPET I: FIGURES OF MERIT FOR TWO MULTIPLE DETECTOR (SINGLE SLICE) AND ONE AREA DETECTOR (MULTIPLE SLICE) SINGLE PHOTON EMISSION TOMOGRAPHIC INSTRUMENTS. Figures of merit are presented for the physical performance of two multi-detector single slice acquisition scanners; (1) a brain scanner, the Cleon-710 and (ii) a whole body scanner, the Cleon-711 ; and one area multiple slice acquisition rotating gamma camera detector, the IGE400T, linked to an Informatek-Simis-3 processor. Phantoms have been constructed for the measurement of sensitivity, slice thickness and reconstructed transverse section resolution. Values have been obtained using three different isotopes: 99 Tcm, 75Se and 13II. The characteristics of these instruments with respect to lesion detection have been investigated by considering the minimum ratios of target-to-background specific activity necessary for visualization. Values have been obtained both for ‘hot’ and ‘cold’ lesions. The effect of energy discrimination and window width on these parameters is demonstrated and discussed. The ability of each instrument to size lesions accurately, both ‘hot’ and ‘cold’, is also reported.
1.
INTRODUCTION
Two fundamental types of instrument have been developed for Single Photon Emission Tomographic (SPET) imaging. Firstly, those developed to image single sections and secondly, area detectors capable of accumulating data from many slices simul taneously. The former have been developed for the specific needs of emission tomography, namely uniformity of sensitivity and resolution. The latter devices have been based upon current gamma camera technology in which the camera has been adapted to perform tomography with the retention of its conventional imaging capabilities. Reconstruction techniques have been developed to compensate for the non-uniformity of resolution with depth for * Work funded by the Sir Jules Thorn Charitable Trust.
243
244
JARRITT et al.
the parallel hole collimators used in these systems, attempting to correct for the inherent lack of uniformity of the area de tector. In the light of these two different approaches, we have evaluated two single slice transverse section emission scanners: (i) the Cleon-710 tomographic brain imager; (ii) the Cleon-711 whole body tomographic scanner, and one area detector, the IGE400T rotating gamma camera linked to an Informatek-Simis-3 computer system. The following physical parameters have been measured: sensitivity, transverse section thickness and resolution within the reconstructed transverse section. Whilst these parameters may be considered sufficient to characterize this instrument ation, small changes within the total system transfer function may influence the reconstructed tomographs. The above mentioned factors have been investigated by considering lesion detectability. Minimum target-to-background ratios of activity required for visualization of 'hot' and 'cold' lesions have been measured. The effect of energy discrimination and window settings upon the contrast ratio obtained for one of these lesions is demonstrated and discussed. The accuracy with which each instrument can size both 'hot' and 'cold' lesions has been investigated. 2. 2.1.
PHYSICAL CHARACTERISTICS The Cleon-710 and 711 Imagers
The two scanners represent a unique approach to the problem of single-photon emission tomography. Each instrument consists of a gantry, a patient couch, computer and operator console. The gantry assemblies represent the major differences between the two instruments. The Cleon-710 brain imager has a field of view of 20 cm and the detector assemblies are rated up to 300ReV. The Cleon-711 body imager has a field of view of 51 cm and the detector assembly is rated up to 370KeV. The method of data collection and image reconstruction has been described previously (1,2,3). 2.2.
The IGE400T Rotating Gamma Camera
This device has been developed from a standard large field of view (LFOV) gamma camera. Rotational capability is obtained by mounting the camera on forks attached to a ring with an in cremental drive motor. The radius of rotation of the camera around the patient can be easily varied by altering the angle between the ring and the forks. The circle of rotation can be divided into 128, 64 or 32 equal increments. The acquisition time at each angle and
IAEA-SM -247/31
245
the number of positions can be set on the camera console, which controls the recording of the information. The data is trans ferred to an Informatek-Simis-3 computer system with 64K words of memory. The recorded information thus consists of 128, 64 or 32 images (ie projections with 360° sampling), stored in 64 x 64 element matrices. In this study, information was recorded utilising a 52-cm diameter rotation circle and 64 projections. A low-energy high resolution parallel hole collimator (H2503AA) was used for 99mjc and the medium-energy standard parallel hole collimator (H2503BA) for 75Se. A standard filtered back projection algorithm was utilized in image reconstruction. 3. 3.1.
DETECTOR CHARACTERIZATION Sensitivity
Sensitivity has been defined as the response of a system to a cylindrical phantom of specified diameter and length with a known concentration of radioactivity (4). Sensitivity values were measured on each device using a cylindrical phantom 20 cm in diameter and 20 cm in length. Values were obtained for 99mjechnetium, 75Selenium, ^ I o d i n e . All of the physical parameters quoted for 75Se have been derived from data acquired as the sum of both peaks for each instrument! The values quoted are for single-section sensitivities expressed as: ______Total counts collected from single section Total scanning time (secs) x radioactive concentration ^uCi/m1) 3.2.
Resolution in the Reconstructed Transverse Section
The resolution in the transverse section was measured using a phantom similar to that described by Budinger et al. (5). A 1-mm diameter line source was placed within a 20-cm diameter, water-filled cylinder. Resolution was defined using two para meters - the Full Width at Half Maximum (FWHM) and the Full Width at Tenth Maximum (FWTM) of the line spread response function in the reconstructed image. The values obtained are average values taken at several radial positions along the line source. Values have not been obtained for "131 1 0n the IGE400T since high-energy collimation is not at present available in our Institute for this instrument. 1 1 C i= 3.70 X 10l o Bq.
JARRITT et al.
24 6
3.3.
The Effective Transverse Section Thickness
In all types of tomographic imagers, photons which arise from sections of the object other than the one being considered are also recorded. The effective transverse section thickness can be defined using the FWHM and FWTM of the response to a point source perpendicular to the section or alternatively the phantom used in the section above can be used and the res ponse measured perpendicular to the line source. Such a phan tom was used in all of our measurements. In the case of the multiple slice machine, the section thickness was measured by considering the response in several adjacent sections, whilst for the single section machine several sections were scanned as the phantom was moved through the image plane. In the case of the rotating camera, the resolution perpendicular to the imaging plane is dependent on the resolution of the collimator along that axis. For the Cleon-710 and 711 imagers, since the response is summed over a number of distances along the collimator axis, a value for the section thickness can only be obtained from the reconstructed images. 3.4.
Lesion Detection
Two sets of measurement were made to evaluate the lesion detection ability of the imagers. The first set involved placing three cylinders in a constant background and varying the specific activity in the cylinders and the background to find the minimum ratio which still enabled visual detection. This was performed for both 'hot' and 'cold' lesions, the background activity always being approximately luCi/ml and 2 million counts were collected in each image. The second set of measurements were taken for a 43-mm diameter cylinder with a fixed target to background ratio as the energy windows were altered. A contrast value defined as: Ave Pixel in Lesion - Ave Pixel in Background Ave Pixel in Lesion + Ave Pixel in Background was measured at each setting. 3.5.
Lesion Sizing
Lesions were simulated using various sized cylinders which were placed within a water-filled phantom for the 'hot' lesions and in a constant activity background for 'cold' lesions. The range of cylinders was chosen with diameters both less than and greater than the FWHM of the detectors. The edge of the recon structed lesion was defined by a maximum pixel count gradient
247
IAE A-SM-247/31
T A B L E I. S E N S IT IV IT Y V A L U E F O R T H R E E TO M O G R A PH IC S C A N N E R S
Sensitivity (counts/s -p C i -ml) Isotope
Energy window (keV)
Cleon-710
99mTc
115-170
16.3K
7SSe
110-170 and 240-320
131j
320-440
Cleon-711 full field
IGE400T
5.7K
1.15K
N/A
17.2K
0.07K
N/A
6.3K
N/A
TABLE II. EFFECT OF ENERGY DISCRIMINATION ON SENSITIVITY RESOLUTION AND SECTION THICKNESS FOR THREE TOMOGRAPHIC SCANNERS
Cleon-711
IGE400T
Section thickness (mm)
Sensitivity (counts/s •/¿Ci -ml)
115-170
16.3K
00
Cleon-710
Resolution (mm)
Energy window (keV)
OO
Imager
16.5
130-170
14.9K
9.2
15.3
115-170
5.7K
26.0
29.0
130-170
3.9K
24.0
2 2 .0
125-160
1.15K
18.0
26.0
135-170
1 .0 K
18.0
17.0
TABLE III. RESOLUTION WITHIN THE RECONSTRUCTED SLICE Resolution in reconstructed slice (mm) Cleon-710
Cleon-711
IGE400T
Isotope 99m Tc 75 Se
131J
FWHM
FWTM
FWHM
FWTM
FWHM
FWTM
9
15
26
41
18
38
N/A
25
48
20
N/A
27
38
43 N/A
248
JARRITT et al.
T A B L E IV. E F F E C T IV E SE C T IO N T H IC K N E SS
Effective section thickness (mm) Cleon-710
Cleon-711
IGE400T
Isotope FWHM
FWTM
FWHM
FWTM
FWHM
FWTM
16
44
29
54
17
43
N/A
33
77
20
46
N/A
45
81 .
99mTc 75 Se
131j
N/A
TABLE V. MINIMUM DETECTABLE TARGET TO BACKGROUND RATIOS: THE BACKGROUND ACTIVITY IS ASSUMED TO BE UNITY Lesion diameter (mm)
‘Hot’ lesion
‘Cold’ lesion
Cleon-710
Cleon-711
IGE400T
Cleon-710
Cleon-711
23
1.5
1.9
2 .1
0.59
0.4
0.31
31
1.5
1 .6
2 .0
0.59
0.4
0.33
42
1.4
1.3
1.9
0.59
0.4
0.33
IGE400T
TABLE VI. EFFECT OF ENERGY DISCRIMINATION SETTINGS ON CONTRAST RATIOS ‘H ot’ lesions
‘Cold’ lesions
Window setting3
A (%)
В (%)
A (%)
Cleon-710
18.4
2 0 .6
13.9
15.9
Cleon-711
23.2
25.3
42.6
47.6
IGE400T
14.7
15.8
23.7
37.5
a Window settings: A: Cleon-710 &711 IGE400T B: Cleon-710 &711 IGE400T
- 1 1 5 - 1 7 0 keV. - 125-160 keV. - 130-170 keV. —135 —170 keV.
В (%)
IAEA-SM-247/31
249
technique and its diameter calculated from the area of this region. The use of cylinders much longer than the FWHM of the section thickness removed any partial volume effect in that direction. 4.
RESULTS AND DISCUSSION
Sensitivity values are presented in Table I. The standard energy discriminator settings are indicated. For the IGE400T rotating cameras the values represent the minimum values derived from a slice of one pixel width based on a 64 x 64 acquisition matrix. In clinical practice, two pixels are added, thus effect ively doubling the sensitivity at the expense of increasing the section thickness (see later section). It should be noted that for the Cleon systems the single section and multiple section sensitivities are the same, since in scanning 'n ‘slices serially, the counts and the scanning time will both be increased by a factor 'n'. For a 20-cm long cylindrical phantom, the multiple section sensitivity of the rotating camera is approximately 30 times that for a single section. Whilst the sensitivity measurement is independent of the reconstruction algorithm, it is highly dependent on the cali bration of the detector with respect to energy discrimination and window width. Varying the window will lead to large changes in sensitivity, although there may not be a concommitant change in resolution as measured by the FWHM and FWTM values for the line spread function. Table II indicates that for 99mjC} varying the energy window leads to changes in sensitivity and section thickness, whilst resolution within the reconstructed slice is unchanged, since it is determined almost solely by the filter function.
Values for resolution within the reconstructed slice are presented in Table. III. Comparison of the two body imagers indicates that the rotating camera has consistently lower values for the FWHM, whilst there is no significant difference between the FWTM values. The reasons for this are not clear but may be related to the different energy resolutions of the two instru ments. It should be noted that the values may not be identical for all directions at all radial positions. Whilst a point source at the centre will yield a circularly symmetrical dis tribution, point sources offset from the centre tend to produce a point spread function which is elongated in the radial direction (measurements from a line source in the plane of the section will mask such artifacts). Effective section thickness results are presented in Table IV. The results show the rotating camera to have superior
250
JARRITT et al.
TRUE S IZ E (л п )
F IG . l.
Sizing de te rm in a tio n f o r ‘h o t ’ and ‘c o ld ’ cylin d e rs on the Cleon-711 w hole-body scanner.
FWHM and FWTM values of section thickness compared with the Cleon-711. For the Cleon-711 imager and the IGE400T, the values change significantly as the window and discrimination settings are altered, although the changes for the Cleon-710 are less marked. These settings, therefore, must be carefully chosen when scanning objects which vary sharply in size in this direc tion. The results for the IGE400T are calculated from single pixel width reconstruction. In clinical practice, two pixels are added, thus the effective section thickness will be increased by approximately 6 mm, the width of a pixel. Several methods have been used to measure the effective section thickness, (a) a point source (1); (b) a line source parallel with the plane of the slice (5); (c) the differential of the edge response function (6). In the case of detectors with large angles of acceptance (Cleon-710 and Cleon-711), these measurements lead to significantly different values. The size and shape of phantoms are thus important. The values presented using method (b) represent a compromise between the values obtained using the alternative techniques. Minimum detectable contrast ratio results are presented in Table V. These values were obtained using standard energy discriminator settings. In all cases 'hot': lesions are more easily detected than 'cold1 lesions. The camera exhibits the worse performance of the three scanners considered.
IAE A-SM-247/31
F IG .2.
251
Sizing d e te rm in a tio n f o r ‘h o t ’ cylinders on the Cleon-710 and the IG E 4 0 0 T tom ographic
imagers.
The effect of varying the energy threshold and window width on sensitivity and resolution has already been demonstrated, its effect on measured contrast ratios, however, has rarely been considered. The results are presented in Table VI. For the Cleon-710 and 711 imagers, the two energy windows chosen repres ent those which have been used in clinical practice. For the IGE400T, the settings correspond to a symmetric and asymmetric 20% window. The scans were recorded using a target-to-background ratio of 2.5:1 for 'hot' spots and 0.2:1 for 'cold' spots. For the Cleon scanners, there is a small decrease in contrast resolution with increasing window depth. This is more marked for 'cold' lesions than 'hot' with the Cleon-711 imager due to the larger field of view and decreased resolution of the colli mators. For the rotating camera, there is a large effect for 'cold' lesions which is partly due to reconstruction artifacts caused by variations in uniformity of response with changing energy window. This effect is not seen with the single section scanners but must be considered with rotating camera type devices. A major advantage of emission tomography is stated to be its ability to enable accurate quantification of radiopharma ceutical distributions. The theoretical and practical aspects of quantitative imaging using tomography have recently been reviewed (7). We have investigated the ability of each instru ment to accurately reflect the size of lesions with the view to obtaining accurate measurement of organ and lesion volumes.
252
JARRITT et al.
A similar trend is seen for all three devices (Figs 1 & 2), ie accurate sizing of 'hot' spots is not possible below the FWHM of the line spread function in the reconstructed slice (although all sizes may of course be detected provided that the target-to-background ratio is sufficient to allow visualization). 'Cold' lesions may be sized down to approximately 0.5 FWHM, below which they cannot be visualized. Throughout the range, accuracy is generally better than 10% for all three tomographs.
5.
CONCLUSION
An understanding of those parameters which are important in the assessment of tomographic imaging devices is necessary if satisfactory comparisons of physical performance are to be obtained. The results from the measurement of sensitivity, resolution and section thickness for the three tomographic imagers have revealed that careful specification of energy discrimination settings is important. For the Cleon-710 and 711 scanners, problems arising from non-uniformity of detector response can essentially be ignored. For gamma camera type imagers, non uniformities derived from a non-linear energy response produce small changes in the system transfer function which are manifest on changes in- contrast resolution. The results from this study indicate that satisfactory operating conditions should not only be defined from sensitivity, resolution and slice thickness measurements, but also from measurements of contrast resolution.
REFERENCES [1] [2] [3] [4] [5]
[6 ] [7]
JARRITT, P.H., ELL, P.J., A new transverse section brain imager for single gamma emitters, J. Nucl. Med. 20 4 (1979) 319. JARRITT, P.H., ELL, P.J., A new transverse section body scanner, Nuc. Med. Commun. 1 (1980) 94. FLOWER, M.A., PARKER, R.O., Quantitative imaging using the Cleon emission tomography system: Recent developments, Radiology (in press). KUHL, D.E., EDWARDS, R.Q., RICCI, A.R., The Mark IV system for radionuclide computed tomography of the brain, Radiology 121 (1976) 405. BUDINGER, T.F., DERENZO, S.E., GULLBERG, G.T., GREENBERG, W.L., HUESMAN, R. H., Emission computed axial tomography, J. Comput. Assist. Tomogr. 1 (1977). FLOWER, M.A., ROWE, R.W., KEYES, W.I., “Sensitivity measurements on single photon emission tomography systems”, Radioaktive Isotopen in Klinik und Forschung, 14 2 (1980). BUDINGER, T.F., Physical attributes of single photon tomography, J. Nucl. Med. 21 6 (1980) 579.
IAE A-SM-247/31
253
DISCUSSION N.G. TROTT: Would you like to comment on the problems of attenuation correction with the Cleon instruments? ' P.H. JARRITT: The problem of photon attenuation does not have a simple solution. However, it has been possible with the latest revisions o f Cleon-710 software to obtain reasonably uniform sensitivity across the field of view. The problems are more severe for the C leon-711 and as yet are not solved. It should also be noted that the problems are less severe than with the rotating camera devices. Software modifications are still under investigation, and improvements will undoubtedly be made.
IAEA-SM-247/32
S P E T II
The clinical role o f single photon emission tomography* P.J. ELL, O. KHAN, P.H. JARRITT, R.G. RADIA Institute o f Nuclear Medicine, The Middlesex Hospital Medical School, London, United Kingdom
Abstract SPET II: THE CLINICAL ROLE OF SINGLE PHOTON EMISSION TOMOGRAPHY. Radionuclide section scanning with standard radiopharmaceuticals — single photon emission tomography (SPET) — has progressed to the point where its influence is felt in the practice of clinical nuclear medicine. Advances in the field of instrumentation and computer hard- and software have made it possible to incorporate tomographic imaging in the routine investigation of a variety of clinical problems. A review is given of the authors’ experience with this new technique in the framework of comparative imaging protocols (radionuclide emission, X-ray transmission, and/or ultrasound whenever applicable) in organs such as the brain, the heart, the liver and the lungs. In the field of cancer detection, SPET has proven to be a superior method with remarkable improvement of lesion detection over and above conventional planar Anger gamma camera imaging (for brain and liver) and X-ray transmission imaging (for the liver). It has raised the overall reliability of clinical reporting of data by as much as 20%. Initial experience in the field of heart and lung imaging points to several but somewhat narrower areas of clinical indications for usage. The ability of cross-section brings with it useful clinical information on depth and improvement in the definition of pathology. In the field of quantification of radionuclide tracer uptake, promising work is in progress. The skull and 99Tcm-MDP have been chosen and normal ranges of uptake established in over 40 individuals in terms of ßC i of tracer per ml of active skull volume. These values are being applied to the diagnostic investigation of a variety of patients suffering from conditions leading to either focal or diffuse but abnormal bone turnover. Initial data of therapeutic protocols are being assessed in patients with osteomalacia, Paget’s disease and other conditions. The observation that patients suffering from skeletal secondaries other than in the skull present with abnormally high tracer uptake values in this organ raises an interesting problem which deserves further investigation.
* Work funded by the Sir Jules Thorn Charitable Trust.
255
256 1.
ELL et al. C L IN IC A L T R IA L S
Three years’ experience with instrumentation designed for single photon emission tomography (SPET) have permitted a number o f areas to be defined and established where radionuclide section scanning appears as a methodology with definite clinical merits [1 ,2 ]. A review is given here o f the experience gained so far in a number o f clinical trials. 1.1. Brain: Emission and transmission imaging protocol Conventional planar imaging with an Anger camera has known limitations. The base o f the brain, the pituitary and posterior fossae are particularly difficult for data interpretation, and depth information is lost, leading to a high percentage o f unreliable reports, in terms o f lesion identification, localization and lateralization. X-ray transmission section scanning is today the best technique for lesion detection in the brain, both in terms o f sensitivity and o f specificity. Technically, it may soon be overtaken by nuclear magnetic resonance imaging (NMR), which offers superior spatial resolution with no radiation exposure risk. Nevertheless, X-ray transmission section scanning is still unable to cope with all the demand, radiation exposure is high and the necessity for contrast imaging reduces its overall safety; NMR will not become generally available within the foreseeable future. In this context, we have compared emission and transmission section scanning techniques in 300 patients (for the emission technique, the Cleon-710 scanner and for the transmission technique the EMI CT5005 or the CGR N D 8000 were used). Studies with and without contrast were always undertaken. Concordant data (identical positive or negative results) were obtained in 242 patients (80% overall). X-ray transmission section scanning yielded a false positive rate of 1 % and a false negative rate o f 6%. Emission section scanning ( " T c m О4 or 99TCm-GH) yielded a false positive rate o f 0% and a false negative rate o f 2.5% for malignant disease and 9% for vascular disease. The higher rate o f false negative results for vascular disease for the emission section scanning technique occurs because o f the ability o f X-ray transmission section scanning to detect old infarction (scar tissue). The rates o f detection for recent vascular disease for both techniques were found not to differ in statistically significant numbers. Emission section scanning proved to be appreciably superior to conventional and planar Anger camera imaging. Five per cent o f all lesions detected with emission section scanning remained silent on the Anger camera. More important, however, was the overall improvement in reporting performance, which was quantified at the 20% level. Emission section scanning o f the brain is now our routine static imaging procedure, totally replacing conventional Anger camera imaging. A one-minute Anger camera-computer dynamic perfusion acquisition still precedes all o f these investigations at the time o f tracer injection.
IAEA-SM-247/32
257
1.2. Liver: E m ission and tran sm ission im agin g p r o to c o l
The investigation o f space-occupying pathology in the liver by means of non-invasive imaging techniques has been traditionally performed with radio nuclide planar scanning (rectilinear scanners and gamma cameras). It soon became apparent that the techniques lacked sensitivity and were liable to a significant false positive rate. In recent years, both ultrasonography and computed X-ray transmission tomography have been extensively applied to the same purpose. Combined liver imaging protocols were initiated in order to establish relative accuracies, sensitivities and false positive and false negative results. A number of these protocols have produced data which have appeared in the literature. Despite the inherent and apparent advantages o f X-ray transmission tomography (increased spatial resolution in particular), and the definite progress obtained with ultrasound grey scales and m odem scan converters, these reports have established that no individual technique appears to have emerged as the significant best method for liver disease detection. It is now understood that in many circumstances, lesions o f significant size can escape detection by X-ray transmission tomography, since similarities in normal and abnormal liver tissue lead to insufficient contrast resolution for visual detection. It is also apparent that a significant percentage of patients end up with poor quality ultrasound examinations, since the presence of fat, air or gas accidentally prevents the penetration o f the ultrasound beam. Because o f these difficulties, in most centres where all three imaging modalities are available, radionuclide planar imaging with a gamma camera is still the initial screening investigation. Among several merits, simplicity and operator independence, cost, speed, safety and reproducibility enhance the general utility o f this method for space-occupying liver lesion detection. In order to establish the role o f single photon emission tomography in the scanning and screening o f patients suspected o f having malignant involvement of the liver, a liver imaging protocol was designed encompassing standard planar radionuclide scanning, radionuclide emission tomography (utilizing either the Cleon-711 body imager or the IGE400T rotating gamma camera and an InformatekSimis-3 data processor) and X-ray transmission tomography with an EMI 5005 scanner. Occasionally, ultrasound data were also recorded, to help clarify difficult diagnostic problem cases. In 36 patients (72% o f total) concordant data were recorded (15 concordant positive and 21 concordant negative results). In 14 patients (28% o f total) discordant data between the three imaging modalities (Anger camera planar and emission and transmission section scanning) were recorded. T о assess the significance o f this group o f patients with conflicting results on different scanning techniques, the following criteria were applied: one year follow-up; histology confirmation, whenever applicable; presence o f clear-cut confirmatory data on at least two different imaging techniques; progression of
258
ELL et al.
disease, as judged by growth o f liver lesions and confirmation through ultra sonography o f the liver. Despite this, it has not been possible so far in 4 patients to arrive at a final conclusion, while in the remaining 10 patients a final diagnosis was achieved. The relative detection sensitivities for three main imaging modali ties were: 80% for planar gamma camera scanning, 80% for X-ray transmission tomography and 90% for radionuclide emission tomography. If it is considered that the four undecided results would be likely to fall into two categories o f correct or incorrect for each imaging modality, the following ranges o f accuracy for each of the techniques are obtained: 72-88% for planar gamma camera scanning, 72-88% for X-ray transmission tomography and 82-98% for radio nuclide emission tomography. From the data o f this study, it was possible to ascertain: that large single lesions are adequately reconstructed and localized via emission tomography; that multiple lesions can be readily detected via section scanning, benefiting from multiplane reconstruction (sagittal, coronal and transaxial) ; that small lesions (o f the order of 8 mm) can be detected in the liver via emission tomography; that even when extensive and diffuse destruction o f liver tissue occurs, emission section scanning is able to reconstruct activity distributions consistent with the true pattern o f the disease; normal Anger camera studies (6-view study) can be transformed into true positive section scans and conversely (although in this series less often) abnormal Anger camera studies can be transformed into true negative section scans. Provided that multiplane image reconstruction is available, it is felt that section scanning o f the liver can replace conventional planar imaging with a high-resolution Anger camera. 1.3. Lungs: 2-D versus 3-D emission imaging protocol Thirty patiens suspected o f suffering from embolic lung disease were included in a comparative study o f planar imaging and tomographic imaging. Standard radiopharmaceuticals and doses were utilized, both for the ventilation studies ( 133Xe) and for the perfusion studies (" T cm-albumin microspheres). Once again, both the C leon-711 and the rotating IGE400T gamma camera were the instruments for data collection. For the gamma camera planar studies, six views were routinely recorded (AP, PA, right and left posterior obliques, and right and left laterals) during perfusion assessment. Wash-in, equilibrium and wash-out scans were recorded with 133Xe in a posterior projection, for ventilation assessment. In terms o f relative merits of each imaging technique, no significant difference in detection sensitivity was found between the planar images and the tomographic studies. Concordant negative studies were reported in 8 patients, and concordant positive studies were reported in 20 patients. While it was not easy to relate perfusion defects to a particular segmental distribution in the
IAEA-SM-247/32
2 59
tomographic study in the transaxial planes (an important feature in the differential diagnosis o f perfusion defects on lung scans), it became apparent that section scanning o f the lungs more closely defined the full extent o f the distribution of the impaired perfusion. This might prove to be clinically useful if a more accurate assessment o f the efficacy o f treatment will be desired. It was also noted that the transaxial plane may not be the optimal plane for image analysis. Initial difficulties with data interpretation may be overcome with reconstruction o f the lung tomograms, in particular in the coronal and sagittal planes. 1.4. Heart: 2-D versus 3-D emission imaging protocol Twenty patients with acute myocardial infarction (A.M.I.) were scanned 4 8 - 7 2 h after the event with " T cm-IDP. SPET shows potential in more accurately delineating size, shape and extent o f the infarction and appears to improve (via improvement in contrast) the diagnosis o f subendocardial necrosis at present a gain o f 5% in sensitivity. 1.5. Tracer uptake measurement SPET is ideally suited to the quantification o f radionuclide tracer uptake in a 3-D image. Recently, it has been stated that practical whole-body tomography in adults is limited to non-quantitative lesion detection [3]. This does not apply to the brain and skull where true quantitation may be feasible. It was therefore decided to establish the feasibility o f " T cm-MDP bone uptake measurement in the skull via radionuclide section scanning o f this organ. The aim o f this pilot study was to utilize the Cleon-710 tomographic brain scanner and establish the feasibility o f defining ranges o f tracer uptake in a normal population, to express this tracer uptake in terms o f juCi o f " T c m-MDP per ml o f active skull volume, and to apply the established normal ranges to the investigation o f a diseased population.1 The following procedure was adopted: For scanning purposes, 15 mCi of " T cm-MDP were given IV and imaging was begun 6 0 - 9 0 min after tracer administration. For each section scan, 4 min o f data acquisition with three tomograms recorded per study. Section scans were obtained 2.5 cm above the EAM plane, slice spacing was o f the order o f 1.25 cm. Slice thickness was 1.3 cm, with images reconstructed in a transaxial plane. A ring source o f uniform activity was utilized to calibrate the scanner. The linearity o f the instrumentation was established by testing its response to increasing and known amounts o f radio activity. It was shown that for a wide range o f activity concentrations (between 0.005 juCi/ml to 5 /¿Ci/ml) the tomographic scanner responded in a linear fashion.
1 1 Ci = 3.70 X 1010 Bq.
26 0
ELL et al.
Three methods were used in the calculation o f tracer skull uptake: Method A consisting o f ring source calibration, no background subtraction and a maximum pixel count; method В consisting o f ring source calibration, 25% background subtraction and an average pixel count; method С consisting o f ring source calibration, variable background subtraction and an average pixel count. For method A, and in 40 normal volunteers, a mean value o f 0.11 juCi/ml for 9^rcm-MDP uptake was obtained (range 0 .0 8 -0 .1 4 ). For method B, and in 11 normal volunteers, a mean value o f 0.06 /¿Ci/ml for " T cm-MDP uptake was obtained (range 0 .0 5 —0.07). For method C, and in 11 normal volunteers, a mean value o f 0.03 juCi/ml for " T cm-MDP uptake was obtained (range 0 .0 2 0.04). It is interesting to note that the normal values recorded in this population have rather narrow ranges. So far, method B, which we feel is at present the most appropriate, was utilized in the calculation of " T cm-MDP skull uptake values in a series o f diseased individuals. Seventeen patients with skeletal deposits other than in the skull, five patients with Paget’s disease, four patients with osteomalacia, three patients post parathyroidectomy, one patient with primary hyperparathyroidism and one patient with thyrotoxicosis were investigated. A preliminary study is in progress whereby patients with osteomalacia are being monitored with a base-line study and then at repeated intervals, on commence ment o f treatment. The data recorded so far clearly shows the feasibility o f this method for tracer quantification and its ability to distinguish a normal from an abnormal population. In all the patients studied so far, much higher tracer uptake values were recorded. This includes both the patients with malignant disease and with metabolic bone disease. A very interesting point in the data is the increased uptake in patients suffering from bony secondaries other than in the skull. Even in this group o f patients, much higher tracer uptake values were obtained ranging from 0.2—0.6 pCi/m l o f " T cm-MDP. It is possible to speculate on the possible reasons for this, and work is in progress to establish the role o f substances such as PTH. The osteomalacia group o f patients were all classical clinical cases o f this condition, caused by dietary vitamin D deficiency. Reduced serum calcium and phosphate levels, a raised serum alkaline phosphatase and parathyroid hormone levels were seen. The uptake ranges varied from 0 .4 —0.7 /iCi/ml o f " T cm-MDP uptake. As predicted for patients with secondary hyperparathyroidism, raised " T cm-MDP skull uptake values were also found. In the serial evaluation of patients with osteomalacia during therapy with 1,25, dihydrodihydroxicholecalciferol, there was initial reduction o f tracer uptake in the skull, followed by an increase to nearly pre-treatment levels and a later slow tapering. Elevated uptake levels were still present six months after commencement o f treatment. It is possible to speculate that one o f the mechanisms involved in the initial reduction o f tracer uptake in the skull is that o f the normalization o f serum calcium, leading to a flattening o f the stimulus to hyperparathyroidism, resulting in the temporary state o f hypoparathyroidism. Three patients with a laryngo-
IAEA-SM-247/32
261
pharyngectomy have been studied. All o f these patients are on supplementary with vitamin D 3, calcium and thyroxine. The two patients who show wide fluctuations in serum biochemistry and poor clinical control also show elevated tracer uptake. The third patient has maintained near normal serum biochemistry on supplements and has normal skull tracer uptake.
2.
CONCLUSION
The clinical value o f single photon emission tomography is shown. It is predictable that radionuclide section scanning techniques will play a greater role in the future routine o f a busy nuclear medicine clinic.
REFERENCES [1] [2] [3]
ELL, P.J., JARRITT, P.H., DEACON, J.M., BROWN, N.J.G., Emission computerized tomography: A new diagnostic imaging technique, Lancet ii ( 1978) 608. ELL, P.J., DEACON, J.M., DUCASSOU, D., BRENDEL, A., Emission and transmission brain tomography, Br. Med. J. 280 (1980) 438. BUDINGER, Th.F., Physical attributes of single photon tomography, J. Nucl. Med. 21 (1980) 579.
DISCUSSION B.D. BOK: Showing results o f a comparison between imaging modalities may be misleading if certain details are not given concerning the m ethodology used for that purpose. With regard to the comparison o f brain TCT and ECT scans, would you give some details concerning: (a) the method o f determining a ‘normal’ or ‘abnormal’ response, i.e. were there one or several observers and did they have any knowledge o f the patient’s history, clinical or other data? (b) the criteria used to establish the ‘final diagnosis’? P.J. ELL: When a speaker is allowed ten minutes to review three years experience with a new technique in four major biological systems (brain, lungs, liver and skeleton), it is simply not possible to go into detail concerning materials and methods. These can be found in Ref. [2] o f the paper. In general, and in the case o f the brain in particular, a final diagnosis was made after a proper history, the findings o f EEG, angiography, surgery and/or follow-up, where appropriate. In our institute, patients are only scanned with full knowledge o f the clinical problem and a presumptive diagnosis.
262
ELL et al.
As far as the liver study is concerned, the materials and methods are given in the paper. D. PAVEL: You quote an improvement o f 5% in sensitivity when emission tomography is compared with planar imaging, but an improvement o f 20% in reporting accuracy. Could this mean that the sensitivity and the reporting accuracy were evaluated by different observers? P.J. ELL: No, this result reflects the known difference between a large percentage o f patients where a doubtful report (planar study) is transformed into a definite category o f either positive or negative (tomographic study). The difference between reporting accuracy and performance also reflects the known improvement obtained with SPET in lesion localization with depth over and above lesion detection in planar imaging (e.g. to distinguish skull from intracerebral deposits, base o f brain, posterior fossae, etc).
IAEA-SM-247/25
CLINICAL RESULTS OF QUANTITATIVE SINGLE PHOTON EMISSION TOMOGRAPHY K.E. BRITTON, B. SHAPIRO, A.T. ELLIOTT Department o f Nuclear Medicine, St. Bartholomew’s Hospital, London, United Kingdom,
Abstract CLINICAL RESULTS OF QUANTITATIVE SINGLE PHOTON EMISSION TOMOGRAPHY. In addition to the traditional skills of pattern recognition in the interpretation of images, it is necessary to add quantitative techniques, particularly in difficult problems, to determine normal and abnormal variation. Single photon emission tomography, SPET, overcomes the problems of tissue background and superficial tissue overlying a suspect lesion. Nevertheless, the goal of absolute quantitation is important in the solution to several clinical problems. The use and success of quantitative SPET in the liver, heart, adrenal and pituitary glands are reviewed.
1.
INTRODUCTION
Because radiopharmaceuticals used in nuclear medicine are never organ specific, every image obtained using a conventional gamma camera or recti linear scanner has a blurring effect due to radioactivity in tissue before or behind the organ under study. Furthermore, for organs o f some size such as liver or brain, superficial structures overlie deeper parts o f the organ and therefore abnormalities deep in the organ may be concealed. These features combine to make it difficult to measure the amount o f radioactivity in a tissue or organ, and thus it is difficult to define what is normal and what is abnormal quantitatively. Therefore only relative, not absolute, measurements are made when conventional imaging is undertaken. Single photon emission tomography overcame these problems affecting quantitation with conventional imaging, so the desire for absolute measurements of organ uptake was again awakened. Unfortunately a new set o f problems, summarized by Keyes [1], also prevents fulfilment o f this expectation. These problems relate to the many sources of signal degradation in single photon emission tomography and include: algorithm noise due to the computer, the analysis program and the filter; statistical noise o f which Poisson noise domi nates; and inconstant noise due to the effects o f attenuation, detector mis alignment, and organ and patient movement. Unfortunately, this disappointment
263
264
BRITTON et al.
TABLE I. QUANTITATIVE SINGLE PHOTON TOMOGRAPHY, SPET, IN ITS CLINICAL CONTEXT Organ
Problems
Principle and SPET solution
Liver
Cancer staging Liver size
Metastatic disease is focal, leaving islands of normal tissue Define normal variation with reference to normal tissue
Heart
Globular organ overlap causes poor quantitation
Sections avoid overlap and allow definition of normal variation with reference to normal tissue
Thyroid
Volume measurement for 1311 therapy
Measure short axes for ellipsoid model [2]
Adrenal
Depth and long axis variation and liver overlap gives poor quantitation
Measure skin adrenal depth. Summate each adrenal activity and a background area through a series of sections. Uptake as fraction of dose
Pituitary
Bone surround and small size
Measure pituitary/background ratio
has led some workers to abandon all types o f quantitation with single photon emission tomography. But such a response is quite inappropriate. The rule that relative measurements are the m ost relevant and reliable in nuclear medicine still applies. The clinical questions that can be answered by quantitative single photon emission tomography are here discussed (Table I).
2.
METHODS
The Tomogscanners II and IIS (J. &P. Engineering, Reading, United Kingdom) are single photon counting systems employing two opposing banks o f 12.5-cm dia., 5-cm deep sodium iodide detectors, six in total in the IIS version. The focusing collimators are designed to give a uniform field and maximum sensitivity at depth. A 40-cm line scan length is usual. The detec tors can be separated from 20 to 76 cm. The total scan is performed in a translate-rotate mode as the scanner rotates through 180° in 6° increments.
IAEA-SM -247/2S
265
The first line scan is displayed on the video screen as a count density profile, and the computer program suggests the scanning speed to choose in order to give an information density of 100 counts per pixel. Images are reconstructed into a 80 X 80 matrix and then interpolated up to 160 X 160 for display. A 64 level grey scale and heated object spectrum and eight colour contour displays are available. Conventional administered doses are used and conventional images are obtained before the patient is transferred to the SPET system. Thus no extra radiation is absorbed by the patient. Typical imaging times are: 4 min per slice for the liver imaged 40 min after injection o f 3 mCi " T c m Sn colloid; 6 min per slice for the heart imaged 40 min after 3 mCi o f 201 TI; 8 min per slice for the pituitary imaged 40 min after injection o f 15 mCi " T c m pertechnetate; and 12 min per slice for the adrenals imaged 40 min after injection of 250 pCi75 Se selenom ethylcholesterol.1
3.
QUANTITATION
3.1. The liver The liver sections are related to the level o f the xiphisternum and typically 2 and 4 cm above and below. The texture o f the liver is uneven in section, and determination o f whether the liver texture is normal or abnormal from an analogue grey scale or heat spectrum display is not to be relied upon. For this reason it is necessary to measure the normal variation o f the liver texture with regard to a reference point such as the most active part o f the liver, e.g. defined as the upper 5 units on a 100-unit scale. Using patients with normal livers and volunteers the normal range o f variation was found to be 20 units on a scale of 100 with a borderline variation o f 5 units. Using a combination o f this quantitative approach and pattern recogni tion, a series of.liver images were assessed in staging patients with cancer o f the large intestine pre-operatively (Table II). The use o f SPET showed more accurate staging than that using a conven tional gamma camera (70% correct in this series) and than that obtained using an EMI whole-body X-ray computed tomography (CT) system. 3.2. The heart Early experience o f SPET o f the heart was promising both clinically [3] and using phantom studies [4], and normal ranges have been established in co-operation with Hisada and Madea o f Kanazawa University, Japan (Table III). 1 1 Ci = 3.70 X 1010 Bq.
2 66
BRITTON et al.
TABLE II. LIVER TRANSVERSE TOMOGRAPHY: SPET AND X-RAY CT SPET
X-ray computed tomography Normal
Abnormal
Total
Correct
Percentage correct
SPET
Normal Abnormal Total
15
4
19
17
3
9
12
12
18
13
31
95
X -ray C T
Correct Percentage correct
15
11
83
TABLE III. NORMAL VARIATION OF THE HEART WITH SPET ON A 100-UNIT SCALE Normal Apex
100-65
Lateral wall Septal wall
Borderline
Abnormal
6 5 -5 0
<50
100—85
85—80
<80
100-85
8 5 -8 0
< 80
The variation o f the inferior wall, which is often sliced transversely at the level o f the apex, appears to vary to a similar extent as the lateral wall. These studies were undertaken with thallium as a prelude to the use o f 123I-labelled fatty acids. 3.3. The adrenal A study o f adrenal depth [5] was shown to vary between 5 and 13 cm, right mean 8.3, left mean 8.5 cm deep. Depth corrections enable accurate uptake measurement o f the dose administered (normal 0 .07—0.30%), shown in Table IV. The overall result in the 70 patients with and without adrenal disorder was 89% correct, but the results were all correct in the 19 patients in whom SPET was used in addition.
IAEA-SM -247/2S
267
TABLE IV. QUANTITATION IN ADRENAL DISORDER USING SELENIUM-75 SELENOMETHYLCHOLESTEROL Number
Unilateral adenoma Aldosteronism Cushing’s syndrome Bilateral hyperplasia Aldosteronism Cushing’s syndrome Remnant hyperplasia Cushing’s syndrome Adrenal carcinoma Non-functioning Functioning
Correct
5 5
0.5 0 .8 8
5 5
14
0.41 0.91
20
0.39
3
3
0 .1
3
1
0.55
0
22
3
53
•Total
Mean (%)
11
47
3.4. The pituitary The calculation o f pituitary uptake was made from regions o f interest, ROI, as a pituitary-to-background ratio, PR, given by PR = (C1 /A )/((C 2 - C 1 )/(B -A )) where C t is the count content o f the pituitary ROI and A is the number o f pixels therein, and C2 is the count content of an ROI extended to include non-pituitary tissue o f size В pixels [6]. The normal variation o f PR is less than 1.25 to 1, the borderline value between 1.25 and 1.27, and the adenoma range is over 1.27 to 1. In a series o f 19 patients with pituitary adenoma proven by air encephalography and two patients without adenoma, X-ray CT and SPET were compared (Table V). All patients had moderate or minor changes in the pituitary fossa, and many had raised hormone levels appropriate to the type o f adenoma, mainly prolactinoma presenting as secondary amenorrhoea. The success o f SPET over X-ray CT and conventional brain imaging is evident.
2 68
BRITTON et al.
TABLE V. PITUITARY TUMOUR IMAGING Tumour
Brain scan
X-ray CT
+
+
Present
6
Not present
0
Correct
6/17
Percentage correct P value
4.
11
8
2
0
SPET +
8
15
2
0
6 2
8/16
15/21
35
50
71
NS
NS
< 0 .0 5
CONCLUSION
It is self evident that one requires from SPET, as from conventional imaging, the best analogue displays that can be obtained so that the traditional skills o f pattern recognition may be applied. Nevertheless it is an essential part o f modern nuclear medicine that the advantages o f quantitation are added, particularly in the determination o f normal and abnormal variation, whose success in difficult problems is illustrated here.
REFERENCES [l ] KEYES, W.I., A practical approach to transverse section gamma ray imaging, Br. J. Radiol. 49 (1976) 62. [2] SHAPIRO, B., RIGBY, L,, BRITTON, K.E., The assessment of thyroid volume with single photon emission tomography, Nucl. Med. Commun. 1 (1980) 33. [3] DYMOND, D.S., STONE, D.L., ELLIOTT, A.T., BRITTON, K.E., SPURRELL, R.A., Cardiac emission tomography in patients using 201 Thallium, a new technique for perfusion scintigraphy, Clin. Cardiol. 2 (1979) 192. [4] ELLIOTT, A.T., SANDISON, G.A., HANSON, M.E., WHITE, D.R., “Quantitative investigation of the performance of emission tomography equipment” , Information Processing In Medical Imaging (DI PAOLA, R., KAHN, N.E., Eds), INSERM, Paris (1980) 257. [5] SHAPIRO, B., BRITTON, K.E., RIGBY, L., ELLIOTT, A.T., Emission tomography of the pituitary, Nucl. Med. Commun. 1 (1980) 150. [6] BRITTON, K.E., SHAPIRO, В., HAWKINS, L.A., Emission tomography for adrenal imaging, Nucl. Med. Commun. 1 (1980) 37.
IAEA-SM -247/25
269
DISCUSSION A.E. TODD-POKROPEK: In your liver scans you did not appear to take into account the ‘partial volume effect’ (the same depth o f a lesion corresponding to different changes in ‘contrast’ as a function o f lesion size). Could you comment on this and suggest why you think that taking this into account might not improve your results? K.E. BRITTON: The results were obtained empirically in that the normal range was that determined in patients or volunteers with normal livers. This empirical approach summates the physical variables, including any partial volume effect and the biological variables to give the ‘decision-aiding ranges’. I would remind you that the liver study was performed pre-operatively, at a stage where métastasés, if present, are likely to be small. It is in this context that the ‘relative quantitation’ approach is, I believe, important. T. JONES: I was interested to hear a comment you made about the physicist’s preoccupation with absolute quantitation. In our experience medical doctors are also concerned with the ability to achieve this. The overall organ function may change and, therefore, regional variations cannot be normalized to an area o f normal tissue uptake. For example, we have found, in certain cases o f dementia, that focal defects o f cerebral function are not evident although overall brain function has decreased. K.E. BRITTON : I wished to emphasize that the defects in equipment and the inability to perform absolute quantitation should not be an excuse for doing nothing. The limitations that are technical result in clinical limitations, but within those limitations, e.g. in large organs with focal disease and small organs such as the pituitary, useful clinical work can still be done using relative quantitation. H.N. WAGNER, Jr.: I feel you are rather too pessimistic with regard to the possibility and usefulness o f absolute quantification. In several institutions, including our own, we are making progress towards achieving measurement of left ventricular volume in millilitres, which is quite a useful asset. Other examples are measurement o f organ volume or lesion volume, the latter being useful in evaluating the effectiveness o f treatment. K.E. BRITTON: I can only say that with less than optimal equipment, relative quantitation is useful in some clinically important contexts, namely staging cancer pre-operatively and in heart attacks, even if it could not answer all liver problems, nor all heart problems. D.E. KUHL: We must recognize that single photon ECT has already provided quantitative determinations o f physiological parameters in the brain. The first determinations o f local cerebral blood volume (in units o f ml o f blood/ 100 g) and o f local cerebral metabolic rate for glucose (in units o f mg of glucose/ 100 g per mm) by emission-computed tomography in man were
2 70
BRITTON et al.
performed using " T c m red cells and 18F-fluorodeoxyglucose scanned by the Mark IV, a single photon ECT device. P. SHARP: Mr. Britton, I was impressed by the fact that you quoted the uptake measurement in pituitary tumours to two decimal places. Have you in fact replicated your measurements to test the reproducibility o f your technique? K.E. BRITTON: The figures to two decimal places do not denote accuracy but represent three significant figures as the measurement is a ratio, 0 .7 —1.25 being the range of patients without tumour determined empirically and over 1.3 being the definite tumour range. I took a cut-off point at 1.28 for this study. The reproducibility o f the measurement technique has been tested by two observers independently with similar results. Further, two patients under went two studies, the results in both cases being in the tumour range.
IAEA-SM -247/61
SCINTILLATION C A M E R A R A D I O N U C L I D E IMAGING WITH L O W - E N E R G Y ISOTOPES ( I O D I N E - 125, C A E S I U M - 1 3 1 ) G. ENDERT, E. KLOSE, E. SCHUMANN Radiological Clinic, Medical Academy Erfurt, Erfurt, German Democratic Republic
Abstract SCINTILLATION CAMERA RADIONUCLIDE IMAGING WITH LOW-ENERGY ISOTOPES (IODINE-125, CAESIUM-131). Anger cameras are normally used in the energy range from 70 keV to 400 keV. The main reason for this limitation is because of the bad intrinsic resolution at lower energies. But there are some clinical applications for camera investigations with low-energy isotopes. The authors therefore modified their camera by increasing the Z-signal amplification. The resolution and sensitivity were measured by comparing an ultra-fine low-energy collimator and an X-ray antidiffusion grid. The improvement of sensitivity from the X-ray grid was about twentyfold, but a decrease of intrinsic resolution was measured (factor: 1 .8 comparing 12SI with 99 Tcm). In clinical investigations the modified camera was used together with the X-ray grid as a special kind of collimator inserted into a collimator changing system. Twenty patients were examined by thrombo-scintigraphy ( 125I-fibrinogen); in 20 patients kidney studies ( 12sI-hippuric acid), and in 10 patients calf-muscle scans ( 131Cs-chloride) were carried out. The advantages of this system are as follows: simple modification of the Anger camera, a high sensitivity and the possibility of dynamic studies. The use is limited to organs near the body surface because of absorption. Another possible application is the demonstration of 1251-cont aminations.
INTRODUCTION Because of physical reasons the Anger camera is usually used in the energy range from 70 keV to 400 keV. At lower energies the intrinsic resolution is worse than the resolution for " T c m . Although there have been many advances in radiopharmaceuticals, the imaging with low-energy isotopes could be a useful complement to the standard methods [1, 2]. In literature a scintillation camera is described as a highly sensitive monitoring device for 12SI-contamination [3]. Another report describes the imaging o f 1251-fibrinogen with an Ohio-Nuclear Model 120 scintillation camera after changing the high-voltage supply [4]. The authors modified their camera (Picker, 3C /12) by increasing the Z-signal amplification and using an X-ray grid as collimator. The scintigram-documentation was performed by using the scintigram-plotter ‘daro 1160’.
271
272 ENDERT et al,
F IG . l.
N o rm a l d is trib u tio n o f 125I-fib rin o g e n in the c a lf vessels. T im e -a c tiv ity curves w ith no d iffe re n ce betw een r ig h t and le ft (Fig. la ).
Hom ogeneous d is trib u tio n 6 h a fte r inje ctio n . C o u n t density : 40 c o u n ts/m in p e r cm 2 (F ig . lb ).
IAEA-SM-247/61
273
METHODS The modification o f the Picker camera 3C/12 was realized by insertion o f a variable resistor at the analogue computer board to increase the amplification. The lower analyser threshold was then 25 keV. It became necessary to correct the windows of isotopes with energies above 45 keV but there was no change in the high voltage supply. At lower energies the intrinsic resolution is affected by the small number of electrons released in the photomultiplier tubes. Despite the application o f photomultiplier tubes with bi-alkali photocathodes, the lower energy limit is about 45 keV [5]. The intrinsic resolution o f collimated line sources was compared with "Tc™ and 125I (2-mm line source, collimator depth 75 mm, no additional scattering material). The values for the full width at half maximum were 8 mm for " T c m and 15 mm for 1251. Similar results in the FWHM were reported in literature with an Ohio-Nuclear Model 120 scintillation camera [4]. A critical point seems to be the choice o f collimator. There does not seem to be a collimator optimally designed for lowest energies at present. Therefore X-ray grids were investigated adapted for the collimator changing system (Picker, Clamshell). The plain source sensitivity was compared with an ultra-fine low-energy " T cm -collimator (Picker, ultra-fine). The X-ray grid had a grid ratio o f 8 : 1 and 40 lines per cm. The improvement in sensitivity comparing these two collimators for 1251 was 20.14 for the X-ray grid. An additional increase in sensitivity was possible by removing the absorbing aluminium cover of the grid [6]. In clinical studies we investigated 20 patients after IV injection o f 7 - 8 MBq of 1251-fibrinogen to study the fibrinogen kinetic in calf veins [7, 8]. Twenty patients were investigated with 1251-hippuric acid. Most of these patients were babies. The dose o f the radiopharmaceutical was 37 kBq per kg (1 pCi per kg). Typical dynamic studies were performed over 30 minutes and for interpretation a data system (Krupp, KANDI DS/C) was used. Ten patients were investigated with 131Cs-chloride to study the calf muscle blood flow [9].
RESULTS AND DISCUSSION The first clinical results showed the usefulness o f the recommended method. But there are some questions on the future o f 12SI- or 131Cs-labelled radiopharmaceuticals in imaging procedures. At present 1251-fibrinogen is widely used in the diagnosis of thrombophlebitis [7]. But it is also in this connection that " T cm-radiopharmaceuticals [8] are under investigation. The same arguments are valid in kidney studies. For the investigation of muscle perfusion [6] 201T1 is a useful isotope. Arguments supporting low-energy isotopes are: high efficiency
274
ENDERT et al.
IAEA-SM-247/61
FIG .2 .
275
Throm bophlebitis o f the left leg. D istinct hyperaem ic reaction
in the left c a lf with a rising tim e-activity curve (Fig.2a). Significant difference in the 1251-fibrinogen activity: 1 h (top left), 3 h (top right),
6h
(bottom le ft) and 24 h (bottom right) and an increase in this
difference (Fig.2b). Angiographic confirm ation o f throm bophlebitis with fresh throm bi in ca lf veins (Fig.2c).
o f specially designed collimators and a high absorption efficiency o f crystals. A disadvantage is a decrease in intrinsic resolution. The technical solution developed is very simple. The authors’ collimator is not optimally designed for highest sensitivity, nevertheless good counting statistics were obtained, in dynamic studies also. A significant lowering o f the applied doses seems to be possible. In the 1251-fibrinogen studies a thrombophlebitis was localized in calf veins. In the patients the thrombophlebitis was diagnosed by the bedside uptake test [7]
ENDERT et al.
276 ■
1** Я
21761 F IG .3 .
.
P flO O O
Intensive accum ulation o f 12SI-fibrinogen in a fresh throm busin a c a lf vein on the right.
and confirmed by contrast-venography, and the kinetics o f fibrinogen during therapy was investigated with the scintillation camera. A normal picture is shown in Fig. 1, and two pathological cases are shown in Figs 2 and 3. In the patient with only one active thrombus in the calf veins an intensive fibrinogen accumulation was measured (Fig.3). In the patient with diffuse thrombophlebitis (Fig.2) a hyperaemic reaction was found in the first thirty minutes after the injection of 125 I-fibrinogen and a further accumulation after 6 and 24 hours. Kidney studies are seldom undertaken with 125I-hippuric acid. One reason could be technical difficulties in isotope renography with small children and especially with babies. But in these patients there is no physical argument to prefer 131I-hippuric acid instead o f 125I-hippuric acid. The sensitivity made it possible to work with the small dose of 37 kBq per kg (1 ;uCi per kg). One typical examination is shown in Fig.4, the child having a body weight of 12 kg. A further lowering o f the dose seems to be possible. A non-invasive approach to peripheral vascular disease was described for thallium accumulation scintigraphy of legs. Orientating investigations o f calf muscles have been undertaken after the injection o f 8 MBq 131Cs-chloride.
IAEA-SM-247/61
277
(a ) .L> I
6
SO
FIG .4 a. Kidney exam ination in a sm all child o f 12-kg body weight. D ynam ic study over 30 min.
6 80
! ií®
3 0 -3 5 v
¡
Z 6581 F IG . 4b.
M IN
P ftO O O 2
Same subject as in Fig.4a. The picture clearly shows the im paired urine flow
on the right after 25 min. Count density: 150 counts/min per cm over the right kidney.
278
ENDERT et al.
F IG . 4c.
Same subject as in Figs 4a and 4b. The m orphologic changes dem onstrate the
urographie examination.
Dynamic studies were carried out after muscle exercise and the time-activity curves measured for more than 20 minutes. The accumulation seems to correlate with the muscle blood flow, but a final assessment o f this method is not yet possible. There appear to be some useful indications for camera investigations with low-energy isotopes at present. The main advantage is the high sensitivity of the collimator-camera system. A further improvement in collimator design might make possible a further lowering in radiopharmaceutical doses.
REFERENCES [1] [2]
R A IK A R , U .R ., G A N A T R A , R .D ., A tten u a tio n o f m onoenergetic p h o to n s from I25I in th y ro id nodules, In t. J. N ucl. M ed. Biol. 5 (1 9 7 8 ) 92. STRAN D, F .E ., PERSSO N , B .R .R ., The dual p h o to p ea k area m eth o d applied to scintillation cam era m easurem ents o f effective d e p th and activity o f in vivo 123Idistrib u tio n , E ur. J. Nucl. M ed. 2 (1 9 7 7 ) 121.
IAEA-SM-247/61
279
T A Y L O R , A., e t al., M onitoring o f 1-125 contam in atio n using a po rtab le scintillation cam era, J. N ucl. M ed. 19 (1 9 7 8 ) 431. ADAMS, R ., e t al., Im aging 1-125 w ith a scintillation cam era, J. N ucl. M ed. 19 (1 9 7 8 ) 1259. A N G E R , H .O ., In stru m en tatio n in N uclear M edicine, A cadem ic Press, New Y ork (1967). K L O SE, E ., E N D E R T , G., R IT T E R , E .P., „K am era-Szintigraphie m it niederenergetischen N ukliden” , R adioaktive Isotope in K linik u n d Forschung 14, H . E germ ann, V ienna (1980). D E N A R D O , G .L ., e t al., A ssessm ent o f conventional criteria fo r th e early diagnosis of th ro m b o p h le b itis w ith the 12SI-fibrinogen u p ta k e test, R adiology 125 (1 9 7 7 ) 765. JO N C K H EE R , M .H., et al., The in te rp re ta tio n o f phlebogram s using fibrinogen labeled " m Tc, E ur. J. N ucl. M ed. 3 (1 9 7 8 ) 233. SIEG EL, M .E., SIEM SEN, J.K ., A new noninvasive approach to p eripheral vascular disease: thallium -201 leg scans, A m . J. R oentgenol. 131 (19 7 8 ) 827.
Poster Presentations DESIGN OF A FAST POSITRON CAMERA HEAD* C. NAHMIAS, D.B. KENYON, E.S. GARNETT, K. KOURIS+ Nuclear Medicine, Section o f Radiology, McMaster University, Hamilton, Ontario, Canada
In general, studies o f regional brain metabolism or brain blood flow require that an image be collected every 4 s, or faster. This is particularly the case in studies o f intracerebral dopamine metabolism for which the authors have designed and tested a positron camera head. A bismuth germanate rather than sodium iodide or caesium fluoride was chosen because o f the higher stopping power of bismuth germanate for 511 keV photons. An added advantage o f bismuth germanate is that this crystal does not require encapsulation allowing for closer packing o f the crystals. Experiments with bismuth germanate crystals o f varying sizes have shown that the face o f the crystal exposed to incident radiation cannot be made smaller than 8 mm X 20 mm. With smaller faces, there is an undue loss o f detection efficiency. Experiments have demonstrated an increase of the order of 20% in detection efficiency for 511 keV photons when the crystal depth is increased from 30 mm to 50 mm. The need for high Z inserts between adjacent crystals was also investigated. It has been shown that the spatial resolution is slightly degraded (o f the order o f 10% o f the FWHM) when 2-mm thick tungsten inserts are omitted. It is suggested that for the high efficiency instruments needed for fast dynamic studies, thin inserts add little to the spatial performance. On the basis o f these results, a prototype instrument has been built and tested comprising two banks o f four detectors each. The completed instrument will comprise 160 bismuth germanate crystals, 10 mm X 30 mm X 50 mm, closely packed on a ring, 26 cm in radius. The detectors will not be encapsulated in aluminium or separated from each other by inserts. It is expected that this positron camera will be at least 20% more efficient than the present generation cameras while main taining the spatial resolution at better than 1 cm (FWHM). *
W ork sup p o rted by the Jo h n A. B auer M em orial F u n d and th e M edical Research
C ouncil of Canada. ^ Present address: D ep artm en t o f Physics, U niversity o f S urrey, G uildford, U nited K ingdom .
281
POSTER PRESENTATIONS
282
INTERDEPENDENCE OF DEAD-TIME LOSSES AND CAMERA INHOMOGENEITY D. LANGE, H.J. HERMANN, Universitäts-Strahlenklinik, Department of Radiology, Division o f Nuclear Medicine, Heidelberg, Federal Republic o f Germany
Because o f the long decay time o f the scintillation light pulses may sum up at high photon fluence rates. False addresses are generated at co-ordinates without any photon impinging at that particular place. The intensity and the distribution of these false addresses depend on the Compton signals. Studies can be made only in scattering medium. A flat circular source o f 1 cm X 22 dia. was filled with about some 200 mCi " T c m. It was placed in a thick water basin 5-cm deep to simulate the clinical scatter condition. Images were stored in 4K matrices for successive periods of 30 min for at least 60 h (10 half-lives). From the images with low activity at the end o f one run the ‘expected images’ at high photon fluence rates can be calculated applying the radioactive law. The expected total count rate of all imaged signals is compared with the measured count rate and the overall correction factor kz is calculated. This is the well-known ‘dead-time’ concept.1 Moreover the images o f the flat source are no longer proportional to the expected images because of a remarkable inhomogeneity at the centre o f the source. Dead-time losses are lowest in the centre and highest for peripheral co ordinates, due to pile-up addresses from Compton signals which add up most frequently to central addresses. Regional correction factors kr are calculated by comparing the expected images with the measured ones. The factor kz is the average o f the regional correction factors kr: /k rr dr z
/rd r
The central inhomogeneity o f the correction factor kr is the most sensitive parameter at high count rates. It seems unreliable to correct images by a calculated correction factor kz deduced from the dead-time concept, if the inhomogeneity exceeds 10%. This limit is different for different cameras and depends on the ‘speed’ o f the system. A ‘fast’ camera produces similar images at high photon fluence rates. Unfortunately the regional distribution o f regional correction factors cannot be calculated for clinical conditions. 1 1 C i = 3.70 X 10IOBq.
POSTER PRESENTATIONS
283
DISCUSSION R.M. ADAMS: In your poster presentation you show clearly considerable pulse pile-up and resulting displaced data and regional variation in data losses at high count rates. However, in a clinical setting, where data losses should be kept below 2 0 —25%, how important are the effects you mention? From one o f the charts shown in your presentation, it appears that with data losses o f 20% there might be a 5% error in the dead-time correction. Five per cent o f 20% is one percentage point and, therefore, quite acceptable. D. LANGE: The pile-up effect will be more significant if a cold area is surrounded by high activity producing false addresses in the cold area. The error produced by normal dead-time correction will be higher than that reported by us. A clinical example in cardiac studies is the period during the first passage o f the tracer when the activity is spread over the lungs, and the heart chambers are rather free o f activity. The error is normalized to the expected true count rate, not to the correction. The corrected loss of 20% leads to an overestimation o f the central count rate by 5% of the total corrected count rate. R.M. ADAMS: Then how do you recommend calculating temporal resolution and the maximum clinically useful counting rate? D. LANGE: We do not calculate temporal resolution or dead time because these mathematical numbers will be spread inhomogeneously over the image. The useful count rate depends on the performance of the camera and on the distribution of activity and scattering material. It must be measured with clinical phantoms.
AUTOMATIC ALIGNMENT OF RADIONUCLIDE IMAGES D.C. BARBER Department of Medical Physics and Clinical Engineering, . Royal Hallamshire Hospital, Sheffield, United Kingdom
The correct positioning of the image o f an organ within the field of view of a gamma camera depends largely on the skill o f the operator. For static imaging the organ can usually be centred using a persistence oscilloscope; for dynamic studies the- organ must be positioned mainly by the use o f anatomical
284
POSTER PRESENTATIONS
landmarks. In addition to variations in position, the shape o f the organ will vary between patients and such patient-to-patient variations in shape and position hamper attempts to analyse numerically radionuclide images. An efficient method has been devised for adjusting the position, linear dimensions, rotational orientation and other possible image distortions (all these adjustments being called by the generic name of displacements) until the image conforms as closely as possible to a standard image o f the same type. The method utilizes a non-linear mapping between the displacements required to standardize an image and the coefficients of a principal components transformation o f that image. The mapping is produced from a population o f displaced standard images and problems o f non-linearity are overcome by an iterative approach. The method has been used to standardize a variety o f radionuclide images including those o f brain, liver and lung. Use of image alignment has relevance to automatic methods o f analysis o f static images and for standardization o f some dynamic studies prior to the use o f pre-drawn regions o f interest.
INFORMATION PROCESSING AND DISPLAY Session 3
IAEA-SM-247/204
Invited Review Paper NEW DEVELOPMENTS IN TECHNIQUES FOR INFORMATION PROCESSING IN RADIONUCLIDE IMAGING R. DI PAOLA Groupe de radiobiologie clinique, Institut Gustave-Roussy, Villejuif, France A.E. TODD-POKROPEK* Faculty o f Clinical Sciences, University College, London, United Kingdom
Abstract NEW DEVELOPM ENTS IN T ECH N IQ U ES F O R IN FO RM A TIO N PRO CESSIN G IN RA D IO NUCLIDE IMAGING. Processing o f scintigraphic data has passed th ro u g h differen t stages in the past fifteen years. A fter an ‘eup h o ric’ era w hen large off-line co m p u ter facilities were used to process very lowquality rectilinear scan pictures, a m uch m ore critical p eriod follow ed the in tro d u c tio n of on-line m in ico m p u ter system s to acquire, process and visualize scintillation cam era data. A selection o f som e o f the available techniques th a t could im prove the e x tra ctio n of info rm atio n from scintigraphic exam inations in ro u tin e is presented. T om ography has been excluded. As exam ples, the d iffere n t techniques o f (a) inhom ogeneity correction of cam era response and (b) respiratory m o tio n corrections are used to show one evolutionary process in the use of co m p u ter system s. F iltering has been fo r a long tim e the m ajor area o f research in scintigraphic image processing. O nly very sim ple (usually sm oothing) filters are w idely distributed. L ittle use o f m ore ‘p ow erful’ filters in clinical d a ta has been m ade, and very few serious evaluations have been published. Nevertheless, the n um ber o f installed m inicom puter and m icroprocessor system s is increasing rapidly, b u t in general perform ing tasks o th er th an filtering. The reasons for this (relative) failure are exam ined. Som e ‘new ’ techniques o f image processing are presented. The com pression o f scintigraphic in fo rm atio n is im p o rta n t because o f th e expected need in the near fu tu re fo r handling o f large num bers o f static pictures as in dynam ic and tom ographic studies. F o r dynam ic in fo rm a tio n processing, the presen t m ethodology has been narrow ed to those techniques th a t’p e rm it the entire ‘d a ta space’ to be m anipulated (as opposed to curve fitting a fte r region o f in te rest d efinition). ‘F u n c tio n a l’ imaging was the first step in this process. ‘F a c to r analysis’ could be the n ex t. The results o btained by various research laboratories are reviewed. * Also at: S.H .F J . D ép artem ent de Biologie, C.E.A ., H ô p ital d ’O rsay, O rsay, France.
287
288
DI PAOLA and TODD-POKROPEK
INTRODUCTION In 1967, an article by Smith and Brill [1] was published reviewing the use o f computers in nuclear medicine. It is interesting to observe the changes that have occurred since that date. It appears that, following a period o f ‘euphoria’ when the use o f a computer was almost self-justifying, we are now in a period where the evaluation o f the costs and benefits o f the techniques used is as important as the development o f the techniques themselves. It is therefore probably not a coincidence that several reviews have recently been published in this area [ 2 - 4 ] and even a (very short) ‘History o f computers in nuclear medicine’ [5]. It seems, however, unreasonable in this survey to limit the discussion o f the techniques themselves to the benefit o f discussing their clinical impact, even though this seems often to be accepted practice [6]. To be sure, it would be possible to discuss the value o f nine point smoothing in a variety o f clinical procedures, but would this be o f value? It is also true that very few systems offer the user, in clinically acceptable conditions, the possibility o f using more complex filters. Nevertheless, whereas an exhaustive review covering all the different processing techniques which have been employed (and still less, those that could or should be used) is excluded, it seems interesting to present several examples o f areas where data processing has been employed in order to obtain a better understanding o f the techniques available, and the problems o f their implementation. These are, in order: correction o f nonuniformity, correction o f respiratory movement, resolution recovery and filtering, data compression, and processing dynamic data. Tomography has been specifically excluded.
CORRECTION OF NONUNIFORMITY Almost as soon as the first cameras were interfaced to computers, to the delight o f the manufacturers, matrix division correction as early as 1967 [7] was presented as a panacea for this problem. Anger himself [8] stated that whereas such a technique only corrects variations o f sensitivity and not those due to spatial distortions (saying that the latter were much smaller than the former), it was quite appropriate to use it. Smith [9] stated, in 1971, that it was in fact more efficient to ask the manufacturer to adjust the camera to optimize uniformity and to provide good maintenance than to rely on computer methods. Nevertheless, he was a proponent o f the same method. Adam, one of the initiators o f the technique, admitted in 1972 that he only used that method since no other method was then available to take the other distortions into account [10]. In the same year, most o f the data processing systems presented at the IAEA Symposium in Monte Carlo boasted use o f the same correction.
IAEA-SM-247/204
289
Although a certain number o f papers were published warning o f the dangers o f the use o f this method [11 —13] it has only been since 1976 that detailed studies o f the response o f gamma cameras and o f the causes o f nonuniformity have been pre sented [1 4 —18]. The important role o f the computer in such studies should be stressed, and included in any ‘balance-sheet’ o f the value o f computers in this domain. Table I shows the most recent publications in this area in an attempt to summarize the different techniques proposed to correct for the various components o f distortion (variations in energy distribution, spatial distortion, variations in sensitivity, variations in the shape o f the point spread function (PSF)). The pioneering work o f Muehllehner in 1973 [19] must be acknowledged. Nowadays, the most common solution adopted by manufacturers appears to be by means o f hardware (a microprocessor) included directly in the (digital) gamma camera head [25, 26]. The use o f this technology apparently permits the reduction o f variations o f uniformity to the order o f 1%, as opposed to 13% reported in 1968 [27]. Perhaps using similar techniques it will also be possible to provide a suitable dead time correction. This area o f research is omitted for reasons o f space, the most recent publication being Ref. [28].
CORRECTION FOR RESPIRATORY MOTION Whereas uniformity appears to be a success story, correction for respiratory motion appears now less satisfactory. Oppenheim [29] used in 1971 a method of calculation o f the centre o f gravity as a function o f time for an ensemble o f images with translation to supçrimpose the median line to obtain a composite static image. Di Paola and co-workers [30] described a more convenient variation o f the same technique using list mode data. Schmidlin [31, 32] also proposed a method using list mode data consisting o f a high pass filter in the frequency domain, which can also be implemented in the time domain using a recursive Butterworth filter [33]. Schmidlin compared various filters in the time domain [34]. However, the use o f the technique in clinical practice (and its value) seems to be limited. Research in this area is currently directed more towards analogue methods [35, 36] based on earlier techniques [3 7 - 3 9 ] but likely to be supplanted by the use of microprocessors. A recent publication [40] is o f interest in its description o f a technique o f using the respiratory signal to provide forms o f ‘gated’ images for pulmonary and other studies. The existence o f important respiratory and movement artefacts seems to indicate that the relentless hunt for resolution (necessarily at the expense of sensitivity), which the various manufacturers have led, is perhaps ill-directed. The problem o f motion artefact in CT was only eliminated by the use o f very fast (e.g. sensitive) devices.
290
TABLE I. RECENT NONUNIFORMITY CORRECTIONS Year
Type
Energy correction
Spatial correction
T odd-Pokropek
[17]
1977
Software
Sliding energy window correction w ith a Z experim ental energy 64 X 64 m atrix indexed by X and Y, fitting the photopeak
Soussaline
[18]
1978
Software (list mode)
Sliding energy window correction with a Z experimental energy 64 X 64 m atrix indexed by X and Y, fitting the photopeak
Calculation o f the true location from X and Y positions using experim ental param eters
Steidley
[20]
1978
Hardware
Digital alignment o f energy peak (64 X 64 table)
Rejection of linearity shifted counts from areas of high co u n t rate according to a 64 X 64 factors table inversely p roportion al to the energy response to flood field data
Kirch
[21]
1978
Algorithm
64 X 64 m atrix of positional shift factors from flood field data
Stoub
[22]
1979
Hardware
Spatial distorsions correction w ith a 64 X 64 array coefficients refined using the gradient o f the flood field data
Knoll
[23]
1980
Hardware
Variable pulse selection by 64 X 64 experim en tal table of spatially dependent energy window
T ranslation o f X and Y positions using a 64 X 64 experim ental table
Blond
[24]
1980
Software (Image intensifier camera, list mode)
Translation of energy value with a 64 X 64 experim ental table indexed by X and Y
Spatial distortion corrections due to the co ordinate com putation principle w ith com puted 256 X 256 table
and TODD-POKROPEK
Ref.
DI PAOLA
A uthor
291
IAEA-SM-247/204
FILTERING OF SCINTIGRAPHIC IMAGES In the beginning (1 9 6 4 -6 7 ) there was a very rapid development o f ‘off-line’ techniques using isocontours [4 1 —44], statistical analysis [41], image subtraction [43], data bounding [45], spatial averaging [43], smoothing [46], dot shifting [48] and resolution correction [4 9 -5 1 ], all with certain clinical success. Certainly the area which has proved the most controversial is that o f ‘resolution recovery’. This area has, o f course, not been the preserve ( ‘chasse gardée’) o f workers in nuclear medicine but rather has been the subject o f extensive publications in other domains [5 2 —55]. It has now become accepted practice to classify those methods designed to reduce statistical fluctuations and to improve resolution dependent on whether or not they are stationary [2, 4, 56]. Table II summarizes published methods with no pretence at being either rigorous or comprehensive. Although little used in the scintigraphic literature, it seems helpful to present the restoration problem using the linear algebraic approach now conventional in digital image processing [7 0 -7 3 ]. Using as a first order approximation a linear space invariant model, the usual two-dimensional discrete convolution summation between an object f(i, j) and a periodic PSF h(i, j) o f size N X N, with a noise term n(i, j), gives an image g(i, j). N -l
N-l
g (i,j)= X ) X ) f(k, l)-h(i-k, j-1) + n(i, j) k=0 k=0 Let column vectors f, g, n be defined for example such that g' = [g(0, 0) g (l, 0) . . . g (N -l, 0) g(0, 1) g ( l, 1) . . . g(N -l, 1) . . . g(0, N -l) g ( l, N -l) . . . g (N -l, N -l)] The image restoration model is now given by g=H f+ n where H is o f dimension N 2 X N 2 partitioned into N 2 matrices where
H=
’ [H0][H N. , ] [ H J [ H 0]
. ..[ H ,] . . [H2]
[HN- i ] [ H N. 2 ] . . . [H0]
and Hj -
hü, 0) hü, N -l) . . • hü, 1) hü, 1) hü, 0) . . • h ü ,2) hü, N -l) hü, N-2) . ■hü, 0)
292
TABLE II.
STATIONARY AND NON-STATIONARY METHODS
W eighted averaging [43] G aussian w eighting [46] D ynam ic 3-dim ensional filtering [47]
Iterative spatial dom áin [ 4 9 - 5 1 , 6 2 ] Single step-frequency dom ain [62] W iener filte r [6 3 ], [64] U nsharp m asking [6 5 ], [66] C onstrained least squares [67]
s t a
t i
о n a r У N
Ö n s
t
a
t
i о n a r У
D ata b ounding [45] Spatial averaging [48] V ariable shape averaging [57] [58] [59] M edian sm oothing [60] A nscom be transform [61]
[56] N on-stationary M etz-filtering [59] [68] Increm ental deconv o lu tio n [69]
and TODD-POKROPEK
R esolution recovery DI PAOLA
S m oothing
IAEA-SM-247/204
293
All the diagonal terms are equal (it is a Toeplitz matrix), and each row is a circular right shift o f the preceding row, so all Hj is a circulant matrix and as such can be diagonalized by a discrete Fourier transform (DFT) computation. H is not only a block Toeplitz but also a block circulant matrix. Using the least-squares Lagrangian method [70], with different objective functions for minimization, it can be shown that the optimal filtering in the case o f the constrained least-squares solution gives an object estimate f = ( H * / H + T Q 'Q r 1H*'gwhere Q is a linear operator on f to minimize o f |Qf|2 subject to |g — H f|2 = |n|2, 7 (= l /Л) is related to the usual Lagrangian Л multiplier and must be computed by iterative techniques, and where * denotes the complex conjugate and ' the transpose. Consider three special cases: (a) Q = Identity matrix I. Here f= (H * 'H + 7 l ) _1H *, g corresponds to the pseudo inverse filtering. Two other similar approaches are: singular value decomposition (SVD) and the Van Cittert method [52]. In addition 7 = 0 corresponds to the inverse filtering with Fourier solution given by F(u, v) = G(u, v)/H(u, v) (b) For Q = a finite difference matrix, with Fourier transform P(u, v). A smoothing constrained filter is obtained, with a corresponding Fourier solution [71]
P( (U,V)
=
G(u, v) H*(u, v) |H (u ,v )|2 + 7 lP (u,v)|2
This solution has been used for scintigraphic pictures by Boardman [67], Q being a Laplacian operator but may in fact be chosen [70] as any order difference from the estimated object. (c) For Q = Ф"1/2- $ i /2 where f and Фп are the co-variance matrices o f the object and the noise, para metric Wiener filtering is obtained: f = (H*' - Н + тФ?1 Ф„)-1 H *'g
294
DI PAOLA and TODD-POKROPEK
If 7 = 1, traditional Wiener filtering is obtained for which H*(u, v) G(u, v) F(u v) — ---------------- ----------- ------------|H(u, v)|2 + Pn(u, v)/Pf(u, v) where Pn(u, v) and Pf(u, v) are the power spectra o f the noise and the object. This filter is probably the most commonly used o f the so-called sophisticated operators. As Andrews and Hunt have stated [70], ‘the use o f the Wiener filter to give “optimal” restoration is part o f the “folklore” o f image processing. . . ’. Another class o f filter combining the advantages o f inverse filtering gives for the estimate o f the object f = (inverse filter)“ •(parametric Wiener filter) 1 'a g where 0 < a < 1. Thus: f = [(H *'H ) _1 H*']a [(H *'H + 7 $ j ‘ ФПГ ! H *']1_0:g and, for a stationary invariant PSF, in the Fourier domain
For 7 = 1 and a = 0, a pure traditional Wiener filter results and for a = 1, an inverse filter. All the intermediate values for a are possible. For example, a.= 1 /2 corresponds to the geometrical mean o f the filters G(u, v)
This corresponds to ‘homomorphic filtering’ [74]. It is simple to add to the constrained least-squares problem an additional constraint o f ‘non-negativity’ (which is certainly appropriate in scintigraphy) such that f > 0. Unfortunately, this does not permit a closed form solution such as given above to be expressed, and non-linear methods must therefore be used [75, 76]. In other words, when the underlying assumptions o f constant mean and covariance stationary random process are not appropriate [72], other methods become advantageous. Cormack and Hutton [69] have proposed an incremental deconvolution method using list mode. From Bayes’ theorem, the probability that the jth scintillation located at r2 in the image originated from pixel rj on the object is given by
IAEA-SM-247/204
295
fj-1 (rJ-hC rj/rj) Pj(rj r2) = -------------------------------£ fj-l (г^'М гг/г!) all ri where fj.j (rj) is the previous estimate o f the object and h the PSF. Then
fj-1
=
fj-1 fr .) + Pj(r,/r2)}
It should be noted that one o f the most exciting filters for static and dynamic studies, Kalman-Bucy, has never, it appears, been used for processing scintigraphic data. According to Andrews and Hunt [70] \ . . it has the potential o f making it the most attractive o f all processing schemes’. Among the reasons suggested for the lack o f use o f these filters, three are commonly given: (a) Lack o f convenience in use and in implementation. (b) Lack o f conviction o f the potential user in their utility. One o f the principal conclusions o f the IAEA intercomparison [77] was that, while filtering appeared highly likely to improve digitized images, there was certainly a loss o f image quality in the digitization process itself, as well as in the display, which these techniques correct only partially. It must be constantly stressed that data processing is totally dependent on the quality o f the data capture. Even a miraculous filter [78] does not permit the recuperation o f information lost as the result o f bad quantification o f the digital image. (c) Manufacturers have little incentive from the user to include such methods, and rarely make the effort. (It may be noted that the wild goose chase o f matrix division uniformity correction was dictated by the users, while the manufacturers, in fact, possessed the necessary information to provide better techniques [19].) What can be done to improve this situation? Three suggestions seem helpful: (a) Implement algorithms requiring only simple (e.g. cheap) systems [7 9 -8 3 ]. (b) Encourage other intercomparisons aimed at producing conclusions in valid conditions comparing different filters. Several recent publications [84—86] show the value o f image processing when it is performed using good data (from the point o f view o f sampling, display and training o f the observers) and with appropriate hypotheses [87]. (c) Encourage the manufacturers to supply interfaces which digitize the input data satisfactorily, and to provide appropriate software.
296
DI PAOLA and TODD-POKROPEK
Before concluding this section, it is appropriate to cite work performed in the domain o f selective structure enhancement, e.g. Lehr and co-workers [88]. Neill and Hutchinson have developed a method o f minimum and maximum convexity [89] which has given very good results in several intercomparisons [77, 78]. Structure analysis [90, 91] using the curvature in scintigrams has also given useful results [77, 91]. The use o f recognition operators has also been described [92].
COMPRESSION OF SCINTIGRAPHIC DATA Compression o f scintigraphic data does not at present seem to be o f great interest to most laboratories working in nuclear medicine in data processing. This is surprising. There has been a considerable development in data compression [ 7 0 - 7 3 , 93] since the initial work o f Shannon [94]. This could be a result of little clinical use o f long-term storage and transmission o f data, the cost in computer time and in implementation and, above all, doubts about the image quality o f compressed data. Unfortunately it is out o f the question to include a complete review o f all the techniques currently available in the rapidly expanding field o f image compression. In their recent classification, Netravali and Limb [95] suggested four major categories for general waveform coding: Pulse code modulation, predictive coding, transform coding and interpolative and extrapolative coding and a further fifth class including all other schemes (and not an empty set! ). Usually these methods are either fixed or adaptive (i.e. data dependent). In the medical field some very simple techniques have already been applied [9 6 —98]. Ram [99] used a gradient encoding method based on a simple gradient operator which allows the determination o f the presence o f an edge and its direction in scintigraphic images; the encoding algorithm suggested for a first pass dynamic heart study saved 99% o f storage by reducing the complete set o f ungated frames to a ‘graph-like’ description, and 88% for a static lung study. One o f the methods which has been o f greatest interest to research workers recently is transform encoding, summarized in Fig. 1. Let the matrix [G] denote the N X N matrix representation o f a scan. It may be written as [G] = [A] [T] [B]'
(1)
where [A] and [B] are assumed to be unitary, i.e. [A] [A]' = [I] = [B] • [B]' Solving for [T], Eq. (1) yields the relation [T] = [A]' [G] [B]
(2 )
[T] is the unitary transform o f the scan [G], [A]' transforming the columns o f
IAEA-SM-247/204
IN V E R S E
FRO M ST O R A G E OR
T RA N SFO RM
T R A N S M IS S IO N
F IG .l.
297
REC O N STRU C TED IM A G E
Transform encoding.
[G], and [B] the lines. Equations (2) and (1) define the forward and the inverse transform. For Fourier and Walsh-Hadamard transforms, [A]' = [A] = [B], and for Haar, slant and cosine transforms, [A] = [B] only. The Jordan decomposition theorem shows that any matrix [G] may be represented by [G] = [ U ][A ] i /2 [V]'
(3)
where [U] and [V] are orthogonal matrices, with [G] [G ]' - [U] [A ][U ]' [G]' [G] = [V] [A] [V]' where [A] is the diagonal matrix o f the eigenvalues o f the symmetric matrix [G] [G]'. The columns o f [U] and [V] are eigenvectors o f [G] [G]' and [G]' [G]. This expansion is a singular value decomposition (SVD) [100] o f the image [G] into its singular vectors and [T] = [U ]'[G ][V ] All the above transforms are independent on image statistics. Another class o f transform is Karhunen-Loève1 [101, 102] (KLT) also known as the Hotelling [103] or principal component transform. Starting from 1 See R osenfeld [73] fo r the difference betw een discrete and co n tin u o u s cases.
298
DI PAOLA and TODD-POKROPEK
Eqs (1) and (2), look for [A] and [В] which decorrelate the rows and columns o f [Т]. The correlation between columns o f [T] is given by E [Tj -Tj ] and between rows by E [(T')i (T)j] where E [T 'T ] = [A c] and E[TT'] = [Ar], E[ ] being the expected value and [A c] and [Ar] two diagonal matrices representing the variances o f the columns and the rows. Using Eq. (2), then [T ]'[T] = [B]'[G ]' [G] [B] [T] [T]' - [A]' [G] [G]' [A] Thus [A c] = [B]'E [G'G] [B] and [Ar] = [A]' E [G G'] [A] The matrices [A] and [B] determined thus diagonalize the covariance matrices o f the rows and columns, which may also be expressed in the form E[GG'] = [KLr] [Ar] [KLr]' E [G 'G ]= [K L C] [ A C][K L C]' These expressions should show clearly the difference between SVD and the KLT since the latter requires the determination o f the covariance matrices. While the term ‘optimal’ must be used with caution in image processing, for data reduction, the KLT can be considered as an optimum transformation when the signal statistics is known since it allows total decorrelation o f the transform coefficients. Unfortunately, unlike the transforms mentioned above, which employ a fixed set o f unitary matrices independent o f image statistics, there is no fast computation algorithm. Correspondence analysis [104] was used by DiPaola and co-workers [33, 105] for the compression o f scintigraphic data in a similar manner to Maraño for videophone images [108]. In correspondence analysis, the initial values gy are transformed by
as opposed to gy — gj as for PCA. For convenience in computing, subpictures are used with subblocks o f from 4 X 4 up to 32 X 32 pixels, eigenelements being extracted from each subpicture.
299
IAEA-SM-247/204
Schmidlin [107] has described a method using PCA and row sampling. In this case the technique reduces to an SVD from Eq. (3) and for rank r < N:
[G] = £ i=l
X¡/2 Uiv'i
The outer product u¡ vj is termed an eigenimage [70]. It may be noted that whereas the use o f the SVD algorithm was originally suggested for image coding in 1970 [ 108], it was only five years later that it was in fact implemented [107, 109, 110]. For all random processes there is no single unique KLT. In the case o f first order stationary Markov processes, the covariance matrix is a Toeplitz matrix whose elements are r ^ ¡ with |r| < 1, and 1 < i, j < N. Even when the eigenvectors are known analytically their computer generation is expensive, and there is no fast algorithm to perform the transform. For r near 9, the discrete cosine transform (DCT) is very close to the KLT [111, 112] and yields better performance than the DFT for r > 0 [ 113]. The DCT o f a real N point sequence v(n) is defined by
N -l
V(k) = £
v(n )c(k ) cos [(2n + 1) 7rk/2N]
к = 0, 1, . . . , N-l
n=0
where c(k) = 1 for к = 0 and c(k) = y/l for к = 1, 2, . . . , N - l. The inverse DCT is defined by
i ” ■* v(n) = — £ V (k) c(k) cos[(2n + 1) 7rk/2N] N ^ k=0
n = 0, 1, . . ., N-l
The DCT coefficients V(k) are real numbers for real v(n). From a practical point o f view the DCT o f an N-point signal can be computed using a 2N-point DFT transform [111]. More recently other methods have been described to compute this transform [114, 115]. For completeness it should be pointed out that according to Jain [116], the DCT is not always the best substitute for the KLT
300
►4
б'
9
s'
\
I /
i о ■“ n
1 I
00 I
DI PAOLA and TODD-POKROPEK
FIG .2 .
Zigzag scanning pattem o f the subpicture in the transform domain.
F IG .3 . Quantization and coding in the transform dom ain: IQ 2 is used from 2 to IL and IQ 1 from I L + 1 to 256. The number o f bits to code the values is different in the regions 2 to IA , IA + 1 to IB and IB + 1 to IL . From IL + 1 to 256 the length o f zero sequences is coded.
TABLE III. COMPRESSION EXAMPLES
Origin
Sam pling
N um ber of bits/pixel
CT scan
256 X 156
1.35
3.3a
9.6
G am m a scan
128 X 128
1.13
13.0 a
12.6
G am m a scan
256 X 256
0.93
4 16.0b
16.0
a Digital (V A X ). b Infórm atele (SIM IS 3) ■
CPU tim e (s)
C om pression rate
IAEA-SM-247/204
301
for first order stationary Markov processes and, for example, very recently, a new transform, the symmetric cosine transform (SCT) has been proposed by Kitajima [117] who claims that SCT is very efficient in bit-rate computational efficiency, residual correlation and the rate-distortion criterion. Using an adaptive DCT, picture compression routines have been implemented for static, dynamic and tomographic gamma scans, X-ray scans, ultrasound pictures, and a number o f other image types. The picture to compress is divided into subpictures o f size 16 X 16. If the sum o f the contents o f the 256 pixels o f a subpicture is lower than a threshold given by the user, it is not processed. The 256 pixels are used according to a zigzag scanning pattern as shown in Fig. 2. The variance distributions have been studied in the transform plane and a quantization method [ 118] o f the coefficients o f these planes summarized in Fig. 3 has been used. Sample timings are given in Table III. Further improvement for multiple image studies could be achieved using interframe transform coding [119, 120]. DYNAMIC STUDIES Isotopic dynamic studies range from simple studies such as the clearance o f an injected product (single exponential model) to the analysis o f complex metabolic functions (multiple compartmental model with difficulties with respect to the validity o f the model itself). It is clear that there has been a considerable development in recent years in functional or parametric images [ 121 ] which permit considerable reduction of the initial data. It seems surprising that Hoehne [122] should state that functional imaging in nuclear medicine has not yet reached the stage o f clinical applicability possible in conventional radiology (ten years later). Nevertheless a great variety o f scintigraphic transforms have proved to be o f clinical use [4]. It is important that any model used be validated (over-simple models must be avoided) and that precautions be taken with respect to poor signal-to-noise ratio [ 123]. A t the moment the most commonly used methods consist o f techniques to determine regions o f interest (often automatically), followed by the determination o f time activity curves and a choice o f (multiple) exponential fitting, deconvolution, Or the use o f other more complex functions derived from various models. The major problem is that o f expressing the physiological features o f interest in the form o f a suitable model, which, contrary to the desires o f certain physicians, requires an appropriate mathematical formulation. Figure 4 shows a flow chart o f the various possibilities that exist at present for a research-orientated group for use in clinical practice. It should also be stressed that in clinical routine a good display is required not only for static images but also for dynamic function studies. An appropriate image compression routine should be useful thereafter. Factor analysis could well be o f considerable value
302
F IG .4 .
DI PAOLA and TODD-POKROPEK
Flow-chart o f the possibilities existing a t present fo r the use o f dynam ic studies
clinical practice.
IAEA-SM-247/204
303
in this area. Storing o f ‘factors’ (especially when limited by space) not only permits a compact description o f the study to be preserved but also enables classification and visualization o f ‘factorial’ images. Whereas such an analysis' might answer most clinical questions, the use o f an interactive model may serve to extract further classification parameters. Like principal component analysis (PCA), factor analysis (FA) tries to reduce an ensemble o f data to a smaller number o f dimensions than originally. However, for statisticians, these two approaches are not identical, even if, as pointed out by Watanabe [124], the common ground o f KLT and FA was found long ago in the diagonalization o f the correlation matrix proposed in 1901 by Pearson [125]. The main differences [126] may be expressed as follows (a) PCA is merely a transformation o f the data, whereas FA supposes an underlying well-defined model; (b) FA is essentially a transformation from the underlying factors to the observed variables, whereas in PCA it is the determination o f the principal components from these variables. In 1974, Barber and co-workers [127] employed PCA for gamma-camera gastric emptying studies, using 32 point time curves over the stomach. The first seven principal components permitted an accurate reconstruction o f the original data. A derived parameter enabled discrimination o f a control group from patients with duodenal ulcers or truncal vagotomy. Similar studies have been performed by Schmidlin and co-workers using correspondence analysis for renograms [128 —131]. This permits the extraction o f few parameters characterizing the curves in detail. Moreover, ‘factor scintigrams’ may be visualized using the largest factors. Oppenheim and co-workers [132, 133] and MacLeod [134] have also used PCA for the analysis o f renograms and Houston [135] has investigated its use for dynamic studies o f kidneys, livers and thyroids, with a comparison with a physiological model. The approaches used by Di Paola and co-workers [33], Bazin and co-workers [136] and Barber [137] are significantly different. In any arbitrary region with the smallest size corresponding to one pixel, the activity time curve is given by S
y¡(t) =
^2
Pik $ k (t) + noise
k=l where s is the number o f unknown physiological factors (or dynamic components), $k (t) the kth factor and the kth coefficient o f the ith region. Each o f these components has a precise functional signification and can be called the principal dynamic component. Suppose that the components Фк and Pik can be determined, each coefficient pjk can be used to generate a functional image reflecting the importance o f the corresponding principal component. The
304
DI PAOLA and TODD-POKROPEK
orthogonal components obtained by the extraction o f eigenelements o f the variance-covariance matrix o f the initial vectors are not necessarily the vectors sought. Since all successive eigenvectors are perpendicular, this implies the presence o f alternating negative components with no physiological significance. The determination o f the fundamental components is followed by the computation o f oblique factors either [136] by a Monte Carlo technique followed by a simplex [138] minimization, or [137] by a direct iterative calculation.
CONCLUSION How is it possible to make use o f the various techniques so far described which seem to require the availability o f a large computing system? In fact, most systems actually available for use in nuclear medicine are classed as mini computers and the number o f microprocessor systems seems to be increasing. As in CT, one solution seems to be the use o f array processors and hardware especially adapted for the algorithms proposed. Considerable progress seems to be being made in this area but which cannot be discussed here for lack o f space. The question o f how much effort is worthwhile in developing processing techniques still needs to be discussed. In 1972, Trott asked [ 139]: ‘ . . . it is now some years since computers were first brought into use in attempts to improve the quality o f gamma camera images. It is still uncertain, however, whether the use o f computers has yet led to improved clinical diagnosis. . . can you yet identify the most promising lines o f development? ’ MacIntyre replied [ 140] that, to him, the most promising areas o f computer application lay ‘in functional studies where regional physiology rather than morphology’ could be studied. Excluding tomography, eight years later this still remains true. But as MacIntyre also stated [ 140]: ‘There has always existed a considerable gap between the refinements o f scanning parameters and an improvement in clinical results’. Optimistically, one can also hope that the use o f image processing techniques, and their evaluation, will still lead to significant progress. It is possible that, like Kuhl in tomography, much o f the research in this area was performed ‘before its tim e’. These techniques seemed o f value in the past since the analogue images then obtained were very poor, and they tended to be discarded with the improvements in camera technology. In part this is due to poor digital image quality resulting from almost universally poor data capture interfaces (even now). It should also be noted that image processing in the general field seems to be o f greatest value when processing images with relatively good signal-to-noise ratio, and one might therefore expect image processing now to be o f increasing rather than decreasing value.
IAEA-SM-247/204
305
It should be noted, finally, that although it is only 13 years since the first gamma camera was interfaced to a computer [141] the use o f data processing for extraction o f parameters and managing dynamic function studies is becoming almost universal. There is still enormous potential for improvement.
ACKNOWLEDGEMENT One o f the authors (R.D.) would like to acknowledge the support o f Professor W.J. Lorenz in developing data compression software using the VAX system o f the Institute o f Nuclear Medicine o f the German Cancer Research Center in Heidelberg, Federal Republic o f Germany.
REFERENCES П [2 [3 [4
[5 [6 [7
[8 [9
[10 [H [12
SMITH, E.M., B R IL L, A.B., Progress with com puters in nuclear m edicine, N ucleonics 25 (1 9 6 7 ) 64. PIZ E R , S.M., TODD -POK ROPEK, A.E., Im provem ent of scintigram s by com puter processing, Sem in. Nucl. Med. 8 (19 7 8 ) 125. CRADDUCK, T .D ., C om puters in nuclear imaging, Med. Instrum . (A rlington) 13 (1 9 7 9 ) 156. TODD -POK ROPEK, A.E., Im age processing in nuclear m edicine, IEEE Trans. Nucl. Sei. NS-27 (1 9 8 0 ) 1080. LOKEN, M.K., W ILLIAMS, L.E., A histo ry o f com puters in nuclear m edicine, Semin. Nucl. M ed., 9 (1 9 7 9 ) 197. LIEBERM A N, D .E., C o m p u ter M ethods: The F undam entals of Digital Nuclear M edicine, The C.V. M osby C om pany, St. Louis (1977). ADAM, W.E., L O R EN Z, W .J., SCH EER, K.E., “ Q uantitative U ntersuchungen m it der S zintillationskam era” , Proc. Sym p. R adioisotope in der L okalisationsdiagnostik (H O FFM A N , G., SCH EER, K.E., Eds), F.K . Schattauer-V erlag, S tu ttg a rt (1 9 6 7 ) 55. A N G ER, H.O ., M edical R adioisotope Scintigraphy (Proc. Sym p. Salzburg, 1968) 2, IAEA, V ienna (1 9 6 9 ) 91 (co m m en t in discussion). SMITH, E.M ., “ O n m ethods fo r norm alizing the response o f the A nger cam era imaging system ” , Proc. Sym p. Q uantitative O rgan V isualization in N uclear Medicine (K ENN Y, P.J., SM ITH, E.M., Eds), U niversity of Miami Press, C oral G ables, Florida (19 7 1 ) 261. ADAM, W.E., M edical R adioisotope S cintigraphy (Proc. Sym p. M onte Carlo, 1972) 1, IAEA, V ienna (1 9 7 3 ) 442 (co m m en t in discussion). HANNAN, W .J., BESSENT, R.G ., Pitfalls in gam m a cam era field unifo rm ity correction, Br. J. R adiol. 47 (1 9 7 4 ) 820. JA NSSON , L.G ., PA R K E R , R.P., Pitfalls in gam m a cam era field unifo rm ity correction, Br. J. R adiol. 4 8 (1 9 7 5 ) 408.
[13 [14
HASMAN, A., G R O O T H E D D E , R .T., G am m a cam era unifo rm ity as a function o f energy and co u n t rate, Br. J. Radiol. 49 (1 9 7 6 ) 718. COHEN, G ., K ER EIA K ES, J.G ., PA D IK A L, T .N ., e t al., Q uantitative assessm ent o f field u n ifo rm ity for gam m a cam eras, Radiology 118 (1976) 197.
306
DI PAOLA and TODD-POKROPEK
[15]
ERBSM ANN, F ., PA TE R N O T, J., PIEPSZ, A., “ Sensibilité e t hom ogénéité” , Proc.
[16]
Sym p. La fixation rénale du m ercure et du DMSA (RA Y N A U D , C., E d.), CEA, Orsay, F rance (19 7 6 ) 28. PA D IK A L, T.N ., A SH A RE, A.B., K EREIA K ES, J.G ., Field flood unifo rm ity correction: benefits or pitfalls? , J. Nucl. Med. 17 (1 9 7 6 ) 653. TODD -POK ROPEK, A .E., ERBSM ANN, F ., SOUSSALINE, F., “ T he non-uniform ity
[1 7 ]
[18]
[ 19] [20] [21] [22] [23]
[24]
[25] [26] [27] [28] [29] [30]
[31]
[32] [33]
of imaging devices and its im pact in q uantitative studies” , Medical R adionuclide Imaging (Proc. Sym p. Los Angeles, 1976) 1, IAEA, V ienna (1977) 67. SOUSSALINE, F ., TODD -POK ROPEK, A.E., RAY NAU D, C., “ Q uantitative studies w ith the gam m a cam era: correction for spatial and energy d isto rtio n ” , Inform ation Processing in Medical Imaging (Proc. 5th Int. Conf. Nashville, 1977) (B R IL L , A.B., PRICE, R., Eds), O RN L/BCTIC-2 (1978) 360. M U EH LLEHNER, G., R adiation imaging device, U.S. P aten t 3 745345, 1973. STEID LEY , J.W ., K EARN S, D.S., H O F F E R , P.B., U niform ity correction w ith the m icro Z processor, J. Nucl. Med. 19 ( 1978) 712. KIRCH, D.L., SHABASON, L., LE FR E E , M.T., STEELE, P.P., A nger cam era non u n ifo rm ity : The source and the cure, J. Nucl. Med. 19 (1978) 712. STOUB, E.W., COLSHER, J.G ., LEE, P.C., M U EH LLEHNER, G., A spatial distortion correction m ethod for gam m a cam eras, J. Nucl. Med. 20 (1979) 608. KN OLL, G .F., BENNETT, M.C., KO RAL, K .F., STRA N G E, D.R., “ Rem oval o f gam m a cam era nonlinearity and nonuniform ities through real-tim e signal processing” , Info rm atio n Processing in M edical Im aging (Proc. 6 th Int. Conf. Paris, 1979) (D I PA OLA, R., KAHN, E., Eds), Les E ditions de l’INSERM , Paris, 88 ( 1980) 187. BLOND, M.T., A U BERT, B., DI PAOLA, R., “ C om puter corrections o f image intensifier cam era response” , In fo rm atio n Processing in Medical Im aging (Proc. 6th Int. Conf. Paris, 1979) (D I PA OLA, R., KAHN, E., Eds), Les E ditions de 1’INSERM , Paris 88 (1 9 8 0 ) 173. ELSCINT LTD, C om m ercial E xhibit, H ospital P hysicist’s A ssociation Meeting, M anchester, June 1980. SEA RLE, Pho/gam m a ZLC. PA TTON , J.A ., ROLLO, F.D ., BRILL, A.B., R ecent developm ents in nuclear m edicine in stru m e n tatio n , IEEE Trans. Nucl. Sei. NS-27 (1 9 8 0 ) 1066. CRAN LEY, K., M ILLAR, R., BELL, Т .К ., C orrection for dead tim e losses in a gam m a cam era/data analysis system , Eur. J. Nucl. Med. 5 (1 9 8 0 ) 377. OPPENHEIM , В.E., A m eth o d using a digital c o m p u te r for reducing respiratory artifact on liver scans m ade w ith a cam era, J. Nucl. Med. 12 (1971) 625. DI PAOLA, R., BAZIN, J.P ., DI PAOLA, M., “ Im provem ent o f gam m a cam era data processing due to sequential acquisition” , In fo rm atio n Processing in Scintigraphy (Proc. 3rd Int. Conf. M .I.T., Cam bridge, Mass., 1973) (M ETZ, C.E., PIZER, S.M., BROW NELL, G .L., Eds), USERD A C O N F-730687 (1 9 7 3 ) 115. SCHMIDLIN, P., PALM TAG, M., CLORIUS, J., “ C orrection o f organ m o tio n by filtering techniques” (Proc. 3rd Sym p. on Sharing o f C om puters Program s and T echnology in N uclear M edicine, Miami, F lorida, 1973), USAEC CO N F-730627 (1 9 7 3 ) 54. SCHMIDLIN, P., E lim inierung von O rganbew egungen in Scintigram m en m it Hilfe eines Filterverfahrens, Nucl.-Med. 12 (1 9 7 4 ) 360. DI PAOLA, R., BERCHE, C., BAZIN, J.P ., “ T raitem ent digital des inform ations scintigraphiques” (Proc. 12th In t. Ann. M eeting o f the Soc. of N uclear M edicine, M unich, 1974), F.K. Schattauer-V erlag, S tu ttg a rt ( 1975) 670.
IAE A-SM-247/204
307
SCHMIDLIN, P., D evelopm ent and com parison of co m puter m ethods fo r organ m otion co rrection in scintigraphy, Phys. Med. Biol. 2 0 (1.975) 465. M cKEIGHEN, R .E., Im proved m eans of correcting m o tio n blurring in scintigraphic images, J. Nucl. Med. 24 (1 9 7 9 ) 353. T U R N E R , D.A ., FO RD H A M , E.W., ALI, A.A., et al., M otion corrected hepatic scintigraphy: an objective clinical evaluation, J. Nucl. Med. 19 (1 9 7 8 ) 142. H O F F E R , P.B., OPPENHEIM , В.E., STER LIN G , M .L., Y A SILLO , N., A sim ple device for reducing m o tio n artifacts in gam m a cam era images, R adiology 103 (1 9 7 2 ) 199. BRÜNING, P .F., CUPERU S, P., W IERING A, D.W., BEEKHUIS, H., A sim ple device fo r the elim ination o f the blurring effect o f respiration on organ scintigraphy, Int. J. Appl. R a d ia t.Is o t.2 4 (1 9 7 3 ) 53. ELK IN S, V.B., M A RTIN , C.B., POLLOCK, I.G ., McCLINTOCK, J.T ., E lectronic device corrected for m otion in gam m a cam era images, J. Nucl. Med. 15 (1 9 7 4 ) 36. LINE, B.R .,COO PER, J.A ., SPICER, K.M ., e t al., R adionuclide cinepneum ography: Flow -volum e imaging o f the respiratory cycle, J. Nucl. Med. 21 (1 9 8 0 ) 219. CH A RLESTO N , D.B., BECK, R .N ., E ID E LB E R G , P., SCHUH, M.W., “ T echniques w hich aid in qu an titativ e in te rp re ta tio n o f scan d a ta ” , M edical R adioisotope Scanning (Proc. Sym p. A thens, 1964) 1, IA EA , V ienna (19 6 4 ) 509. KAW IN, B., H U STO N , F .V ., COPE, C.B., Digital processing/display system for radioisotope scanning, J. Nucl. Med. 5 (1 9 6 4 ) 500. SCH EPERS, H., W IN K LER, C., “ An a u to m a tic scanning system , using a tape p e rfo ra to r and co m p u te r tech n iq u es” , M edical R adioisotope Scanning (Proc. Sym p. A thens, 1964) 1, IA E A , V ienna (1 9 6 4 ) 321. LA U G H LIN , J.S., K ENN Y, P.J., CO REY, K .R ., et al., “ L ocalization and to ta l body high-energy gam m a-ray scanning studies in cancer p a tie n ts” , M edical R adioisotope Scanning (Proc. Sym p. A thens, 1964) 1, IA E A , V ienna (1 9 6 4 ) 253. BROWN, D.W., Digital com puter analysis and display o f th e radioisotope scan, J. Nucl. Med. 5 (1 9 6 4 )8 0 2 . SPRAU , A.C., TA U X E, W .N., CHA APEL, D.W., A co m puterized rad io iso to p e scan-data filter based on a system response to p o in t sources, M ayo Clin. Proc. 41 (5 8 5 ) 1966. Y U ILL E, D .L., A new approach to the sm oothing o f d ynam ic nuclear m edicine data, J. N ucl. Med. 19 (1 9 7 8 ) 836. P IZ E R , S.M., V E T T E R , H.G ., The problem o f display in the visualization o f radio isotope d istributions, J. Nucl. Med. 7 (1 9 6 6 ) 773. DI PA OLA, R ., A LB A R ED E, P., PA TA U, J.P ., TUBIANA, M., C orrection de la réponse d ’un d é te cte u r collim até: A pplication à la scintigraphie, C.R. Séances Acad. Sei. 2 6 3 (1 9 6 6 ) 1160. IINUM A, T .A ., N A GAI, T ., R epetitive co rrectio n for a finite resolving pow er o f the collim ator in scintiscanning, Int. J. A ppl. R adiat. Isot. 18 (1 9 6 7 ) 325. IIN U M A , T .A ., N A G A I, T ., Im age re sto ratio n in radioisotope imaging system s, Phys. Med. Biol. 12 (1 9 6 7 ) 501. VAN C IT T E R T , P.H., Z um E influß der S paltbreite au f die Intensitätsverteilung in Spektrallinien: Part II, Z. Phys. 69 (1 9 3 1 ) 298. HUBBELL, J.H ., HAYW ARD, E., T ITU S, W .F., E nergy and angular d istrib u tio n o f X-rays scattered in lead, Phys. Rev. 1 0 8 (1 9 5 7 ) 1361. SK A RSG A RD , L.D ., JO H N S, H .E., G R E E N , L .E .S., Iterative response correction for a scintillation sp ectro m eter, R adiat. R esearch 14 (1 9 6 1 ) 261. F R IE D E N , В.R ., “ Im age enhancem ent and re sto ra tio n ” , P icture Processing and Digital Filtering (H U A N G , T .S., E d.), Springer, Berlin (1 9 7 5 ) 177.
308 [56]
[57]
[58]
[59]
[6 0 ] [61]
[62]
[6 3 ] [64] [65] [66] [67]
[68]
[6 9 ]
DI PAOLA and TODD-POKROPEK METZ, C.E., PIZ E R , S.M., N onstationary and n o nlinear scintigram processing, 2nd Int. Conf. on D ata H andling and Im age Processing in Scintigraphy, H anover, 1971 (unpublished). G U STAFSSO N, T., TODD -POK ROPEK, A.E., Design and application o f filters w ith variable shape, 2nd Int. Conf. on Data H andling and Im age Processing in Scintigraphy, Hanover, 1971 (unpublished). PIZ E R , S.M., CO R R E IA , J.A ., CH ESLER, D.A., METZ, C.E., “ R esults o f non linear and non stationary image processing” , In fo rm atio n Processing in Scintigraphy (Proc. 3rd Int. Conf. M .I.T., C am bridge, Mass., 1973) (M ETZ, C.E., PIZ E R , S.M., BROW NELL, G .L., E ds), U SERD A C O N F-730687 (19 7 3 ) 93. GU STAFSSO N, T., PIZ E R , S.M., “ N on-stationary M etz-filtering” , In fo rm atio n Processing in S cintigraphy (Proc. 4 th Int. Conf. O rsay, 1975) (R A Y N A U D , C., TODD-POKROPEK, A .E., Eds), CEA, Orsay (1 9 7 5 ) 56. BELL, P.R ., DO U G H ERTY , J.M ., N onlinear image processing m ethods, IEEE Trans. Nucl. Sei. NS-25 (1 9 7 8 ) 928. PISTO R, P., H O NER, J., WALCH, G., “ A two-stage nonlinear filter fo r processing o f scintigram s” , In fo rm atio n Processing in Scintigraphy (Proc. 3rd Int. Conf. M .I.T., Cam bridge, Mass., 1973) (M ETZ, C.E., PIZ E R , S.M., BROW NELL, G .L., Eds), USERD A C O N F-730687 (1 9 7 3 ) 82. METZ, C.E., A M athem atical Investigation o f R adioisotope Scan Im age Processing, D octoral Thesis, Univ. o f Pennsylvania, A nn A rbor, Mich. Univ. M icrofilm s, Publ. No. 7 0 - 1 6 , 1 8 6 (1 9 6 9 ). HU NT, W.A., M EDER, H .G ., PISTO R, P., et al., Digital Processing o f Scintigraphic Images, IBM Tech. R ep., F.R . G erm any, 70.03.001 (1970). M EDER, H.G., HU NT, W.A., LO R E N Z , W .J., Sam ple rate resolution and deconvolu tion filters in digital radioscintigraphy, J. Nucl. Med. 11 (1970) 346. C O R FIELD , J.R ., D evelopm ent o f a fast image enhancem ent filter for ro u tin e use, 13. Int. Jahrestagung der G esellschaft für N uclearm edizin, Copenhagen, 1975. HOUSTON, A.S., An a tte m p t to optim ize tw o zero sum m ation filters for use in radioisotope scintigraphy, Int. J. Nucl. Med. Biol. 3 (19 7 6 ) 111. BOARDMAN, A.K., “ C onstrained optim isation and its application to scintigraphy” , In fo rm atio n Processing in M edical Im aging (Proc. 6 th Int. Conf. Paris, 1979) (DI PAOLA, R., KAHN, E., Eds), Les E ditions de 1TNSERM, Paris, 88 ( 1980) 603. BIZAIS, Y., DE LARM IN AT, Ph., LEM ORT, J.P., TODD -POK ROPEK, A.E., «R estauration d ’images scintigraphiques par filtres ad ap tatifs à su p p o rt borne», these Proceedings, IA E A -SM -247/144 (P oster Presentation). CORM ACK, J., H U TTO N , B .F., In crem en tal d e convolution, Phys. Med. Biol. 25 (1 9 8 0 ) 339.
[70]
ANDREWS, H.C., HU NT, B .R ., Digital Im age R esto ratio n , Prentice-H all, Englew ood Cliffs, New Jersey (1977).
[71 ]
G O N ZA LEZ, R.C., W INTZ, P., Digital Im age Processing, Addison-W esley, Reading, M assachusetts (1977). HA LL, E.L., C om puter Im age Processing and R ecognition, A cadem ic Press, New Y ork (1976).
[72] [73] [7 4 ]
RO SEN FE LD , A., KAK, A.C., Digital Picture Processing, A cadem ic Press, New Y ork (1976). STOCHKAM, T.G ., Jr., CANNON, T.M ., IN G EB R E T SE N , P.B., Blind deconvolution through digital signal processing, Proc. IEEE 63 (19 7 5 ) 678.
IAEA-SM-247/204
309
V A R O U TA S, P.P., N A R D IZ Z I, L .R ., STO K ELY , E.M ., Digital image processing applied to scintillation images from biom edical system s, IE EE Trans. Bio-Med. Eng. BME-24 (1 9 7 7 ) 337. C O R D IER , S., BIRA U D , Y., CHA M PA ILLER, A., VO UTAY, M., A study o f the application o f a deconvolution m ethod to scintigraphy, Phys. Med. Biol. 24 (1 9 7 9 ) 577. IN T E R N A T IO N A L ATOM IC E N E R G Y AGENCY, “ IAEA C o-ordinated R esearch Program m e on the In tercom parison o f C om puter-A ssisted Scintigraphic T echniques” , 2nd and 3rd Progress R ep o rts, M edical R adionuclide Im aging (Proc. Sym p. Los Angeles, 1 9 7 6 )1 , IAEA, V ienna (1 9 7 7 ) 571, 585. R o u n d T able on Im age F iltering in Scintigraphy, 1st Conf. on C ontrast E nhancem ent T echniques in D igital Scanning, Brussels, 1970 (unpublished). B ERCHE, C., DI PA OLA, R ., Analyse frêquentielle et filtrage des im ages scintigraphiques: III. F iltres linéaires invariants de correction de réso lu tio n , J. Biol. Med. N ucl. 38 (1 9 7 4 ) 8. SEBERN, M .J., H O R G A N , J.D ., MEADE, R.C., e t al., M inicom puter en h ancem ent of scintillation cam era images using fast F o u rie r transform techniques, J. Nucl. Med. 17 (1 9 7 6 ) 647. CANTONI, V ., DE LO TTO , I., FA V IN O , A., V A LE N Z IA N O , F ., “ An image re sto ratio n technique: A pplications to nu clear m edicine im ages” , Digital Im age Processing (Proc. G I/N T G F achtagung, M unich, 1977) (N A G E L, H.H ., E d.), Springer, Berlin (1 9 7 7 ) 162. KAN, R .C ., H O PKIN S, G .B., Edge en h ancem ent o f ECG-gated cardiac images using directional m ask, Radiology 127 (1 9 7 8 ) 525. MEADE, R .C ., BAM RAH, V.S., H O R G A N , J.D ., et al., Q uantitative m ethods in the evaluation o f thallium -201 m yocardial perfusion images, J. Nucl. Med. 19 (19 7 8 ) 1175. R A I, G.S., H A G G ITH , J.W ., FENW ICK, J.D ., JAM ES, O., Clinical evaluation of co m puter processing o f liver gam m a cam era scans, Br. J. R adiol. 52 (19 7 9 ) 116. HO USTON , A.S., SH ARP, P .F ., T O FT S, P.S., D IF F E Y , B.L., A m u lticentre com parison o f co m p u ter assisted image processing and display m eth o d s in scintigraphy, Phys. Med. Biol. 24 (1 9 7 9 ) 547. HO USTON , A .S., M acLEOD, M.A., An in tercom parison o f co m p u ter assisted image processing and display m ethods in liver scintigraphy, Phys. Med. Biol. 24 (1 9 7 9 ) 559. NOW OTNY, R ., The applicability o f deconvolution m ethods in scintigraphic picture processing, Phys. Med. Biol. 24 (19 7 9 ) 1291. L EH R , J.L ., PA R K E Y , R.W., HARLOW , C.A ., e t al., C o m p u ter algorithm s for the detection o f brain scintigram abnorm alities, R adiology 97 (1 9 7 0 ) 269. N E IL L, G.D .S., HU TCH INSO N, F ., C om puter d etectio n and display o f focal lesions on scintiscans, Br. J. Radiol. 44 (1 9 7 1 ) 962. AMMANN, W.W., V A K N IN E, R ., “ S tru ctu re analysis. A new m eth o d for evaluating scintigram s” , In fo rm atio n Processing in Scintigraphy (Proc. 4 th In t. C onf. O rsay, 1975) (R A Y N A U D , C., TODD -POK ROPEK, A.E., E ds), CEA, Orsay (1 9 7 6 ) 66. VA K N IN E, R ., AMMANN, W.W., L O R E N Z , W .J., Investigation and application of p a tte rn recognition techniques to m edical pictu re data (scintigram s), K ern tech n ik 19 (19 7 7 ) 1. K A HN, E., DI PA OLA, R., BAZIN, J.P ., “ An interactive recognition program for lesion d e te cta b ility in scintigraphic images” , In fo rm atio n Processing in M edical Imaging (Proc. 6th In t. Conf. Paris, 1979) (DI PA OLA, R., KAHN, E „ E ds), Les E ditions de 1’INSERM , Paris, 88 (1 9 8 0 ) 613. PR A TT , W .K., Digital Im age Processing, Jo h n Wiley, New Y ork (1978).
310 [94] [95 [96
[97 [98
[99
[100 [101 [102 [103
[104; [105
[106 [107
[108 [109
[110
DI PAOLA and TODD-POKROPEK SHANNON, C.E., W EA VER, W., The m athem atical th eo ry o f com m unication, Bell Syst. Tech. J. 2 7 (1 9 4 8 ) 379. N E T R A V A LI, A.N., LIMB, J.O ., P icture coding: A review, Proc. IE EE 68 (1 9 8 0 ) 366. FA RRO W , A .S.J., R U T O V ITZ , D,, TU CK ER, J.M ., “ Television data com pression system for cytology chrom osom e analysis” , M edical Images: F o rm a tio n , Perception and M easurem ent (Proc. 7 th L.H. Gray C onf., Leeds, 1976) (HA Y, G.A ., Eds), The In stitu te o f Physics and Jo h n Wiley (1 9 7 6 ) 309. WATT, A., BOTTOM LEY, D., The com pression o f m edical ultrasonic d a ta , C om put. Biom ed. Res. 11 (1 9 7 8 ) 2. TA Y A K A , K., PA TEL, R ., “ D ata com pression fo r efficient storage o f digitized X-ray p ictu res” (Proc. 6th Conf. on C om puter A pplications in Radiology and C om puter/ A ided Analysis o f R adiological Im ages), IE E E , N o. 79-CH 1404-3C (19 7 9 ) 32. RAM , G., A general purpose com puterized display and analysis system fo r image processing in nuclear m edicine, C om put. Program s B iom ed. 10 (1 9 7 9 ) 245. GOLUB, G .H ., REIN SCH , C., Singular value decom position and least squares solutions, N um er. M ath. 14 (1 9 7 0 ) 403. KA RHUN EN , K., Ü ber L ineare M ethoden in der W ahrscheinlichkeitsrechnung, Ann. Acad. Sei. F enn., Ser. A I 37 (1947). LOEVE, M., «F onctions aléatoires du second ordre», Processus stochastiques et m ouvem ent brow nien (P. LEV Y , E d.), H erm ann, Paris, F rance (1948). H O TELLIN G , H., Analysis o f a com plex o f statistical variables in to principal com ponents, J. E duc. Psychol. 24 (19 3 3 ) 4 17, 498. в BEN ZECRI, J.P ., e t collaborateurs, L’Analyse des données 2 , D unod, Paris (1973). DI PA OLA, R ., PEN EL, C., BAZIN, J.P ., BERCHE, C., “ F a c to r analysis and scintigraphy” , In fo rm atio n Processing in S cintigraphy (Proc. 4 th In t. C onf. O rsay, 1975) (RA Y N A U D , C., TODD -POK ROPEK, A .E., Eds), CEA, Orsay (1 9 7 6 ) 91. M ARAÑO, P., A pplication d e l’analyse factorielle des correspondances à la com pression des signaux d ’images, A nn. T elecom m un. 27 (1 9 7 2 ) 163. SCHM IDLIN, P., RO SEL, F ., “ A pplication o f fa cto r analysis for scintigraphic picture processing” , In fo rm atio n Processing in S cintigraphy (Proc. 4 th In t. C onf. O rsay, 1975) (RA Y N A U D , C., TO DD -POK ROPEK, A .E ., E ds), CEA, O rsay (1 9 7 6 ) 80. KEN N ED Y , J.D ., C LA RK , S.J., PA RKYN , W.A., Jr., Digital Im agery D ata C om pression T echniques, M cDonnel Douglas C orp., H untin g to n Beach, R ep. MDC-GO-402 (1 970). ANDREW S, H.C., PA TTER SO N , C.L., O uter p ro d u c t expansions and th eir uses in digital image processing, Am. Math. Mon. 1 (1 9 7 5 ) 1. OTSU, N ., Picture re p re se n tatio n by m arginal eigenvectors, Bui. E lectrotech. Lab. 39 (1 9 7 5 ) 910.
[111
AHM ED, N., N A TA R A JA N , T ., RAO , K .R ., D iscrete cosine tran sfo rm , IE EE Trans. C om put. C-25 (1 9 7 4 ) 90.
[112
AHM ED, N ., RA O , K .R ., O rthogonal T ransform s fo r Digital Signal Processing, Springer, Berlin (1975). HA M IDI, M., PE A R L , J., C om parison o f the cosine and F o u rier transform s o f Markov-1 signals, IEEE Trans. A coust., Speech Signal Process. ASSP-24 (1 9 7 6 ) 428. CHEN, W ., FR A LIC K , S .C., A fast c o m p u ta tio n al algorithm fo r the discrete cosine transform , IEEE Trans. C om m un. T echnol. COM-25 (1 9 7 7 ) 1004. NARASIM HA, M .J., PETERSO N , A.M., On the c o m p u tatio n o f th e discrete cosine transform , IEEE Trans. C om m un. T echnol. COM-16 (1 9 7 8 ) 934. JA IN , A.K., A sinusoidal fam ily o f unitary transform s, IE EE Trans. P attern A nalysis and M achine Intelligence PAMI-1 (1 9 7 9 ) 356.
[из; [114 [115
[116
IAEA-SM-24 7/204 117] 118] 119] 120] 121]
122]
123] 124]
125] 126] 127]
128]
129]
130] 131]
132]
133]
134] 135]
311
K ITA JIM A , H., A sym m etric cosine transform , IEEE Trans. C om put. C-29 (1 9 8 0 ) 317. BACCHI, H., C onsidérations sur le choix des tran sfo rm atio n s orthogonales e t sur leurs perspectives, A cta E lectron. 19 (1 9 7 6 ) 299. N A TA R A JA N , T .R ., AHMED, N ., On interfram e tran sfo rm coding, IE EE Trans. Com m un. T echnol. СОМ-25 (1 9 7 7 ) 1323. R O E SE , J.A ., PR A TT , W.K., Interfram e cosine transform image coding, IEEE Trans. C om m un. T echnol. COM-25 (1 9 7 7 ) 1329. KA IH A RA , S., N A TA R A JA N , Т .К ., W AGNER, H.N., Jr., M A YNARD, C.D., C o nstruction o f a fun ctio n al image from regional ra te constants, J. N ucl. Med. 10 (1 9 6 9 ) 347. H O EH NE, K .H ., BOEHM , M., NICO LA E, G.C., “ The processing o f X-ray image sequences” , Advances in Digital Im age Processing (Proc. Int. Sym p. Bad N euenahr, FR G , 1978) (STU CK I, P., E d.), Plenum Press, N ew Y ork ( 1979) 147. PIZ E R , S.M., TOD D -PO K RO PEK , A .E., “ Sm oothing of fu n c tio n al images” , Proc. 1st W orld Congress of N uclear M edicine, T okyo (1 9 7 4 ) 161.. W ATANABE, S., “ K arhunen-L oève expansion and factor analysis: T heoretical rem arks and applications” , Trans. 4 th Prague Conf. Inform . T heory, Statist. Decision F unctions and R andom Processes, Prague (1 9 6 5 ) 635. PEARSON, K., O n lines and planes o f closest fit to system s o f p o in ts in space, Philos. Mag. 6 ( 1 9 0 1 ) 559. M A RDIA, K .V ., K EN T, J., BIBBY, J.M ., M ultivariate Analysis, A cadem ic Press, L ondon (1979). BARBER, D.C., D U TH IE, H .L., HOW LETT, P .J., W ARD, A.S., “ Principal com ponents. A new a pproach to the analysis o f gastric em ptying” , D ynam ic Studies w ith R adio isotopes in M edicine 1974 (Proc. Sym p. Knoxville, 1974) 1, IA EA , V ienna (1 9 7 5 ) 185. SCHM IDLIN, P., CLO RIU S, J., KUBESCH, R ., D R EIK O R N , K., “ E valuation of dynam ic studies by m eans o f fa cto r analysis” , M edical R adionuclide Im aging (Proc. Sym p. Los Angeles, 1976) 1, IA EA , V ienna (1977) 397. SCHM IDLIN, P., “ F a c to r analysis o f sequence scintigram s” , In fo rm atio n Processing in M edical Im aging (Proc. 5th In t. Conf. Nashville, 1977) (B R IL L , A.B., PRICE, R ., Eds), N ational T echnical In fo rm atio n Service, O RN L/BCTIC-2, Springfield, Virginia (1 9 7 8 ) 472. SCHM IDLIN, P., Q uantitative evaluation and imaging of functions using p a tte rn recognition m eth o d s, Phys. Med. Biol. 24 (1 9 7 9 ) 385. SCHM IDLIN, P., C LO R IU S, J., L O R E N Z , W .J., “ P a tte rn recognition in renography” , In fo rm atio n Processing in M edical Im aging (Proc. 6 th In t. Conf. Paris, 1979) (D I PA OLA, R ., KAHN, E., Eds), Les E ditions de l’INSERM Paris, 88 (1980) 335. OPPENHEIM , В.E ., “ Id en tificatio n o f renogram s by fa cto r analysis” , In fo rm atio n Processing in M edical Im aging (Proc. 5th Int. C onf. Nashville, 1977) (B R IL L , A.B., PR IC E , R ., E ds), N ational T echnical In fo rm atio n Service, ORN L/BCTIC-2, Springfield, Virginia (1 9 7 8 ) 481. OPPENHEIM , В., A PPLED O RN , C., “ F u n ctio n al renal imaging using fa cto r analysis” , In fo rm atio n Processing in M edical Im aging (Proc. 6 th In t. Conf. Paris, 1979) (DI PA OLA, R., KAHN, E., Eds), Les E ditions de l’INSERM , Paris, 88 (1 9 8 0 ) 321. M acLEOD, M.A., “ A com parison o f 3 display m eth o d s o f assessing renal fu n c tio n ” (Proc. Coll. In t. Biosigma, Paris, 1978) 2 (AM RAM, Y., E d.) (1 9 7 8 ) 37. HO USTON , A.S., “ C lassification o f dynam ic fu n c tio n studies in nu clear m edicine. D oes featu re e x tractio n help? ” , In fo rm atio n Processing in M edical Im aging (Proc. 6th Int. Conf. Paris, 1979) (D I PA OLA, R., KAHN, E ., E ds), Les E ditions de 1’INSERM , Paris, 8 8 ( 1 9 8 0 ) 381.
312 [136]
[1 3 7 ] [138] [1 3 9 ] [140]
[141]
DI PAOLA and TODD-POKROPEK BAZIN, J.P ., DI PA OLA, R., GIBAUD, B., et al., “ F a c to r analysis o f dynam ic scinti graphic d ata as a m odelling m ethod: An application to the d etectio n o f m étastasés” , In fo rm atio n Processing in Medical Im aging (Proc. 6th Int. Conf. Paris, 1971) (D I PA OLA, R ., K A HN, E. , Eds), Les E ditions de l’INSERM , Paris, 88 (1 9 8 0 ) 345. BA RBER, D.C., T he use of principal com ponents in the q uantitative analysis o f gam m a cam era dynam ic studies, Phys. Med. Biol. 2 5 (1 9 8 0 ) 283. OLSSON, D.M., A sequential sim plex program fo r solving m inim ization problem s, J. Qual. T echnol. 6 (1 9 7 4 ) 53. T R O T T, N .G ., Medical R adioisotope S cintigraphy (Proc. Sym p. M onte Carlo, 1972) 1, IAEA, V ienna (1973) 441 (co m m en t in discussion). M acIN TY RE, W .J., “ A dvances in data-processing techniques” , Medical R adioisotope Scintigraphy (Proc. Sym p. M onte Carlo, 1972) 1, IA EA , V ienna (1 9 7 3 ) 421 (and com m ent in discussion, 441). AMMANN, W., ADAM, W .E., LO REN Z, W .J., SCH EER, K .E., „D er Einsatz eines C om puters in der S zintigraphie” (V ortrag 12. K ongress der Ges. f. Radiologie, Berlin, 1967), Radiobiol. R adiother. 9 (1968) 703.
IAEA-SM-24 7 /1 28
MAXIMUM ENTROPY RECONSTRUCTIONS IN EMISSION TOMOGRAPHY M.C. KEMP Department o f Applied Mathematics and Theoretical Physics, University o f Cambridge, Cambridge, United Kingdom
Abstract MAXIMUM EN TRO PY RECO N STRU CTIO N S IN EM ISSION TOM O GRAPHY. The m axim um e n tro p y m e th o d (MEM) is applied to the problem o f re co n stru c tio n from pro jectio n s in em ission tom ography. Its use leads to considerable im provem ents over con ventional m ethods, especially w hen the d ata are poorly sam pled o r noisy. A series o f exam ples using sim ulated d a ta are given w hich show how the MEM re co n stru c tio n depends o n the num ber of p h o to n s d ete cte d and the n um ber and angular range o f th e projections. T he results of experim ents w ith liver p h a n to m d a ta are presented, and the problem o f determ ining the a tte n u atio n w hen th e body shape is un k n o w n discussed.
1.
INTRODUCTION
The problem o f reconstructing a function in two dimensions from a finite number o f sampled projections, central to emission tomography [ 1 ], is very similar to inverse problems in image restoration and radio interferometry. The maximum entropy method (MEM) is a powerful and general technique for solving such problems given incomplete and noisy data [2]. It has been applied with success to deconvolution problems in optical and gamma-ray astronomy [3, 4] and to radio interferometry [2]. Conventional reconstruction methods are o f two types. The first which includes Fourier and filtered back-projection algorithms [5] gives an explicit formula based on the analytic solution for complete noise-free data. The effects of noise can be reduced by ad hoc filtering methods at the expense o f resolution, and incomplete sampling produces artifacts. The second type are the iterative methods where a reconstruction is sought which either fits the data exactly, as in ART [6], or is within the expected fluctuations, as in the iterative least-squares method [7]. These methods give non-unique solutions and tend to amplify the effects o f noise on the data. In our method we consider the set o f all reconstructions consistent with the data taking into account the known uncertainties due to noise. Of these we
313
314
KEMP
choose the one with the largest configurational entropy. The maximum entropy solution is the one which contains the minimum information and is still consistent with the data. It is thus likely to contain the minimum o f false structure (artifacts, noise) and we can be sure that there is evidence in the data for all features which are seen. There may be other structure present in the object but a better experiment is required to reveal it. This strategy gives reconstructions with very desirable properties. The solutions are positive, unique and have the maximum resolution and dynamic range allowed by the geometry and quality o f the data while retaining a low noise level. They are perhaps the most easily interpreted reconstructions.
2.
THE METHOD The data d¡ are related to the object structure Fj by the relation di = OijFj + ni
(1)
The matrix element 0¡j is the contribution of the j th object pixel to the i th projection element. The noise on the data is represented by nj. In our case noise is principally due to photon counting statistics and we take the n¿ to be independent Gaussian random variables with variance af = d¿ + 1
(2)
We seek a reconstruction f such that the entropy S = -S P jlogP j Pj = fj/Zfj
(3)
is a maximum subject to the constraint that the fj are consistent with the data. The constraint used here is based on the x 2 statistic although others may be used [3]. We require that X2 = 2 ( O ijfj - d i)2/ff1?
(4)
has its expectation value N, the number of data points. There have been other formulations of MEM where the entropy expression is Zlogfj [8, 9], or where the data are fitted exactly [9, 10]. We believe that they are inappropriate to this kind o f problem [2], for example, if the data are fitted exactly then one is almost certainly introducing structure due purely to noise. Further, these methods lead to considerable computational difficulties on problems of useful size.
IAEA-SM-247/128
315
The constrained optimization, equations (3) and (4), is performed using an iterative algorithm from J. Skilling which is very robust and normally converges in 6 —10 iterations. Each iteration involves 6 transforms (multiplications by Oy or its transpose) and a number o f scalar product operations. It is therefore a factor of 5 0 -1 0 0 slower than filtered back-projection and a factor o f 5—10 slower than other iterative methods. The forward operation (E q .(l)) is done in a similar way to that described by Budinger [ 1] using a long file to store the weighting factors. Attenuation for a known body shape is included by computing the path length through the body for each Oy. Note that as we are not restricted to non-overlapping parallel projections, a realistic spreading collimator response can easily be included. Free parameters, such as scaling factors for each projection angle or a series describing deviations from the assumed body shape, may also be introduced. These simply require minimizing a modified form o f Eq.(4) at each iteration. This has proved useful in correcting for unknown offsets.
3.
NUMERICAL EXPERIMENTS
To illustrate our method we have reconstructed a number o f test phantoms from data generated in the computer. The basic shape used is shown in F ig.l. digitized on a 64 X 64 grid. The elliptical ‘body’ has major and minor axes 2a = 38 cm, 2b = 24 cm and the two hotspots are of radius 2 cm with centres 5 cm apart. The ‘camera’ is 42 cm in diameter and counts are recorded in 64 bins. Different values o f background A b and hotspot A2, A3 intensities were used, and an appropriate quantity o f random noise added to each data point. All reconstructions were done on a 64 X 64 grid. For comparison we also present filtered back-projection reconstructions using the algorithm o f Shepp and Logan [11]. We use this method simply to provide a convenient benchmark, and do not claim it to be the best o f the con ventional methods in any given situation. Indeed, with a very limited dataset the performance o f different methods depends markedly on the object structure [12]. Attenuation compensation was done by applying the hyperbolic sine correction [7] to the geometric mean of opposing projections. 3.1. Uniform source with hotspots The test phantom (F ig .l) with relative intensities 1.0, 1.2, 1.5 was used to investigate the reconstruction o f a large uniform source and, at the same time, detect small regions o f enhanced activity. The first two columns of Fig.2 are MEM and conventional reconstructions from 16 projection angles uniformly spaced over 0 —180°. No attenuation was
316
FIG .2 .
KEMP
Reconstructions o f a large uniform source w ith two hotspots by M E M (colum ns 1
and 3) and filtered back-projection (colum ns 2 and 4). Regions o f negative intensity in the filtered back-projection reconstructions have been suppressed.
IAEA-SM-247/128
317
10s
II
P h o to n s detected
О О
TABLE I. NOISE LEVELS (rms/mean) OF THE RECONSTRUCTIONS SHOWN IN Fig.2. ¡1 =
0.15 cm 1
MEM
FILBP
MEM
FILBP
0.03
0.23
0.04
0.26
3.2 X 10s
0.02
0.08
0.03
0.11
106
0.01
0.05
0.02
0.08
3.2 X 106
0.01
0.04
0.01
0.08
included. The total number of photons detected are 10s (top), 3.2 X 105, 106 and 3.2 X 106. The low noise level on the MEM reconstructions without loss of resolution is striking. The third and fourth columns show reconstructions from data calculated using an attention coefficient л = 0.15 cm-1 . Thirty-two angles over 360° were used with the same photon counts as before. Note that the noise in the maximum entropy reconstructions decreases towards the centre of the phantom. This occurs because the experiment is insensitive to this region (only 10% o f the photons emitted are detected), and the entropy criterion gives a smooth re construction if there is no evidence for structure at the centre. This is in marked contrast to conventional methods (column 4 and also [7]) where larger fluctuations occur at the centre because they affect the fit to the data very little. The noise levels in the reconstructions are given in Table I where the figure quoted is the rms/mean intensity over the uniform part o f the phantom. 3.2. Small number o f views The effects o f reducing the number o f projection angles is shown in Fig.3. In the first two columns we show MEM and conventional reconstructions of the test phantom with intensities 0., 5., 10. from 16, 8, 4 and 2 angles in the range 0—180° without attenuation. The number o f photons detected was 3.2 X 106. The last two columns show the effect o f including background emission using intensities 1., 5., 10. We see that with as few as 4 angles MEM gives a reasonable estimate o f both the shape and magnitude o f the background whereas the filtered back-projection method barely shows its presence.
318
KEMP
« #
ж ш т т я т п т т Щ Ё т я ш т ж
F IG .3 .
* #
’
* »
'
* #
,C
' *♦
Reconstructions from a lim ited number o f projections by M E M and filtered back-
projection.
3.3. Limited angle range To give an indication o f how MEM might be useful with devices such as the 7-pinhole and rotating slant hole collimators, we show in Fig.4 the effect of reducing the angular range o f the projections. Data were calculated for 16 projections o f the test phantom with intensities 1., 5., 10., without attenuation, and for 3.2 X 106 detected photons. The total angular range was 180, 90, 60, and 45° centred on a line perpendicular (column 1) and parallel to the line joining the two hotspots. We consider that these reconstructions are extremely good considering the nature o f the data.
IAEA-SM-247/128
319
F IG .4. M E M reconstructions from a lim ited angular range.
A related area where MEM might be useful is the case where the camera is smaller than the object so that not all points on the object contribute to each projection. Experiments with X-ray CT data (S.F. Gull, personal communication) have shown that, although the reconstruction in the poorly sampled area is degraded (as expected), artifacts are not produced in the completely sampled area as is the case with conventional algorithms.
4.
EXPERIMENTS WITH LIVER PHANTOMS
The encouraging results o f the tests just described led us to try our method on actual gamma-camera data, kindly provided by the Nuclear Medicine Depart ments of Guys and the Middlesex Hospitals. The data discussed here were taken
KEMP
320
F IG .5 .
M EM reconstructions o f a liver phantom slice.
with a G.E. Maxicamera II rotating gamma camera using a " T cm-filled liver phantom in a rubber body with д = 0.15 cm-1 . To calculate the effects of attenuation, the body shape was approximated by an ellipse with 2a = 38 cm, 2b = 24 cm. The dataset consisted o f 64 equally spaced views digitized on 64 X 64 grid. Approximately 106 photons per slice were collected. A MEM reconstruction o f one slice using all 64 angles is shown in Fig.5 (top). The phantom contains two cold lesions at this level, one o f diameter 3 cm, the other in the upper left hand corner of diameter 2 cm. A conventional reconstruction
IAEA-SM-247/128
321
F IG . 6. Filtered back-projection reconstruction o f liver phantom slice made from two adjacent projection slices.
of a double-thickness slice, done using a method similar to that described in Section 3 but with a different filter giving more smoothing, is shown in Fig.6. From a comparison o f the two reconstructions with the known structure o f the phantom, it appears that MEM gives a more accurate and sharper overall shape, although the lesions are better detected by the conventional algorithm. Tests on simulated data o f similar geometry show that MEM is capable of much better reconstructions from data of this quality. Indeed the reconstruction in Fig. 5 does not fit the data to within x 2 = N; instead the algorithm stopped at x 2 = 5N. Inspection of the residuals (individual components o f x 2 ) shows large systematic deviations across the data which are due to a discrepancy between the actual and assumed body shapes. It is not therefore surprising that the lesions are not well detected since their presence in the reconstruction only affects x 2 by a small amount. Since the quality o f our reconstruction depends on these systematic effects rather than the photon statistics or angular sampling, we investigated the effect of reducing the number of projections (resulting also in a reduction in the total photon count). The second and third reconstructions in Fig.5 are from 32 and 16 angles. There is little difference between these and the 64 angle case. The fourth reconstruction was done from 32 angles over a range o f 180° around the side o f the body where the liver is closest to the surface. This is also very similar to the others, the main difference being the loss o f resolution at the centre caused by absorption. The large lesion is in fact better detected, presumably because errors in the body shape around the far side of the body are no longer important. We are currently investigating the attenuation problem. In the case where there is background emission from the whole o f the absorption region as in most clinical studies, the body shape can be determined by methods similar to those described by Budinger [7]. A line source placed around the body is another possibility. We have tried fitting the body shape using a number o f free para meters as described in Section 2. This goes some way towards solving the problem but a completely satisfactory method has not yet been found.
322 5.
KEMP
CONCLUSIONS
The application o f the maximum entropy method to emission tomography should lead to considerably improved reconstructions, especially in the case where the data are very noisy and the sampling incomplete. Experiments with camera data have given encouraging results, although they have highlighted the problem o f determining the shape o f the body to allow accurate attenuation calculations. The only inherent disadvantage of MEM is that it is much slower than other methods. We note, however, that much o f the computation is very suitable for implementation on array processing devices. Although we have concentrated on transaxial tomography, the same techniques are applicable to other forms o f tomographic imaging in nuclear medicine and elsewhere.
ACKNOWLEDGEMENTS The author is m ost grateful to S.F. Gull and J. Skilling for discussion on maximum entropy, to V.A. Brookeman, N.J.G. Brown, P.J. Ell and P.H. Jarritt for discussions on emission tomography and the Radio Astronomy Group o f the Cavendish Laboratory, Cambridge, for the use o f their Nord 10/50 computer.
REFERENCES [ 1] B U D IN G ER, T .F ., G U LLB ER G , G .T., T hree-dim ensional reco n stru ctio n s in nuclear m edicine em ission imaging, IE EE T rans. N ucl. Sei. NS21 3 (1 9 7 4 ) 2. [2] G U LL , S.F., D A N IE LL , G .J., Im age re co n stru c tio n from incom plete and noisy data, N ature (L o n d o n ) 272 (1 9 7 8 ) 686. [3] BRYA N, R .K ., SK ILLIN G , J., D econvolution by m axim um e n tro p y , as illustrated by application to the je t o f M 87, M on. N ot. R. A stron. Soc. 191 (1 9 8 0 ) 69. [4] SK ILLIN G , J., STRO N G , A.W., BEN N ETT, K ., M axim um e n tro p y image processing in gam m a-ray astronom y, M on. N o t. R. A stron. Soc. 187 (1 9 7 9 ) 145. [5] ROW LAND, S.W., “C o m p u ter im plem entation o f image reco n stru c tio n fo rm ulas”, Im age R e co n stru c tio n from P rojections (H ERM A N , G .T ., E d .), Springer-Verlag, Berlin (1 979). [6] G O R D O N , A., A tu to ria l o n A R T , IEEE T rans. N ucl. Sei. N S21, (1 9 7 4 ) 78. [7] B U D IN G ER, T .F ., G U LLB ER G , G .T ., HUESM AN, R .H ., “E m ission co m puted to m o g ra p h y ”, Im age R eco n stru ctio n from P ro jectio n s (H ERM A N , G .T ., E d .), SpringerVerlag, B erlin (1979). [8] W ERN ECK E, S.J., D ’A D D A R IO , L .R ., M axim um e n tro p y image re co n stru c tio n , IEEE T rans. C om put. C-26 (1 9 7 7 ) 351.
IAEA-SM- 247/128
323
[9]
PONSONBY, J.E .B ., A n e n tro p y m easure fo r p artially polarized ra d ia tio n and its app licatio n to estim ating radio sky p o larizatio n d istrib u tio n s from incom plete ‘aperture synthesis’ d ata by the m axim um e n tro p y m e th o d , M on. N ot. R. A stron. Soc. 163 (1 9 7 3 ) 3 69. [10] M IN ERBO , G., M ENT: A m axim um e n tro p y algorithm for reco n stru ctin g a source from p ro jec tio n data, C o m p u t. G raphics Im age Process. 10 (1 9 7 9 ) 48. [11] SHEPP, L.A ., LO GA N, B .F., The F o u rie r reco n stru c tio n o f a h ead section, IEEE Trans. Nucl. Sei. NS21 3 (1 9 7 4 ) 21. [12] GILLM AN , G ., M ACLEOD, I., R e co n stru c tio n o f X-ray sources from pen u m b ral images, C om put. G raphics Im age Process. 11 (1 9 7 9 ) 227.
DISCUSSION A.E. TODD-POKROPEK: Could you tell us how sensitive your reconstruction technique is to the attenuation correction itself? Attenuation correction in single photon emission computed tomography is, after all, dependent on the activity distribution, so that a transmission measure or use o f an ellipse is not adequate. This might explain your difficulties in fitting the liver phantom data. M.C. KEMP: Our method is no more sensitive than others to attenuation. The correction we apply is distribution-dependent, and errors are due only to inaccuracy in the assumed body shape and the value of the attenuation coefficient itself.
IAEA-SM-247/76
LONGITUDINAL AND TRANSVERSE DIGITAL IMAGE RECONSTRUCTION WITH A TOMOGRAPHIC SCANNER D.R. PICKENS, R.R. PRICE, J J . ERICKSON, J.A. PATTON, C.L. PARTAIN, F.D. ROLLO Vanderbilt University Medical Center, Division o f Radiological Sciences, Nashville, Tennessee, United States o f America
Abstract LO N G ITU D IN A L AND T R A N SV ER SE D IG ITA L IMAGE RECO N STRU CTIO N WITH A TOM O GRAPHIC SCANNER. A Siem ens G am m asonics PH O /CO N -192 M ultiplane Im ager is interfaced to a digital co m p u ter fo r the purpose of perform ing tom ographic re co n stru ctio n s from the data collected during a single scan. D ata from th e tw o m oving gam m a cam eras as well as cam era position in fo rm a tio n are sent to th e c o m p u te r by an interface designed in th e a u th o rs’ la b o ra to ry . B ackprojection re co n stru c tio n is im plem ented by th e com puter. L ongitudinal images in wholebody fo rm at as w ell as sm aller form ats are reco n stru cted fo r up to six planes sim ultaneously from th e list m o d e d ata. Transverse reco n stru ctio n s are d em onstrated fo r 201T1 m yocardial scans. P ost-reco n stru ctio n deconvolution processing to rem ove th e blu r a rtifac t (characteristic o f focal plane tom o g rap h y ) is applied to a m ultiplane p h a n to m . D igital d a ta acquisition of data and re co n stru c tio n o f images are practical, and can extend the usefulness of the m achine w hen com pared w ith th e film o u tp u t.
1.
Introduction
The Siemens Gaimnasonics PHO/CON-192 Multiplane Imager is a focal plane tomographic scanning system designed to image both relatively small regions such as the heart as well as the entire body. The commercial instrument produces an output which displays twelve longitudinal tomographic image planes on transparent film from each scan. The interfacing of this instrument to a computer system for the purpose of producing digital tomographic images is described in this paper. 2.
System Description
The mechanical hardware of the PHO/CON-192 is relatively complex. The machine consists of dual-opposed, highperformance, small-field scintillation cameras having 325
326
PICKENS et al.
F IG .L
Diagram o f Siemens Gammasonics PHO/CON-192 M ultiplane Imager.
intrinsic full width at ¡half maximum resolutions of 0.6 cm which are mounted on a large 'C' frame (Fig. 1). One camera is positioned over a stationary bed while the other is located below the bed. Under control of the internal logic, the frame and cameras are moved around the bed in such a way that the cameras describe a rectilinear raster. The vertical position of each camera can be adjusted manually, but is not changed during scanning. Each camera has its own focused collimator. The PH0/C0N internal control for positioning and raster generation is all digital and is, therefore, easily accommodated by an external interface which has been designed and installed by our laboratory personnel. Position information indicating the location of the cameras relative to the stationary bed is supplied to the external interface. Camera coordinate information is applied as analog X and Y signals to the interface for conversion to digital form. Information is provided to indicate which camera is producing the current analog data and from which one of the two available energy windows the event came. The interface digitizes the x and y camera information into seven bits each and combines them into a single 16 bit data word with flag bits indicating upper/lower camera and isotope 1/isotope 2. The scanner produces up to 3200 data points per scan line which can be reduced up to a factor of 8 by a resolution switch on the interface. The bed position is encoded into the lower twelve bits of two 16 bit data words, one for x position and one for y position. During data collection, the interface produces a data stream consisting of a flag word of zeros, an x bed coordinate word, a y bed coordinate word, and camera words associated with the current bed coordinates. Ä Digital Equipment PDP-11/34 computer receives the data from the PHO/CON interface and
IAEA-SM-247/76
FICTITIOUS LINE-----
LONGITUDINAL PLANE F IG .2.
327
CRYSTAL FACE-
TRANSVERSE PLANE—
Backprojection o f events through longitudinal and transverse planes.
stores them on removable disk cartridges. Reconstruction of the tomograms from this list mode data occurs after completion of the scan. 3.
Method of Backprojection
The method of backprojection [l] provides for the reconstruction of tomographic planes of information from data collected using a position-sensitive detector with a focused collimator that is moved in a rectilinear raster. If for each detected event in the crystal a fictitious line is drawn through the event on the crystal and the focus of the collimator, the event can be projected back along its ray path, the fictitious line. This ray path intersects all possible tomographic planes. Thus for a particular plane of interest, each ray path associated with each event passes through the plane at some position in the plane represented by a coordinate pair (Fig. 2). If all ray paths are summed in the plane of interest at their intersection coordinates, an approximation of the actual distribution in that plane is formed. Since all ray paths intersect the plane, contributions from other planes to the plane of interest increase the background and produce the well-known blur artifact characteristic of this method. There are several major advantages to using the computer to perform the backprojection. Since the raw data are collected, reconstructions can be performed at any time for several planes and can be repeated easily. Also, any plane can be produced at any distance from the face of the collimator. The current versions of the reconstruction
328
PICKENS et al.
F IG .3 .
Diagram o f plastic volume phantom.
programs can produce up to six images simultaneously from the upper or lower camera in 64 X 64 pixel or 128 X 128 pixel formats. In addition special large field reconstruction of individual planes is available in 400 X 256 or 256 X 99 formats. The computer also can provide transverse images from data collected during a single scan with the PHO/CON. In transverse section reconstruction the fictitious line once again is drawn through the crystal coordinates of a detected event and through the focal point of the collimator (Fig. 2). The plane intersected by the ray path is orthogonal to that of the data collection and to the long axis of the scanner. Summation of the intersections of ray paths with each plane produces a transverse image, a reconstruction mode not available from the film output. While the vertical resolution is inferior to the horizontal resolution, the results are especially interesting in certain types of scans, notably in Tl-201 scans of the heart, and can provide information not readily apparent in standard longitudinal scans. 4.
Results
Several phantom studies have been performed to test the digital collection and reconstruction system. A threedimensional volume phantom was used to demonstrate both longitudinal and transverse section reconstruction. The phantom (Fig. 3) is plastic and contains four individually filled volumes surrounded by a larger annular volume. Data were collected with the phantom lying on its side. Reconstruction of three longitudinal sections through the phantom (Fig. 4) shows a sharp image at levels one and three with a less distinct image due to scatter and off plane information at level two. The annulus which had somewhat less activity in it than the inner volumes is also visible.
IAEA-SM-247/76
329
CAMERA LI L2 ■L3
TI
T2
T2
TI F IG .4 .
Tomographic scans o f volume phantom.
PLANE 1 TOP DETECTOR
PLANE 2
PLANE 1 PLANE 2 LOWER DETECTOR F IG . 5.
PLANE 3
Whole-body 67Ga citrate soft tissue scan.
330
PICKENS et al.
T
# F IG . 7.
Technetium-201 transverse tomograms o f the m yocardium.
Transverse plane reconstruction through the phantom is performed on data collected from the same scan. A section through the phantom (Fig. 4) at plane one shows only the annulus. At transverse plane two, which is through the four inner volumes, a continuous region of increased activity is seen. In the transverse sections the combination of low resolution and scatter cause the inner volumes to blur together vertically. This agrees with the results seen with the the longitudinal sections. The annulus is dimly visible. Several types of whole-body longitudinal scans using different isotopes such as Tc-99m or Ga-67 can be collected
331
IAEA-SM-247/76
BACKPROJECTI ON
□ DECONVOLUTION
F IG .8.
M ultiplane number phantom after deconvolution.
and reconstructed using the digital system. Figure 5 shows a Ga-67 citrate whole-body scan. Three planes are shown for each camera, although any number of planes can be reconstructed. Gallium-67 is used in visualizing soft tissue abnormalities. In this study there are indications of several areas of increased uptake. The PHO/CON with digital reconstruction and post reconstruction processing is being evaluated for scanning the myocardium using Tl-201. The patient normally is imaged in an LAO orientation so that the long axis of the heart points directly at the lower detector, with the apex nearest the collimator face. Fig. 6 shows eight sections at one centimeter plane separations. The inferior portion of the heart is at the bottom of the image. This patient study shows a known inferior aspect lesion extending from midway in the ventricle through the apex in the lower right hand image. Transverse section reconstruction of the myocardial images (Fig. 7) shows sections through the left ventricle proceeding from the interior aspect at the upper left hand image and moving at 0.7 cm steps superiorly to the lower right image. The myocardial wall surrounding the left ventricular cavity is visible clearly. Variations in image density at the lower portions of the inferior aspect images may be part of the lesion that is seen in the longitudinal scans.
332
5.
PICKENS et al.
Post-Reconstruction Processing
Many types of post-reconstruction processing are possible. A bandpass digital Butterworth filter is used on all myocardial images. Digital filters can improve contrast and remove background noise but do not address the problem of blur contributions from other planes (inherent in focal plane tomography). Chang, Macdonald and Perez-Mendez [2] have attempted to reduce these blur contributions by a method of multiplane simultaneous deconvolution. In this technique, the multiplane impulse response of the system, the point source response, is deconvolved with the images reconstructed by backprojection, but corrupted by blur components, to yield a better estimate of the true distribution on each plane. In noise-free simulations the method works for many planes simultaneously. In practice, with real data, much less satisfactory results are achieved. Low pass pre- and post process filtering of images and point source responses is needed to reduce high frequency noise which can cause artifacts that obscure the images. A number phantom was scanned on the PHO/CON and reconstructed for three planes along with the corresponding point source response. The results are shown in Fig. 8. The background blur contributions are reduced in this example at the expense of loss of uniformity in the images. This latter problem is due to frequency space oscillation in the data and is caused by background noise. The algorithm shows promising results but presents a computational burden. 6.
Conclusions
Digital data acquisition and reconstruction of PHO/CON images are practical and can extend the usefulness of the machine when compared to the film output from the instrument. Several advantages of computer-processing are immediately evident: 1) any arbitrary plane can be reconstructed from list-mode data at any time; 2) transverse images can be produced, a reconstruction mode not available otherwise; 3) post-reconstruction image processing including digital filtering and deblurring can be performed. REFERENCES [1 ]
A N G E R , H .O ., “T om ographic gam m a-ray scanner w ith sim ultaneous read o u t o f several planes” , F u n d a m en ta l P roblem s in Scanning (GO TTSCH ALK, A., BECK, R.M ., Eds), Charles C. T hom as, Springfield (1 9 6 8 ) 195.
[2]
CHANG, L .T ., MACDONALD, B., PEREZ-M EN DEZ, V ., IEEE T rans. N ucl. Sei. 23 2 (1 9 7 6 ) 568.
IAEA-SM-247/76
333
DISCUSSION E.F. CROCKER: D o you as yet have any clinical evidence to suggest or confirm that post-scan selection o f a longitudinal plane increases lesion detectability or the diagnostic confidence level? D.R. PICKENS: We have preliminary informal comments from physicians in our clinic which suggest that the computer processing with post-scan selection o f longitudinal and transverse planes is very useful for diagnostic purposes. A study is under way which compares PHO/CON images o f the myocardium with standard camera views and 7-pinhole views. We hope that this study will give us a better idea o f the usefulness o f the digital processing. E.F. CROCKER: Do you envisage a potential application to regional quantification o f organ function? D.R. PICKENS: We have not made any quantitative studies yet, but intend to work in this area. It must be remembered that some o f the scans, such as those of the heart, are heavily processed and will require particular attention to retain their quantitative information.
IA E A -S M -247/127
DEVELOPMENT OF THE COMPUTERIZED ANGER TOMOGRAPHIC SCANNER FOR POSITRON IMAGING E.A. SILVERSTEIN, E.W. FORDHAM, A. CHUNG-BIN, T. WACHTOR Rush University Medical Center, Chicago, Illinois R.W. ATCHER* Chemistry Division, Argonne National Laboratory, Argonne, Illinois, United States of America
Abstract DEVELOPMENT OF THE COMPUTERIZED ANGER TOMOGRAPHIC SCANNER FOR POSITRON IMAGING. An Anger multiplane tomographic scanner (PHO/CON) was modified for imaging of positron emitters by coincident detection of the positron annihilation radiation. The coinci dence electronics, in the form o f NIM modules, was added externally to the system. Additions and modifications were made to the internal electronics of the PHO/CON to allow the acquisition of positron image data by the existing acquisition system. Positron mode or normal, single photon mode operation is selectable by switches. Operation with a fast coincidence resolving time (2 r) of 20 ns is possible with coincidence losses of 13%. The sensitivity was 3.6 X 103 counts/m in per ( i d with the detectors 60 cm apart for a point source midway between the detectors. A FWHM of 1.5 cm was obtained for the point spread function. Application of the instrum ent for animal and hum an studies is in progress.
1.
INTRODUCTION
A natural extension of our previous work [1 ] on interfacing the Anger tomo graphic scanner [2] (PHO/CON)1 to a digital computer is to add the capability of imaging positron emitters by coincident detection of the two positron annihi lation photons. The PHO/CON used in its normal single photon mode of operation consists of two vertically opposed scintillation cameras fitted with focusing
* Present address: D epartm ent of Radiology, Peter Bent Brigham Hospital, Boston, Massachusetts, United States of America. 1 Model 2, made by Searle Radiographies, now Siemens Gammasonics, under the trade name of PHO/CON.
335
ZU ZL
PULSE SHAPINQ
TIMINQ SCA
PULSE 8HAPINQ
TIMINQ 8CA
336
E X T E R N A L NIM BIN
I N T E R F A C E BIN О HZ Ш (О о íg
COINC.
POSITRON LOGIC BOARD
CONTROL INTERFACE BOARD
DIGITAL INTERFACE BOARD
о о
I______
LINE DRIVERS AND RECEIVERS
____ i :______ OTHER LOQIC SIGNALS
XU.YU
XU.YU
XL.YL
XU.YU XL.YL
S3
COMP« MULTIPLEX CONTROL
DETECTOR CENTRELINE LOCATION
XD.YD XYLfM SCAN
SLAVE BUFFER SYNC
---► COMP ---te- HOLDp--* TO/FROM ACQUISITION SYSTEM
XL.YL SLAVE DERAND. BUFFER iXYUM.SCAN
SHAFT ENCODER DATA FIG.L
Block diagram o f electronics o f PHO/CON modified fo r imaging o f positron emitters.
J
et al.
POSITION PULSE COMPUTATION
MASTER DERAND. BUFFER
SILVERSTEIN
ENERGY ANALY8IS
XU.YU
DEFEAT U/L MUTUAL EXCLUSION
IAEA-SM -247/127
337
MULTICHANNEL
FIG.2. Data acquisition system.
collimators which move in a scanning pattern in a horizontal plane. With positron emitters, by removing the collimators and placing the detectors in coincidence (Fig.l), a tomographic geometry similar to that of the normal PHO/CON can be realized with the advantages of greater geometrical efficiency and freedom from septal penetration problems. Positron emitters such as 52Mnm having high-energy gamma rays (1.43 MeV) that are extremely difficult to collimate can be imaged with a system o f this type. A positron camera consisting o f two opposing scintillation cameras has been constructed [3, 4] and exemplifies many of the properties of this type of system.
OJ OJ
со
л FIG.3. Oscilloscope photo, made with a one second exposure, showing XU versus XL from the upper and lower derandomizers respectively. A 5 ßCi point source o f 6SGa was used, placed midway between the detectors which are 40 cm apart. The centre photo shows the source centred, and the left and right photos show the effect o f decentring the source slightly along the scan line (X direction). The bright spot in the centre is caused by insufficient blanking o f the display.
£
IAEA-SM -247/127
339
The scanning capability o f our PHO/CON based system is felt to be a distinct advantage. Scanning eliminates the variation of the angular acceptance from each source point as a function of position [3], wherein the angular acceptance changes from a maximum at the centre of the camera face to zero at the edge. A scanning system is also capable o f continuous whole-body coverage. The longitudinal (coronal) nature of the sections produced is considered to be useful for this type of application. The cost of the PHO/CON, even with the addition of positron electronics, is quite moderate for a positron imaging system. The instrument is also available for use in its single photon mode. It has been found to be an effective imaging instrument and has found extensive application for brain scanning and 67Ga imaging.
2.
ELECTRONICS
Most of the modifications made to the PHO/CON to allow operation with positrons are electronic in nature and will be described with reference to Fig.l. (The symbols and terminology used follow that of the vendor.) The electronics were modified to allow the following modes of operation, selectable by switches: (a) (b)
Operation in the positron mode with data collected by the acquisition system. Single photon operation with recording of the image on the system’s multi format imager along with simultaneous acquisition, if desired, o f the image data by the data acquisition system (Fig.2).
2.1. Operation in the positron mode A valid positron event consists of two simultaneous events lying within the PHO/CON 20% energy analyser windows, centred at 510 keV. The position signals (XU, YU) and (XL, YL), from the upper and lower detectors respectively, appear at the inputs to the multiplex control board. This board has modifications which defeat the normal PHO/CON logic and thus allow an upper - lower event pair to coexist. While this event processing has been occurring the energy pulses ZU and ZL are brought to the external NIM electronics bin where they enter timing single channel analysers set at 510 keV with a 20% window. The coinc. pulse from the fast coincidence unit then enters the positron logic board, where it is suitably shaped, delayed, and logically combined with other signals. These signals are the GMUNB signals from the PHO/CON energy analysers which prevent certain artifacts from being generated. The output is called MUUB for modified UUB, which is the usual signal causing events to be entered into the derandomizer.
340
SILVERSTEIN et al.
Since a positron event consists of two x,y pairs instead of the single pair from a single photon event, it was necessary to add the additional slave derandomizer, wired so that data enter and leave it in synchronization with the master derando mizer. Thus, when a MUUB signal occurs, (XU, YU) and (XL, YL) enter the derandomizer buffers, if the buffers have room. When data are ready at the output of the derandomizers, the signal S3 is produced which starts data acquisition by the computer and causes the output stage to hold the signals until acquisition is completed. The acquisition system then gives the signal COMP which enables the derandomizer to accept another event. The digital interface board contains latches which read and hold the shaft encoder pulses (giving detector position) when the signal HOLDj3 is received from the acquisition system. The control interface board handles other logic signals. 2.2. Operation in the single photon mode In the single photon mode either an upper or lower detector event is available at the output o f the single derandomizer. A signal, U/L, indicates whether the event was in the upper or lower detector, and the signal A1/A2 indicates which of the dual analysers for each detector accepted the event. These signals are shown entering the data acquisition system in Fig.2.
3.
DATA ACQUISITION SYSTEM
The data acquisition system is shown in Fig.2. It consists entirely of standard peripheral components. For positron operation data are acquired in a three word (3 X 1 6 bits) time independent list mode format for each event. The first word contains digital shaft encoder data giving the position of the detectors, XD, YD. Spare bits are available to provide additional capabilities. The second and third words hold the co-ordinates of the event on the detectors: (XU, YU), (XL, YL), with each co-ordinate being digitized to a precision of 8 bits. The acquisition system is completely configured by software. The acquisition software for positron operation is described below. The single photon software is very similiar to this, lacking the acquisition of a second pair of detector position signals, but making use of the signals U/L and A1/A2. 3.1. Acquisition software, positron mode The presence of a valid event at the output of the derandomizers is signalled to the computer by S3 which actuates the external A/D start (Fig.2) to start conversion of XU. When conversion is complete, an interrupt is generated to the event service routine which operates at the highest priority level. Conversions of
IA E A -S M -2 4 7 /127
341
FIG.4. Phantom consisting o f four Ga sources spaced 4 cm apart in a plane. The sources are 4-mm dia. discs each having an activity o f approximately 0.2 pCi. Detectors are 20 cm above and below the phantom plane. The horizontal direction is along the scan line (X) direction. Reconstructions o f the same data are made at two image plane heights: (left) image plane in plane of phantom, (right) image plane 4 cm above phantom. The study was acquired without scanning the detectors.
YU, XL and YL are then successively made. While waiting for the conversions (20 ps), other operations are performed: the detector centreline positions XD and YD are read in, and all data are packed into the data words. The data are stored in memory in a double buffer which is transferred to disc when full. When processing of the event is complete, the COMP signal is sent to the master derando mizer and the computer exits the event service routine. The signal SCAN, which indicates whether or not the PHO/CON is scanning, and the signal XYLIM, which indicates the end o f a scan line, are used to control data acquisition.
4.
DATA ANALYSIS AND DISPLAY
The data discs are processed on a PDP-11/45 computer system and the results are displayed on this system’s high resolution display system or are put on to magnetic tape and displayed on a nuclear medicine Gamma-11 computer system. Data analysis at present is by back projection of the image events on to the desired image plane. Use of other reconstruction and deblurring techniques will be introduced as needed.
5.
RESULTS AND DISCUSSION
The coincidence timing spectrum was measured with a time-to-pulse height converter and found to have a full width at half maximum (FWHM) of 21 ns
342
SILVERSTEIN et al.
and a full width at one tenth maximum o f 38 ns. With the fast coincidence resolving time (2 r) set at 20 ns, coincidence losses are 13%. This is somewhat better than the FWHM measurement and indicates improvement after careful tuning of the electronics. The sensitivity was 7.5 X 103 counts/min per pCi with the detectors 40 cm apart for a point source in the midplane, or 3.6 X 103 counts/min per pCi for a 60 cm detector separation.2 An instructive and useful display is shown in Fig.3. A centred point source should give a straight line of unity slope, whereas decentring the source should displace the line. This display was very helpful for adjusting the electronics, verifying correct operation, and for tracking down many artifacts. Spatial nonlinearity is evident in these images. The scattered points outside the lines are not from accidental coincidences, but have an unexplained origin. Images of a point source phantom are shown in Fig.4. The FWHM of the sources when in focus is seen to be 1.5 cm. Reconstruction in a plane removed from the source plane gives a more blurred distribution. Improvements in the system remain to be made. Graded absorbers over the detectors and cylindrical shields extending from the detectors are being added to minimize the loading of the electronics from undesired radiation including that from outside the field of view. The event acquisition time should be shortened from its present value of 170 ¿¿s. Addition of more or faster ADCs is a solution, but optimization of the acquisition software could help. A large amount of spatial distortion in the detectors is seen when an orthogonal hole phantom is used with " T cm. A software correction of this should improve the spatial resolution.
6.
CONCLUSION
The addition of coincidence electronics has given the Anger tomographic scanner the capability of imaging with positron emitters. The system has demon strated the expected level of performance, enabling studies to begin with animal and human subjects.
ACKNOWLEDGEMENTS The authors wish to thank the following: A.M. Friedman, of Argonne National Laboratory, for the loan of the NIM electronics and for the provision of radio nuclides; G.V.S. Rayudu of our Nuclear Medicine Department for help in prepa
2
1 Ci = 3.70 X 1010 Bq.
IA EA-SM -247/127
343
ration of phantoms and radioactivity measurements; Digital Equipment Corporation for consulting support and for the loan of the components of the data acquisition system; Siemens Gammasonics for continued material and consulting support, along with extra service calls resulting from the conversion of the electronics; and, finally, M. Danneil of Siemens Gammasonics whose extensive knowledge of the inner workings of the PHO/CON circuitry has enabled us to coerce this complex instrument into performing a function for which it was not designed.
REFERENCES [1 ] SILVERSTEIN, E.A., FORDHAM, E.W., et al., Interfacing a PHO/CON tomographic scanner to a digital com puter, J. Nucl. Med. 18 (1977) 640. [2] ANGER, H.O., “Tomographic gamma-ray scanner with simultaneous readout of several planes”, Fundam ental Problems in Scanning (GOTTSCHALK, A., BECK, R.N., Eds), Thomas Books, Springfield, Illinois (1968) 195. [3] MUEHLLEHNER, G., BUCHIN, M.P., DUDEK, J.H., Performance parameters of a positron imaging camera. IEEE Trans. Nucl. Sei. NS-23 1 (1976) 528. [4] MUEHLLEHNER, G„ ATKINS, F., HARPER, P.V., “Positron camera with longitudinal and transverse tomographic capabilities” , Medical Radionuclide Imaging (Proc. Symp. Los Angeles, 1976) 1, IAEA, Vienna (1977) 291.
IA EA-SM -247/103
EMISSION-COMPUTERIZED TOMOGRAPHY FOR VISUALIZATION OF ORGAN VOLUMES K.J. VIKTERLÖF Radiophysics Department, Regionsjukhuset, Örebro B. SÖDERBORG Department of Hospital Physics, Södersjukhuset, Stockholm, Sweden
Abstract EMISSION-COMPUTERIZED TOMOGRAPHY FOR VISUALIZATION OF ORGAN VOLUMES. Estimation of size, shape and localization of organs are im portant medical problems. Emission-computerized tom ography was mainly developed to improve the ordinary multiple view methods of nuclear medicine, but it also seems to be useful for visualization of volumes and localization of organs. In this study a General Electric 400 T gamma camera was used that was discontinuously rotated around the patient in 64 stationary positions. The computer, Digital Equipm ent PDP 11/34 (Gamma 11), was supported by a software program (SPETS), designed by A. Israelsson and S. Larsson, Karolinska Sjukhuset, Stockholm. When measure m ents were taken the radionuclide distributions in 32 parallel slices were reconstructed, orthogonal to the 64 views from which they were obtained. The software program allowed a simple esti mation to be made of the area of the organ in any of the slices, and the volume of the organ was calculated by means of numerical integration. By mèans of variable radiopharmaceutical kits and 99Tcm-labelling, it was possible to visualize most of the actual organs. The low radiation dosage from nuclear medicine investigations makes the technique useful also in applications when calculation o f the size o f the volume is of medical interest. A technique that makes it possible to visualize organs may also constitute a basis for 3-D therapy dosage planning with quite sufficient resolution in all planes, even if not as good in the transversal section as in pictures derived from conventional transmission-computerized tom ography (CT) technique.
The estimations of size, shape and localization of organs are important medical problems. Periodic documentation o f changes in organ volumes are of special interest in current management o f diseases. Traditionally such determinations are made by using radiographs in several views, The use of ultrasound and nuclear medicine techniques in combination with a manual planimeter has also been recommended in modern literature.
345
346
VIKTERLÖF and SÖDERBORG
FIG.l. Schematical cut-away drawing illustrating the resolution in the three main planes in emission-computerized tomography.
When the transmission-computerized tomography (CT) was introduced the determination o f the size and shape of different organs in different slices was largely improved. The real visualization o f the organ volumes was in this technique dependent on a sequence o f transverse scans. In the same way the calculation of the organ volumes was based on addition of the separate areas in the selected slices. The accuracy o f the whole procedure thus was very much dependent upon the number of slices used and the resolution in the planes perpendicular to the transverse planes. The CT-technique, being characterized by a very high resolution in the transverse scan and poorer resolution in the other planes principally is poor in determining organ volumes. The drawback of the poor resolution in the other planes can be compensated for by means of an increased number of slices. To have a sufficiently high resolution, however, it is necessary to have distances o f only some millimetres instead of the normal 1 or 2 cm. From the radiation protection point of view such a technique is normally not recommended. In general, both the visualization and estimation of organ volumes - to be correct in a real 3-D sense - should be based on a technique where the whole organ is registered by means o f cubical volume elements. In this respect the emission-computerized tomography offers a better basis for adequate determination. As illustrated schematically in Fig. 1 there is about the same resolution in all three main planes when using an emission tomography technique.
IA EA-SM -247/103
347
Techniques for emission tomography using conventional gamma cameras have been developed during the last few years, for example, by Larsson and co-workers1 in Stockholm. The authors o f the present report used a General Electric 400 T gamma camera for the measurements. The camera was discontinuously rotated around the patient in 64 stationary positions. The computer, Digital Equipment PDP 11 /34 (Gamma 11 ), was supported by a special software program (Single Photon Emission Tomography Software — SPETS) designed by A. Israelsson and S. Larsson, Karolinska Sjukhuset, Stockholm. The field o f the gamma camera (400 mm) was divided into 64 equal parts of about 6 mm, which was the live size represented of a pixel. When measurements were taken the radionuclide distribution in 32 parallel slices was reconstructed, orthogonal to the 64 views from which the slices were derived. These 32 transversal slices were also displayed in 64 X 64 matrix pictures. The distance between two slices was about 12 mm, but the volume which was measured in one slice had a thickness of about 20 mm. Unlike the CT-technique, the sagittal and frontal slices that could be reconstructed from these 32 transversal slices are displayed with a comparable resolution in a 64 X 64 matrix. The intermediate values are calculated by the software using interpolation between the surrounding slices. The same technique is often used when the computerized pictures are displayed in the 128 X 128 or 256 X 256 matrix. Thus, the software o f Gamma 11 makes it possible to estimate the areas of the organ in any of the 32 slices, and the volume of the organ can be calculated in a rather correct 3-D manner by means of numerical integration. To test the quality of the depiction the line spread function was measured in different planes. At first the original line spread function of the gamma camera (with collimator) was estimated by a densitometrical method when zooming and blow-up were maximal. The object was a radiation source consisting o f a bottle of " T cm behind a 1-mm-thick slit of lead. The picture of the line source was then taken by means of the Gamma 11 system. It was possible to use different matrix settings 32 X 32, 64 X 64 and 128 X 128, including the total circular field of the gamma camera. A change in the matrix settings can be considered as a method of zooming, decreasing the size of the picture element by a factor of 2. Moreover, the size o f the picture element could be altered by zooming the picture of the gamma camera from one to three in six steps. By means of these different settings it was possible to change the size of the picture element in the combined gamma camera and Gamma 11 system from 0.96 mm to 25 mm, giving 36 different values. The real size o f the element was measured directly from the observed distance between two point sources surrounding the slit.
1 LARSSON, S., Development and studies on SPETS (Single P hoton Emission Tomography Software), Acta Radiol., Suppl. (to be published).
348
VIKTERLÖF and SÖDERBORG
In a separate phantom study how much the FWHM was altered by the mathematical reconstruction of tomography was tested. The phantom was a cylindrical Perspex container filled with water and with a thin tube with "T cm mounted near the centre of the cylinder. Measurements were taken in a normal gamma camera picture and in a tomographic reconstruction. Both measurements were taken with the same size of the picture element, and the FWHM as well as a total line spread function was registered. Furthermore, tomographic reconstructions were performed in planes orthogonal to the primarily reconstructed tomographic pictures. The presentation of these reconstructed pictures on the screen was the same as for the primarily reconstructed pictures achieved by interpolation. The size of the matrix elements in the doubly reconstructed pictures is double the size of the primarily constructed ones. The FWHM of the gamma camera with a collimator, as measured on photo graphic film, was found to be 4.1 mm. The digital pictures were evaluated directly by the computer system. The values of the FWHM, Z, varies with the size of the picture element, I, and the FWHM for the gamma camera, G. The connection between Z and I was found to be very adequately described by Z2 = I2 + G2. Thus, there is little advantage in using pixel sizes much smaller than the FWHM of the gamma camera (Fig.2).
IA EA-SM -247/103
349
The estimated FWHM in a conventional picture by a gamma camera was found to be 15.7 mm whereas the reconstructed one in the transverse slice was 16.2 mm and the reconstructed one in the frontal slice was found to be 21.8 mm. Most of the difference is caused by the pixel sizes. In conclusion, it is argued that it has been proved that the emission tomography technique is well suited for adequate determination of organ volumes simply by adding the number of pixels which contain a selected number o f counts. The total number of cells is pro portional to the organ volume. The single-photon emission tomography system using a gamma camera mainly utilizes " T cm-labelled radiopharmaceuticals. By means of standard kits it is thus possible to demonstrate the most interesting organs, both as single and as combined organs. This means that normal bones, lungs, kidneys, bladders, livers, spleen, etc., can be visualized and their volumes can be quantified. With access to such a technique for estimating size, shape, localization and volumes of organs, it seems quite obvious that it could have interesting applications for 3-D dose planning in radiotherapy. It has been shown above that such a technique gives a sufficient resolution in all main planes, even if not as good as the CT-technique in the transverse section, nevertheless on the whole it establishes a better basis for the 3-D determination of the target volume and surrounding inhomogeneities. In fact, in physical dose planning at present mostly performed by means of computer calculations, correct information concerning the extensions and locations of inhomogeneities, such as bones and particularly the lungs, is of the greatest importance. One of the ideas behind the design of emission-computerized tomography with the gamma camera has been that all examinations should be done with ordinary activity doses, and with a total examination time equal to or shorter than for the equivalent normal gamma camera examinations. It could be supposed that this technique could be well adapted for problems related to determining organ volumes and information on patients adequate for 3-D dose planning. Software programs for simpler definition of the organ volumes as well as for the better presentation of the results are under development.
DISCUSSION H.N. WAGNER, Jr.: How difficult is it to gate the 400 T system for cardiac volume studies? K.J. VIKTERLÖF: No major difficulties are encountered, except that a longer acquisition time is required.
IA EA-SM -247/43
CONSTRUCTION AND CLINICAL APPLICATION OF COMPLEX UTILITY PROGRAMS IN THE SEGAMS-80 SYSTEM E. MATÉ, J. CSIRIK Laboratory of Cybernetics, University o f Szeged L. CSERNAY Department of Nuclear Medicine, Medical University of Szeged Á. MAKAY Department of Computer Science, University of Szeged, Szeged, Hungary
Abstract CONSTRUCTION AND CLINICAL APPLICATION OF COMPLEX UTILITY PROGRAMS IN THE SEGAMS-80 SYSTEM. SEGAMS-80 is a system for processing isotope-diagnostic pictures easily and safely for physicians. The functions built into the system form a tree-structure. In certain stages of processing, tables completed according to the medical point of view show identification and a short description of the actual performable functions. The functions available allow an inter active performance of diagnostic processes for different purposes. Interactivity is undesirable while processing routine examinations, since the functions to be performed, their sequence and parameters could be identical in all cases. SEGAMS-80 makes it possible to construct complex programs, to put them into the system and execute them. During the complex program the desired functions are automatically executed. The operator’s interference is needed only where the author of the complex program has stated that it is necessary from the medical aspects. Experience gained with several SEGAMS-80 systems has shown that they can be successfully used in isotope diagnostics, w ithout requiring any training in computing techniques from the physicians. A schematic description is given of the structure of SEGAMS-80 together with a detailed account of how to construct complex utility programs.
During the last few years minicomputers connected with gamma cameras have played an ever increasing role in the daily routine work of isotope diagnostic laboratories. To obtain and process data by the Szeged Gamma Camera System (SEGAMS) and its further developed variant, SEGAMS-80 has been processed. The nearly two hundred elementary functions built into the system allow the physician to perform diagnostic work to evaluate widely the results of various examinations. 351
352
MÁTÉ et al.
TABLE I. PICTURE PROCESSING TABLE (IN ENGLISH)
PT
PARAMETERS TO PRINTER
PI
PRIMARY PICTURE
KT
PICTURE REFLECTION
FO
PICTURE ROTATION
EX
EXPANSION
UC
UNIFORMITY CORRECTION
sw
WEAK SMOOTHING
SI
SMOOTHING WITH VARIABLE CONFIGURATION
SS
STRONG SMOOTHING
BG
BACKGROUND SUBTRACTION
OS
OPTIMAL SMOOTHING
WA
WORKING AREA
SP
STEPPING THROUGH THE FRAMES
C4
MINI-PICTURES ON THE TV
DT
DISPLAY TABLE
RE
RETURN TO THE PRECEDING TABLE
FIG.l. Schematic structure o f data processing.
When composing and building up the system, one of the most important aims was that the system should not presuppose a knowledge of computing technique in the user; tables, messages and questions should direct the work of the user. To ensure this, the functions were grouped, put into tables and these tables were arranged in a hierarchy, corresponding to the strategies and tactics of the physicians evaluating the work. The physician carries on the evaluating procedure through a dialogue with the alpha numerical display connected to the computer. The short names and attributed identifiers of functions to be activated at a given point of processing are written in tables on the display. Processing goes
353
IAEA-SM -247/43
TABLE II. PICTURE PROCESSING TABLE (IN GERMAN)
PI
UNBEARBEITETES BILD
PT
DRUCKEN DER PARAMETER
KT
GESPIEGELTE DARSTELLUNG
FO
BILDROTATION
EX
NORMIERUNG
UC
INHOMOGENITAETSKORREKTUR
SW
LEICHTE GLAETTUNG
SI
GLAETTUNG NACH WAHL
SS
STARKE GLAETTUNG
OS
OPTIMALE GLAETTUNG
BG
UNTERGRUND SUBTRAKTION
SP
HANDGESTEUERTE BILDFOLGE
WA
ARBEITSBEREICH
DT
DISPLAY TABELLE
C4
MINIBILDER AUF FARBDISPLAY
RE
ZURUECK ZUR VORIGEN TABELLE
on through typing the two-letter identifier, chosen from the table, and performing the corresponding function. Functions regarded necessary for processing a given picture are summarized in Table I. Other tables (e.g. ROI-treatment, curve treatment, etc.) are o f a similar character. The schematic building of the structure composed out of the tables is illustrated in Fig. 1. Some functions always take place in the same way (e.g. EX, SW, etc.) whereas others depend on the parameters which the function obtains from the user, through the dialogue carried out on the alpha numerical display. For example, a possible form o f the simple BG (background subtraction) is: BG STANDARD BACKGROUND?
Y
PERCENTAGE (MAX. 99): 10 READY?
N(10% low background subtraction)
READY?
N (another 10% low background subtraction)
READY?
Y (function performed)
The system is composed out of relocated programs, subroutines and detached program libraries by means of the so-called generator program. To insert a newer function is quite easy after preparing programs realizing the functions. On trans lating the text library, the original system in English can easily be rewritten into any other desired language by attaching the new text library. This solution ensures that the user could use his mother tongue in the course of evaluation (Table II).
354
MATÉ et al.
The complete interactivity of the system, the writing of tables, orientation questions and messages supply many advantages when evaluating individual examinations. On the other hand, interactivity in a standard evaluation of routine investigations is undesired, since then the functions to be performed, their sequence and production of their parameters are the same in all cases. Therefore SEGAMS-80, besides performing routine examinations in a pre-defined manner, makes it possible to attach so-called complex processing programs put together by adding SEGAMS functions to the pre-defined ways of processing. When the complex program goes on the desired functions are performed automatically, the operator has to intervene only when it is thought important from the physician’s point of view by the writer of the complex program. How is a complex program put together? We take from the catalogue an examination, the data acquisition and data processing of which is required to be standardized. The data acquisition is built into the system in a pre-defined way. The evaluation of the system is carried out by operating manually and the characters are typed in the course of the evaluation; the series of characters given to the individual functions are closed with the symbol ; and at the end is written $. For example: EX SW ; BG Y ION Y ; EX ; g The series of signs thus obtained can be regarded as a very simple complex program, which can be built into the system at the corresponding point and can be performed later at any time. The program example described above first expands the picture shown on colour television, then performs a weak smoothing, then carries out a 10% low background subtraction and expands the picture again. The result of the complex program will be this processed picture. A complex program of this type does not allow any interactive intervention, and does not contain branches, and it has the disadvantage that in the run the television screen flickers, an automatic dialogue going on the display without the user being able to follow it attentively. To eliminate these inadequacies and to open newer possibilities various directing functions can be written into the complex program. This information makes possible: Producing interactive parameters. Interactivity until the performance of the given function. Free branch from the complex program (any desired function can be inter actively called, but it is important that the direction should be given to the complex program on the table as supposed by the latter). Writing the message of the complex program on the display. Unconditional handing over of direction to any point in the complex program.
IAE A-SM -247/43
355
Conditional handing over of direction on the basis of Y or N answer of a question put by the complex program (IF . . . GO TO). Conditional handing over of direction on the basis o f Y or N answer given to any of the functions (IF . . . GO TO). There is no sign on the colour television. There is again sign on the colour television. The system does not write on display. The system writes again on display. The system deletes the text on display. Indicating the approximate time of the run of the complex program. For a better orientation commentaries can be added anywhere in the complex program. When building in the complex program a list of the program is completed, the system carries out a syntactic control, and shows the possible errors and misprints. It inhibits writing syntactically incorrect programs. With the process described above many simple and more complicated complex programs have been developed in the authors’ institute and elsewhere using the SEGAMS software system in everyday use. Even users lacking basic computer technique training can develop effective evaluating programs. In routine clinical practice the most frequent are evaluation programs showing correspondingly processed pictures o f static brain and liver scinti-programs from four directions, dynamic camera renograms prepared with 99Tcm-DTPA and 131I-Hippuran, and brain-blood circulation examinations carried out with "T cmpertechnetate. When carrying out dynamic programs, apart from producing and showing the resulting pictures and curves, the complex program performs the calculation and printing of the necessary numerical parameters. The time neces sary for the evaluation by manipulation manually is shortened to one fifth, and one tenth by summarizing the task into a complex program. The evaluation of a camera renogram, in an interactive way, manually usually takes 12 —15 min, whereas with a complex program it is shortened to 3 —4 min. In our opinion and according to our experience, this very simple and applicable method is of importance in particular for users with less experience in computing techniques and for nuclear diagnostic laboratories in developing countries. Naturally, some special processing desires may arise, which could not be met by performing SEGAMS functions. To solve such problems a FORTRAN pro gramming possibility has.been envisaged. SEGAMS has been attached to FORTRAN as a periphery; to solve part-tasks in SEGAMS, the FORTRAN program must send a text to SEGAMS which, interpreted and performed as a complex program, can yield the desired result. Such a solution basically means that the subroutine
356
MÁTÉ et al.
library of FORTRAN has been completed with all the functions of SEGAMS. In this way programming the special problems in FORTRAN has also become simpler.
IAEA-SM -247/100
AN ELECTRONIC IMAGE PROCESSING DEVICE FEATURING CONTINUOUSLY SELECTABLE TWO-DIMENSIONAL BIPOLAR FILTER FUNCTIONS AND REAL-TIME OPERATION B.D. CHARLESTON, F.H. BECKMAN, M.J. FRANCO, D.B. CHARLESTON The Franklin McLean Memorial Research Institute, University o f Chicago, Chicago, Illinois, United States of America
Abstract AN ELECTRONIC IMAGE PROCESSING DEVICE FEATURING CONTINUOUSLY SELECTABLE TWO-DIMENSIONAL BIPOLAR FILTER FUNCTIONS AND REAL-TIME OPERATION. A versatile electronic-analogue image processing system has been developed for use in improving the quality of various types of images with emphasis on those encountered in experi mental and diagnostic medicine. The operational principle utilizes spatial filtering which selectively controls the contrast of an image according to the spatial frequency content of relevant and non-relevant features of the image. In cases where the relevant feature of an image is obscured by noise, the noise can be reduced or eliminated by selectively lowering the contrast of inform ation in the high spatial frequency range. In other cases, edge sharpness can be enhanced by accentuating the upper midrange spatial frequencies. Both methods of spatial frequency control may be adjusted continuously in the same image to obtain maximum visibility of the features of clinical or research interest within an image. A precision video camera is used to view medical diagnostic images, either prints, transparencies or CRT displays. The ou tp u t of the camera provides the analogue input signal for both the electronic processing system and the video display of the unprocessed image. The video signal input to the electronic processing system is processed by a two-dimensional spatial convolution operation. The system employs charged-coupled devices (CCDs), both tapped analogue delay lines (TADs) and serial analogue delay lines (SADs), to store inform ation in the form of analogue potentials which are constantly being updated as new sampled analogue data arrive at the input. This inform ation is convolved with a programmed bipolar radially symmetrical hexagonal function which may be controlled and varied at each radius by the operator in real-time by adjusting a set o f front panel controls or by a programmed microprocessor control. Two TV m onitors are used, one for processed image display and the other for constant reference to the original image. The completed version will allow reasonably accurate quantification of the filter function. The working prototype has a full-screen display m atrix size of 200 picture elements per horizontal line by 240 lines. The matrix can be expanded vertically and horizontally by adding modified redundant circuitry and changing the clocking rates, allowing for the processing of images which contain greater detail with finer structure. In addition, the processing array can be expanded to contain many more ‘rings’ or radii.
357
358 1.
CHARLESTON et al.
INTRODUCTION
The prototype of an electronic image processing system is described which was designed to circumvent the inconveniences, obstacles and expenses which most investigators have encountered in past image enhancement studies. For example, little progress has been made over the past few years in enhancement studies of radionuclide images because of the cumbersome, incon venient and expensive methods required for generation and evaluation of clinical images before and after processing. The authors have had considerable experience with coherent optical spatial filtering systems [1,2], digital image processing and electronic analogue image processing with both black-and-white and colour TV display systems [3 -5 ]. Digital image processing requires the use of sophisticated and expensive computer hardware, with considerable software development for support, especially where iterative operations are required for dynamic display of variable filter functions. Digital processing requires that the original image either be obtained initially in digital form or be translated into digital form for computer processing. In addition, the service of a computer operator is usually required to generate a display series. Coherent optical spatial filtering systems require support from highly skilled technical operators, darkroom photographic techniques and precision opto mechanical equipment including liquid gates. The systems are tedious to operate. Each filter function (usually in the form of a film transparency with a specific concentric density exposure pattern) is inserted in the transform plane with extreme precision. The operation is cumbersome and slow. Coherent optical spatial filtering systems are valuable tools for investigators because they have exquisite spatial resolution characteristics, but they cannot be considered as practical on-line clinical diagnostic aids. The authors’ previous studies of image enhancement with TV systems were useful, but limited. Sharp filtration suffered from phase shift limitations, and the various adjustable parameters such as brightness, contrast and grey-scale level interlocked and overlapped so that it was difficult to establish quantitatively what process or combination of adjustments produced a successfully enhanced image. Evaluation of the efficacy of any clinical image enhancement operation should involve the clinical diagnostician, whose visual-mental training offers the best practical judgement. Involving a busy clinician in a complicated or cumbersome study program with delayed results has always been a drawback to image enhancement studies. In order to circumvent these obstacles to the study of image processing and enhancement, and to make it more convenient for the clinical diagnostician to participate in the evaluation of the process, the authors proposed selected design parameters for a practical instrument. Ideally, the instrument should:
IAEA-SM -247/100
1. 2. 3.
4. 5. 6. 7.
359
Operate in real-time, on-line in the clinics. Provide continuously selectable two-dimensional bipolar filter functions. Be capable of viewing and processing images on film (transparencies or photographs) of varying size, on motion picture film and on CRT displays. Be simple to operate, with ease of programming by either manual or microprocessor control. Provide continuously variable programming controls which allow operator interaction with the dynamic display. Provide simultaneous viewing of processed and unprocessed images; and Provide a dynamic quantitative display of the filter function values introduced.
Our present working prototype meets all of these criteria with the exception of (7). However, the linear slide potentiometer controls provide a rough indication of relative filter function values (positive or negative) by their position on the control box. The system is designed to accept the quantitative display feature and microprocessor; both will be added. A physician with even a cursory knowledge of spatial filtering functions can operate the system after minimal training. Once he is familiar with the system, evaluation of enhanced regions of interest within a clinical image should be straight forward. Dynamic changes within the processed image can constantly be compared visually with the unprocessed image display; the operator offers a degree of control which tends to limit the probability of ‘accidental’ or excessive enhancement, which can generate artifacts.
2.
TECHNICAL DESCRIPTION
A precision video camera views medical diagnostic images, which may be prints, transparencies or CRT displays. The output of the camera provides the analogue input signal for both the electronic processing system and the video display of the unprocessed image. Image data are entered into the processing system serially from the output of a conventional raster scan high-resolution television camera. This standard television signal is amplified and split into AC and DC signal paths. The DC portion of the signal bypasses the image-processing stages and is returned to the processed image at the output of the system. This DC stripping and restoration procedure prevents the background intensity of an image from affecting the processing operation, while maintaining the background level for direct comparison of processed and unprocessed images.
360
CHARLESTON et al.
F I G .L Sampling structure-input sample-and-hold (highly exaggerated). Letters designate a horizontal line, numbers designate pixel on a line.
The AC portion of the signal is amplified and low-pass-filtered to prevent aliasing by subsequent sampling. The video signal is now measured by a video rate absolute-value amplifier and averaging filter. The analogue potential from this circuitry is compared with the output of an identical circuit measuring the average value of the processed video signal. By comparison and display of input and output averages, the operator gains some control over the contrast differential of input and output images. The comparator circuitry at present drives two lightemitting diodes which indicate whether the output average is less than or greater than the input average. Scaling of the filter functions is then possible to compen sate for differences in image averages. The input video signal is sampled by a high-speed sample-and-hold circuit which develops an accurate and uniform pixel configuration to equalize the sampling discrepancies found in the two types of charge transfer devices used in the processor. Figure 1 indicates the sampling structure of the output of the sample-and-hold circuit, showing a small portion of the image in a sampled format. The numbering convention used throughout the illustrations corresponds to a specific pixel developed from the image. By staggering of pixels on a line-to-line basis, radially symmetric filter functions can be implemented at a higher spatial resolution than is obtained with an in-line vertical structure.
361
IAEA-SM -247/100 INPUT TO С .CD. CHAIN FROM -* SAMPLEB HOLD
к E
8
B. D INPUT TO —►—►8 S A D. N02'1 1. INPUT TO S.A.D.N0.3
INPUT TO S.A.D. N0.4
INPUT TO S.A.D. N0.5
FIG .2.
1.
D D D D 7 8
В В В В В 8 8 7 7 6
BL« NK А INTERVAL 8
—►
I +
BLA NK С с С С INTERVAL 8 8 ! / B.
OUTPUT 0F-FIRST SAD.
А А А 8 7 7
D D D D D D D D D D D 6 5 5 4 4 3 3 2 2
С С С b b b
BLANKNiS INTERV AL
С BLANK№G 1 1 INTERV AL
С С
С С С С С С b 4 4 ó ó ¿
¿
В В В В В В В В В В В 6 5 5 4 4 3 3 2 2 1 1
BLANKNб INTERVAi.
А А А А А А А А А А А 5 5 4 4 3 3 2 2 1 1
А
1'
6 6
9LANKNG INTERVAL
t
OUTPUT OF LAST S.A.D.
Illustration o f a charge-coupled serial analogue delay line configuration. N ote first
in-first out structure.
ГТТ1 EXT a
T~rr ■ T I,I (
J J X W* ------1
i i inii....m:r'l------
■■rillllllHIIHiHHtWty*
h ex a g o n al FILTER
ARRAY XX
I ---- T
ЭЕ
г...:ч i i L-1J
х Я :
□ ET
п жf i l
m iïiiW ïiT iiT
I I I I I I I I
3r T “ -, xzri
ЦП
n
IX Tr -
IX
IX Д, ,1 л,~гх л~г т iX
F IG .3.
.
l"îrT|T|,T.;Ùijnl5
'77
i= F
TX т
а
X I
Hexagonal filter array shown within delay array.
Г' IN
S.A.D. I
BLANK ING E - 8 E - 7 E -6 E - 5 E - 4 TAPHO.1 TAPN0.2 TAPN0.3 TAPN0.4 TAPN0.5 TAPN0.6 TAPN0.7
BLNK. S.A.D 2
D G S.A.D. 3
D -8
D -7
D -6
D -5
D -4
D -3
TAPHO.1 TAPN0.2 TAPN0.3 TAPN0.4 TAPNO. 5 TAPN0.6 TAPNO.7 L i < ÿ -----------------------------
BLANK ING r C - 8 C - 7 C - 6 C - 5 C - 4 TAPNO.1 TAPN0.2 TAPNO.3 TAPN0.4 TAPNO.5 TAPN0.6 TAPN0.7 1 4_ f ^
D S.A.D.4
BLNK. B -8 B -7 B - 6 B - 5 B -4 B -3 TAPN0.1 TAPN0.2 TAPN0.3 TAPN0.4 TAPNO.S TAPN0.6 TAPN0.7
S.A.D. 5
4 k S --------------------------------BLANK ING A -8 A - 7 A -6 A -5 A -4 TAPNO.1 TAPN0.2 TAPN0.3 TAPNO.4 TAPNO.5 TAPN0.6 TAPN0.7 TAPPED ANALOGUE DELAY LINE ARRAY
CCD. SERIAL ANALOGUE DELAY CHAIN I I__________________ I
FIG.4. Simplified system block diagram.
X-INPUT ANALOGUE MULTIPLIER XxY OUTPUT N0.2 Я У Y-INPUT DESIRED WEIGHTING ANALOGUE X-INPUT MULTIPLIER У \_____ » b o N0.1 I Y INPUT DESIRED WEIGHTING
OUTPUT
IA EA-SM -247/100
363
The output from the sample-and-hold circuit enters a cascaded chain of chargecoupled devices (CCDs) operating as a chain of analogue shift registers. Each CCD in the chain stores a complete horizontal line. The output of each horizontal line storage element delays the sampled video by one horizontal line relative to the previous output. In the present system, there are 33 CCDs which form the chain; therefore, 33 consecutive horizontal lines are stored within these shift registers at any given moment. Figure 2 illustrates a chain of CCDs at an arbitrary instant in time, with pixels corresponding to those of Fig. 1 stored within the chain. Each input to the CCDs and the final output of the chain enter a different type of charge transfer device, called a tapped analogue delay line (TAD). These devices shift analogue potentials in a manner similar to the CCDs, and also provide an output from a non-destructive sensing of the analogue potential in each consecutive storage site. The outputs of the TADs access every pixel in an entire region o f the input image simultaneously. The number of pixels in the region accessed is determined by the number of tap outputs available on each TAD, multiplied by the number of TADs used. This region will be called the ‘delay array’. Because simultaneous access to a pixel and its surrounding pixels is provided, the foundation has been laid for real-time image processing. In the prototype system, the central tap output from the delay array forms the central point of an approximately radially symmetric filter function. A hexagon was chosen for the approximation and is shown in the delay array of Fig.3. The hexagonal filter array outline in the region of Fig.3 identifies 74 distinct radial distances from the central tap output. Outputs with identical radial distances are summed together. These 74 sums and the central tap output enter 75 separate wide-band four-quadrant multipliers. Each multiplier has a bipolar weighting coefficient input that is adjustable to the contribution desired at that radial distance. At present, these weighting coefficients are derived from a linear interpolation of values programmed by means of slide potentiometers. There are 17 slide potentiometers which directly program the major radii extending from the central output tap to the points of the hexagon. The linear interpolation operation generates a weighting coefficient that is dependent upon the weighting of the main radii and the relative radial location between two consecutive major radii. The product of all tap outputs and the associated weighting coefficient multipliers are summed, as shown on the left in Fig.4. This signal now has the DC (background level) component re-introduced. A delayed composite video sync signal is added for output to a conventional television monitor. The sync signal is delayed by one half the total delay array. To generate the desired filter function, the corresponding impulse response is directly programmed by adjustment of the slide potentiometers.
364
CHARLESTON et al.
This two-dimensional spatial filtering system is a discrete time system and, as such, lends itself to Z-transform analysis techniques which have been extensively developed for use in systems engineering. The system accepts bipolar impulse response functions, thus allowing for differentiation of the signal which permits image sharpening for edge detection and edge enhancement of image detail. The maximum clocking frequency of present commercially available charge-transfer tapped analogue delay lines limits the resolution of the present system. For improved resolution, use of multiplexed redundant TADs and related downstream circuitry would make it possible to implement a real-time system approaching 1000 lines of resolution. Although the authors chose a hexagonal filter array to prove the feasibility of real-time two-dimensional image processing, other filter configurations are possible. By interfacing the computational power of a microprocessor to control individual weighting coefficients and summing modes, one can implement a wide variety of symmetric and non-symmetric filter functions.
3.
CONCLUSION
A system with such versatility could facilitate the search for specific patterns or shapes within an image. The prototype described here has been used in the investigation of nuclear medicine images, ultrasound images and chromosome images, with generally favourable results. This is the first use of charge-coupled devices (CCDs) for two-dimensional bipolar image processing, and the first realization of analogue television spatial filtration without phase shift limitations. The major electronic breakthrough was made when CCDs were utilized at TV rates; this implies that any TV camera image can be processed directly by autocorrelation techniques, filtration, matched filtration, adaptive filtration, grey-scale compression or expansion, etc. Although the system was initially designed to handle nuclear medicine images, it can easily be expanded to handle high-resolution images such as aerial photographs, radiographs and anything ‘seen’ by a TV camera and lens.
REFERENCES [1] BECK, R.N., HARPER, P.V., CHARLESTON, D.B., YASILLO, N.J., “Optical processing of radionuclide images”, Biomedical Sciences Instrum entation 6, Instrum ent Society of America, Pittsburgh, Pennsylvania (1969) 241. [2] BECK, R.N., HARPER, P.V., CHARLESTON, D.B., YASILLO, N.J., Improvement of image quality of photoscans by spatial filtering, J. Nucl. Med. 8 (1967) 286 (abstract only).
IAEA-SM -247/100
365
[3] CHARLESTON, D.B., BECK, R.N., ÉIDELBERG, P.E., SCHUH, M.W., “Techniques which aid in quantitative interpretation of scan d ata” , Medical Radioisotope Scanning (Proc. Symp. Athens, 1964) 1, IAEA, Vienna (1964) 509. [4] CHARLESTON, D.B., “ An analog approach to scan readout data m anipulation and enhancem ent” , Fundam ental Problems in Scanning (GOTTSCHALK, A., BECK, R.N., Eds), Charles C. Thomas, Springfield, Illinois ( 1968) 275. [5] CHARLESTON, D.B., BECK, R.N., WOOD, J.C., YASILLO, N.J., “A versatile instrum ent for photoscan analysis which produces color display from black and white photoscans” , Radioactive Isotopes in the Localization of Tumours (McCREADY, V.R., TAYLOR, D.B., TROTT, N.G., CAMERON, C.B., FIELD, E.O., FRENCH, R.J., PARKER, R.P., E d s), Heinemann, London (1969) 56.
DISCUSSION A.E. TODD-POKROPEK: Could you please give us an idea of the cost of the system and whether such a system is more cost effective than a general-purpose array processor? B.D. CHARLESTON: A mass-production unit might cost about US $20 000. This system is extremely versatile as it is capable of handling all types of images. This versatility combined with its real-time capability suggests that it is more cost effective. R. DI PAO LA: Can your system handle dynamic studies? B.D. CHARLESTON: The capability of the system to handle dynamic studies is subject to the same restrictions as conventional NTS television systems, i.e. 30 frames per second.
IA EA-SM -247/26
PRACTICAL APPLICATION OF DECONVOLUTION TECHNIQUES TO DYNAMIC STUDIES C.C. NIMMON, T.Y. LEE*, K.E. BRITTON, M. GRANOWSKA, S. GRUENEWALD Department of Nuclear Medicine, St. Bartholomew’s Hospital, London, United Kingdom
Abstract
PRACTICAL APPLICATION OF DECONVOLUTION TECHNIQUES TO DYNAMIC STUDIES. The stability of the impulse retention function (IR F) as derived by deconvolution of activity-time (А/T ) curves obtained from a dynamic study, is dependent both on the nature of the input function and on the degree of statistical noise present on the raw data. For the type of input function obtained in first pass circulation studies the derived IRF is potentially unstable. Methods have been investigated which seek to achieve a satisfactory solution by the incorporation of a priori knowledge of the system expressed in mathem atical form as a set of constraints. The best recovery of a theoretical IRF is achieved by the incorporation of a combination of smoothness, m onotonicity and non-negativity constraints using a regularization method applied in the frequency domain by means o f the fast Fourier transform. In the example o f the cerebral blood flow study using a non-diffusible tracer, the m ethod has been evaluated with respect to errors arising both from statistical noise and from preprocessing of the raw data. The resultant expected error on the mean transit time (MTT) has been found to be slightly lower than the variation found in patient reproducibility studies. For an input function which is expressible as a sum of exponential terms, the IRF derived using uncon strained least-squares m ethods has been shown to be relatively stable. In the example of the renogram study, it has been found th at the m atrix algorithm m ethod yields good estimates of the MTT. In order to facilitate interpretation of the transit time spectrum (TTS), a model is proposed in which TTS are calculated corresponding to two zones whose principal con tributions arise from the cortical and medullary regions of the kidney respectively. After reduction of the noise com ponent by means of a correlation technique, com bination of the TTS leads to enhanced resolution of modes originating in the separate zones. Preliminary studies with this m ethod indicate a possible advantage over m ethods which seek to resolve a bimodal distribution from a single TTS.
* Present address: D epartm ent of Clinical Physics and Bioengineering, West of Scotland Health Board, Glasgow, United Kingdom.
367
368 1.
NIMMON et al.
INTRODUCTION
Over the last decade, much progress has been made in the understanding of the consequences of the statistical noise term in the convolution equation and in the development o f strategies required to obtain satisfactory solutions. It has been pointed out [1—3] that unconstrained least-squares methods, which minimize the squared deviation W2 between the reconvoluted retention curve and the original data, do not necessarily lead to a solution with a low value of S2, the squared deviation between the derived and the true impulse retention function (IRF). Of particular importance has been the recognition of the existence of two categories of the deconvolution problem. In the first category, class A, the dispersion S2 remains finite and converges for increasing number of terms in the sampled data sequence, and in the second, class B, the dispersion increases. For a given random noise process the stability of the derived IRF has been shown to depend on the nature of the input function [ 1]. In class A problems, the effects of the noise term may be regarded as an oscillatory perturbation on the true solution. Further, some characteristics of the true solution may be quite accurately determined without the use of sophisticated noise reduction techniques. In the class В problems, however, meaningful solutions can only be determined through the incorporation of constraints based on a priori knowledge about the particular system under study. It has been demonstrated (see Appendix 2) that an input function which can be approximated as a sum of exponential terms leads to a class A problem. An important example is the renogram study, in which the blood clearance activity/ time curve (А/T) is sampled at intervals of 10 or 20 s. A more general input function, as represented by a gamma function, leads to a class В problem. An example of the latter is the first pass cerebral blood flow study using a non-diffusible tracer [4], in which the input function is the aortic А/T curve sampled at intervals of 0.3 s. The aims of this study were firstly to examine the use of constrained leastsquares methods in determining values of mean transit time (MTT) in the cerebral flow study, and, secondly, to investigate whether such methods when applied to analysis of the renogram would yield improved reliability in the measurement of indices of the transit time spectrum (TTS).
2.
CEREBRAL BLOOD FLOW STUDY: CLASS В DECONVOLUTION PROBLEM
The deconvolution is carried out between the input obtained from a probe positioned over the aortic arch and first pass А/T curves corresponding to selected regions-of-interest on the vertex view. Using a multiple input linear model, the
IAEA-SM -247/26
369
regional MTT may be calculated from the IRF by means of an area/height relationship [4]. 2.1. Selection of optimal deconvolution method 2.1.1. M ethod o f analysis Deconvolution methods were chosen from the following three types: (a) Matrix methods. This type utilize the expression of the least-squares (LS) solution in terms of an inverse matrix. In general, incorporation of constraints leads to either an iterative or a regularized matrix technique, (b) Fourier trans form methods. This type makes use of the fact that deconvolution is equivalent to division in the frequency domain. Frequency filtering is readily included in these methods and the need for matrix inversion is bypassed, (c) Numerical function minimization methods. From the many available [5], simplex and stochastic methods were chosen as examples of direct search techniques. The mathematical formulation of the basis of the LS methods used is given in Appendixes 1 and 2. Further mathematical details of all methods have been given by Lee [6]. The following three types of constraint based on a priori knowledge have been studied: (a) Non-negativity. A negative value of the IRF has no physical interpretation. (b) Monotonicity. This constraint considers the IRF to have a single peak and to behave monotonically in the segments pre- and post-peak. This is equivalent to the assumption that all of the spike input originating at the aortic arch reaches the brain before the minimum cerebral transit time, and in addition that there is no backflow of tracer. (c) Smoothness. The squared second difference of the IRF was used as an index of smoothness. Qualitatively, a degree of smoothness may be expected on account of the complex composite nature of the cerebral circulation. Five combinations of the constraints (a), (b), (c) were implemented as follows using the particular deconvolution methods (A) to (I): (1)
No constraint. (A) Matrix algorithm; (B) Orthonormal polynomials [6]; (C) Fast Fourier transform (FFT). (2) Non-negativity. (D) Iterative matrix [7]; (E) Simplex [8]; (F) Stochastic [9]. (3) Non-negativity and monotonicity. (G) Stochastic. This is an extension of method (F) in which the gradient of the IRF is parameterized and segmental search techniques used.
370
NIMMON et al.
•D ) ITERATIVE MATRIX
CF) STOCHASTIC
FIG.L Results o f comparison o f the deconvolution methods (A) to (I), (a) Unconstrained methods; (b) Non-negativity constraint; -(c) Monotonicity and smoothness constraints. W1 is the squared deviation between the reconvoluted retention curve and the original data; T is the computing time in minutes.
(4)
(5)
Smoothness. (H) FFT regularization method. Based on the method developed by Phillips [10], Twomey [11] and Hunt [12], this version together with method (I) differs in the method o f choice of the regularization parameter (see Section 2.2). All three constraints. (I) FFT regularization method.
IAEA-SM-247 /2 6
[ J 3 J naHOTDHXCITY /
371
SHaOTHNSSS CONSTRAINTS
FIG.l (cont.j
The performance of methods (A) to (I) was tested using a theoretical IRF generated from a gamma function formula (see Appendix 3). Methods (G) and (I) were compared using a set of IRFs with MTT values between 5 and 12 s. Simulated А/T curves were constructed by convolution with an input curve from a patient study and Poisson noise was added. 2.1.2. Results The results obtained in the comparison of methods (A) to (I) are shown in Fig. 1, together with the corresponding values of W2 and the computing time, T min, required. (A Varían 620/L100 was used for all computation.) As expected, all the unconstrained solutions show a rapidly oscillating component. Application of constraint (a) alone shows insufficient noise suppression. In this group, the methods showed variation in the solution obtained, the stochastic method appearing to give the best result. This may. be partly explained by the particular iterative strategy used along with range limiting on subsequent iterations, and illustrates that optimization techniques may contain ‘hidden’ constraints peculiar to a given method. It is observed that application of constraint (b) alone may allow a falsely high peak value. Application of constraint (c) alone yields an oversmoothed solution . with loss of the shape of the upstroke. Method (I) shows best recovery of the shape of the theoretical curve.
372
NIMMON et al.
FIG.2. Selection o f solutions using the FFT regularization method. D2 is the squared second difference o f the IRF. Solutions A, В correspond to selection (i) by a minimum in W2, and (ii) by a relaxed monotonicity constraint respectively.
Comparison of methods (G) and (I) over a range of MTT values resulted in the following regression equations: (G) Tc = 1.04 + 0.77T; r = 0.99 (I) Tc = 0.19 + 0.98T; r = 0.99 where Tc, T are the calculated and theoretical values of MTT respectively, and r is the coefficient of linear correlation. Although good correlation is shown by both methods, method (G) tends to under-estimate progressively the value o f MTT with increasing MTT. The method (I), incorporating all three constraints, was selected for further evaluation on the basis of these results. 2.2. Selection of regularization param eter
In the regularization method, a family of solutions are obtained using Eq.(l 1) in Appendix 2, each solution corresponding to a value of the multiplier j . For a given set of А/T curves, normalized to a constant maximum height, a range of the multiplier can be chosen so as to encompass a minimum in W2 as illustrated in Fig. 2. For a height of 120, the range 10s < у < 108 was found to be suitable. The solution corresponding to this minimum value of W2 (as selected by method (H)), may be oversmoothed as in the example shown. The criterion adopted in method (I) was to accept the minimally smoothed solution which satisfies a relaxed monotonicity constraint. Monotonicity violation of up to 1, 10, 25%, is tolerated
IA E A -S M -24 7/26
373
in the segments max to l/4max, l/4m ax to l/8max, 1/8 max to 3/80max, post peak respectively. Residual trailing negative values in the solution due to truncation effects are set to zero, and it is found that monotonicity can then be enforced on the selected solution without significant change in the value of MTT obtained. 2.3. Evaluation of the selected method 2.3.1. M ethod A series of 63 different IRFs were generated using the method given in Appendix 3. MTT values were in the range of 5 to 14 s. Evaluation was carried out separately with respect to errors arising from (a) bias and truncation effects consequent on using a discrete FFT; (b) Poisson distributed noise on the measured data (noise/signal ratios 0.11, 0.066, 0.047); (c) necessary preprocessing of the raw data: in particular, extrapolation of the leading edge of the input function in order to remove contributions due to pickup from activity in the superior vena cava and the pulmonary artery. Further added contribution from these sources during the recirculation phase may lead to deviation of the reconvoluted curve from the original data Proportional reduction of the magnitude of the recirculation phase of the input was used to correct for this effect. In addition to the simulation study, reproducibility has been investigated in a group of patients presenting with cerebrovascular disorders. 2.3.2. Results The calculated MTT values were linearly correlated with the theoretical values and the results are tabulated in Table I. It is seen that errors due to preprocessing of the data may be comparable to those due to statistical noise. As a first approximation, the resultant combined errors can be assumed to be given by addition. Resultant error bounds expected for values within the same patient study are shown in Fig.3. A difference in regional MTT values of 1.0 s is required for significance at the 95% confidence level. For comparison between patients possible variation in preprocessing increases this estimate to 1.4 s. In the patient study, the mean difference in MTT between two consecutive measurements is 0.3 s, S.D. of 0.98 s (2.04 s for 95% level), sample size 36. The slightly increased variation found in the patient study would be consistent with changes in physiological factors. 2.4. Discussion The results of the comparison of methods indicate that good recovery of the IRF is achieved by the introduction of a combination of constraints. In particular, the use of a relaxed monotonicity criterion to determine the optimum
374
NIMMON et al.
TABLE I. LINEAR REGRESSION COEFFICIENTS (A0, A ¡) AND STANDARD ERROR ESTIMATE (S.E.) OF MTT OBTAINED FROM EVALUATION OF THE FFT REGULARIZATION METHOD3 Type of error simulated
A0
Bias and truncation
0.48
Noise on g(t): Noise/signal ratio 0.11 0.07 0.05
rb
S.E. (s)
0.92
0.99
0.09
0.78 ■ 0.41 0.34
0.91 0.95 0.96
0.97 0.99 0.99
0.41 0.29 0.23
Noise on f(t): Noise/signal ratio 0.11 . 0.07 0.05
. 0.06 0.03 -0 .0 1
1.00 1.00 1.00
0.99 0.99 0.99
0.11 0.09 0.07
E xtrapolation of f(t) Increased slope Decreased slope
0.42 -0 .3 1
0.95 1.02
0.99 0.99
0.32 0.08
0.03 0.13 0.28
0.96 0.92 0.88
0.99 0.99 0.99
0.06 0.08 0.16
-0 .1 3
1.03
0.99
0.18
Recirculation phase correction
Preparation of g(t)
a Y = A0 + A[X, where Y ,X are the calculated and theoretical values of MTT respectively. ^ r = coefficient of linear correlation corresponding to the regression.
degree of smoothing has been demonstrated. The technique is currently being evaluated in the measurement of cerebral blood flow in cerebrovascular disease [13].
3.
RENOGRAM STUDY: CLASS A DECONVOLUTION PROBLEM
Linear model After correction of the IRF for the vascular component, the MTT is given by the area under the remaining IRF, divided by the ‘plateau’ height. The TTS may be formed by multiplying the gradient of the IRF by - 1 . In addition to the
12 T (SECS)
--->
FIG.3. Results o f error simulation for regional mean transit times obtained in the same patient study. Tc is the calculated value o f MTT (each scale division represents 1.5 sj, and T is the theoretical value. The line o f identity is also shown and the difference between two regional values required for significance is given by ET = 0.96 s.
measurement of MTT, the interpretation of the TTS in terms of a modal structure is relevant to the measurement of the distribution of intrarenal blood flow between the two nephron populations in the kidney [14]. 3.1. Methods Both determination of MTT and modal resolution of the TTS have been studied using a group of simulated renograms derived from a bimodal TTS distribution. Gaussian-shaped modes were used with standard deviations equal to 20 s, a separation of 100 s, and with the contribution from the second mode M2 equal to 0, 10,..., 100% of the total spectrum. Retention functions were constructed corresponding to extraction efficiencies of 15 and 80%, simulating DTPA and Hippuran renograms respectively. For each value of M2, thirty patterns of Poisson distributed noise were added with a noise-to-signal ratio o f 0.06. Two deconvolution techniques were used: (A) matrix algorithm, and (J) stochastic method. Using the latter method the three constraints of non-negativity, mono tonicity and smoothness were applied. The smoothness constraint was restricted to the downstroke of the IRF in order to retain the natural high frequencies in the leading edge. The function minimized was equal to (W2). (1D1 )n^3; n = 0, 1 ,.., 5, where 1D1 is the sum of the absolute values of the sequence representing the second difference of IRF.
376
NIMMON et al.
FIG.4. Modal resolution o f a theoretical bimodal TTS using matrix algorithm and stochastic methods. Also shown is the result o f applying a 1-2-1 smoothing algorithm to the matrix solution.
3.2. Results 3.2.1. Measurement o f M T T For MTT values in the range 320 s to 420 s, good linear correlation was obtained with both methods. The regression equations are (A) Tc = 26.6 + 0.90 T; r = 0.82 (J) Tc = 4.72 + 0.86 T; r = 0.99 where Tc, T are the calculated and théoretical values of MTT respectively. The standard error of the estimated values from methods (A), (J) is 5.5 and 1.3% respectively. These results indicate that good estimates of MTT can be obtained using a simple unconstrained method, and are in agreement with other published data [15, 16]. The expected standard error may be reduced by the use of a constrained method at the expense of computing time. 3.2.2. Modal résolu tion An example of the results obtained by the two methods is shown in Fig.4. Using method (A) the resulting amplitude of noise on TTS is greater than the signal. Performing 1-2-1 smoothing reduces the oscillations but was found insufficient to enable reliable modal recognition. The method (J) yields qualitative recognition of the bimodal pattern for M2 between 20 and 80%.
IAEA-SM -247/26
377
Merging of the modes consequent on applying even minimal smoothing makes quantitative separation rather subjective. Nevertheless modal estimates were obtained within 15% of the theoretical value. These results are compatible with the simulated 131I-Hippuran probe renogram studies reported by Brown and co workers [17] using a gradient search algorithm. It is concluded that whereas constrained methods may enable demonstration of possible modes, problems of separation and quantitation remain. 3.3. Modal resolution using a correlation technique 3.3.1. Theory The TTS solution, y(t), as given by the matrix algorithm, can be considered to be composed of a fraction, c, of a single modal solution x(t) and an uncorrelated noise component. The fraction с may be expressed in terms of the cross-correlation function IPxy as f18]: + T
2
y (t-r ) x(t) ФХ у ( т )
r = —T
С(т )=----------------------------------------------------------------------- = —
+т 2 * О)
фхх(0)
t= -T
where both y(t) and x(t) have been adjusted to yield a zero mean value over the data range used. The maximum value of i//xy(r) gives the maximum probable amount of the mode present and the corresponding value of тgives the modal transit time. Further, if the signal y(t) is composed of two signals Уд(1), Ув№ containing the modes A, В respectively, so that: y(t) = k y A(t) + yB(t) then the cross-correlation function (CCF) of the signal yg(t) is given by ^yBx = Ф ух~кФ у А х
05)
Hence, if the signal у contains two modes A ,В and a signal yA containing mode A alone is available, then the percentage contribution of the second mode B, M2 may be obtained from the relation: t ^ l m a x X 100% г и г -
i
№ yB* ]max + М ^ у д х ] max
(1 6 )
- (14)
378
NIMMON e ta l.
FIG.5. Resolution o f bimodal spectra using the correlation technique. Results are shown for theoretical curves with M2 = 30%. The CCFs formed by cross-correlation of the TTS resulting from the matrix algorithm with a gaussian mode with a = 50 s, are normalized by comparison o f their leading edges. Subtraction yields a value for M2 of 27%.
• »Tc"
• *>Tcm
-- 50 s
(b
(
с
S.D.
(%)
a (s)
o (sí
FIG. 6. Results o f simulation studies using the correlation technique, (aj Linear regression for calculated and theoretical values o f M2 between 10 and 90%. Each point plotted corresponds to the mean o f 30 simulations, (b) Chi-squared values formed between the regression lines obtained similar to the example shown in (a) and the line o f identity for a range o f the S.D. a o f the gaussian mode used in cross-correlation. (с) Values o f S.D. o f M2 obtained from 30 simulations for each value o f a.
IA EA-SM -247/26
379
FIG. 7. An example o f selection o f ROI corresponding to outer, middle and inner zones of the kidney.
3.3.2. Method The TTS results obtained using the matrix algorithm were cross-correlated with gaussian modes x(t) with a = 10, 2 0 ,..., 60 s. Values for M2 were calculated by normalization and subtraction of the CCF obtained with M2 = 0. The stages in the method are illustrated in Fig.5. 3.3.3. Results The results of the simulations are shown in Fig.6. It was found that the best correlation with the theoretical values is obtained by choosing a value of 50 s for a, equal to one half of the separation of the modes. The corresponding regression equation is: (Ma)c = -1 .0 0 + 0.99 M2; r = 0.99 The S.D. of (M2)c, the calculated second mode, is 6.8 and 10.2% of the total spectrum for the simulated Hippuran and DTBA renograms respectively. 3.3.4. Application o f correlation m ethod Regions of interest may be drawn as shown in Fig.7, delineating three zonal regions, outer, middle and inner respectively. The outer zone is drawn using a mean time functional image [16] and excludes any contribution from the calyceal system. The inner zone is drawn to encompass the renal pelvis. The level of separation of the outer and middle zones is indicated diagrammatically in Fig.8.
380
NIMMON et al.
CORTICAL NEPHRON
FIG.8. Diagrammatic illustration o f the level o f separation o f the outer and middle zones.
It is seen that the outer zone may be expected to contain the cortical nephron group (CN). The middle zone, in addition to including the calyces and an overlying cortical component, will also contain the loops of Henle of the juxtamedullary nephron group (JM). It may be predicted that output from the outer zone will be unimodal, the distribution due primarily to CN transit. Further, this mode should not be significantly altered in passage through the collecting ducts in the middle zone. The expected delay in the JM group will occur mainly in the middle zone leading to a combined bimodal output. Normal calyceal transport is expected to be within 10 s, which is the sampling time. Thus, using this model, TTS from the outer zone is considered to contain a single mode, and the TTS from the middle zone is considered in addition to contain a potential second mode. The correlation analysis described above can then be applied.
IAEA-SM -247/26
4.
381
RESULTS
Preliminary application of this technique to 123I-Hippuran renograms obtained from a small group of patients with no evidence of hypertension yielded a mean value for M2 of 17.7% with a S.E. of 1.1%. These results are consistent with the expected intrarenal blood flow distribution in normal subjects. In a group of 10 mildly hypertensive patients, a corresponding mean value for M2 of 28.2, 5.E. = 2.5%, has been obtained. The difference of the means of these two groups is significant with p < 0.001. In addition, comparison of values obtained for the left and right kidneys results in a good linear correlation with r = 0.89. 4.1. Discussion In comparison with techniques which seek to resolve a bimodal distribution from a single TTS, the correlation method leads to clearer quantitative resolution of modes and avoids the use of iterative algorithms. Incorporation of data obtained from the outer and middle zones enables the technique to be used even in patients with mild pelvic retention of the tracer. Problems due to retention in a single calyx may be overcome by inclusion of the calyx in the inner zone. The results obtained show that the correlation method is a promising approach in the practical quantitation of the TTS.
5.
CONCLUSIONS
In both the examples studied, solutions to the deconvolution problem have been achieved by the addition of a priori information. Although this is of a general and non-parametric type, the validity of application may depend crucially on the validity of the underlying model. The constraints applied will vary with the system studied and there is probably no optimal deconvolution technique applicable to the solution of every problem. The use of simulation studies plays an important role both in comparing the performance of available methods and in the estimation of the expected error in the chosen method. Knowledge of the expected error and its dependence on the quality of a particular set of data are vital requirements for the investigation to have predictive diagnostic value. Such error estimates provide a strong base on top of which clinical evaluation may proceed to determine the contextural usefulness of the technique. Following this strategy, we believe that deconvolution techniques offer the facility of measuring physiologically significant indices in dynamic studies.
382
NIMMON et al.
Appendix 1 (a) MATHEMATICAL FORMULATION OF THE LEAST-SQUARES METHODS The convolution equation may be written as i g ' ( i ) = 2 f(i—j)h(j) + e(i) j= 0
(1)
where the sequences g’(i); f(i); i = 0, 1, ... , K -l, are formed by the sampled values of the organ and input А/T curves at times iT respectively, T being the sampling time interval. The sequence e(i) is the noise contribution to the measured sequence g'(i), and h(j), j = 0, 1,... , N -l, represents the IRF. In the following formulation g'(i) is replaced by g(i), where g(i) = g'(i)—e(i). The exact sequence e(i) is unknown and only non-unique estimates for h(j) can be determined. Using matrix notation, E q.(l) can be rewritten in the form g” = F Í r
(la)
where«- represents a column vector (and -*■ represents a row vector). The K X1 vector g, and the NX 1 vectorÍT have elements g(i), h(j) respectively, and F is a KXN matrix with elements FjjXFjj = f(j—1), for j > i; and Fj¿ = 0, for j < i. For К > N, Eq.(la) represents an overdetermined set of linear equations. The LS solution minimizes the square of the Euclidean norm of the expected residual vector W : W = E (g - F h), and is given by: h = F" g
(2)
where F" = (FTF)-1 FT
(3)
is the generalized inverse of the matrix F. (The superscript —1 denotes the inverse and T denotes transposition.) In general, LS methods either seek an approximation to the matrix F_ which is frequently 4—qp^ ill-conditioned, or concentrate on numerical minimization of the norm W1W. As an example o f the first approach, calculation of F may be simplified by the expansion of h, f in terms of the orthonormal Laguerre poly nomials [6]. In the second approach_the use of direct search techniques avoids the calculation of derivatives of WTW.
IAEA-SM-247/26
383
The matrix algorithm sidesteps the calculation of F" by conveniently choosing К equal to N. For this special case, the solution to Eq.(la) may then be calculated using the recurrence relation j- 1
h(j) = (gCi) - 2
Fjkh(k))/Ffi,j = 0, 1,... , K"1
(4)
k=0
This method is particularly sensitive to errors in the first non-zero value of f(i), as illustrated in the example shown in Fig.l.
(b) INCORPORATION OF CONSTRAINTS USING A REGULARIZATION METHOD If the constraint can be represented by the inequality (Ch)T(Ch) < A,
(5)
wh^re С is a matrix operator, then following a Lagrange method [ 1] a minimum of WTW subject to condition (5) may beexpected to lead to a minimum of the function L = WTW + 7(Ch)T(Ch)
(6)
where у is an undetermined multiplier. This equation can also be applied to minimizationof(Ch)T(Ch) subject to the constraint WTW < B. Equation (6) leads to the family of solutions ÍT= (FTF + 7CTC)-1 FT g"
(7)
A priori information about the form of h is then required to select the preferred solution. In this study, the method is used to implement a second difference smoothness constraint and two criteria for selection of у are used: (1) a minimum for у > 0. (2) M r a minimum for у = M, the least value of у for which h satisfies a relaxed monotonicity constraint.
384
NIMMON et al.
Appendix 2 FORMULATION IN THE FREQUENCY DOMAIN
The sequence x(i) may be expressed in the frequency domain using the discrete Fourier transform Q -l
X (m )= 2 x(i)exp(-27rjim/Q); m,i = 0, 1,... , Q - l i= 0
(8)
where j = y / —l. The frequency corresponding to m is m/QT where T is the sampling interval. In order to reduce effects of the abrupt termination of the data sequence, a cosine window term is added to x(i). Further, to allow for the aperiodic nature of the sequence, zero valued terms are added and the value of Q is chosen so that Q > 2K + N —1, and Q = 2n with n an integer. The latter condition facilitates the use of the fast Fourier transform algorithm (FFT) [19]. The convolution equation is: G(m) = H(m) F(m) + E(m) and H(m) = (G(m) - E(m))/F(m)
(9)
where H(m), G(m), F(m), E(m) are the Q-term Fourier transforms of the sequences h(i), g'(i), F(i) and e(i) respectively. For F(m) < E(m), noise amplification results and possible instability of the solution will occur. A measure of the stability is given by the dispersion, S2 [2]
•sb=— 2 ----------------4 Q2 m = 0 F*(m )F(m )
do)
where the noise sequence has been approximated by a normal distribution with variance a2. (The superscript * denotes the complex conjugate.) The behaviour of the dispersion for an increasing number of terms Q in the sampled sequence (and hence for increasing frequency) is dependent on the nature of the input function. Deconvolution problems may be divided into two classes, A, B, depending on whether the dispersion converges or diverges as Q is increased. The class A may be expected to be relatively stable. (1)
o2 For the input function f(t) = exp(—kt), Sq = — 2 V Q2
(2)
For the input function f(t) = t2 exp(-kt),
о У
q
2
(m2 + k2) and converges.
2 (m2 + k2)2 and diverges.
385
IA E A -S M -247/26
(1) and (2) are important examples o f input functions encountered in dynamic studies.
INCORPORATION OF CONSTRAINTS Provided that the constraint on h can be expressed as a convolution of a sequence c(i), the regularization method described above can be applied in the frequency domain. The solution corresponding to Eq.(7) is F*(m) G(m) (F*(m)F(m) f
7
C*(m)C(m))
where C(m) is the Q-term transform of the constraint sequence c(i). The sequence c(i) = 1, - 2 , 1, for i = 1, 2, 3; and c(i) = 0, for i > 3, convoluted with h, yields the second difference and is the constraint sequence used in this study.
Appendix 3
THEORETICAL IRFs Two groups of theoretical IRFs were used for simulation in the cerebral blood flow study. (1) Gamma functions given by the formula h(t) = Ck(t —10)“ exp(—(t —t0)/j3)
(12)
where a set of the parameters C^, a, j3, were chosen to give values of MTT in the range 5 to 14 s. (2) The expected IRF h(t) was generated from the equation h(t) = hj(t) * h2(t) where * denotes convolution.
(
(13)
386
NIMMON et al.
The function h ^ t) represents the spread of the spike aortic input on arrival at the base of the brain, and was chosen from group (1). The function h2(t) is given by
OO h2 (t) = A =
hj(t) dt, for t < t2 0
where h3(t) is a member of group (1).
REFERENCES [ 1] TURCHIN, V.F., KOZLOV, V.P., MALKEVICH, M.S., The use o f mathematical-statistics methods in the solution of incorrectly posed problems, Sov. Phys.-Usp. 13 (1971) 681. [2] HUNT, B.R., Biased estimation for nonparam etric identification of linear systems, Math. Biosci. 1 0 (1 9 7 1 )2 1 5 . [3] GAMEL, J., ROUSSEAU, W.F., KATHOLI, C.R., MESEL, E „ Pitfalls in digital com putation o f the impulse response of vascular beds from indicator-dilution curves, Circ. Res. 32 (1973) 516. [4] BRITTON, K.E., NIMMON, C.C., JARRITT, P.H., GRANOWSKA, M., LEE, T.L., McALISTER, J.M., “Cerebrovascular disorder: assessment w ith radionuclides”, Advanced Medicine 13 (197 7 )4 4 4 . [5] GILL, P.E., MURRAY, W. (Eds), Numerical Methods for Constrained Optim ization, Academic Press, London (1974). [6] LEE, T.Y., Cerebral blood flow by radioisotope technique: experimental and clinical aspects, PhD Thesis, University of London, 1980. [7] TURCHIN, V.F., TUROVTSEVA, L.S., Restoration of optical spectra and other non negative functions by the statistic regularization m ethod, Opt. Spektrosk. 36 (1974) 280. [8] NELDER, J.A., MEAD, R., A simplex m ethod for function minimization, Comput. J. 7 (1965) 308. [9] BRITTON, K.E., NIMMON, C.C., LEE, T.Y., JARRITT, P.H., GRANOWSKA, M„ GREENING, A., McALISTER, J.M., “Carotid and cerebral blood flow”, Information Processing in Medical Imaging, Oak Ridge National Laboratory Press (1978) 499. [10] PHILLIPS, B.L., A technique for the numerical solution of certain integral equations o f the first kind, J. Assoc. Comput. Mach. 9 (1962) 84. [11] TWOMEY, S., The application o f numerical filtering to the solution o f integral equations encountered in direct sensing measurements, J. Franklin Inst. 279 (1965) 95. [12] HUNT, B.R., The inverse problem of radiography, Math. Biosci. 8 (1970) 161.
IAE A-SM -247/26
387
[13] BRITTON, K.E., NIMMON, C.C., GRANOWSKA, M„ “Total and regional cerebral blood flow: A new quantitative non-invasive m ethod for cerebrovascular disease”, these Proceedings, IAEA-SM-247/24. [14] BRITTON, K.E., BERNADI, M„ WILKINSON, S.P., BROWN, N .J.G., PEARCE, P.C., JENNER, R., “Validation of a noninvasive m ethod for estimating intrarenal plasma flow distribution”, Radionuclides in Nephrology, Georg Thieme Press, Stuttgart (1980) 204. [15] LAWSON, R.S., A m ulticentre comparison of techniques for deconvolution of the renogram, 7th annual meeting, 1979, British Nuclear Medicine Society. [16] BRITTON, K.E., NIMMON, C.C., WHITFIELD, H.N., KELSEY FRY, I., HENDRY, W.F., WICKHAM, J.E.A., “The evaluation of obstructive nephropathy by means of parenchymal retention functions”, Radionuclides in Nephrology, Georg Thieme Press, Stuttgart (1980) 164. [17] BROWN, N.J.G., BRITTON, K.E., GOW, N.M., ELL, P.J., The precision and reproducibility of the frequency distribution of renal transit times measured with 131I-Hippuran, Nucl.-Med. 16 Suppl. (1979)507 . [18] LANGE, F.H., Correlation Techniques, Iliffe Books, London (1967) 37. [19] BRIGHAM, E.O., The Fast Fourier Transform, Prentice-Hall, New York (1974).
DISCUSSION M.C. KEMP: May I ask you whether you have considered the maximum entropy method, which automatically gives positivity and optimum smoothing? What do you use as the convolving function in view of the fact that the input curve is itself experimentally determined and noisy? Lastly, do you consider that background subtraction problems are sufficiently understood to make deconvolution techniques useful in practice? C.C. NIMMON: For the cerebral flow simulation studies, the theoretical impulse retention functions were chosen from a set of gamma functions. This form has been suggested as an approximation to experimentally observed curves following a bolus injection into the internal carotid artery. The preprocessing of the data including background subtraction leads to errors comparable with those arising from the statistical noise. We are most interested in considering the application of the maximum entropy method to one-dimensional deconvolution. In general, we feel that addition of justifiable a priori knowledge, which of course may be specific to the particular system studied, will lead to improved recovery of the retention function and, in our example of the cerebral flow study, monotonicity is an important criterion. W. MÜLLER-SCHAUENBURG: There are two aspects to the transport of tracer through the kidney: (1) continuous transport which suits the deconvolution method; and (2) the discontinuous movements of the renal pelvis (dyskinesia) which conflict with the general principle of deconvolution. Do you have any other ideas on how to handle the pelvic movements, apart from cutting off the
388
N IM M O N et al.
curve before the pelvic movements (if you evaluate the curve originating from the inner zone of the kidney)? C.C. NIMMON: In the normal kidney the frequency of peristalsis is of the same order as the sampling interval (10 to 20 s) and so does not significantly affect the deconvolution. We have observed renogram studies with transient tracer retention in the renal pelvis. In these patients use of the activity time cuves from the outer and middle zones only has been found to remove the discontinuous effects seen in the total renogram curve.
IA EA-SM -247/11
EXPERIENCE WITH A MOBILE DATA STORAGE DEVICE FOR TRANSFER OF STUDIES FROM THE CRITICAL CARE UNIT TO A CENTRAL NUCLEAR MEDICINE COMPUTER T.D. CRADDUCK, A.A. DRIEDGER Victoria Hospital Corporation and the University of Western Ontario, London, Ontario, Canada
Abstract EXPERIENCE WITH A MOBILE DATA STORAGE DEVICE FOR TRANSFER OF STUDIES FROM THE CRITICAL CARE UNIT TO A CENTRAL NUCLEAR MEDICINE COMPUTER. The introduction of mobile scintillation cameras has enabled the more immediate provision of nuclear medicine services in areas rem ote from the central nuclear medicine laboratory. Since a large num ber of such studies involve the use of a com puter for data analysis, the concurrent problem of how to transm it those data to the com puter becomes critical. A device is described using hard magnetic discs as the recording media and which can be wheeled from the p atien t’s bedside to the central com puter for playback. Some initial design problems, primarily associated with the critical timing which is necessary for the collection of gated studies, were overcome and the unit has been in service for the past two years. The major limitations are the relatively small capacity of the discs and the fact that the data are recorded in list m ode. These constraints result in studies having poor statistical validity. The slow turn-around time, which results from the necessity to transport the system to the departm ent and replay the study into the computer before analysis can begin, is also of particular concern. The use o f this unit has clearly dem onstrated the very im portant role that nuclear medicine can play in the care of the critically ill patient. The introduction of a complete acquisition and analysis unit is planned so th at prom pt diagnostic decisions can be made available within the intensive care unit.
1.
INTRODUCTION
'
A n u m b e r of nucl e a r m e d i c i n e procedures have b e e n developed w h i c h offer considerable p o tential in the diagnosis and con tinuing care of patients located in the critical care, intensive care and coronary care units. One of the more obvious applications has b e e n the use of pyrophosphate scans to confirm o r deny the presence of m y o c a rdial infarct upon admiss i o n to the coronary care unit. Equally important is the use of lung scanning to detect the presence of pulmonary
389
390
CRADDUCK and DRIEDGER
embolus and b r a i n scanning in cases of suspected subdural hematoma. The advent of gated cardiac studies has led to a demand for these studies in patients suffering from cardiac distress. W h e n patients are subject to intensive care there may be s everal monitors, intravenous lines and a ventilator present, as w e l l as various attendant personnel. To move such a pati e n t f r o m the intensive care environment to a normal n u c l e a r m e d i c i n e department becomes a m a jor operation invol v i n g logistics problems and, not least of all, a m a j o r risk to the patient. It was in response to this p r o b l e m that mobile scintil lation cameras w e r e d e veloped and it is in such areas that they have had their greatest application. Such devices can, however, only fulfill the need for analog images and unless some auxiliary computing power is a v ailable it is not possible to o b t a i n the various determinants of cardiac function which first pass and gated e q u i l i b r i u m studies can otherwise provide. One solution w h i c h has b e e n offered to this p r o blem is to b u i l d a computer s y s t e m into the m o b ile unit. This has the adv a n t a g e that the camera/computer becomes one unit thereby facilitating movement from one location to another. It also implies that results can b e made available at the bedside w h i c h can be very important w h e n rapid answers are required in the case of the very sick patient. There are also some inherent d isadvantages with such a system. The first is that of expense w h e n a computer is al ready available in the nuclear m e d i c ine department itself. A n o t h e r is that if the m o b i l e s y s t e m differs from the 'fixed' system, it is difficult to ensure that the results from each are comparable, and to archive studies appropriately so that later review and comparison are possible. Indeed, communi cation b e t w e e n the two systems m a y e ven be difficult w h e n they are of the same type. Y e t another difficulty may lie in the performance of the s c i n t i llation camera itself w h i c h then w o u l d tend to preclude choice of the computer w i t h w h ich it is combined. 2.
T R A N S F E R OF D A T A F R O M REMOTE LOCATIONS
2.1. Cable Dowsett and Roberts [1] and Elliott et al.[2] have both d e scribed systems using coaxial cables and involving the t r ansmission of the analog signals f rom the scintillation camera. S a nton et al. [3] described a sy s t e m in w h i c h the signals w e r e digitized and then transmitted v i a twin-paired lines.
IA EA -SM -247/11
391
The m a j o r limita t i o n associated w i t h such systems is the lack of m o b i l i t y w h i c h is inherent. Indeed, the systems des cribed by Dowsett and Roberts and b y Santon et al.both involved 'fixed1 (i.e. non-mobile) s cintillation cameras. The system d escribed b y Elliott et al. was designed for a mobile unit, but it could only b e used in those locations to w h ich cable connections had b e e n made. A similar system described by Feiglin et al. [4] is restricted to acquiring data from only two locations remote from the nuc l e a r medicine department. A second m a j o r constraint of the fixed wire system is that control of the computer must be transferred to the remote location and this can severely restrict or interrupt the smooth flow of w o r k w i t h i n the d epartment itself. In the case of the last s y s t e m ment i o n e d above, it is n e c e s s a r y to locate techno logists at both ends of the s ystem and use a telephone link in order to acquire data. 2.2.Ma g n e t i c M e d i a The alternative to direct transm i s s i on of signals is to record them in some fashion for subsequent transfer to the computer site. Two forms of m a g netic m e d i a are available - magnetic tape and ma g netic disks (both floppy and h a r d ) . At least one commercial m obile camera was offered w i t h a magnet i c tape drive. This drive was a d vocated to be 'computer c o m p a t i b l e ' , but this term only implies that the drive is similar to that used on computers (9-track) and records the data digitally. Those data may be 'readable' by another m a g netic tape drive on a computer but the format w i l l need to be interpreted if they are to b e understood by that c o m p u t e r ’ s software [5]. In addition to the problems of direct compatibility is the fact that m a g n e t i c tape is inherently slow and a significant loss of data occurs w i t h such a system. Several disk systems h a v e b e e n developed. One utilizing floppy disks required that the central computer be from the same m a n u f a c t u r e r and that a floppy disk drive b e added to it in order to read the disks. Apart from the extra expense involved, the use of floppy disks tended to limit the capa bilities of the data acqu i s i t i o n m odule both in terms of speed and overall count capacity. We chose to use a s y s t e m i n c o r p orating hard disks (one fixed, one removable) beca u s e of the greater speed and capacity of such a system. This disk drive unit may be docked w i t h the scinti l l a t i o n camera to form a single unit. It may also be transported separately. In ord e r to interface w i t h the com puter, a separate input channel was required that w o u l d accept
392
CRADDUCK and DRIEDGER
FIXED COUNT RATE(50k cps) CLOCKED AT 10 ms
111
111
111
lili
OUTPUT
CLOCK
/
FIXED FREQUENCY
10 ms
FIG. 1. The data from the scintillation camera arrive at the disk recording device randomly in time and are recorded in list (serial) mode with 10 ms timing marks interspersed. The out put was originally clocked at 10 ms intervals with the counts played back at a rate o f 50 kHz each 10 ms.
the digital signals from the disk drive and bypass the analogto-digital converters, y et give the appearance that the disk sy s t e m was a s cintillation camera. Using this approach, it was not necessary to w r i t e auxiliary software to accommodate the n ew device.
3.
D I S K DRIVE D A T A ACQ U I S I T I O N
The disk s y s t e m 1 embodies two disks on the same spindle. Each has a capacity of 5 megabytes or 2.5 m i l l i o n counts for a total of 5 mil l i o n counts. One disk is removable, the other is fixed and w h e n o v e r f l o w occurs from the removable disk, the extra counts are recorded on the fixed disk. Due to the loca tions in w h i c h our s ystem was to be used and the difficulties of transportation, the capacity was effectively limited to that of the removable disk, several of w h i c h could b e recorded upon for different views or different patients before transfer to the nuc l e a r m e d icine department a nd subsequent playback. The data are recorded in list mode and the full EKG signal is also recorded together w i t h timing marks (FIG.l). Thus, one m ay record static studies, dynamic studies (up to 50 к counts p er second) and studies containing the EKG so that subsequent gate s y n chronization could be accomplished.
1 Scintistore by Searle Radiographies.
393
IAEA-SM -247/11 ORIGINAL
1 0 0 -,........................................................... ' « , • • • •• 80 *2 § 3 60Ф >
...
. .
.
.
.
•
.
.
ф 40 ■
tt
20 0 ^.......'T 60
1I
I
I
11I
I
120
180
240
300
360
I.......4 420
480
Time (ms)
FIG.2. Plot o f counts versus time for a constant count rate input and a simulated EKG synchronizing the computer acquisition each 480 ms. Note the beating effect which took place at approximately 50 ms intervals.
The data could be played b a c k into the computer either in real time or at a compressed rate of 50 к counts per second. The latter is only useful for static images since the computer needs to u se its own clock for t iming dynamic and EKG syn chronized studies. One p r o b l e m is therefore immediately evident - any one study takes at least twice as long to per form w h e n it m u s t first b e recorded and then played b a c k in real time. 4.
SOME D E S I G N P R O BLEMS
D uring initial acceptance testing a source of fixed counts was pl a c e d i n front of the scin t i l l a t i o n camera and a simu lated E K G of u n i f o r m rate (75 p er minute) was recorded w i t h the data f r o m the camera. These data w e r e subse q u e n t l y p layed b a c k into the computer as a gate synchr o n i z e d acquisition. The result should have b e e n a constant n u m b e r of counts p e r frame throughout the R- R interval. This w as not the case (FIG.2). Instead, a curve demons t r a t i n g repetitive 'beating' resulted. Every fourth or fifth frame contained some 50% or less than the expected value It b e c a m e evident that this p h e n o m e n o n was a result of the me t h o d used to replay the data. Two buffers are used. Each 10 ms timing m a r k e r o n the disk caused a change of bu f f e r and the data f r o m the previous 10 ms w e r e then output at a rate of 50 к counts per second. Thus data w e r e output in bursts each 10 ms at a rate o f 50 kHz.
394
CRADDUCK and DRIEDGER
COMPUTER INPUT CLOCKED AT 6 ms
INTERVALS
FIG.3. Schematic timing diagram indicating cause o f beating effect shown in Fig.2. Using a frame rate o f 6 ms as an example, the counts which were recorded in 10 ms intervals became redistributed in time. Note that a 6 ms framing rate is chosen only to simplify the diagram. A similar situation would obtain for longer frame times.
The computer, on the other hand, clocks the input at 1 ms intervals and in this partic u l a r instance, framed the data at 17 ms per frame. Under these circumstances it is possible for some frames to collect m o r e and some to collect less than the proper number of counts for that interval of time. Figure 3 illustrates the situation w h i c h could exist for 6 ms time frames and demonstrates that a redistribution of counts with respect to time can occur. An interval of 6 ms was chosen to simplify the illust r a t i o n but the principle is equally applicable at other frame rates. It should also be not e d that no synchronization of clocks on the two devices (disk s ystem and computer) is made and 1 ms on one device is not n e c e ssarily 1 ms on the other. The first attempt by the design engineers to solve this pr o b l e m involved two m i n o r modifications. The clock rate on the disk s ystem was i n creased to 1 kHz and the output rate was decreased from 50 kHz to 20 kHz. Thus, data were no w output in 20 к count per second bursts once every millisecond. Some improvement was evident, but the count versus time curve for the same test situation still demonstrated a be a t i n g effect ( F I G . A ) . Desp i t e the digital input port b y -pas s i n g the analog-todigital converters, the disk s y stem m a y have 'looked' like a s c i n t i l l a t i o n camera f r o m the point of v i e w of the computer,
395
IA E A -S M -2 4 7 / 1 1
INTERIM SOLUTION
lOO-i . • • •
*
•• •
' »^
. ••
•••
^
^*
•
80
С
*
J
60-
и
40-
«> Ю
a.
20
0
----------1----------1---------- 1----------1---------- 1---------- 1----------1----------1 60 120 180 240 300 360 420 480 Time (ms)
FIG.4. Plot o f counts versus time for a constant count rate input and a simulated EKG after the interim modification o f increasing the clock rate from 10 ms to I ms and decreasing output rate from 50 kHz to 20 kHz. The beating effect evident in Fig.2 is still present but is less pronounced.
but it failed to 'behave' like one. The counts w e r e being transmitted to the computer in chronological order but failed to b e a r the correct time relationship to one another. The actual rate of o u tput must simulate as closely as possible the rate of input w i t h as little as 1 ms time resolution if gate synchronized studies are to be faithfully reproduced. To achieve this o bjective a second m od i f i c a t i o n was engineered by the manufacturer. Figure 5 illustrates in schematic f o r m the m a n n e r of operation. A variable frequency oscillator is used to control the rate at w h i c h counts are output, so that those events w h i c h occurred during a 1 ms time interval are spread e venly over each 1 ms instead of arriving at the computer in 20 kHz bursts. The variable frequency oscill a t o r ensures that the data are clocked out at 1 ms intervals at a rate p r oportional to the nu m b e r of counts input in each 1 ms interval. The resulting curve of counts versus time for the fixed source test situation is n o w con stant w i t h i n s t a t istical limits (FIG.6) and the disk system was n o w assumed to mim i c the output of the scintillation camera faithfully. 5.
SOME O P E R A T I O N A L PROBLEMS
Reference has already b e e n m a d e to the limited disk capa city of this system. Since it is nece s s ary to transport the
396
CRADDUCK and DRIEDGER
SWITCH
INPUT CLOCKED AT 6 me INTERVALS
FIG.5. Schematic diagram o f final solution in which the counts collected in each 1 ms interval control the rate at which those counts are output to the computer. Thus, to within a time resolution o f 1 ms, the counts are played back with the same time relationship as they are acquired.
FINAL SOLUTION
lOO-i.
80
60
20
60
120
180
240
300
360
420
480
Time (ms)
FIG.6. Plot o f counts versus time for a constant count rate input and a simulated EKG after the final modification (Fig.5) had been made.
w h o l e unit to the computer, the second non-removable disk b e comes v i rtually useless. It is p o s sible to copy the data from the fixed disk to the removable disk and back again so that a study w h i c h 'overflows' onto the fixed disk can be completely recovered. This procedure is, however, time con suming and finds little favour in a busy department. The fact that disk cartridges can take up to five minutes to reach full speed is also not conducive to such transfers.
IAEA-SM-247/11
397
The limited disk capacity implies that gate synchronized studies lack sufficient counts to b e statistically sound. This p r o b l e m is further aggravated w h e n patients demonstrating arrhythmias are studied. It is o ur policy to reject those data following R - R intervals that are m o r e than ±10% of the m ean interval and, further, to reject a study w h ere more than 20% of the R - R intervals are so rejected. W h e n the data are already sparse this type of rejection can cause a study to be uninterpretable. W e have usually resorted to transferring the data as a gated list study and then selecting those R-R intervals w h i c h w e w i s h to include in the study. This p r o c e dure usually provides somewhat more reliable results, but is, again, time consuming and p e rformed only w h e n absolutely necessary. The n umber of disks can b e c o m e b u r d e n some and on busy occasions yet another cart for transport of the disk cart ridges has been brought into action. The labour involved in moving all these units is n o t inconsequential. Another p r o b l e m w h i c h occas i o n a l l y manifests itself relates to the EKG. W h e n a patient is connected to a ventilator, it appears that some, as yet unexplained, ground loop current is generated and this causes a r epetitive shift of the E K G b a s e line. This b a s eline shift is rarely demonstrated on the normal E K G monitor, only on that one associated w i t h the s cintillation camera and disk system. W h e n replayed for input into the computer, the shifting b a s e l i n e causes the gate signal to be corrupted and erroneous R - R intervals are detected by the computer. 6.
CLINICAL USEFULNESS
The m obile scin t i l l a t i o n camera and disk recording system w e r e first put into oper a t i o n on a clinical basis in September 1978. D uring the twelve months ending M a rch 31,1980, a total of 551 studies or an average of a little more than two studies per day w e r e performed u s ing this system. The d i s tribution of these studies is shown in Figure 7. Lung scans include b o t h per f u s i o n and v e n t i l a t i o n studies, b r a i n scans include CSF studies and liver scans include scans of the gall bladder. It should b e emphas i z e d that the dynamic heart scans as listed represent one injection of radioactive tracer agent and several studies m a y be involved. The cardiac output can be derived from the first pass data and the ejection fraction and wall m o t i o n may also b e obtai n e d repetitively w i t h therapeutic interventions taking place. Thus the numbers given in Figure 7 represent less than half the total n u mber of dynamic heart studies performed.
398
CRADDUCK and DRIEDGER (A uS
12 Ш t< < Л û 5. z X¿ 250 I» 220 -ОЙ I 200 150 ■
149
STUDIES PERFORMED USING MOBILE SCINTILLATION CAMERA APRIL 1 1979 - MARCH 31 1980
§ э H <л < 3 z =J 125
100 50
35
C E Ш >
z о
m
10 FIG. 7. Spectrum o f nuclear m edicine procedures perform ed using m obile scintillation camera over a 12-m onth period. The figures represent adm inistrations o f radioactive tracer agents and, as a consequence, the figure fo r dynam ic heart scans represents less than h a lf the actual num ber o f studies p erform ed because a t least tw o procedures, and frequently m ore, fo llo w each injection.
One other interesting fact w h i c h became evident w h e n these numbers w e r e reviewed was that p y r o p hosphate heart scans were the most popular studies during the first six-month period. They w e r e quickly overtaken by the dynamic heart studies as the v a l idity of the latter studies was established. The extraordinarily low n u m b e r of thallium heart scans is a r eflection of the p a t t e r n of practice, since the mobile s y s t e m is used for investigations involving acutely ill patients. Out-patients or in-patients w i t h chronic heart conditions are imaged in the nuclear medicine department on standard equipment. This p a t t e r n of practice is, w e believe, contrary to that w h i c h scin t i l l a t i o n camera manufacturers must have assumed w o u l d prevail w h e n the decision was taken to re duce crystal thickness to one- q u a r t e r inch in order to improve statial resolution. The efficiency for 201^1 does not change a great deal but the efficiency for 99mTc is reduced by about 17%, w h i c h in our application, w h e r e we are already photon limited, becomes a significant factor. 7.
CONCLUSION
The use of a m a g netic disk r e cording system for transfer of data from a m o b i l e s c i n t i llation camera to a central nuclear m e d icine computer has enabled the practice of nuclear medicine to be extended to include the critically ill.
IAEA-SM-247/11
399
There w e r e some initial d esign problems associated w i t h the short time resolution w h i c h gated synchronous acquisition demands. These w e r e overcome b y design changes and the system fulfilled its specifications. The delays engend e r e d by transfer of the system to the nuclear medicine department and transfer of data into the computer are considerable, and the extra w o r k l o a d causes scheduling problems at the computer. Our present plans are to purchase a complete computer a c q uisition and display system w h i c h w i l l be m o u n t e d on w heels a nd travel to the bedside. The p hysician w i l l then h a v e the facility for immediate analysis and reporting so that patients' problems can be dealt with e xpedi tiously and therapeutic measures commenced sooner.
REFERENCES [1] [2] [3] [4]
[5]
DOWSETT, D.J., ROBERTS, K., Long-distance transmission of analog gamma camera signals, J. Nucl. Med. 15 10 (1974) 896. ELLIOTT, A.T., BROWN, N.J.G., BURNS, D., ELL, P.J., Operating a nuclear medicine computer system with a remote mobile gamma camera, Phys. Med. Biol. 23 5 (1978) 981. SANTON, L.W., PRATO, F.S., ASPIN, N., Long-distance transmission of digital scintillation signals, J. Nucl. Med. 17 5 (1976) 394. FEIGLIN, D., SPIERS, W.E., COCKS, N., Private communication and presentation at Spring meeting, 1980, of Eastern Great Lakes Chapter of the Society of Nuclear Medicine, giving details of cabling from two remote locations to central computer. KOWALSKY, W.P., HOLODNY, E.I., VAN HEERTUM, R.L., SCHWARTZ, G., Low data loss acquisition from a gamma camera with subsequent computer analysis, J. Nucl. Med. 19 11 (1978) 1256.
DISCUSSION A.E. TODD-POKROPEK: A lthough we published a description o f analogue data transmission, as listed in R ef.[l ] o f y o u r paper, we have abandoned this technique at University College H ospital and find digital transm ission m uch m ore reliable and, in fact, cheaper. Could y ou tell us why you do n o t directly process the list m ode data on the disc w ith y o u r central m achine rath er than, presum ably, waste rather a lot o f tim e by playing the d ata back through the central gamma camera interface? It would also solve y o u r tim ing problem s. T.D. CRADDUCK: T hank you for y o u r com m ent. I am n o t surprised th a t you have found digital transm ission m ore reliable. We were unable to process the list m ode data directly because, although the figures shown m ay give the im pression th a t the disc packs used in the m obile
400
CRADDUCK and DRIEDGER
system are com patible w ith the central com puter, th a t is n o t in fact the case. The discs are sectored differently, are run in quite different drives and are in a different form at. Y ou are correct in assuming th a t reading the list m ode data directly would have elim inated the timing problem s and w ould have avoided replay of the data in real tim e. It w ould, however, have involved considerable software developm ent and would not have avoided entirely the ‘tu rn around’ tim e needed to feed results back to the critical care unit. E.F. CROCKER: We have ru n analogue data over a 200-m route from intensive and coronary care units for the past tw o years. The system has been entirely satisfactory. In view o f the problem s th a t you have encountered and w ith the benefit o f hindsight, would y ou still use the same m ethod? T.D. CRADDUCK: Yes. In 1977 we would certainly have preferred to acquire a m obile com puter system capable o f b o th acquisition and analysis and also com patible w ith our central system . However, such a system was not available. In retrospect we m ight have been wiser to configure a mobile system ourselves, b u t I suspect th a t in doing so we would have run the risk o f voiding th e w arranty on the disc drives an d /o r losing field service coverage because the drives w ould have been subject to rugged m ovem ent. O ur experience has certainly been valuable and the use o f the disc storage system has enabled us to dem onstrate the value o f nuclear m edicine procedures in the critical care unit, thereby providing ju stification fo r the m obile com puter now on order. The overriding factor in electing to use a m obile system rath er than installing fixed cables was th a t the latter restrict the locations in which one can use the m obile scintillation camera, and we w anted to retain the flexibility th a t a m obile data acquisition system would provide.
IAEA-SM-247/86
THE FUTURE OF FUNCTIONAL IMAGING
Success or failure? W.J. MACINTYRE, S.A. COOK, R.T. GO Cleveland Clinic F oundation, Cleveland, Ohio, United States o f America
Abstract THE FUTURE OF FUNCTIONAL IMAGING: SUCCESS OR FAILURE? A brief survey of functional or parametric imaging in radionuclide studies has shown numerous investigations since 1966 applied to both blood flow and function of the heart, lungs, kidneys and brain. Although originally hampered by limitations in data acquisition, data processing and display, later advances have overcome most of these problems and made this methodology now available to most laboratories. In spite of this, no application of functional imaging has been accepted as a routine clinical procedure at present. This is not to say that individual reports have not found functional imaging superior to conventional techniques. Such results have already been reported for functional lung imaging and regional ventricular ejection fractions. However, neither of these procedures has become the method of choice for the field as a whole. It is felt that the unique information provided by functional imaging should be re-examined so as to determine what parameters are of such critical diagnostic value to provide sufficient additional information to the clinician, so as to justify the time and effort expended to acquire and process this information.
1.
INTRODUCTION
F u n c t i o n a l o r p a r a m e t r i c imaging o f an org an r e f e r s t o t h e reg io n al d i s p l a y o f a parameter t h a t i s r e l a t e d to the physio l o g i c f u n c t i o n o f an o r g a n , r a t h e r t h a n a d i s p l a y o f an an at omi c c o n f i g u r a t i o n v i s u a l i z e d by such mechanisms as x - r a y a t t e n u a tio n , ultrasound re fra c tio n or radionuclide deposition. In most c a s e s , t h e p a r a m e t e r s e l e c t e d i s t ime d e p e n d e n t and i n v o l v e s t h e r e c o r d i n g o f s e q u e n t i a l images f o r i t s measurement . The f u n c t i o n a l image was i n i t i a l l y p ropo se d i n o r d e r t o combine t h i s e n t i r e t i m e s e r i e s i n t o a s i n g l e d i s p l a y and a l s o t o i d e n t i f y r e g i o n a l v a r i a t i o n s t h a t a r e n o t r e a d i l y a p p a r e n t by v i s u a l co mpar ison o f s e q u e n t i a l i mages. Th i s t e c h n i q u e was i m me d ia t e l y a d a p t a b l e t o r a d i o n u c l i d e methodol ogy s i n c e f a s t s e q u e n t i a l r e c o r d i n g from s c i n t i l l a t i o n cameras c o u l d be r e a d i l y a c h i e v e d and d i g i t a l r e c o r d i n g o f r a d i o n u c l i d e imaging has been a v a i l a b l e sin ce th e m id-1960's. 401
402
MACINTYRE et al.
F IG .l. Upper: Three-dimensional displays o f &6R b accum ulation in the dog as constructed by L ove and co-workers. D ecreased areas are shown follow in g ligation o f the anterior coronary artery. Lower: Early isom etric com pu ter displays o f the data above. Circled area shows the region o f the left ventricle affected by ligation.
F u n c t i o n a l imaging o f t h e myocardium was r e p o r t e d as e a r l y as 1966 by Love, Smi th, M i t c h e l l and P u l l e y [ 1 ,2J who measured Д?
fifi
th e accumulation of К or Rb i n t h e myocardium w h i l e main tain in g a sta b le a r t e r i a l concentration of the radionuclide by c o n t i n u o u s i n f u s i o n . The i n c r e a s e i n c o u n t i n g r a t e w i t h t i me was r e g a r d e d a s l i n e a r d u r i n g t h e e a r l y a s c e n t [ 3 , 4 ] and t h e s l o p e o f t h e r e g r e s s i o n l i n e was c o n s i d e r e d t o be p r o p o r t i o n a l t o t h e r a t e o f c l e a r a n c e o f t h e r a d i o n u c l i d e from
IAEA-SM-247/86 NORMAL RESPIRATION
Г-ШТЕЯСЕРТ
PERFUSION
403 PANTING
Г-I N T E R C E P T
PERFUSION
F IG .2. Isometric displays o f relative perfusion (lower figures) and clearance slopes o f intra venously injected 133X e dissolved in saline. Figures at the left were recorded from the subject at rest and the figures at the right recorded during accelerated breathing (panting). Values are normalized so that absolute differences are not visualized.
the blood by the myocardium. This parameter represented the relative perfusion of the myocardium and a regional display of the accumulation on a 40 x 40 matrix, Figure 1, thus demon strating a functional image of relative regional myocardial flow. Computer graphics were somewhat primitive at that time and Love frequently constructed solid, three-dimensional models in which the height of each pixel was proportional to the slope of accumulation or intercept. An early computer equivalent of the isometric display is shown in the lower section of Figure 1. Later procedures were more frequently based on a single bolus technique rather than measurements derived from a con tinuous infusion. Single bolus methods were introduced in 1969 to determine the parameter, Л , for regional washout curves of the exponential form, Coe_xt. Functional images representing the washout of 133Xe from the lung were obtained by measuring the regional exponential clearance constant or transit times [5, 6,7,8], thus obtaining images of regional ventilation. Back extrapolation of the exponential clearance curve yields the value Co, which represents the maximum deposition before clear ance and, thus, provides an index of perfusion. Isometric images of these parameters are illustrated in Figure 2.
404
MACINTYRE et al.
F IG .3. Isometric displays o f the peak accumulation (perfusion on lower figure) clearance slopes (middle) and the product o f the two parameters (top) o f the clearance o f™ T c m-DTPA from a functioning and non-functioning kidney.
1 33
Similar washout curves for Xe injected into the upper thigh of a dog were obtained in order to reflect regional blood perfusion [9]. Regional rate constants of the accumula tion of chloromerodin were also determined [9] by solving for the exponential constant, k, assuming an exponentially rising curve of the form Co (l-e-kt). Following these early investigations, many other examples were reported. Most of these involved lung studies, primarily with gases [10,11] or renal studies, primarily with radiohippuran and involving analysis of many segments of the renogram curve [12,13] and some with analysis of regional renal blood flow [14] by use of radioxenon.
IAEA-SM-247/86 2.
405
PROBLEMS OF FUNCTIONAL IMAGING
It is surprising that in its fourteen-year history, none of these applications of functional imaging have been accepted as a routine clinical procedure^ Several reasons may be / responsible for this lack of enthusiasm. ,‘ У-
2.1. Computer Cost Originally it was felt that few institutions possessed the requisite computer facilities required for this type of analysis. A. 1970 review of the computer re- •quirements for functional scanning stated that a large main, storage, an on-line interface, ’ a graphic terminal and a physician-oriented command languagé were desirable for the interactive communication for functipnaT imaging [15]. It . was considered that such interaction was necessary since, from . the large amount of complex information potentially available ■ from a functional scan, only a small part may ;be :utiTized.) :>■ Initially, these requirements were considered both expensive and difficult to achieve, thus delaying the in corporation of functional imaging into routine studies. This limitation can hardly be considered a factor in the last few years, however, since camera interfacing, small on-line com puters, microprocessors, operator interaction and improved displays are readily available from commercial suppliers.
2.2. Selection of Critical Parameters In spite of the effort expended in previous years for renogram analysis, this procedure is being used less and less. Thus, while a functional image may provide an excellent por trayal of regional variation of a segment or parameter of a renogram, the incorporation of this information into a clinical decision is, at present, not being performed routinely [16]. It is also possible that the parameter selected is difficult to measure. The accumulation slope proposed by Love et al., as an index of relative myocardial perfusion required the maintenance of a reasonably stable concentration of the radionuclide in the arterial blood. This technique also re quired a high precision rectilinear scanner so that at the end of a three-to-six minute cycle the scanner probe was in the exact position to start the next cycle.
406
MACINTYRE et al.
FIG. 4. Functional images derived from a tim e series o f normal kidney (left) and a kidney with a nephrotic syndrom e displaying relative values o f the tim e fo r each region to reach half o f the m axim um concentration o f injected contrast material. Y ellow and green den ote slower blood velocity. (From Ref. [1 7 ].)
2 . 3 . C o m p e t i t i v e Mechanisms F i n a l l y , one must c o n s i d e r i f t h e same o r b e t t e r i n f o r m a t i o n can be o b t a i n e d from a l t e r n a t i v e measurements o r a n a l y s e s , e . g . i n f o r m a t i o n d e r i v e d from a f u n c t i o n a l image o f 133xenon c l e a r a n c e ti me f o r t h e l ung may be no b e t t e r as an i ndex o f v e n t i l a t i o n t h an a v i s u a l i n s p e c t i o n o f a s e r i e s o f washout images. In a s i m i l a r s e n s e , a f u n c t i o n a l s ca n o f t h e k i d n e y , as shown i n F i g u r e 3, may g i v e an e x c e l l e n t r e p r e s e n t a t i o n o f t h e s a t i s f a c t o r y r a d i o n u c l i d e d e p o s i t i o n i n bo t h ki dneys (l o w e r f i g u r e ) b u t t h e c l e a r a n c e s l o p e o f t h e r i g h t ki dney i s c o m p l e t e l y f l a t ( mi ddl e f i g u r e ) , c o n c l u d i n g t h a t t h e r i g h t ki dney can a c cu m ul at e t h e r a d i o n u c l i d e b ut c a n n o t r e l e a s e i t . Thi s i n f o r m a t i o n , however, i s u s u a l l y r e a d i l y a p p a r e n t from a s e q u e n t i a l s c a n , t h u s a f u n c t i o n a l image may g i v e l i t t l e additional information.
IAEA-SM-247/86 F R A H f c . ТО- M A X АКНД» и h h № и
0 a 0 о и 0 0 e 0 9 Я Я 0 0 0 0 0 0 » 9 9
0 0 0 0 0 0 0
0 0 0 0 г ?■ и
h
Г
? M
M
0 0 00 00 00
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 3 8 4 7 5 05a < t9 39 3a3 «fB ~ 4 l 3 "¿ S 3 0 0 0 3 6 3 8 4 1 4 9 4 3 3 5 3 0 ffe M ] ] ) ] 0 0 3 8 3 8 3 8 3 4 3 2 3 l 3 0 f b 5 4 4 33 ) 3 042423634303
0 0 0 0
407
0 0 8 0 0 ïb ? b 3 b 3 b S 5 B 0 0 03030373737373/33 0 038474746383737 39391; 0 i««14744S¿493737343*t
0 0 0 0 0 5 J S IS « I4 3 0
И 0 0495049351480 0 0 03 33 93 935 1« ! 0 0 03339393 0 03437173 0 03436269 0 034HSS3 B 0341
0 6 9 в7 в9 9 0 9 490т09г9£ё9в9в979665б «4 0в1е7б994Ч 59з91904г9£9геча9ьбМ 5г<1ч 0 0
.
. 079в7в99«9?9Т9Г909г9Ё9гв95049494949 0 07 98 7899297*7979992925290*646484949 0 07 98792949697979*923450494646484848 В 094949696969588504949494646464848 0 09 6«**«9б9б в850494949494646454843 ______ 0 0 9 0 * * * * 9 6 8 4 5 4 5 0 3 0 4 9 4 9 4 9 4 7 4 5 4 5 4 5 4 3
0 0 0 0 0 076«09ea49*SZ 50SeSe«*4947454545 0 e 0 0
e
0 0 0 0 0 0 в 7 ь Т Ь Ь 5 525 Ц 3 05 04 9 4 9 4 7 4 6 4 5 4 5 0 0 0 0 0 0 0 0 0 09653515048464747474544 0 0 0 0 0 0 0 0 0 0 0 04848404646464644 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0
0 0 0 P 0 0 0 0 0 0 0
Id
0 0 0 0 0 0
0 0 0 0 0 0
0 0 0 0
INPUT 3 - 6 R T H E A R T 7-11 Ш 15-18 LT H E A R T 2 4 - 2 6 FIG.5. Functional images o f the thorax show ing the num ber o f 0.5 second sequential fram es in which the m axim um counts were recorded. This tim e-to-m axim um image allow s the various chambers, blo o d vessels and lungs to be id en tified by observing the sequential tim e passage o f injected material.
I t must a l s o be c o n s i d e r e d t h a t o t h e r t y p e s o f measurement s may pr od u ce f u n c t i o n a l s c a n s t h a t e x h i b i t a g r e a t e r r e s o l u t i o n than ra d io n u c l id e tech niq u es. A fu n c tio n a l image o f t h e ki dney as prod uc ed by a c o m p u t e r i z e d r a d i o g r a p h i c t e c h n i q u e by Höhne e t al [1 7] i s shown i n F i g u r e 4. Thi s image was c o n s t r u c t e d from a s e r i e s o f fr ames f o l l o w i n g i n j e c t i o n o f n o n - r a d i o a c t i v e c o n t r a s t m a t e r i a l , and t h e p a r a m e t e r d i s p l a y e d i s t h e t ime f o r t h e c o n t r a s t m a t e r i a l t o
408
MACINTYRE et al.
GAS EXCHANGE POTENTIAL
INHALATION
FIG.6. Isom etric scan o f perfusion (Q intercept), clearance rates, and gas exchange poten tial (Q X slope) derived from the clearance o f an intravenous injection o f i33X e dissolved in saline. Upper figure show s clearly the reduction o f lung fu n ction in the le ft lung. B o tto m figure represents ventilation from inhalation o f 133X e gas and show s much less involvem ent o f the affected lung. N o te sim ilarity o f inhalation scan w ith the perfusion images.
IAEA-SM-247/86
409
N O R M A L P A T IE N T 1 0 % C H A N G E F R O M R E S T IN G
.
4 0 % C H A N G E F R O M R E S T IN G
R E S P IR A T IO N T O P A N T I N G
R E S P IR A T IO N T O P A N T I N G
APEX
APEX
R T.
LT.
LT.
R T.
BASE
BASE
A B N O R M A L P A T IE N T 1 0 % C H A N G E F R O M R E S T IN G
5 0 % C H A N G E F R O M R E S T IN G
R E S P IR A T IO N T O P A N T I N G
R E S P IR A T IO N T O P A N T I N G
APE*
APEX
LT.
R T.
BASE
LT.
R T.
BASE
■ = IN C R E A S E X = DECREASE
FIG. 7. Upper: R a tio o f clearance slopes o f su bject in Fig.2 showing relative increase a t all sites o f the lung w ith accelerated breathing. L ow er: R atio o f clearance slopes o f a p a tie n t show ing appreciable increase in clearance o f xenon on the norm al right lung during panting, b u t an actual decrease o f clearance by the abnorm al left lung.
41 0 MACINTYRE et al. FIG .8. U pper (A): R estin g p a tte m o f relative cerebral b lo o d flo w derived from clearance o f 133X e from the brain. Average b lood flo w is green w ith 20% low er values designated by shades o f blue and rates up to 20% above the mean show n b y shades o f red. L eft and right hem i spheres displayed. L ow er: D eparture in flo w rate from hyperfrontal resting p attern (above) w ith processing o f inform ation from sensory stim uli. A t left (B) the su bject fo llo w e d a m oving o b ject b y eye, with increases in clearance rate in the visual association cortex, fro n ta l e ye field, and supplem entary m o to r area. A t right, (C) the su bject listened to spoken w ords resulting in activation o f the a u d ito ry cortex and W ernicke’s area. (From R ef. [2 7 ].)
IAEA-SM-247/86
411
reach one-half of its maximum accumulation per pixel. The resolution produced by relative density recording on the 256 x 256 matrix is obviously superior to the radionuclide methodology. 3.
ADVANTAGES OF THE FUNCTIONAL IMAGE
Although the functional image may not have reached the state of routine clinical use, it has in many cases produced information not available from other sources. 3.1. Presentation of Objective Data The functional imaging technique that probably comes closest to being accepted as a routine clinical application is the time-of-maximum imaging proposed by Jones et al.[18] to identify specific heart chambers. A typical image is shown in Figure 5 in which each pixel displays the frame number in which the intravenously injected activity reaches maximum. This technique is now used by many groups [19,20,21] to identify heart chambers or lung regions from which dilution curves are to be recorded. The advantages include not only an objective selection of the area without operator intervention but also a quality control method of ensuring accurate area selection (i.e. eliminating pixels that represent superimposition of more than one chamber). 3.2. Combinations of Multiple Parameters Although many functional images reflect only one functional parameter, it is very simple to utilize several parameters in various combinations. As previously stated, the early work on lung clearance of 133Xe assumed that the clearance of the initial 60% of the curve following an intravenous injection of ТЗЗхе dissolved in saline could be represented by a single exponential clearance function of the form Co e”x . The parameter, Co, represented the relative perfusion, Q; and the parameter, (the semilogarithmic slope of clearance) was considered indicative of ventilation, V. It had been pointed out by Dollery et al in 1962 [22] that the washout curve described by this technique is related to the regional ventilation since the disappearance rate is determined primarily by the rate of clearance of the gas from the alveoli. The product of the relative activity of gas delivered, Q, multiplied by the rate of clearance of that
412
MACINTYRE et al.
activity, л, thus represents the regional gas exchange and designates the relative function of each site of the lung. In a study of twenty-eight cystic fibrotics [23] the regional gas exchange parameter gave the best correlation with the severity of the disease as determined by conventional techniques [24] . The parameters of perfusion and clearance of injected radioxenon are shown on the middle two isometric plots of Figure 6 as measured on a subject with cystic fibrosis. It may be noted that the lower image of ventilation, as obtained by inhalation, approximates very closely the above image of perfusion. The regional V/Q image calculated from the inhaled and perfused xenon in this case would be rela tively normal, and in this series it frequently failed to re flect the severity of the disease. The top figure, which is the regional gas exchange parameter calculated as the product of the perfusion area and slope areas, shows a marked decrease in the function of the left lung which is not readily appreciated on the other parameters. This applied as an it provides a function that
index of lung function has also been successfully index of lung cancer resectability [25] since direct measurement of the extent of loss of can be anticipated by surgery.
3.3. Pattern Recognition In the evaluation of regional organ function, it is usually of greater interest to the clinician to ascertain the areas that are not responding normally to an external stimulus rather than to have attention diverted to normal variation of function within the organ. One way to obtain such a representation is to use each organ as its own standard or baseline and measure the deviation from this baseline as a function of some external process on the original function or distribution. Attention would then be directed not to the normal differences of regional function or distribution, themselves, but to the regional changes in the function when the external stimulus or perturbation has been applied. Thus, in Figure 2, it is difficult to visualize any appreciable difference in the clearance slopes from a subject breathing at rest and panting. In studies performed in 1971 [26] it was shown that the matrix distribution of the ^ЗЗхепоп clearance slopes could be superimposed, as illustrated in
IAEA-SM-247/86
413
Figure 7, to demonstrate this pattern change. These patterns can also be quantified by assigning various symbols or colors to incremental changes so that the alterations can be readily identified. A similar type of difference pattern illustrating the variation of cerebral blood flow from the resting non-stimulated state as a function of visual, auditory or motor stimuli has recently been obtained [27] by Lassen, Ingvar and Skinh(jj. These investigators have produced functional images of relative cerebral flow by measuring the clearance rate of '^3xe from the superficial cerebral cortex below the detector array. Figure 8A shows the pattern of relative cerebral flow averaged from 26 subjects at rest. Figure 8B shows a pattern reflecting the de parture of clearance from the resting state obtained when the subject followed a moving object with his eyes. In this case, the visual association cortex in the rear of the brain shows an increase in the xenon clearance rate along with the frontal eye field and the supplementary motor area in the upper part of the frontal lobe. In Figure 8C, the subject listened to spoken words and it is seen that the auditory cortex in the tem poral lobe shows an increase in flow as does the adjacent area which mediates understanding of speech (Wernicke's area). Another type of pattern that has recently been ad vanced is the regional heart wall motion displays developed by Adam, Bitter and their co-workers [28,29] . A number of parametric images were proposed including differences between counts in end-systolic and end-diastolic phase, count diff erence between the end-systolic phase and the end of the fastfilling phase, contraction velocity, relaxation velocity, and the amplitude and phase of the first Fourier element from each pixel of a multiple gated ejection fraction study. In Figure.9A, a functional image of the amplitude of the first Fourier component is shown and in Figure 9B, the image of the phase angles of this component is illustrated. Note the excellent separation between the phase angle of the region of the atria and the ventricles. 4.
CONCLUSION
It is apparent that the areas in which functional imaging have been successful are, as the name implies, areas in which physiology rather than morphology is of interest. Secondly, it appears that to be truly meaningful, the functional image must do more than merely compress a time series into a single frame. It must either enhance a parameter that, because of
414 MACINTYRE et al.
FIG. 9. F unctional images o f wall m otion o f the heart (regional ejection fraction ) from m ultiple ga ted studies. Time variation o f the counting rate a t each p ix el was represented by a Fourier series w ith the phase o f the first Fourier com pon en t shown on the right and the am plitude o f the first co m p o n en t shown on the left. N o te the excellent separation on the phase angle display betw een the atria and ventricles. (Figure obtain ed from A dam and Bitter, University o f Ulm.)
IAEA-SM-247/86
415
small increments occurring in a noisy environment, could not be readily perceived, or it must devise an entirely new para meter that would not be otherwise available. It may be as simple as a difference scan or as complex as a Fourier dis play; but in either case, it must bring out information not previously apparent. Lastly, since it is functional, it must usually be correlated with other functional studies. In many cases, we are accustomed to viewing the images produced by nuclear medicine techniques as definitive answers denoting either positive or negative results. In many cases, functional images must be interpreted not as positive or negative, but merely as descriptive of the function of that region. The image may be consistent with a pathologic process or it may require correlation with other information to identify the presence or absence of abnormality. In any event, the successful use of functional imaging will require a much greater involvement of the interpreter with in-depth knowledge of the patho-physiology of the organ being studied, a con dition that may be attractive only to the more challengeoriented nuclear medicine specialist.
ACKNOWLEDGMENTS The authors are grateful to the many investigators and colleagues who have contributed examples of functional images for this report, including Drs W.E. Adam, F. Bitter, J.H. Gallagher, K.H. Höhne, S.R. Inkley, Y. Ishii, N.A. Lassen, W.D. Love, E. Roth, and R.O. Smith and who have all contributed so much to the field of functional imaging in general.
REFERENCES [1] LOVE, W .D., SM ITH, R .O ., M ITCH ELL, L.M ., PU LL E Y , P .E ., L ocating areas o f reduced co ronary blood flow b y ex ternal m easurem ent o f m yocardial K42 clearance, C irculation 34 Suppl. 111 (1 9 6 6 ) 160. • [2] LOVE, W .D., PU LL E Y , P.E ., M apping regional coronary blood flow by ex ternally m oni toring o f m yocardial R b 86 and K42 clearance, Clin. Res. 14 (1 9 6 6 ) 83. [3] SMITH, R .O ., LOVE, W .D., LEH AN, P.H ., HELLEM S, H .K ., D elayed c o ronary blood flow d ete cte d b y c o m p u te r analysis o f serial scans, Am . H eart J. 8 4 (1 9 7 2 ) 670. [4] SMITH, R.O ., BEN N ETT, K .R ., SU ZU KI, A., LEH A N , P.H ., H ELLEM S, H .K ., A traum atic evaluation o f m yocardial revascularization procedures w ith 43K, R adiology 114 1 (1 9 7 5 ) 99. [5] M A CINTYRE, W .J., IN K LEY , S.R ., F u n c tio n a l lung scanning w ith 133X enon, J. Nucl. Med. 10 6 (1 9 6 9 ) 355.
416
MACINTYRE et al.
[6] DE ROO , GO RIS, M., VAN DER SCH U EREN , G., COSEM ANS, J., B IL LIE T, L., GY SELEN, A., C om puterized dynam ic scintigraphy o f the lungs, R espiration 26 6 (1 9 6 9 ) 408. [7] LOKEN, M.K., M EDINA, J.R ., L IL L E H E I, J.P ., L ’H E U R E U X , P., KUSH, G.S., EB ER T , R .V ., Regional pu lm o n ary fu n c tio n evaluation using X enon 133, a scintillation cam era, and com p u ter, R adiology 93 6 (1 9 6 9 ) 1261. [8] M A CINTYRE, W .J., INK LEY, S.R ., ROTH, E ., D RESCH ER, W.P., ISH II, Y ., Spatial recording of disappearance co n stan ts o f 133X enon w ashout from the lung, J. Lab. Clin. Med. 7 6 (1 9 7 0 ) 701. [9] K A IH A RA , S., N A TA R A JA N , Т .К ., M A YNA RD, C.D., W AGNER, H.N ., Jr., C onstru ctio n of a fun ctio n al image from spatially localized ra te constants o btained from serial cam era and rectilinear scanner d a ta , R adiology 93 6 (1 9 6 9 ) 1345. [ 10] BU RD IN E, J.A ., M URPHY, P.A ., ALAG ARSA M Y, V., R Y D ER , L.A ., C A R R , W.N., F u n c tio n a l pulm onary imaging, J. N ucl. M ed. 13 12 (1 9 7 2 ) 933. [11] A L PE R T , N.A ., McKUSICK, K.A ., C O R R E IA , J.A ., SHEA, W., BROW NELL, G.L., PO TSA ID , M.S., Initial assessm ent o f a sim ple fu n ctio n al image o f v entilation, J. Nucl. Med. 17 2 (1 9 7 6 ) 88. [12] W IENER, S.N., BORK AT, R .F ., FL O Y D , R .A ., F u n c tio n a l imaging: A m eth o d of analysis and display using regional rate constants, J. N ucl. M ed. 15 2 (1 9 7 4 ) 65. [13] G ELFA N D , M .J., A L L EN , F.H ., G R E E N , M.V., JO H N STO N , G.S., BA ILEY, J.J., “ F u n c tio n a l m apping o f renal tran sit o f 131-I-iodohippuran” , Proc. T hird Sym posium on Sharing o f C o m p u ter Program s and T echnology in N uclear M edicine, Miami, F lorida, 1973, O ak Ridge, T ennessee (1 9 7 3 ) 100. [14] ISH II, Y., KAWAM URA, J., M UKAI, T., TA K A H A SH I, M., T O R IZ U K A , K., F u n c tio n a l imaging o f intrarenal blood flow using scintillation cam era and co m p u ter, J. N ucl. Med. 16 1 0 (1 9 7 5 ) 899. [15] M A CINTYRE, W .J., D R ESC H ER , W.P., IN K LEY , S.R ., “T he fu tu re o f functional scanning and its dependence on com puter-assisted analysis” , Q uantitative Organ V isualization in N uclear M edicine (K EN N Y , P .J., SM ITH, E.M ., E ds), U niversity of M iami Press, C oral G ables, F lorida (1 9 7 1 ) 865. [16] BUREAU O F RA D IO LO G ICA L H EA LTH (F D A ), M edically oriented data system (MODS) survey o f nuclear m edicine procedures, P ersonal com m unication, 1978. [17] HÖ HNE, K.H ., BÖHM, M., N IC O LA E, G.C., “ T h e processing o f X-ray image sequences” , Advances in Digital Im age Processing, Plenum Press (in press). [18] JO N E S, R .H ., SABISTON, D.C., Jr., BATES, B.B., M O RRIS, J.J., AN D ERSO N , P.A.W ., GO O D RICH , J.K ., Q uantitative radionuclide angiocardiography for d e te rm in a tio n of cham ber to cham ber cardiac tran sit tim es, Am . J. C ardiol. 30 (1 9 7 2 ) 855. [19] G O R IS, M .L., BAUM, D., W ALLING TO N, J., K R ISS, J.P., N uclear angiocardiography: A uto m ated selection o f regions o f interest fo r th e generation o f tim e-activity curves and param etric image display and in te rp re ta tio n , C lin. Nucl. M ed. 1 3 (1 9 7 6 ) 99. [20] G A LL A G H E R , J.M ., FO U A D , F.M ., COOK, S.A., M A CINTYRE, W .J., A u to m atic ROI selection fo r first transit nuclear cardiology, I.E .E .E . Trans. Nucl. Sei. NS-27 1 (19 8 0 ) 513. [21 ] BIT TE R , F ., ADAM, W.E., KAMPM ANN, H., M EY ER, G., W ELLER, R ., “ A utom ated selection o f areas o f interest in dynam ic studies and cam era-cinem atography o f the h e a rt” , Proc. F ifth Sym posium on Sharing o f C o m p u ter Program s and T echnology in N uclear M edicine, Salt Lake C ity, U tah , 1975, O ak R idge, T ennessee (19 7 5 ) 48. [22] D O LLER Y , C .T., HU GH-JONES, P., MATTHEW S, C.M .E., Use o f radioactive xenon for studies o f regional lung fu n ctio n : A com parison w ith oxygen-15, Br. Med. J. 2 (1 9 6 2 ) 1006.
IAEA-SM-247/86
417
[23] IN K LEY , S.R ., ST E R N , R.C., DO ERSH U K , C .F., FIS H E R , B., M A CIN TY RE, W .J., A com parison o f 133 X enon techniques w ith conventional m easurem ents o f lung dysfunction in tw enty-eight cystic fibrotics, Clin. Res. 20 (1 9 7 2 ) 807. [24] IN K LEY , S .R., M A CINTYRE, W .J., Relative effectiveness o f 133X enon techniques in evaluation o f regional fu n c tio n in cystic fibrosis, J. Nucl. Med. 13 6 (1 9 7 2 ) 439. [25] BAUGHM AN, R.P., IN K LE Y , S.R ., M A CINTYRE, W .J., Regional lung fu n c tio n as an in dicator o f lung cancer resectability, Clin. Res. 25 (1 9 7 7 ) 625A . [26] M A CINTYRE, W .J., IN K LEY , S.R ., R O T H , E „ PR IT C H A R D , W., ISH II, Y „ C om puter rep resen tatio n o f regional organ fu n c tio n w ith external stim uli, J. Nucl. Med. 12 6 (1 9 7 1 ) 379. [27] LASSEN , N.A ., IN G V A R , D.H ., SK IN H 0J, E ., Brain fu n c tio n and blood flow , Sei. Am . 239 4 (1 9 7 8 ) 62. [28] ADAM, W .E., TARKOW SKA, A., B IT TE R , F ., STAUCH, M., G E F F E R S , S., E quilibrium (gated) radionuclide ventriculography, Cardiovasc. R adiol. 2 (1 9 7 9 ) 161. [29] BIT TE R , F ., ADAM, W .E., G E F F E R S , H., W ELLER, R ., E LLEBRU CH , H., N uclear m edicine: Synchronized steady state h e art investigations (in press).
DISCUSSION H.N. WAGNER, Jr.: I should ju st like to com m ent th at examples o f functional images in clinical use today are regional stroke volume or ejection fraction; other examples are phase and am plitude F ourier images. I expect m ore will be seen in the future. W J. MACINTYRE: Exam ples o f phase and am plitude F ourier images are shown in Fig.9 o f the paper, and I agree th a t this application is im portant now and will becom e even m ore so in the future. K.E. BRITTON : I am surprised th a t y ou consider functional imaging has n o t been in routine use. In Europe it has been used for m any years and the approach has been to sort o u t the problem in physiological and patho-physiological term s first and then design the imaging o f the appropriate function. W.J. MACINTYRE: Well, it w ould seem th a t although functional imaging o f various types is perform ed in m any o f the large centres, it has n o t really becom e a p art o f com m unity hospital h ealth care, w ith the exception, perhaps, o f the functional images used to select regions o f interest, as illustrated in Fig.5 o f the paper. I would like to see an extension o f these techniques to all levels o f nuclear medicine. R J . DI PAOLA: In a recent sym posium (Ref. [ 17 ] in y o u r paper), Höhne stated th a t functional imaging in nuclear m edicine had n o t y et reached the stage o f clinical applicability offered by conventional radiology. Can you com m ent? W.J. MACINTYRE: The high inform ation density available from digital radiography and conventional radiography will always provide spatial resolution superior to nuclear imaging (Fig.4 o f th e paper). When this resolution is n o t required, the radionuclide techniques should have equal applicability.
Poster Presentations OPTIMIZED DISPLAY O F DIGITIZED SCINTIGRAPHIC DATA B.F. HUTTON, J. CORMACK D epartm ent o f Nuclear Medicine, Royal Prince Alfred Hospital, Cam perdown, Sydney, New South Wales, Australia
A technique has been devised for optim izing the display o f scintigraphic data by minimizing the inform ation lost in transferring data via display directly to an observer. Loss o f inform ation is m inim ized by evaluating and m inimizing the mean pixel u ncertainty (m .p.u.) for the image. In previous w ork evaluation o f m.p.u. has been simplified by assuming th a t the display levels used are discrete and associable. This m ay n o t necessarily be correct, particularly when a large num ber o f display levels are used. A m ore accurate estim ation o f m .p.u. therefore requires th at perceptual factors be taken into account. The perceptual characteristics of a display m ay be described by the conditional probability m atrix Pj^ whose elem ents p (l/k ) give the probability th a t, given a display level k, an observer will recognize it as level 1. In the case o f discrete distinguishable levels this is simply a u n it m atrix; however, in the m ore general case the m atrix m ust be evaluated experim entally. An observer is presented with a random ly selected display level and asked to choose w hich o f the com plete set o f available levels best m atches the displayed level. The m atrix Plk is thus obtained by expressing the frequency with which an observer perceives a particular level when level к is displayed, as a fraction o f all occasions when level к is displayed. Q uantitative inform ation transfer param eters may also be derived directly from the m atrix w hich may be used to characterize the perceptual efficiency of the display. In the absence o f noise the uncertainty as to the source value o f a pixel perceived as a particular level m ay be evaluated. Hence an expression for m .p.u. m ay be derived which depends only on the level w idth betw een boundaries selected, th e discernibility m atrix Рц. and the histogram o f pixel counts. A ttem pts have been m ade to assess the technique using localization receiver operator characteristic (LROC) curves. A series o f 100 brain scans, in which a lesion had a 50% probability o f occurrence, were presented to observers. Results indicated a slight decrease in lesion detectability when perceptual characteristics 419
420
POSTER PRESENTATIONS
were included rather than assuming discrete levels. This result is understandable since images were optim ized for the com plete image rather than the suspected region. The LROC analysis therefore did n o t provide an adequate means of assessment since it was insensitive to im provem ent in detail in areas o f the image o th er than the lesion site. A dditional w ork is therefore necessary to assess this technique further. DISCUSSION P. SHARP: Would you say how y our discernibility m atrix is used to characterize the perceptibility efficiency of displays, and how the m atrix would apply to colour-coded images? B.F. HUTTON; The discernibility m atrix is obtained by perform ing a perception experim ent for a particular display/observer com bination and describes the uncertainty regarding the observer’s recognition o f a particular displayed level. The inform ation transfer o f display levels to an observer m ay thus be calculated and used as a param eter to com pare betw een different display/observer com binations. The ideal display w ith discretely distinguishable levels will have no loss o f inform ation, whereas in the norm al situation there will be a m easurable loss. The situation w ith colour is m ore com plicated. A lthough colours may be perfectly discrete, it m ay be difficult to select colours which are perfectly associable (i.e. which can be placed w ithout error in order o f m agnitude) w ithout allowing a considerable learning process. We have n o t yet assessed the loss o f inform ation to be expected from the use o f colours which are n o t associable. Spatial effects such as contouring, which is particularly evident w ith the use of colour, is a furth er com plicating factor.
EFFECT OF DISPLAY MATRIX SIZE ON THE QUALITY O F DIGITIZED IMAGES P. SHARP, R. CHESSER D epartm ent o f Bio-Medical Physics and Bio-Engineering, University o f A berdeen, Aberdeen, United Kingdom
Whereas the use of a com puter-interfaced-display system has the advantage th at an interactive process can be used to optim ize image display, the digitization of the image is often regarded as impairing image quality. The use of a large
POSTER PRESENTATIONS
421
num ber o f small picture elem ents (pixels) minimizes any degradation o f spatial inform ation, b u t will also reduce the num ber o f counts per pixel and so increase statistical noise. Initially the effect o f pixel size on the detectability of a single ‘h o t’ abnorm ality, o f 2-cm FWHM, in a uniform background o f radioactivity, was investigated. A bnorm ality detectability was m easured by the m ethod o f constant stimulus. If the to tal background counts was k ep t constant, i.e. imaging tim e was constant, then varying the pixel size from 2 mm to 10 mm had no effect on abnorm ality detectability. However, if the m ean counts per pixel were kept constant then detectability im proved w ith decreasing pixel size. This latter process is equivalent to interpolating the image data. Simply detecting a 'h o t sp o t’ is n o t a good test o f the degradation o f spatial inform ation by image digitization. In a second experim ent, a test-p attem was used in which h alf o f the abnorm alities (again 2-cm FWHM) had a ‘cold’ centre of 7-mm FWHM. Varying pixel size from 2 mm to 6 mm had little effect on the observer’s ability to discrim inate betw een those abnorm alities w ith a ‘cold’ centre and those w ithout, b u t an increase from 6 mm to 8 mm caused the num ber correctly identified to drop from 70% to 25%. Interpolating from this 8-mm pixel to a 4-mm one produced no significant im provem ent in the identification rate. The effect o f pixel size on detectability w hen norm al image structure is present was tested by adding the ‘h o t’ abnorm ality to norm al brain images. A 6-mm pixel produced a small drop in image quality, but no difference was found betw een 2-mm and 4-mm pixel sizes. Whereas the use o f small pixels gives the impression o f im proving the quality o f images, m easured perform ance actually varies only very slowly over a wide range o f pixel sizes.
DISCUSSION A.E. TODD-POKROPEK: Y our conclusion in y our poster presentation th at there is no need to interpolate coarse and grainy images to obtain the same detectability o f lesions is rath er im portant, if true. Could you state the viewing conditions th a t the observers o f yo u r images used? Could the results be explained by the fact th a t observers used, for instance, visual m inim ization? What are the confidence limits o f your conclusion, e.g. have y ou m ade a statistical test o f the null hypothesis o f there being no difference before and after interpolation? P. SHARP: As little inform ation exists on optim um viewing conditions, and since clinical practice places few restrains on viewing conditions, we allowed the observers to choose their own viewing distance. When grey-scale displays were used, the observer could also vary screen brightness and contrast via analogue controls. On m odern system s it is unnecessary to pre-define the grey scale. The
422
POSTER PRESENTATIONS
data showing th a t interpolation did not improve image quality were highly significant. However, fu rth er w ork m ust be perform ed w ith o th er phantom s before we can be certain th a t this result is generally true.
NON-UNIFORMITY AND N ON-ST ATION A R IT Y IN EMISSION TOMOGRAPHY A.E. TODD-POKROPEK*, S. ZUROWSK1, F. SOUSSALINE S.H.F.J. D épartem ent de Biologie, C.E.A., H ôpital d ’Orsay, Orsay, France
All tom ographic reconstruction algorithm s assume th a t the values to be m easured on projections are constantly proportional to line integrals through the object to be reconstructed and th at the ‘geom etry’ such as the position of the axis of ro tatio n is know n accurately. In emission tom ography, one com m on cause of the failure of a first constraint occurs as a result o f attenuation. O f the other types o f failure, three are considered here, being results of 1. V ariations in the sensitivity (uniform ity) o f the detector. 2. D ifferential non-linearity of the detector. 3. Integral non-linearity (spatial distortion). These effects have been studied, by sim ulation, by use o f phantom data, by perturbing real data, and by m aking actual m easurem ents on a GE 400 T rotating gamma camera system linked to an Inform atek Simis 3 com puter. F or a given area o f changed (non-uniform ) sensitivity o f the detector, a ring artifact is created in the tom ogram w ith the following properties. The am plitude o f the artifact is proportional to the inverse o f the distance betw een the area of changed sensitivity and the axis o f ro tatio n o f the detector, thus potentially creating very large artifacts on this axis. The am plitude is also dependent, again inversely, on the square ro o t o f the area o f changed sensitivity (for a constant size o f object and fractional change in sensitivity). Thus this artifact is m ore easily visible w ith those cameras w hich have tended to be optim ized in term s o f resolution and for which fluctuations in ‘u n ifo rm ity ’ tend to be m ore rapid across the field o f view. U niform ity correction is thus norm ally essential. The m athem atics o f the process o f artifact creation and the influence o f uniform ity correction have been evolved. * Also at: D ep artm en t o f M edical Physics, F a c u lty o f Clinical Sciences, U niversity College, L o ndon, U nited K ingdom .
POSTER PRESENTATIONS
423
Increasing sampling, fo r exam ple o f the angular sampling, or the num ber of points along a projection, leads initially to an increase in the quality o f the tom ographic reconstruction. It m ay be observed th at, after a certain point, no further increase in quality is obtained, and indeed, as the sampling is further increased, so the images som etim es becom e worse. To this end a general treatm ent of sampling, and the closely linked area o f noise propagation, has been developed. One phenom enon which then has been investigated is the effect of differential non-linearity (prim arily o f the data capture system ) on the quality o f the re construction, and on the signal-to-noise ratio. The effect o f such non-linearity has been investigated and found to be surprisingly im p o rtant in th at the relatively high frequency noise which is injected on to the projections can cause a serious degradation o f the tom ographic signal-to-noise ratio. Noise pow er as a function of spatial frequency, for various filters and p erturbations, has been predicted as being filter-shaped, and confirm ed in practice. Tom ographic reconstruction is therefore in general a high frequency noise amplifier. The use o f various window functions (Hamming, H ann, Parzen, etc.) has been studied as a function o f their abilities to reduce such ‘high frequency artifacts’. Integral non-linearity has been investigated as such and, at the signal-to-noise ratios found for conventional emission tom ography, n o t found to be troublesom e. In sum m ary, the influence o f various types o f non-uniform ity and non-stationarity has been studied, and a m athem atical basis for predicting their effects on tom ographic reconstructions developed. The predictions th a t have been m ade have been tested prim arily by sim ulation, b u t also, as far as possible, confirm ed with physical m easure m ent on a rotating gamma cam era tom ograph.
IN FLU EN CE O F NON-UNIFORM RESOLUTION ON IMAGE QUALITY AND QUANTITATION IN POSITRON-EMISSION COMPUTED TOMOGRAPHY (PCT) E.J. HOFFM AN, SUNG-CHENG HUANG, D.L. PLUMMER, M.E. PHELPS, D.E. KUHL L aboratory o f Nuclear Medicine and R adiation Biology, UCLA School o f Medicine, Los Angeles, California, U nited States o f America
In tw o-dim ensional scintigraphic imaging, uniform resolution is a prim ary factor in judging th e quality o f imaging devices. N on-uniform ity also has a serious im pact on image quality in positron-em ission com puted tom ography (PCT).
424
POSTER PRESENTATIONS
Sources o f resolution non-uniform ities th a t are o f prim ary concern in PCT design are: (1) intrinsic variation o f resolution as a function o f depth betw een coincident detectors and (2) additional resolution variation th a t is introduced in m ultiplane systems when image planes betw een true detector planes are approxi m ated by collecting data from coincidences betw een detectors in neighbouring planes. To assess intrinsic resolution variation, m easurem ents o f resolution as a function o f depth betw een coincident detectors were obtained for a series o f detectors, in terd etecto r distances, scattering m edia and energy discrim inator threshold settings. M easurements included b o th image resolution and axial resolution. Experim ental LSFs were used as in put for sim ulation o f effect o f resolution variation on quantitative studies in PCT systems th at em ployed such detecto r geometries. In q u an titatio n in PCT, pixel values in images should be proportional to true isotope concentrations. However, for objects smaller than twice system resolution, image pixel values will be low er than true values. In addition, resolution variations cause variability in pixel values o f the same object in different positions in th e field o f view. The following conclusions were reached: 1. If intrinsic resolution variation o f ± 10% is required, only the central 30% o f space betw een tw o coincident detectors can be used. 2. Errors in q u an titatio n o f isotope u ptake in smaller organs are seriously increased w ith smaller distances betw een coincident detectors. 3. Resolution non-uniform ities can cause serious errors in q u an titation and im a g e i n t e r p r e t a t i o n in P C T .
4. PCT-system designs w hich com prom ise resolution uniform ity to achieve high sensitivity and high resolution, will also have seriously com prom ised their ability for quantitative m easurem ents consistent w ith their apparent high resolution and sensitivity.
425
POSTER PRESENTATIONS RESTAURATION D ’IMAGES SCINTIGRAPHIQUES PAR FILTRES ADAPTATIFS A SUPPORT BORNE Y. BIZAIS*, Ph. de LARM INAT+, J.-P. LEMORT*, A.E. TODD-POKROPEK** *
L aboratoire de biophysique, F aculté de m édecine, N antes
*
ENSM, L aboratoire d ’autom atique, N antes
** C entre F rédéric Joliot, Orsay, France
On considère l’image scintigraphique (v) comm e la m esure de la projection plane (u) de la distribution radioactive d ’une structure biologique. Elle est dégradée par un flou h et un b ru it add itif poissonnien b: v=hXu+b C ette transform ation n ’est pas linéaire, car h dépend des positions respectives du d étecteu r et de la source, et b est égal au niveau du signal. Par ailleurs, le filtre linéaire optim al est le filtre de Wiener: f X (h X 0 UU X h + 0 bb) —h X uu = 0
(1)
où 0uu(r) = m 2exp ( - I r | /T ) est la fonction d ’autocorrélation du signal idéal et m 2 le m om ent d ’ordre 2 du signal, e t 0bb = m la fonction d ’autocorrélation du b ruit qui est blanc, de m oyenne nulle et linéairem ent indépendant du signal. En posant p = m /m 2 et 4'Uu = 0 u u /m 2, l’équation (1) s’écrit: fX (h X v^uu X h + p 2) - h X ф и и -0
(2)
où seul p dépend de l’image à traiter. L’équation (2) perm et de calculer f sur u n support borné, — soit en calculant la F F T inverse de la transform ée de F ourier analytique de f échantillonnée à des pas variables1, — soit en m inim isant le m em bre de gauche avec la contrainte: f = 0 | x > L ou y > L.2 1 BIZA IS, Y., Thèse de 3èm e cycle, 1979, ENSM, N antes. 2 LEM ORT, J.P., BIZAIS, Y., de LARM IN AT, Ph., Use o f fin ite m em ory W iener filter to im prove th e quality o f scintigraphic images, accepté p o u r publicatio n dans E ur. J. Nucl. Med.
426
POSTER PRESENTATIONS
On p eut ainsi constituer une fois p our toutes une batterie de filtres (f¡) pour différentes valeurs de p (pj). L’algorithm e de restauration consiste alors en chaque point: —à calculer m et m 2 sur un m édaillon, puis p, — à choisir le filtre f¡ de sorte que | Pj—p | soit minimal, — â convoluer spatialem ent fj et un médaillon-image centré sur ce point. Ce filtrage s’adapte donc en fonction du niveau de bruit et de la variance locale de l’image, p erm ettan t ainsi de traiter des images présentant de fortes variations d ’intensité. A titre d ’exemple, quelques résultats sont proposés.
RADIOPHARMACEUTICALS Sessions 4 and 4a
IAEA-SM-247/20S
Invited Review Paper NEW DEVELOPMENTS IN RADIOPHARMACEUTICALS FOR IMAGING A Review* G. SUBRAMANIAN, J.G . McAFEE D epartm ent o f Radiology, U pstate Medical Center, Syracuse, New York, U nited States o f America
Abstract NEW DEVELOPMENTS IN RADIOPHARMACEUTICALS FOR IMAGING: A REVIEW. During the last four years several new radiopharmaceuticals have been introduced for radionuclide imaging. More than 30 compounds have been evaluated for use in hepatobiliary studies of which 2, 6 diethyl-HIDA and 2,6 diisopropyl-HIDA have been found to be the two best agents of choice. A series of new diphosphonate bone imaging agents have been introduced, and some of them show improved concentrations compared with older agents in experimental bone lesions. These are yet to be evaluated in clinical trials. The introduction of 99Tcm-Diars as a possible replacement for 201T1 in myocardial imaging is a good start. However, a simpler method of preparation for this complex is required. The 99Tcm-red-cell-labelling technique has become common even in community hospitals, and this radiopharmaceutical also has enhanced the use of cardiovascular studies in nuclear medicine. The commercial availability of n lIn-oxine has solved only one end of the problem in labelling white cells and platelets for clinical use. Simpler, individual cell isolation methods for neutrophils, lymphocytes and monocytes are required. Sterile kit methods for these separations, which are easy to perform in community hospitals, are preferable to current techniques.
TECHNETIUM -99m-LABELLED HEPATOBILIARY AGENTS Since the in tro d u ctio n o f " T c m-labelled HIDA (hepatoim inodiacetic acid, 2,6 dim ethyl acetanilido im ino diacetate) by Loberg [1] in 1975, considerable progress has been m ade by m any investigators to im prove upon the original HIDA. Several new IDA derivatives have been synthesized and evaluated in experim ental animals [2—15] and clinical evaluations [16] and com parisons betw een different agents [17—19] have been perform ed. H unt and co-workers [20] recently reported on several new 99Tcm-labelled hepatobiliary agents based on benzamidazole-IDA. * This work was supported in part by USPHS grant No.GM-23033.
429
430
TABLE I. D ISTRIBU TIO N OF TECHNETIUM-99m-LABELLED HEPATOBILIARY AGENTS IN RABBITS FO R COMPOUNDS SHOWN IN FIG . 1 P ercentage d o se in w h ole organs a 5 m inutes
15 m inutes
60 m in u tes
C om pound B lood
Liver
G .I.T .+ G.B.
Urine
Blood
Liver
G .I.T .+ G.B.
U rine
Blood
A niline
21.9
23.1
13.6
2.1
8.2
11.3
41.5
19.1
5.5
4.4
52.9
23.2
D im ethyl D iethyl D iisopropyl
13.2 7.4 7.6
31.1 37.2 53.7
31.3 29.5 22.0
0.5 1.8 1.1
3.5 2.6 5.8
9.8 21.9 44.5
69.7 56.9 35.4
5.2 3.6 1.6
1.3 2.2 2.4
1.0 1.8 5.2
78.8 81.1 80.2
10.3 7.2 4.8
T rim eth y l
7.9
32.3
29.5
1.8
3.4
10.6
68.0
3.3
2.5
1.0
81.2
8.1
P-ethyl P-isopropyl P-butyl
14.1 5.3 12.4
29.6 43.3 62.7
29.8 33.6 7.7
1.8 0.2 0.9
3.8 3.2 2.1
22.1 24.0 42.9
54.0 56.8 44.4
5.0 2.9 0.4
1.2 1.1 1.9
2.4 3.6 10.5
72.3 83.0 78.7
16.5 4.8 0.9
P -ethoxy P -butoxy
24.5 6.0
18.1 45.1
20.1 31.1
0.9 0.4
7.7 2.6
6.9 20.0
54.4 64.5
13.2 1.6
3.2 1.7
1.3 4.7
62.0 81.9
22.4 2.4
O -butoxy
9.1
15.9
55.3
2.6
6.3
13.4
62.9
2.1
1.7
1.0
83.7
6.9
G .I.T .+ G.B.
Urine SUBRAMANIAN and M cAFEE
3 F ro m Ref. [8] (rep ro d u c e d by perm ission).
Liver
IAEA-SM-247/205
431
The various derivatives o f HIDA have been prepared by altering the sub stituents on the H ID A ’s arom atic ring. In 1976 our laboratory reported on the preparation and biodistributions o f three new derivatives [2] and several other new com pounds were reported subsequently [5—7]. The synthesis o f these com pounds was accom plished by tw o different m ethods. In the original m ethod [ 1], 2,6 dim ethyl aniline was reacted w ith chloroacetyl chloride to pre pare the chloroacetanilide. This com pound was then reacted w ith im inodiacetic acid to form the HIDA. Burns [7] introduced a new one-step m ethod in which the aniline is directly reacted w ith acetic anhydride and nitrilotriacetic acid to produce the same com pound. Fields [5] also followed this procedure to prepare several new HIDA derivatives. Van Wyk [9] prepared five isomers o f dim ethyl HIDA to study their biodistribution differences. M olter [10, 11] reported on the synthesis o f 10 more new HIDA derivatives and their evaluation in experi m ental animals. The benzam idazole-IDA com pounds were synthesized [20] by the alkylation of the IDA dim ethyl ester w ith chlorom ethyl benzam idazole (w ith suitable sub stituents) followed by mild alkaline hydrolysis to remove the m ethyl group from the IDA. The chlorom ethyl benzam idazoles were prepared from th e corres ponding su bstituted phenylene diamines. The IDA derivatives were labelled with " T c m using stannous chloride as the reducing agent. Fritzberg [14] recently evaluated the quality control systems for these labelled com pounds. Structural evaluations o f " T c m -HIDA derivatives were perform ed by Callery and co-workers [3] and Loberg and co-workers [13]. Their results show th at " T c m -HIDA is a distinct radiopharm aceutical different from 14C-labelled-HIDA and tw o molecules o f HIDA are usually attached to one atom o f " T c m . During the analysis o f several " T c m-HIDA derivatives by HPLC, Fields and co-workers [5] found m ultiple peaks w ith several com pounds. The chemical identification o f these peaks is unknow n at this time. B iodistribution o f these derivatives was perform ed in mice, rats, rabbits, dogs and baboons by different investigators. Some o f the results are shown in Tables I—IV and the structural form ulae can be seen in Figs 1 and 2. There seems to be a big difference in th e biodistribution betw een rats and rabbits. The results obtained in rabbits were fairly reproducible in higher mammals and in norm al patients. The rat results unfo rtu n ately cannot be extrapolated to humans. However, for intercom parison o f agents rat studies may be helpful. From these studies o f HIDA derivatives several conclusions can be reached: (1) As the lipophilicity o f the com pound increases, the hepatic clearance rate decreases and urinary excretion is minim ized. (2) F o r those com pounds w ith fast clearance from the liver, th e renal clearance is higher. (3) The position o f substituent groups plays an im p o rtan t role in the biodistribution o f these agents. Of the BMIDA (benzam idazole IDA) com pounds, the brom ine-substituted derivative seems to be the best agent. This, however, has yet to be verified in higher mammals and in clinical trials.
432
TABLE II. PARAM ETERS OF THE ANIMAL EXPERIMENTS, THE DISTRIBUTION CO EFFICIENTS AND THE PROTEIN-BINDING RATE OF THE IDA-DERIVATIVES STUD IED a D istribution in organs of ra t (n = 3) in % o f the applied dose 30 m in p.i.
Liver kinetics (rab b it)
Derivative Liver
84.1 78.8 61.6 62.3 78.5 86.8 83.3 85.4 80.4 8 5 .5 ' 53.0 62.1 81.0 80.6 64.0 53.8 53.3 81.8
5.14 4.47 5.80 4.69 2.80 2.83 2.97 0.96 2.16 1.96 6.40 5.90 3.30 2.28 4.40 4.80 8.30 1.57
Bladder w ith c ontent 7.74 5.62 6.20 16.31 9.70 8.44 2.76 0.13 4.89 0.79 20.10 8.40 2.90 5.74 15.60 21.50 18.40 1.79
T h y ro id gland
1 ml blood
R est of body
0.01 0.01 0.01 0.008 0.005
0.13 0.18 0.32 0.15 0.09 0.13 0.32 0.05 0.20 0.19 0.14 0.13 0.13 0.17 0.10 0.16 0.15 0.07
6.94 8.14 17.20 10.50 6 .40 6.56 9.43 3.11 9.0 7.17
0.006 0.009 0.004 0.01 0 .009 0.008 0.01 0.01 0.01 0.007 0.01 0.01 0.005
10.70 9.90 6.40 9.70 8.10 12.40 11.1 4 .70
a F ro m Ref. [11] (rep ro d u c e d by perm ission). b H ere the pH o f th e stock solution was 6.5; the labelled injection solution was distinctly tu rb id .
Tm ax
T l/2
(m in)
(m in)
5.7 6.8 4.6 3.7 6.2 3.9 6.3 7.4 4.2 6.8 3.6 5.0 4.2 2.6 5.5 7.5 7.6 14.3
9.5 15.7 15.3 12.7 14.7 11.7 27.6 41.0 18.4 38.7 12.8 9.9 10.4 7.7 15.3 31.5 25.4 48.7
0.038 0 .14 0.42 0.38 0 .94 1.74 4.35 >32 2.17 4.66 0.003 0.19 0.075 0.037 0.096 0 0 0
(%)
10 10 12 14 11 11 50 >90 10 48 8 5 < 1 1 18 5 3 0
and M cAFEE
0.83 2.53 9.30 2.40 1.60 1.97 4.01 12.47 2.80 4.66 1.60 1.30 1.60 1.70 1.90 2.70 3.10 3.80
K idneys
P ro te in binding ra te a fte r 3 h in cu b atio n
SUBRAMANIAN
2,6-D im ethyl-ID A 2,6-D iethyl-ID A 2,6-D iisopropyl-I DA 4-M ethyl-IDA 4-E thyl-ID A 4-Isopropyl-ID A 4-n-B utyl-ID A 4-n-Pentyl-ID A 4-t-B utyl-ID A 4-Phenyl-IDAb 4-M ethoxy-ID A 3,5-D im e thyl-ID A 2,4,6-T rim ethyl-ID A 2,4,5-T rim ethyl-ID A 4-Fluoro-ID A 2,4-D ifluoro-ID A 2,5-D ifluoro-ID A 2,3,4,5,6-Pentafluoro-ID A
In testin e s
D istrib u tio n coefficient o c ta n o l/ w ater
433
IAEA-SM-247/20S TABLE III. DISTRIBUTION OF TECHNETIUM -99m-LABELLED BENZIMIDAZOLYL-IDA DERIVATIVES P ercentage d o se in organ a t 1 h (m eans ± S.D. o f 3 to 6 anim als) Blood
S ub stitu en t
Liver
K idneys
G .I.T .+ G.B.
U rine
0.8 ± 0.08 1.8 ± 0.5
84.0 ± 2.9 68.6 ± 6.0
6.0 ± 0 .8 6.7 ± 1.0
10.3 ± 0.5
2.2 ± 0 .1
74.6 ± 2.2
5.1 ± 0.7
8.7 ± 0.3
16.4 ± 0.8
10.4 ± 0.4
14.3 ± 1.0
26.6 ± 1.5
R ats R abbits
0.7 ± 0.1 3.8 ± 1.4
1.3 ± 0.2 3.3 ± 0.3
0.3 ± 0.0 0.5 ± 0.2
92.3 ± 2.2 82.5 ± 0.3
1.7 ± 0 .4 1.5 ± 0 .7
R ats R abbits
1.2 ± 0.3 1.4 ± 0.5
' 2.5 ± 1.5
1.6 ± 0.2 0.8 ± 0.1
92.1 ± 2.5 86.8 ± 1.4
6.1 ± 2.6 1.5 ± 0.4
D i-m ethyl
R ats R abbits
0.9 ± 0.1 5.1 ± 1.3
1.7 ± 0.3 7.2 ± 0.9
n-butyl
R ats
5.1 ± 0.4
Di-me-BIMpropyl
R ats
Chloro
Brom o
3.6 ± 0.9
TABLE IV. DISTRIBUTION OF TECHNETIUM-99m-DIMETHYL- AND Br-BIMIDA IN JAUNDICED RATSa Percentage d ose in organ a t 1 h (m eans ± S.D. fo r 3 rats'0 Sub stitu en t
Blood
(a) D im ethyl
9.0 ± 0.7
(b) Brom o
9.3 ± 2.6
Liver
K idneys
G.I.T.
Urine
7.6 ± 1.3
3.9 ± 0.8
11.4 ± 5 .3
34.2 ± 1.7
10.9 ± 2.5
2.8 ± 0.2
47.7 ± 2.0
5.2 ± 4 .0
a F rom Ref. [20] (reproduced by perm ission). b R ats adm inistered CC14 (50 m g/kg b o d y w eight) for 14 days. Serum bilirubin levels (a) 1.8 ± 0.3 mg%; (b ) 1.3 ± 0.08 mg%.
It was suggested by Loberg th a t th e " T c m -labelled-HIDA is excreted intact into th e bile. This conclusion was reached on the apparent finding th a t when bile containing excreted " T c m-HIDA was injected into another animal, the activity distribution obtained was similar to th a t o f the original com pound. M olter [10] dem onstrated th a t the " T c m-HIDA is broken down at least partially in the urine. It is n o t yet know n w hether any conjugation occurs when " T c m -HIDA is excreted into the bile.
SUBRAMANIAN and McAFEE
434
Rl /С Н Х О О Н R ^ 4 O > N H C 0 - C H 5- N < 2 2 x C H 2COOH R6 Compound
Rl
M
Rfi.
A NILINE
H
H
H
CH3
H
CH3
2 , 6 .D ie th y l
C2H5
II
C2H5
2 , 6 , Di iso p ro p y l
C3H7
H
C3H7
p-Ethyl
H
C2H5
H
p -Iso p ro p yl
H
C3H7
H
p-butyl
H
C4H9
H
p-ethoxy
H
C4H50
H
p-butoxy
H
C4H90
H
2 , 6 , Dimethyl
0-butoxy
F IG .l.
C4H90
H
H
S u bstitu ted acetanilide derivatives o f N -substitu ted IDA.
Compound
Ri
dim ethyl
CH3
CH3
1
n-butyl
C4H9
H
1
d i m ethyl-BIM -propyl
CH3
CH3
3
chi oro-
Cl
H
1
Br
H
1
'bromo
FIG.2.
R2
S u bstitu ted derivatives o f benzim idazolyl-ID A.
n
435
IAEA-SM-247/205
PO3H2 i H- С- H I PO3H2 M DP _„
Р03Н2 H- С - Br I Р03н2 Br M DP
PO3H2 CH3 - NH- С - CH3 I PO3H2 NHEDP
PO3H2 i СНЗ - NH- С -H I PO3H2 NM M DP
PO3H2
P03H2 I N- С- H I P03H2 DM AD
PIO3H2 HO- С- OH
i
РО3 Н2
< r y CH2 - С- H
H -
I С I
- OH
P03H2
P 03H2
BM DP
MHDP
PIO3H2 C2H5 - С- OH I PO3H2 HPO
P O3H2 1 H2N- CH2 - CH2 - С- OH I Р03н2 APD
i
P 03H2
I с =0 I
P03H2 COP
I
P03H2
OHM DP PI03H2 CH3 - С- CH3 P03H2 IPDP
FIG.3. Structural form ulae o f diphosphonates evaluated [ 3 1 - 3 4 ] .
Several clinical evaluations and intercom parison o f " T c m-HIDA complexes have been reported in th e literature [16—19]. According to H ernandez and Rosenthal [17], " T c m -labelled-2,6 diisopropyl-H ID A and 2,6 diethyl-HIDA are the tw o best agents to use for the diagnosis o f hepatobiliary disorder.
NEW BONE IMAGING AGENTS In the last sym posium on radionuclide imaging conducted by the IAEA in 1976 at Los Angeles, a comprehensive survey reported [21] on the preparation and clinical evaluation o f " T c m-labelled bone imaging agents. Since th en clinical evaluation and intercom parisons have been m ade betw een " T c m -labelled m ethylene diphosphonate and o th er bone agents [22—26]. Bone m etabolism studies [27] and localization o f nonosseous lesions [28] were perform ed with " T c m-MDP. Recently Bevan and co-workers [29] reported on a new bone imaging agent, " T c m-labelled h y droxy m ethylene diphosphonate, an analogue o f MDP (Fig.3) w ith biological characteristics similar to those o f MDP. In experim ental animal acute m yocardial infarcts, " T c m -HMDP showed considerably higher Concen tratio n in the cardiac lesions com pared w ith th a t o f other diphosphonates and pyrophosphate. However, later clinical trials [30] indicated th a t " T c m-labelled pyrophosphate is superior to HMDP fo r acute m yocardial infarct imaging.
436
TABLE V. DISTRIBUTION O F TECHNETIUM-99m-DIPHOSPHONATES AND STRONTIUM-85 IN RABBITS S im u ltan eou s s tu d y o f te ch n etiu m -9 9 m /stro n tiu m -8 5 ra tio s 3 h p o s t in jection MDP [8]
APDP [6]
NM MDP [6]
DM ADP [9]
1,1 EDP [4]
HP DP [3]
NM EDP [3]
0.375
0.285
0.272
0.330
0.344
0.4 5 2
0.675
0.167
0.153
0.142
0.169
0.145
0.182
0.338
Fem ur
0.785
0.772
0.778
0.516
0.485
0.553
0.446
Ave. bóne
0 .816
0.820
0.818
0.557
0.542
0.616
0.465
Callus m aterial
0.948
1.122
0.988
0.880
0.897
0.786
-
Callus m a t./tib ia
1.78
1.79
1.86
2.57
2.45
2.15
-
C allus/tibia
1.34
1.19
1.36
1.92
1.78
1.81
1.58
and M cAFEE
Blood Muscle
SUBRAMANIAN
O rgan
IAEA-SM-247/205
437
Wang [31, 32] recently evaluated a variety o f new diphosphonates analogous to MDP w ith substitutions on the m ethylene carbon (Fig.3). These evaluations were carried o u t in rats w ithout any internal standards (85Sr or 18F ) to com pen sate for th e uneven quantitative distribution o f bone imaging agents in individual animals o f different age and weight. Hence, th e intercom parison o f the data of these new bone agents becom es difficult to evaluate. Nevertheless, these studies show th a t o f all the com pounds studied to date b o th MDP and HMDP seem to be superior bone imaging agents. U nterspann [33] reported several new diphosphonates w ith am inom ethane substituents o f bis phosphonic acid. F o u r o f these com pounds showed higher concentration than HEDP in the rat bone. F u rth er clinical evaluations have not been follow ed up. Localization o f " T c m -labelled bone imaging com pounds in norm al bone is a desirable characteristic o f the radiopharm aceutical. However, the concentration o f these agents in bone lesions is m ore im portant. Any com parison or evaluation o f a new bone agent should also include bio distribution studies in animals w ith experim ental bone lesions before they should be recom m ended for clinical trials. Our laboratory recently evaluated [34] six diphosphonates in rabbits with experim ental bone lesions. The com pounds were synthesized by m ethods outlined in the literature [35, 36] and verified by standard analytical m ethods (elem ental analysis, IR and NMR). These com pounds were labelled w ith " T c m using freeze-dried kits containing stannous tin, and quality controls were perform ed before use. Bone lesions in the rabbits were m ade by drilling tw o 1/ 8-in burrholes side by side in the tibia [37, 38]. One week to 10 days later these animals were injected w ith " T c m -diphosphonate and 85Sr (as an internal control), and three hours after injection th e rabbits were sacrificed, the m ajor organs were sampled and assayed. From the norm al tibia a section o f norm al bone, an exact equivalent to the area o f biopsyhole lesion on the o th e r tibia, was taken and assayed for inter com parison. The actual callus m aterial was also removed and assayed separately. The results are shown in Table V. The values reported are " T c m / 8sSr ratios for each sample and these are the only meaningful data suitable for intercom parison betw een " T c m agents. The callus m aterial/tibia ratios were obtained from individual Т с/Sr ratios for each sample. These results indicate th a t o f the six new diphosphonates evaluated, APDP, NMMDP, DMADP and 1,1-EDP seem to con centrate in bone lesions at equivalent or higher levels than MDP, and therefore m ay w arrant fu rth er evaluations in clinical trials.
NEW MYOCARDIAL IMAGING AGENTS Thallium-201 is the current agent o f choice for imaging the m yocardium . However, this nuclide is subject to several lim itations which include high cost and
438
SUBRAMANIAN and McAFEE I-CH2(CH2)ftC02H (n-6-26)
125
01 ARS
I-CHjCCH^j-C^C-CCHsVCOjH
1251-U)-I0D0
FATTY ACIDS
NH ■CHj-NH-C-Ж. н3с-ссн2)7-те-сснг)7-со2н m-IODO RENZYL GUANIDINE
Нз«сн2)ю-те-сснг)4-со2н 15-p-BROMO P H E N Y L PFNTADECANOIC ACID
FIG.4.
Te-LABELLED FATTY ACIDS
Chemical structure o f radiopharm aceuticals discussed.
low gamma energy (6 9 —80 keV X-rays) w ith resultant poor quality gamma camera images o f th e heart. Therefore, a good substitute for 201T1 is desirable. Over the years there has been considerable interest in the developm ent o f fa tty acids labelled w ith radiohalogens, the m ost notable being 123I-labelled co-iodo fatty acids. These radiopharm aceuticals also have lim itations. The deiodination, resulting in high blood levels and short half-life in the m yocardium , is among their drawbacks. During th e last few years considerable effort has been devoted by m any investigators to prepare an im proved and m odified radioiodinated fatty acid (Fig.4). R ecently O tto and co-workers investigated several 125I-labelled oo-iodo fa tty acids for m yocardial localization. T hey varied the chain lengths from 6 —26 carbons and included a triple bond in one o f them . The biodistribution o f these com pounds in rats showed m oderately encouraging results. Their best com pound had a carbon chain length o f 20 and concentrated in the m yo cardium equivalent to th a t o f 201Tl, but w ith considerably higher blood level than th e latter. The triple-bond-containing com pound fared poorly in this evaluation. Because o f th e deiodination problem s associated with oj-iodo fatty acids Stocklin [40] evaluated a 15 carbon chain fatty acid attached to a p-brom ophenyl group containing a brom ine radiolabel. This com pound concentrated well in the m yocardium o f experim ental animals, and its disappearance rate from the heart was also slower than oth er regular fa tty acids. Bromine-75 ( T 1 / 2 98 min) seems to be th e desired nuclide. Machulla and Stocklin [41 ] also evaluated several fatty acids labelled w ith UC, 34Clm , 77Br and 123I. Their results indicated th at o f all the above labels, b o th th e 123I and n C com pounds showed im proved localization in the m yocardium .
IAEA-SM-247/205
439
Knapp and co-workers [42] evaluated several 123Tem-labelled long chain fatty acids for m yocardial localization and studied the effect o f heteroatom position and to tal chain length on biodistribution. Significantly higher concentrations in the heart were achieved w ith com pounds w ith a carbon chain length greater than 15. The concentrations in th e rat heart at 5 min and 60 min for C-18-chain length were 2.1% and 3.1% o f injected dose respectively. In this com pound the 123Tem atom was attached in betw een 8 and 9 carbons. The heart-to-blood ratios for this com pound ranged from 16 at 5 m in to 9 at one hour. Even though 123Tcm-labelled fatty acids are promising agents, the physical characteristics o f 123T em make them less desirable as a 201T1 substitute. O ther com pounds o f different m olecular configurations have been tried as m yocardial localizers by several workers. These include quarternary am m onium salts w ith iodom ethyl group attach m en t [43,44], and derivatives o f b-adrenoreceptor blockers labelled w ith radioiodine [45—47] and phenyl alkyl am ine [48] labelled w ith an iodine radionuclide ( 123I is preferred for imaging). The last com pound is also know n for its localization in the brain. Many research laboratories are actively engaged in developing a " T c m-labelled m yocardial localizing agent to replace 201TI for obvious reasons. R ecently Deutsch [49] reported a cationic complex o f " T c m labelled w ith Diars (Fig.4) (O-phenylene bis dim ethyl arsine) which showed prom ise for m yocardial imaging. Technetium -99-labelled, m ulti valent Diarsine com plexes were prepared by Fergusson and N yholam as early as 1959 [50, 51 ]. However, D eutsch is th e first to prepare a " T c m-labelled Diars and study its biodistribution in experim ental animals. (Diars is a commercially available easily volatile and toxic chemical. It has to be carefully handled, only in properly ventilated hoods.) The labelling m ethod involves reduction o f pertechnetate in concentrated HC1 or HBr and reacting w ith an ethanolic solution of Diars. Some heating is helpful. The " T c m-Diars com plex is extracted into m ethylene chloride and th e excess u n b o u n d Diars is removed by passing through a small alum ina colum n. The " T c m com plex is then eluted w ith dilute ethanol in virtually ‘carrier free’ form. B iodistribution o f " T c m -Diars in rats showed m yocardial concentration o f 2 -2 .5 % o f injected dose at 10 min w ith a biological half-life o f 2 h in the heart. The heart/b lo o d concentration ratio rem ained high (10 to 20) and considerable activity was seen in th e liver. These biodistributions are similar to those obtained with 201 TI. M yocardial images obtained in dogs starting at 5 m in post injection showed excellent visualization o f the heart. F o r all practical purposes, this com pound should be considered as a beginning in the developm ent o f 99 Tcm -labelled m yocardial imaging agents. More interesting com pounds are in the offing, we hope.
440
SUBRAMANIAN and McAFEE
BLOOD CELL LABELLING WITH RADIONUCLIDES In this section a sum mary is given o f the progress made in the last few years on harvesting, labelling and utilization of blood cells for nuclear imaging studies.
Technetium -99m -labelled red cells R ecent interest in cardiovascular nuclear medicine has resulted in considerable progress in th e labelling o f red cells. New and simpler m ethods have been developed for labelling red cells w ith " T c m . Pavel [52] first introduced in vivo labelling o f red cells and obtained good labelling efficiency. His m ethod consisted o f injecting the patient w ith non-radioactive stannous pyrophosphate (bone imaging agent kit) containing at least 1 mg o f stannous tin and injection of " T c m -pertechnetate 30 m inutes later. The m igrated stannous ions inside the red cells labelled the " T c m . The labelling efficiency rem ained as high as 95%. How ever, in later clinical studies others have noticed considerable patient-to-patient variation in the labelling efficiency, and on an average, one can obtain a ‘usable’ in vivo labelling yield o f only 8 0 —85% o f th e injected pertechnetate. Availability o f a good red cell label p rom pted interesting and im portant applications [53] in cardiac cham ber imaging and heart wall m otion studies. Using Pavel’s technique, Armas and co-workers [54] developed a simplified red cell labelling m ethod for splenic imaging. This procedure involved withdrawing 5 - 1 0 ml o f whole blood from the patient after injection o f cold stannous py ro phosphate (containing at least 1 mg tin), incubating the plasma-free cells w ith pertechnetate and denaturing these cells by heating for 10 min at 50°C, similar to a m ethod described by Sm art [55]. Availability o f these labelled cells prom pted further clinical use o f labelled cells for diagnosing gastrointestinal bleeding [56]. F or use in situations in which the bleeding is slow, Winzelberg [57] suggests the use o f ulIn-oxine-labelled red cells. Technetium -99m -labelled red cells have also been used for the m easurem ent o f red cell mass in infants [58]. Sm ith [59] first proposed a red cell labelling kit for in vitro labelling o f plasma-free cells w ith " T c m . In this m ethod consistently greater than 95% labelling efficiency was obtained. It is still the preferred m ethod o f labelling red cells in our laboratory. C olom betti [60] reported a similar in vitro labelling kit using stannous pyrophosphate. Jones [61] com pared four stannous com pounds for in vivo labelling o f red cells w ith " T c m in dogs. O f these com pounds, Sn-DTPA seemed to be the best agent. There was a direct correlation betw een th e q u an tity o f tin adm inistered (irrespective o f the chelating com pound) and labelling efficiency. He determ ined for good labelling yields Sn+ ion concen tratio n o f > 10 jwg/kg body weight should be used. This value incidentally agreed w ith the quan tity o f tin used by Pavel [52].
IAEA-SM-247/205
441
A t the present tim e the in vitro labelling [59, 60, 62] o f red cells seems to be a b etter m ethod because one can determ ine the labelling efficiency before the labelled cells are injected to the patient. Also, in quantitative cardiac studies a consistently high labelling efficiency is highly desirable. L eukocyte labelling with in dium -lll-oxine Since the in tro d u ctio n o f lu In-oxine for leukocyte labelling [63, 64], " T c m-oxine and o th e r " T c m radiopharm aceuticals [65—69] have also been suggested for the same purpose. However, the excellent cell labelling characteristics o f ulIn-oxine make it a b etter cell labelling agent. Several studies have been reported on the preparatio n and labelling o f leukocytes each w ith a m inor m odifi cation in m ethodology [70—72]. O ur laboratory recently evaluated the different m ethods o f labelling techniques in harvesting blood cells, and reported the results at a sym posium [73]. The finalized procedures for leukocytes have also been evaluated in dogs w ith acute inflam m atory lesions. Preparation o f in diu m -oxine
A to tal o f 1.5 mCi carrier-free m In-chloride (0.75) is m ixed w ith 50 pg o f oxine in 50 //1 o f ethanol followed by 0.2 ml sodium acetate solution and mixed well. The m In-oxine com plex is extracted in to 1 ml o f m ethylene chloride and evaporated to dryness. The dry com pound is then redissolved in 50 jul o f ethanol using a vortex m ixer. Preparation o f in diu m -oxine sulphate
Ducassou [75] recently suggested the use o f w ater soluble oxine sulphate instead o f oxine. This m ethod was m odified as follows: Tris buffer solution pH 7.4, 0.2M strength containing 2.6 m g/m l NaCl was used for the cell labelling m edium . A to ta l o f 50 jug o f oxine sulphate in 50 /ul w ater was m ixed w ith 0.75 ml o f m In-chloride (1 —1.5 mCi, 0 .1 —0.05M HC1), and an equal volume of Tris buffer was added resulting in a final pH o f 7.0 —7.4 w ith an osm otic pressure o f 320 m O sm .1 Leukocyte separation from whole blood can be accom plished by sedim enta tion using h ydroxy ethyl starch [76] or m echanical élutriation [77] techniques. The form er technique provides only m ixed cell suspensions w ith m uch red cell contam ination (1:1 w hite cell to red cell). The élutriation procedure using a Beckman J-21C centrifuge w ith a JE-6 elu triato r ro to r provided a ‘clean’ neutro phils preparation w ith only 2—5% erythrocytes and 3 -4 % m onocytes or large lym phocytes [74]. 1 1 Ci=3.70X 101OBq.
442
TABLE VI. ABSCESS LOCALIZATION AT 24 h BY TISSUE ASSAY3 (m eans ± sta n d a rd error o f th e m ean) 67G a-citrate
m In-oxine leu k o cy tes E lutriated cells
N o. o f dogs
6
HES cells
4 0 0 ± 142
389 ±
% Dose X 103/g
E. coli
543 ± 102
A bscess/blood ratio
78
423 ±
79
483 ±
64
385 ±
74
101 ±
21
21
4
±
7.3
8.2 ± 3.6
8.9 ±
2.4
5.5 ± 1.0
13.7 ± 4.3
11.4 ± 2.5
jo in t
299 ± 76
489 ±
chem .
108+
38
95 ±
20
E. coli
147 ±
29
103 ±
19
125±
81 ±
21
149 ±
23
108±
2444 ± 977
2575 ±
142
2 509 ± 471
182
±70
286
3 0 9 4 ± 323
72
± 19
7.4 ± 1.6
2 2 9 0 ± 649
100
±31
12.2 ± 3.7
chem . E. coli
3293 ± 602
2895 ±
jo in t
1468 ± 6 1 5
3112 ± 1086
1.7 ± 0.72
4.9 ±
1.7
18
2.4 ±
0.76
1.2 ± 0.21
18
3 .4 +
1.1
2 .4 ± 0.45
chem .
72 ± 615
217 ±
103
144±
55
E. coli
95 ±
14
253 ±
127
174 ±
65
jo in t
86 ± 46
287 ±
142
177±
72
25
±9.1
8.4 ± 17
±
10.7 ± 4.4
2.3 ± 0 .3
2.5
2.0 ± 0.4
7.1
6.5 ± 3.2
a
F ro m Ref. [7 4 ] (rep ro d u c e d by perm ission). A t th e tim e o f sacrifice, the chem ical and E. coli abscesses w ere 48 hours old, and the jo in t lesions 26 hours old.
and M c A F E E
ra tio
12
394 ±
129
jo in t A bscess/m uscle
12 83
SUBRAMANIAN
chem .
% D ose/1% b o d y w t.
C om bined series e lu triated & HES cells
6
Abscess c o n ce n tra tio n
lllIn-chloride
IAEA-SM-247/205
443
The cells separated by either m ethod were labelled w ith b o th inIn-oxine preparations w ith high labelling yields by simple mixing and incubating for 10—15 m in. Final centrifugation was used to rem ove the unbound U1ln. The viability o f this labelling technique has been assessed in dogs with inflam m atory lesions [74]. T hree types o f lesions were induced in mongrel dogs. A chemical abscess was made by injecting turpentine emulsion, and a bacterial abscess by inoculation o f E. coli into thigh muscles o f dogs. A third type o f lesion was made by injection o f sodium urate crystals (10—15 p m ) suspended in saline in to knee joints. Blood from the dogs w ith inflam m atory lesions was collected 24 h later, and the leukocytes isolated and labelled with luIn-oxine as described above and reinjected to the same dogs. The next day the animals were sacrificed and organs assayed. Samples from the three types of lesions were collected and counted. In all dogs 67Ga-citrate was simultaneously injected to serve as a control. Whole blood and w hite cell bound radioactivities were assayed over a 24-h period. The results are shown in Table VI and Figs 5 and 6. There was no significant difference in the localization o f elutriated cell or mixed cell suspension after HES sedim entation. The abscess-to-muscle and abscess-to-blood ratios were also similar. In addition, use o f oxine sulphate instead o f oxine also did n o t sign ifican tly change the biodistribution. A t present, the clinical applications o f m In-labelled leukocytes are progressing slowly and are used routinely only in m ajor m edical centres [78, 79]. It is to be hoped th at the commercial availability o f preprepared ready-to-use ulIn-oxine in buffer solutions (at least in E urope) will speed up widespread use o f this im portant technique for abscess localization. L abelling o f p la te le ts an d ly m p h o c y te s R a d io a c t i v e - l a b e l l e d p l a t e l e t s c a n b e u s e f u l in lo c a liz in g a v a r i e t y o f v a s c u la r
lesions [8 0 - 9 4 ] including throm bosis and atherogenesis. Platelet aggregation and m icrothrom bosis are im p o rtan t also in m yocardial infarction, transient ischemic attacks, in glom erular and vascular renal diseases and in renal failure from shock, burns o r toxaem ias. They may also localize in developing haem atogenous m etas tasis where th e platelet aggregation on the tu m o u r cells may be essential. Because o f this im portance o f platelets, several investigators attem pted to label these cells w ith gam m a-em itting radionuclides IUIn and " T c m . In the past few years, m any m ethods have been proposed for harvesting and labelling platelets. Scheffel [91 ] recently reported a useful day-to-day workable procedure. In this m ethod a small volum e o f blood, 4 0 - 4 5 ml, is m ixed w ith 6 ml ACD (NIH-A) and spun at 225 g fo r 15 m in to isolate the PRP. However, if platelets are m eant to be used in kinetics studies one should follow the recom m endations o f th e ICSH [92] and use 500 m l o f blood for isolating a large num ber o f cells. R ecently Heyns [87] dem onstrated th a t by following the ICSH m ethod
444 and M c A F E E
Dose
in Whole
Blood
SUBRAMANIAN
FIG.5. R ecovery o f activity (A ) in canine w hole b lo o d fo r m In-oxine-labelled elu triated neutrophils and m ixed cells after HES sedim entation , com pared w ith 61Ge and nlIn-chloride; (Bj R ecovery in circulating cells (whole blood minus plasma a c tiv ity). R e p ro d u c e d from Ref. [ 74] b y permission.
445
IAEA-SM-247/205 io ’° ♦О *-
Ö
. •
о
♦
x Ю'
,I0‘ 111 In -o x in e • elutriated neutrophils О leukocytes öfter H E S □ leukocytes after H E S ; oxine in saline ф sulp hate ; elutriated cells x whole blood * p la s m a + saline ▲ é lu triatio n after labelling
lo£
111 In - chtoride ¿ t r a n s f e r r in
10
% FIG. 6.
20
30
40
50
Dose in C irculating Cells at
R elation betw een cell-bound in In a ctivity in the blood a t 4 hours and abcess: muscle
concentration ratio a t 2 4 hours fo r E. coli abcesses in dogs. R eprodu ced from R e f [74] by permission.
excellent in vivo platelet viability could be obtained in hum ans. He determ ined th at 72 + 16% o f th e injected inIn-oxine-labelled platelets rem ained in the bloodpool w ith a survival tim e o f about 9 d. This is one o f the best results obtained to date w ith m In-labelled platelets in hum ans. There seem to be disagreem ent on the exact suspending m edia for labelling. Some prefer to use T yrode-album in buffer solution and others like to use an ACD-containing plasma. Sinzinger [94] recently evaluated the use o f prostocyclin (PG I2 , 25 ng/ml) during th e labelling step and found no significant change in labelling rate o r yield. A good m ethod w ould be to use small volumes o f blood (50 ml) and obtain an in vivo survival tim e equivalent to th a t o f 51Cr-label (9 d). W ith these developm ents m In-platelets also may find diagnostically useful applications in radionuclide imaging. Considerable interest has been directed tow ards labelling lym phocytes w ith a suitable radionuclide and especially w ith u lIn-oxine [95—99]. In this sym posium recent research w ork on the subject is presented by G oodw in and co-workers.2 2 GOODW IN, D.A., HECKM AN, J.R ., F A JA R D O , L .F ., CALIN, A ., PR O PST, S.J., DIAM ANTI, C.I., “ K inetics and m igration o f in d iu m -1 1 1-labelled hum an lym phocytes” , these Proceedings, IA EA -SM -247/95.
446
SUBRAMANIAN and McAFEE
There seems to be a need to develop a simpler isolation technique th an th e FicollH ypaque m ethod for harvesting b o th T & В cells separately, and keeping them viable after labelling for in vivo imaging and kinetic studies. This goal, if achieved, will present a new tool in understanding th e mechanisms involved in the im m unolo gical functions o f lym phocytes. Sin and Silvester [100] recently reported on the use o f n lIn-acetylacetone as an o th er com plex o f indium for cell labelling. Mathias and co-workers [101] re-examined these procedures by com parative studies w ith indium -oxine in dogs. T heir results showed considerable agreem ent betw een indium -acetylacetone and indium-oxine-labelled cells for b o th neutrophils and platelets. The quality control o f prepared indium -oxine solution is difficult to perform , and there is no simple system available except for the back extraction technique in w hich indium -oxine is extracted into organic solvents from aqueous solutions. A chrom atography m ethod proposed by Tosch and co-workers [102] did n o t w ork well in our laboratories. F u rth er developm ent in this area is needed.
ACKNOWLEDGEMENTS The authors wish to th an k G. Gagne, T. Feld, E. H ofm ann, C. Z apf and M. R oskopf for their technical assistance.
REFERENCES [1]
[2]
[3]
[4] [5]
[6]
[7]
LOBERG, M.D., COO PER, M., H A RVEY, E., e t al., D evelopm ent o f new radio pharm aceuticals based on N -substitution o f im inodiacetic acid, J. Nucl. Med. 17 (1 9 7 6 ) 633. WISTOW, B., SUBRAM ANIAN, G., M cA FEE, J.G ., et al., A n evaluation o f Tc-99m labeled HIDA derivatives as hepatobiliary agents in experim ental anim als, J. Nucl. Med. 17 (1 9 7 6 ) 545 (abstract). CA LLERY , P.S., FA ITH , W.C., LOBERG, M.D., et al., Tissue d istrib u tio n of Tc-99m and C-14 labeled N (2,6 dim ethylphenyl carbam oylm ethyl) im ino diacetic acid, J. Med. Chem. 19 (1976) 962. WISTOW, B.W., SUBRAM ANIAN, G., VAN HEERTU M , R.L., et al., A n evaluation of Tc-99m labeled hepatobiliary agents, J. Nucl. Med. 18 (1977) 455. FIE L D S, A .T., PO R T E R , D.W., C A L L ER Y , P.S., e t al., Synthesis and radiolabeling of Tc-99m radiopharm aceuticals based on N -substituted im inodiacetic acid: effect of radiolabeling conditions on radiochem ical pu rity , J. Labelled C om pd. 15 (1 9 7 8 ) 387. SUBRAM ANIAN, G., M cA FEE, J.G . H END ERSON, R.W., T he influence o f structural changes on b io d istribution o f Tc-99m labelled N -substituted IDA derivatives, J. Nucl. Med. 18 (1977) 624. BURNS, H.D., W ORLEY, P., W AGNER, H.N ., “ Design o f tech n etiu m radiopharm a ceuticals’, T he C hem istry o f R adiopharm aceuticals, M asson Publishing, New Y ork (1977) 269.
IAEA-SM-247/205
447
SUBRAM ANIAN, G., M cA FEE, J.G ., H EN D ERSO N , R.W., et al., “ T he influence of stru ctu ral changes on bio d istrib u tio n o f Tc-99m labeled N -substituted IDA derivatives, N uklear M edizin (W OLDRING, M., SCHM IDT, H .A .E., Eds), 15th In tern a tio n a l A nnual M eeting o f the Society o f N uclear M edicine, G roningen, S ept. 1977, F.K . Schattauer-V erlag, Stuttgart-N ew Y ork (.1978) 136. VAN WYK, A .J., FO N R IE , P.J., VAN ZYL, W.H., et al., Synthesis o f five Tc-99m HIDA isom ers and com parison w ith T c-99m H ID A , E ur. J. Nucl. Med. 4 (1 9 7 9 ) 445. M O LTER, M., KLOSS, G., “ Studies of the pharm acokinetics o f various Tc-99m IDA derivatives” , A bstracts o f T hird In tern a tio n a l Sym posium on R adiopharm aceutical C hem istry, 16—20 Ju n . 1980, St. Louis, Missouri, M O LTER, M., KLOSS, G., “ P ro p erties o f various IDA derivatives” , A bstracts of T hird In tern a tio n a l Sym posium on R adiopharm aceutical C hem istry, 16—20 Jun. 1980, St. Louis, M issouri, 56. WISTOW, B.W., SUBRAM ANIAN, G., GAGNE, G., et al., E x p erim en tal and clinical trials of new T c-99m labeled hepatobiliary agents, R adiology 128 (1 9 7 8 ) 793. LOBERG, M.D., F IE L D S , A .T., Chem ical stru ctu re of Tc-99m labeled N (2,6 dim ethylphenyl carbam oylm ethyl) im ino diacetic acid (Tc-HIDA), In t. J. A ppl. R adiat. Isot. 29 (1 9 7 8 ) 167. FR JTZ B E R G , A .R ., HUCKABY, D., “ D evelopm ent and results of ro u tin e quality co n tro l p rocedures for T c-99m im inodiacetate h ep ato biliary agents” , R adiopharm a ceuticals II (Proc. Sym p. Seattle ), T he Society of N uclear M edicine, New Y ork (1979) 545. K O UTOULIDS, C., C H IO TELLIS, E ., LYM BERIS, C., A bsorbed dose estim ation of some 99mT c-hepatobiliary agents, Eur. J. Nucl. Med. 4 (1 9 7 9 ) 441. RYAN, J., CO O PER, M., LO BERG , M.D., et al., T echnetium -99m labeled N -(2,6 dim eth y lp h en y l carbonoyl m eth y l) im inodiacetic acid (T c-99m H ID A): A new radiopharm aceutical for h ep ato b iliary imaging studies, J. N ucl. M ed. 18 (1 9 7 7 ) 997. H ERN A N D EZ, M., RO SEN TH A L, L., A crossover stu d y com paring th e kinetics of Tc-99m labeled D iethyl and D iisopropyl-ID A , Clin. N ucl. Med. 5 (1 9 8 0 ) 352. H ER N A N D EZ, M., RO SEN TH A L, L., A crossover study com paring th e kinetics of Tc-99m labeled D iisopropyl and Р-B utyl IDA analogs in patien ts, Clin. N ucl. Med. 5 (1980) 159. TA A V ITSA IN EN , M., K O RH O LA , O., RIIH IM A K I, E., et al., T c-99m dieth y l HIDA cholescintigraphy in th e differential diagnosis of jaundice, Scand. J. G astroenterol. 14 (1 9 7 9 ) 567. HUNT, F.C ., W ILSON, J.G ., M A DDA LENA , D .J., “ S tru c tu re activity relationships for Tc-99m benzim idazoyl m eth y l im ino diacetic acid hepato b iliary radiopharm aceuticals” , A bstracts o f T h ird In tern a tio n a l Sym posium on R adiopharm aceutical C hem istry, 16—20 Ju n . 1980, St. Louis, Missouri. SUBRAM ANIAN, G., M cA FEE, J.G ., BLAIR, R .J., THOM AS, F .D ., “ R adiopharm a ceuticals for bone and bone m arrow imaging: A review” , M edical R adionuclide Im aging (Proc. Sym p. Los Angeles, 1976) 2, IA E A , V ienna (1 9 7 7 ) 83. SUBRAM ANIAN, G., M cA FEE, J.G ., BLA IR, R .J., et al., T echnetium 99m m ethylene dipho sp h o n ate — a superior agent for skeletal imaging, J. N ucl. M ed. 16 ( 1975) 744. RO SEN TH A L, L., ARZO M N A N IA N , A., LISBON A, R., et al., A longitudinal com parison o f th e kinetics o f T c-99m MDP and Tc-99m HEDP in hum ans, CUn. Nucl. Med. 2 ( 1 9 7 7 ) 233. K ELLEY , R .J., C H ILTO N , H.M., HACKSHAW , B.T., et al., C om parison o f Tc-99m p y ro p h o sp h a te and T c-99m m ethylene d iphosphonate in acute m yocardial infarction, J. Nucl. Med. 21 (1 9 7 9 ) 402.
448
SUBRAMANIAN and McAFEE
[25]
SINGH, A ., U SHER, M., C om parison o f T c-99m m ethylene dip h o sp h o n ate w ith Tc-99m p y ro p h o sp h ate in the detection o f acute m yocardial infarction, J. Nucl. Med. 1 8 (1 9 7 7 ) 790.
[26]
RUDD, T .G ., A LLEN , D .R., SM ITH, F.D ., Tc-99m labeled m ethylene diphosphonate and h y d roxyethilidine d iphosphonate — biologic and clinical com parison, J. Nucl. Med. 2 0 (1 9 7 9 ) 821 (Concise com m unication). CAN IGGIA , A., VATTIM O, A ., K inetics o f T c-99m -tin-m ethylene: dipho sp h o n ate in norm al subjects and pathological conditions: a sim ple index of bone m etabolism , Calcif. Tissue. In t. 30 (1 9 8 0 ) 5. DESAI, A., EYM O NTT, M., A LA VI, A., e t al., 99mT c MDP u p tak e in nonosseous lesions, Radiology 135 (1 9 8 0 ) 181. BEVAN, J.A ., T O FE , A .J., FRA N CIS, M.D., et al., “Tc-99m h y droxym ethylene d iphosphonate (HM DP): A new skeletal imaging agent” , R adiopharm aceuticals II (Proc. Sym p. S eattle), T he S ociety of N uclear M edicine, New Y ork (1 9 7 6 ) 645. WAKAK, M., C H ILTO N , H.M., HACKSHAW, B.T., et al., C om parison o f Tc-99m py rophosphate and T c-99m hyd ro x y m eth y len e d iphosphonate in acute m yocardial infarction, J. Nucl. Med. 21 (1 9 8 0 ) 203. WANG, T .S .T ., M O JD EH I, G .E., FAWWAZ, R .A ., et al., A study o f the relationship betw een chem ical stru ctu re and bone localization o f Tc-99m -Sn diphosphonic acids, J. Nucl. Med. 2 0 ( 1 9 7 9 ) 1066. WANG, T .S.T ., FAWWAZ, R .A ., JO HNSON, L .J., et al., Bone seeking pro p erties of Tc-99m carbonyl diphosphonic acid, d ihydroxy-m ethylene diphosphonic acid, J. Nucl. Med. 21 (1 9 8 0 ) 767. UN TERSPANN, S., E x perim ental exam inations o f the suitability of organoam inom ethane bis phosphonic acid for bone scintigraphy by m eans o f T c-99m in animals, Eur. J. Nucl. Med. 1 (1 9 7 6 ) 15 1. SUBRAM ANIAN, G., M cA FEE, J.G ., FE L D , T., et al., A n evaluation o f five new d iphosphonates for skeletal imaging, J. Nucl. Med. 21 (1 9 8 0 ) 15 (abstract). PL O G ER , V.W., SCH IN D LER, N., W OLLMANN, K., et al., H erstellung von 1-am inoalkan 1,1 diphosphonsauren, Z .A norg. Allg. Chem. 389 (1 9 7 2 ) 119. GRO SS, V.H., C O STISELLA , B., GNAUK, T ., e t al., D erivatives der am inom ethane bis phosphonsaure, J. P rakt. Chemie. 318 1 (1 9 7 6 ) 116. CLARK, C.R., M ORGAN, C., SO NSTEGARD , D.A., et al., T he effect o f biopsyhole shape and size on bone strength, J. Bone Jo in t Surg. 5 9 A (1 9 7 7 ) 213. BU RSTEIN , A.H., C U R R E Y , J., FR A N K E L , V.H., et al., B one strength, J. Bone Jo in t Surg. 54A (1 9 7 2 ) 1143. OTTO, C.A., BROWN, L .E ., W IELAND, D .J., et al., “ S tru ctu re-d istrib u tio n study of
[27]
[28] [29]
[30]
[31]
[32]
[33]
[34] [35] [36] [37] [3 8 ] [39]
[40]
[41]
[42]
1-125 co-Iodo-fatty acids” , T hird In tern a tio n a l Sym posium on R adiopharm aceutical C hem istry, 16—20 Jun. 1980, St. Louis, M issouri (1 9 8 0 ) 43 (ab stract); See also J. Nucl. Med. 21 (19 8 0 ) 14 (abstract). STOCKLIN, G., COHEN, H.H ., CO EN EN , M .F., et al., 15-(p-Brom ophenyl-) p entadecanoic acid: A new p o ten tial agent for m yocardial imaging, J. Nucl. Med. 20 (1 9 8 0 ) 58 (abstract). MACHULLA, H .J., STOCKLIN, G., K U PFE R N A G E L, C.H., et al., C om parative evalu ation o f fa tty acids labeled w ith C -l 1, Cl-34m , Br-77 and 1-123 for m etabolic studies o f the m yocardium , J. Nucl. Med. 19 (1 9 7 8 ) 298. KNAPP, F .F ., Jr., BU TLER, T .A ., AM BROSE, K .R ., e t al., “M yocardial u p tak e of 123mTe labeled fa tty acids: E ffects of h etero atom position and to ta l chain length” , T hird In tern a tio n a l S ym posium on R adiopharm aceutical C hem istry, 1 6 -2 0 Jun. 1980, St. Louis, M issouri (1 9 8 0 ) 175 (abstract).
IAEA-SM-247/205
449
H U ANG , C.C., FRIED M A N , A.M., RAY UDU , C .R .S., et al., E valuation o f iodom ethyl trim e th y l am m onium 1-131 for m yocardial imaging, J. Nucl. Med. 1 9 (1 9 8 0 ) 717. H U ANG, C.C., FR IED M A N , A.M., RAY UDU , C .R .S., et al., “ R adiohalogenated q u a rte rn ary am m onium s as a p o ten tial m yocardial imaging agent” , T hird In tern a tio n a l Sym posium on R adiopharm aceutical C hem istry, 1 6 - 2 0 Jun. 1980, St. Louis, M issouri (1 9 8 0 ) 265 (abstract). JIA N G , V.W., GIBSON, R .E ., RZ ESZO TA R SK I, W .J., et al., R adioiodinated derivatives of b -adrenoreceptor blockers for m yocardial imaging, J. Nucl. Med. 19 (1 9 7 8 ) 918. W IELAND, D.M., WU, J.L ., BROWN, L .E .L ., et al., R adiolabeled adrenergic neuronblocking agents: A drenom edullary imaging w ith 131I-Iodobenzyl guanidine, J. Nucl. Med. 21 (1 9 8 0 ) 349. HA NSO N, R .N ., DAVIS, M.A., COHEN, Z., C ardioselective radiolabeled b-adrenergic antagnostics as p o ten tial m yocardial agents, J. Nucl. Med. 21 (1 9 8 0 ) 58. W INCHEL, H.S., BALDW IN, R.M ., LIN , J.H ., L ocalization o f 1-122 io dophenyl alkylam ine in rat brain, J. Nucl. Med. 21 (1 9 8 0 ) 68 (abstract). DEUTSCH, E ., GLAV AN, K.A ., FE R G U SA N , D.L., et al., D evelopm ent of a Tc-99m m yocardial imaging agent to replace Tl-201, J. N ucl. Med. 21 (1 9 8 0 ) 56 (abstract). FER G U SSO N , J.E ., NYHOLM , R .S., New oxid atio n states o f technetium , N ature, L ondon (A pril 1959) 1039. FER G U SSO N , J.E ., NYHOLM , R .S., P entavalent technetium , C hem istry ind. (M arch 1960) 347. PA VEL, D.G., ZIM M ER, A.M., PA TTER SO N , V.N., In vivo labeling of red blood cells w ith Tc-99m : A new approach to blood p ool visualization, J. Nucl. Med. 18 (1 9 7 7 ) 305. LA RSEN , S.M., HA M ILTO N, G.W., et al., K it labeled Tc-99m red blood cells (Tc-99m RBC’s) fo r clinical cardiac cham ber imaging, Eur. J. N ucl. Med. 3 (1978) 227. ARM AS, R .R ., T H A K U R , M .L., GO TTSCH ALK, A., A sim plified m eth o d o f selective spleen scintigraphy w ith Tc-99m labeled ery th ro cy tes: clinical applications, J. Nucl. Med. 21 (1 9 8 0 ) 413. SM ART, R.C., T he use o f T c-99m labeled h eat dam aged red cells for th e quantitative m easurem ent o f splenic function, E ur. J. Nucl. Med. 1 (1 9 7 6 ) 211. W INZELBERG , G.G., McKUSICK, K.A., STRAU SS, H.W., E valuation o f gastro intestinal bleeding by red cell blood cells labeled in vivo w ith Tc-99m , J. Nucl. Med. 2 0 (1 9 7 9 ) 1080. W INZELBERG , G .G ., CASTRO NAV , F.P., CALLAHA N, R .J., et al., In -1 11 oxine labeled red cells fo r d e te ctio n o f sim ulated low er gastrointestinal bleeding in an anim al m odel, R adiology 135 (19 8 0 ) 455. L INDERKA M P, О., e t al., Tc-99m labeled red blood cells for the m easurem ent of red cell mass in new born infants, J. Nucl. Med. 4 (1 9 7 9 ) 187 (Concise com m unication). SM ITH, T .D ., RIC H A RD S, P., A sim ple k it for the prep aratio n o f Tc-99m labeled red blood cells, J. Nucl. Med. 17 (1 9 7 6 ) 126. COLOM BETTI, L .G ., “ A simple technique for ‘in v itro ’ labeling o f e ry th ro c y te s” , N uklear M edizin (SCHM IDT, H .A .E ., W OLDRING, M., Eds), 15th In tern a tio n a l A nnual M eeting of the Society o f N uclear M edicine, F.K . Schattauer-V erlag, S tuttgart-N ew Y ork (19 7 8 ) 171. JO N ES, A.G ., DAVIS, M.A., U REN , R .F ., et al., “ C om parison o f fo u r stannous com pounds for in vivo red cell labeling w ith T c-99m ” , N uklear M edizin (SCHM IDT, H .A .E., W OLDRING, M., Eds), 15th In tern a tio n a l A nnual M eeting o f the Society o f N uclear M edicine, F.K . Schattauer-V erlag, Stuttgart-N ew Y ork (1978) 175.
450 [62] [63] [64] [65]
[66] [67]
[68] [69] [70]
[71] [72] [73]
[74] [75]
[76]
[77]
[78] [79]
[80] [81]
SUBRAMANIAN and McAFEE H EN D ER SH O FT, L .R ., et al., A n u ltra filtratio n technique for labeling red blood cells w ith Tc-99m , E ur. J. Nucl. Med. 4 (1 9 7 9 ) 187. M cAFEE, J.G ., T H A K U R, M .L., Survey o f radioactive agents fo r in v itro labeling of phagocytic leukocytes: soluble agents, J. Nucl. Med. 17 (1 9 7 6 ) 480. SEGAL, A.W., A R N O T , R .N ., T H A K U R, M .L., et al., In d iu m -1 11 labeled leukocytes for localizations of abscesses, L ancet 13 (19 7 6 ) 1056. SUBRAM ANIAN, G., M cAFEE, J.G ., GAGNE, G., e t al., “ Tc-99m oxine: a new liphophilic radiopharm aceutical for labeling leukocytes and p latelets” , N uklear M edizin, Proc. 14th In tern atio n al M eeting o f the Society o f N uclear M edicine, Berlin, 1 5 -1 8 Sep. 1976. UCHIDA, T., VINCENT, P.C., In vitro studies o f leukocyte labeling w ith Tc-99m , J. Nucl. Med. 1 7 (1 9 7 6 ) 730. PA PIERN IA K , C.K., BOU REY , R.E., K R ETSC H N ER , R .R ., e t al., T echnetium 99m labeling o f hum an m o n o cy tes fo r ch em otactic studies, J. Nucl. Med. 17 ( 1976) 988. UCH IDA, T ., NEM OTO, T ., Y U I, T., et al., Use of T c-99m as a radioactive label to study m igratory p a tte rn s o f leucocytes, J. Nucl. Med. 20 (19 7 9 ) 1197. L IN H A R T, N., ВОК, B., GO U G ERO T, M., et al., Tc-99m labeled hum an leukocytes, an in vitro fu nctional study, A cta. H aem atol. 63 (1 9 8 0 ) 71. ZA K H RIEH , B., T H A K U R , M .L., MALECH, H .L., et al., In d iu m -1 11 labeled hum an p o lym orphonuclear leukocytes: viability, random m igration, chem otoxis, bacterial capacity and u ltra stru ctu re , J. Nucl. Med. 20 (1 9 7 9 ) 741. T H A K U R , M .L., L A V EN D ER, J.P ., A RN O T, R.N ., et al., In d iu m -1 11 labeled autologous leukocytes in m an, J. Nucl. Med. 1 8 (1 9 7 7 ) 1012. COLEM AN, R .E ., W ELCH, D., Possible pitfalls w ith clinical imaging of I n -1 11 labeled leukocytes, J. Nucl. Med. 21 (1 9 8 0 ) 122. M cA FEE, J.G ., “ T echniques o f harvesting platelets and neu tro p h ils and labeling w ith In -1 11 o xine”, In d iu m -1 11 L abeled N eutrophils, Platelets and L ym phocytes (T H A K U R , M.L., GO TTSCHALK, A., E ds)(Proc. Sym p. 1979), T rivariah Publishing Co., New Y ork (in press). M cA FEE, J.G ., GA GNE, G., SUBRAM ANIAN, G., et al., D istribution o f In -1 11 oxine labeled leukocytes in dogs w ith acute in flam m atory lesions, J. Nucl. Med. (in press). DUCASSOU, D., N O VEL, J.P ., BREN D EL, A., “ Le m arquage des elem ents figures du sang p u r la indium rad io actif - m éthodologie - résultats - indications” , R adioaktive Iso to p e in K linik u n d F orschung 13 1 (1 9 7 8 ) 91. ROY, A .J., FR A N K L IN , A., SIMMINS, W.B., et al., A m ethod o f separation of granulocytes from norm al hu m an blood using h y d ro x y e th y l starch, Preparation Biochem . 1 (1971) 197. JEM IO N EK , J.F ., CO N TR ER A S, T .J., FR E N C H , J.E ., et al., T echnique fo r increased granulocyte recovery from hum an w hole blood by counterflow centrifugation elutriatives: I. in vitro analysis, T ransfusion 19 ( 1979) 120. W EIBLAN, B .J., FO STRUM , L., M cCULLOUGH, J., Studies o f the kinetics o f In -1 11 labeled granulocytes, J. L ab. Clin. Med. 9 4 (1 9 7 9 ) 246. SCH ROTH, H .J., M U LLER, K.P., BERBERICH , R., et al., T he applications of radioactive labeled leukocytes for the p ro o f o f inflam m ations, Eur. J. Nucl. Med. 4 (19 7 9 ) 359. T H A K U R, M.L., WELCH, M .J., JO IST, J.H ., e t al., In -1 11 labeled platelets: studies on p rep aratio n and evaluation o f in vitro and in vivo functions, T hrom b. Res. 9 (1 9 7 6 ) 345. GOODW IN, D.A., BUSHBERG, J., D O H ERTY , P.W. e t al., In -1 11 labeled autologous platelets for location o f vascular th rom bi in hum ans, J. Nucl. Med. 19 (1 9 7 8 ) 626.
IAEA-SM-247/20S [82]
451
[93]
KN IG H T, L.C., PRIM EA U, J.L ., SIE G E L , B.A., e t al., C om parison o f In -1 11 labeled platelets and iodinated fibrinogen for the d etectio n o f deep vein throm bosis, J. Nucl. Med. 1 9 (1 9 7 8 ) 391. RIBA, A.L., T H A K U R , M.L., GO TTSCH ALK, A., Im aging e x perim ental infective e ndocarditis w ith In -1 11 labeled blood cellular com ponents, C irculation 59 (1 9 7 9 ) 336. RIBA, A.L., T H A K U R, M.L., GO TTSCHALK, A., Im aging e x perim ental coronary artery throm bosis w ith I n -1 11 labeled platelets, C irculation 60 ( 1979) 762. DAVIS, H .H ., SIEG EL, B.A., SHERM AN, L.A ., et al., S cintigraphic d e te ctio n o f carotid atherosclerosis w ith indium labeled autologous platelets, C irculation 61 (1 9 8 0 ) 982. HEATON , W.A., D A VIS, H.H ., W ELCH, M .J., et al., In d iu m -1 11: a new radionuclide label for studying hum an platelet kinetics, Br. J. H aem atol. 42 (1 9 7 9 ) 613. HEYN S, A., DUP., LO T TE R , M.G., BAD EN HORST, P.N ., et al., K inetics, d istribution and sites of destru ctio n o f In -1 11 labeled hum an platelets, Br. J. H aem atol. 44 (1980) 269. WISTOW, B.W., GROSSM AN, Z.D ., M cA FEE, J.G ., e t al., L abeling o f platelets w ith oxine com plexes o f T c-99m and In -1 11: I . I n v itro studies and survival in th e rabbit, J. N ucl. Med. 19 (1 9 7 8 ) 483. W ILKINSON, A .R ., HAW KER, R .J., HAW KER, L.M., In -1 11 labeled canine platelets, T hrom b. Res. 13 (19 7 8 ) 175. SC H EFFE L , U., TSAN, M., M cIN TY RE, P.A ., L abeling o f hum an p latelets w ith In -1 11 8-hydroxyquinoline, J. Nucl. Med. 20 (1 9 7 9 ) 524. IN T E R N A T IO N A L COM M ITTEE F O R STA N D A RD IZA TIO N IN HEM A TO LO GY, R ecom m ended m eth o d s fo r radioisotope p late le t survival studies, B lood 50 (1 9 7 7 ) 1137. M cA FEE, J.G ., “ Im p o rtan ce o f cell labeling techniques”,In d iu m -111 L abeledN eutrophils,
[94]
P latelets and L y m p h o c y tes (T H A K U R , M .L., GO TTSCH ALK, A., E ds)(Proc. Sym p. 1979), T rivariah P ublishing C o., New Y ork (in press). Z IN Z IN G E R , H., A N G E LB E R G E R , D., KOLBE, H ., “ M ethodological study of
[83] [84] [85]
[86] [87] [89]
[90] [91] [92]
[95] [96]
[97] [98] [99]
[100] [101]
[102]
In -111 o xine hum an platelet and leu k o cy te labeling” , T hird In tern a tio n a l Sym posium on R adiopharm aceutical C hem istry, 1 6 - 2 0 Ju n . 1980, St. L ouis, M issouri (1 9 8 0 ) 69 (abstract). SAHA, G.B., SCH ELL, E .T ., F A R R E R , P.A ., L abeling o f ly m p h o cy tes w ith m ercury 197, J. N ucl. Med. 1 8 (1 9 7 7 ) 70. GOBUTY, A.H., ROBINSON, R .G ., BA RTH, R .F ., O rgan d istrib u tio n o f T c-99m and C r-5 1 labeled autologous peripheral blo o d lym p h o cy tes in rabbits, J. N ucl. M ed. 18 (1 9 7 7 ) 141. FR O ST , H., FR O ST , P., WILCOX, C., e t al., L y m p h node scanning in sheep w ith In -1 11 labeled lym phocytes, In t. J. Nucl. Med. 6 (1979) 90. FR O ST , P., SM ITH, J., FR O ST , H., T he radiolabelling o f ly m p h o cy tes and tu m o r cells w ith in d iu m -111, Proc. Soc. E xp. Biol. Med. 157 (1 9 7 8 ) 61. PO NTES, J.E ., FR O S T , P., PO K O R N Y , M., et al., G am m a cam era imaging o f renal allografts using In -1 11 oxine labeled autologous lym phocytes, Invest. Urol. 17 6 (1980) 451. SINN, H ., SIL V E ST E R , D .J., Sim plified cell labeling w ith In -1 11 acetylacetone, Br. J. Radiol. 5 2 (1 9 7 9 ) 758. MATHIAS, C .J., H EATON , W.A., WELCH, M .J., et al., “ C om parison of oxine and acetylacetone as chelates fo r cell labeling w ith In -1 11 cell labels” , T hird In tern a tio n a l Sym posium on R adiopharm aceutical C hem istry, 16—20 Ju n . 1980, St. Louis, M issouri (1 9 8 0 ) 271 (abstract). TOSCH, K.W., H O RN ICK , J.S ., GEN N A RO , G.P., R a diochrom atography o f indium ( lu In) oxine, J. R adioanalyt. C hem . 53 (1 9 7 9 ) 345.
452
SUBRAMANIAN and McAFEE DISCUSSION
P.H. COX: The " T c m-Diars com plex m entioned is extrem ely interesting because it is to my knowledge the first positively charged technetium complex. Does it really behave as a potassium analogue? Moreover, because o f its lipophilicity, does it pass th e blood brain barrier? G. SUBRAMANIAN: This new com pound is still under investigation, and we do n o t know a great deal about the mechanisms involved in the biodistribution o f " T c m-Diars. P.H. COX: With reference to y o u r rem arks concerning the developm ent o f bone scanning reagents, I w onder w hether these will produce any significant increase in th e detection o f m etastatic lesions. The tum our:bone ratios are still low, about 2:1 in fact, and we have already seen th at MDP misses m any lesions associated w ith bone m arrow and trabecular bone. ■G. SUBRAMANIAN : At the present tim e, o f the com pounds reported, tw o - NMDP and 1,1-E D P - show 50% im provem ent in lesion : norm al bone ratios in animals. Any im provem ent in this area would be welcomed. E. PETURSSON: Living in a country rather far away from any supplier o f radioisotopes or radiopharm aceuticals, I have quite often found th at a 67Ga-citrate com pound ordered for a patien t w ith a suspected abscess is not delivered until the abscess has resolved or the patient has been operated on! The same would obviously apply for U1ln. I am therefore very interested in the possibility of labelling leukocytes w ith " T c m . Is it possible to use " T c m-labelled leukocytes to search for abscesses, and, if so, do you have to use oxine o r is it sufficient to reduce the " T c m in the same way as when labelling red blood cells? G. SUBRAMANIAN: Technetium -99m can be used for labelling white cells. With regard to alternatives to oxine, a paper recently published in Europe suggested pre-incubation o f the w hite cells w ith pyrophosphate and then re-incubation with pertechnetate. In vitro studies showed th at the cells were viable, the only draw back being th e labelling efficiency, which was only 30%. A second paper reported the use o f technetium p h y tate for labelling w hite cells, showing very good results for th e imaging o f abdom inal abscesses. The m ajor draw back in using " T c m is th a t th e labelling efficiency never exceeds 2 5 —45%, the higher lim it being achieved w ith oxine. Whereas th e oxine com plex is very stable, the same cannot be said, at least on the basis o f th e existing data, for the technetium p h ytate and pyrophosphate complexes.
IAEA-SM-247/49
SOLID-PHASE LABELLING
An improved method for the preparation of technetium-labelled radiopharmaceuticals P.H. c o x D epartm ent o f Nuclear Medicine, R otterdam sch R adio-Therapeutisch In stituut, R otterdam , The N etherlands
Abstract SOLID-PHASE LABELLING: AN IM PROVED M ETHOD F O R THE PR E PA R A T IO N OF TECH NETIUM -LA BELLED RAD IOPH ARM A CEU TICA LS. T he m ost w idely used radionuclide fo r th e p re p ara tio n o f radiopharm aceuticals for in vivo use is 99Tcm w hich has favourable rad iatio n characteristics com bined w ith a high degree o f chem ical reactivity. In general, sodium p e rte c h n e ta te is added to a reactio n vial containing a m ixture o f ligand and reducing agent. T h e tech n etiu m is reduced to th e 4 o r 5 valency state and com bines w ith th e ligand to form a com plex chelate. T he reducing agent o f choice at present is stannous chloride alth o u g h several o th e r reagents have been used successfully. It has alw ays been assum ed th a t th e stannous ion was in co rp o rated in to th e tec h n etiu m carrier com plex as an essential p a rt o f th e chelate. T o o b tain op tim u m labelling several m illigrams o f stannous tin are included in th e labelling m ix tu re, and this has th e disadvantage th a t the fo rm atio n o f ox id atio n p ro d u c ts m ay o ccu r w hich affects th e biological d istrib u tio n o f the labelled p ro d u c t. R ecent studies have indicated th a t th e stan nous ion is n o t in co rp o rated into tech n etiu m com plexes b u t w ould appear to be necessary to catalyse th e reactio n and act as electron donors. T he elim ination o f th e stannous ion from th e injection m ix tu re w ould have p o ten tial advantages fo r th e p a tie n t. W ith this in m ind a solid-phase labelling tech n iq u e has been evolved in w hich the p e rte ch n e tate ligand m ix tu re is exposed to an insoluble reducing agent, such as stannous sulphide w hich acts as an electro n d o n o r to com plete com plex fo rm atio n b u t w here o nly sufficient tin ions are m obilized as are necessary for the reactio n so th a t the p a tie n t receives m icrogram q u a n titie s o f tin in place o f milligrams.
INTRODUCTION The use o f gam m a-em itting radionuclides, bound to ligands having specific pattern s o f biological distribution and m etabolism , as in vivo diagnostic agents has becom e w idespread during the last fifteen years. The m ost widely used radionuclide to date has been " T c m because o f its favourable radiation character istics com bined w ith a high degree o f chem ical reactivity. In general, sodium pertech n etate solution is added to a m ixture o f the carrier and a soluble reducing agent. T he technetium is reduced to the 4 or 5 valency state and form s a com plex w ith the carrier which is stable fo r some hours. 453
454
СОХ
The m ost widely used reducing reagent to date has been stannous chloride although several others such as copper sulphate, iron ascorbate, zirconium and titanous chloride have been reported. It has always been assumed th a t th e stannous ion was incorporated into the technetium carrier com plex as an essential p art o f the chelate. R ecent w ork [1 ], however, has suggested th a t the stannous ions are n o t incorporated into the chelate bu t are necessary as electron donors to catalyse the reaction. This finding implies th a t the stannous ions have no direct influence on the biological activity o f the technetium com plex and th at indeed the p atient receives an unnecessary injection o f betw een 1 and 5 mg o f tin per study. The elim ination o f excess concentrations o f red uctant from the injectable radiopharm aceutical has a num ber o f potential advantages. In the first place the oxidation o f technetium to form colloidal tin com pounds [2] would be reduced. The labelling o f blood com ponents in vivo in the presence o f excess stannous ions would also be greatly reduced. This effect can be observed for several days after the adm inistration o f a technetium tin com plex, and may lead to high background activity if fu rth er studies w ith technetium complexes are carried o u t w ithin this period.
SOLID-PHASE REDUCTANTS A solid-phase red u ctan t is a reducing agent th a t is present in an insoluble form. When exposed to an aqueous solution it acts as an electron donor in the presence o f an oxidizing agent. In the presence o f p ertechnetate the technetium will be reduced to th e 4 o r 5 valency state which can then react w ith the ligand. The reducing agent will only be ionized in sufficient quantities to com plete the reaction provided the pH is n o t acid. Am ong the reductants already in use in soluble form to prepare technetium com plexes it is possible to obtain insoluble salts th a t could be used as solid-phase reducing agents. The exam ple w hich will be dem onstrated in this presentation is stannous sulphide. An interesting approach to this problem has already been presented in the p a te n t literature [3], a discussion o f which will serve to illustrate a num ber of poten tial advantages o f the m ethod described here. In this p aten t, pertechnetate is first absorbed on to a colum n o f reducing agent consisting o f metallic iron, tin or m agnesium, at which stage it becomes reduced. The reduced technetium is then eluted from the colum n by passing over it an acidic eluant solution containing the ligand. The eluted technetium com bines w ith the ligand and the solution is then passed over a cation exchange resin to absorb reduced uncom bined technetium ions and excess ions o f th e reducing agent, at which tim e the pH adjusts to physiological levels.
IAEA-SM-247/49
455
This ingenious m ethod is open to a num ber o f criticisms. In the first place the elution o f reduced technetium and red u ctan t from the colum n will only take place at an acid pH th a t will lim it th e choice o f ligand. Secondly, the elution from a colum n in w hich the red u ctan t is a m etal will result in a considerable solubili zation o f m etal when th e eluant is acid, leading to an excess o f reductant ions which defeats th e object o f the exercise. This is inherently recognized by the fact th a t a second colum n o f ion exchange resin is incorporated into the system. Whereas this cation resin will rem ove excess redu ctant, unbound reduced technetium will still pass because these com plexes are prim arily negatively charged ions.
TIN SULPHIDE In this study a sim pler system has been evaluated using insoluble salts of reducing metals and, as a specific exam ple, stannous sulphide. Stannous sulphide is less reactive than pure tin, and is therefore less pH sensitive, b u t proved to be adequate as an electron donor. When pertech n etate solution is passed over a thin colum n o f stannous sulphide pow der all the activity binds to the reductant and can only be eluted at acid pH. If, however, a m ixture o f pertechnetate and ligand, a t a neutral pH is passed over th e colum n, direct labelling occurs and the eluate contains only labelled com plex. T he labelling procedure takes betw een 30 seconds and one m inute to com plete, and th e com plex can be prepared so th at it is ready for injection im m ediately after preparation. In this way samples of glucoheptonate, MDP, diethyl ida, DMSA and album in have been prepared which appear to be identical in th eir biological behaviour w ith currently available labelling kits but where the concentration o f stannous ions is in the microgram range per litre instead o f milligrams per m illilitres.
THE PREPARATION AND CHARACTERISTICS O F MDP COMPLEX A practical illustration o f the technique can be given by describing the preparation o f the MDP com plex. S tannous sulphide was prepared in fine suspension by precipitating stannous chloride solution w ith sodium sulphide solution. The suspension was collected on a 200-nm m em brane filter and washed by filtering w ith physiological saline. A cake containing approxim ately 400 mg o f stannous sulphide was form ed. A solution o f 10 mg sodium MDP in 1 ml physiological saline was m ixed w ith 0.5 ml sodium p ertechnetate containing 2 mCi " T c m. T he m ixture was filtered u nder suction through the filter bed o f stannous sulphide, and 0.3 ml o f th e filtrate was injected intravenously into a series o f adult Wistar rats. O ne h o u r post injection accum ulation o f the com plex in the
СОХ
456
skeleton and kidneys was observed com parable in quality w ith control studies prepared using comm ercial MDP reagent. No thyroid or stom ach uptake was observed.1 Thin-layer chrom atogram s showed no change in quality when m easured up to 3 hours post preparation.
CONCLUSION This rep o rt describes an alternative labelling technique for the preparation o f technetium com plexes for in vivo nuclear m edical investigations. The use o f a solid-phase red u ctan t reduces the am ount o f reducing agent injected and offers a num ber o f technical advantages: 1. 2.
3. 4.
U ncom plexed reduced technetium rem ains bound to the reductant. By using a solid-phase red u ctan t the preparation o f freeze-dried stannous chloride/ligand labelling kits is elim inated w ith consequent reduction in costs and enhancem ent o f shelf life o f ligand. Re-usable nature o f reduction element. The labelling reaction occurs at neutral pH so th a t pH-sensitive materials m ay be labelled.
REFERENCES [1] De K IE V IE T, W., “ R ecent developm ents in tech n etiu m chem istry” , presented at the 18th A nnual M eeting o f the S ociety o f N uclear M edicine, Nürnberg, S eptem ber 1980. [2] PE T TIT , W.A., DELAND , F .H ., PEPPER, G.H ., et al., C haracterisation of tin tech n etiu m colloid in technetium labelled album in preparations, J. N ucl. Med. 19 (1 9 7 8 ) 387. [3] BARA K, M., W INCKEL, H.S., R adiopharm aceutical G e n era to r K it, U.S. P a ten t 3749556 (1971).
DISCUSSION G. SUBRAMANIAN: As y ou know, similar work on solid phase labelling o f technetium com plexes was reported earlier in the literature by Sew atkar and co-workers ( 15 th A nnual M eeting o f the Society o f Nuclear Medicine, Groningen, Sep. 1977), by Basmadajian and co-workers, and by the NEN B oston G roup at the 2nd International Sym posium on R adiopharm aceutical Chem istry held in O xford. A t th a t tim e it was shown th at the q u antity o f tin leaking o u t o f the 1 1 Ci = 3.70 X 1010 Bq.
IAEA-SM-247/49
457
system varied in p ro p o rtio n to the stability constant o f the chelate and the quantity o f chelate. Did y ou observe this phenom enon in yo u r studies? P.H. COX: W ithin th e concentration ranges o f ligand used, the concentration o f tin eluted was in the 2 0 —30 pmol/1 range. I do n o t know w hether this varied w ith ligand concentration. G. SUBRAMANIAN : W hat are the reduction media you have used in your studies, o th er th an SnS, Sn-ferrocyanide, Sn-ferricyanide and o th er com pounds reported previously? P.H. COX: We have used b o th m etal and insoluble salts o f tin, iron, titanium , and copper. A ny reducing m etal could be used. K. KRISTENSEN: You m entioned in y o u r oral presentation th at the stability was acceptable for five hours. Did you notice any difference in stability o f com pounds labelled w ith " T c m by y o u r m ethods and com pounds labelled by the stannous chloride m ethod? P.H. COX: We observed no difference at these ligand concentration levels. It will be o f some interest to exam ine this at low er ligand concentration levels where this technique can be fully exploited. J.-L. M ORETTI: When labelling MDP by y o u r m ethod and removing tin from the preparation, did you im prove th e bone : soft tissue ratio? P.H. COX: O ur initial im pression is th a t this is so but there will have to be m ore extensive studies in b o th animals and m an in order to obtain a statistically reliable confirm ation.
I A E A - S M - 2 4 7 / 152
TECHNETIUM-99m-BIDA A p o t e n t i a l c h o le s c in tig r a p h ic a g e n t f o r ic te r ic p a tie n ts J. WEININGER, J. TRUMPER Soreq Nuclear Research Centre, Yavne E. LUBIN, M. COHEN Beilinson Hospital, Petah Tikvah T. SADEH, M. JUSZINSKY Soreq Nuclear Research Centre, Yavne, Israel
Abstract TECHNETIUM-99m-BIDA: A POTENTIAL CHOLESCINTIGRAPHIC AGENT FOR ICTERIC PATIENTS. Technetium-99m-BIDA (N-(p-butylcarbamoylmethyl) iminodiacetic acid) was developed and evaluated for its efficacy as a cholescintigraphic agent for icteric patients. Extensive in vivo studies in mice show a similar biodistribution for 99 Tcm-BIDA and " T c m-HIDA, but the excretion from the hepatobiliary system is much slower for the first complex. A clinical trial in 30 adult jaundiced patients and in 13 new-born babies with obstructive jaundice has been performed. From our clinical experience it seems that 99 Tcm-BIDA has the advantage over " T c m-HIDA of a better liver concentration even in patients with high bilirubin levels. This advantage may be somewhat lost by its slower excretion rate into the gut. The detection of biliary atresia in the new born was very acceptable with both 99 Tcm-BIDA and " T c m-HIDA. There was, however, a difference between the groups in the visualization of the liver: it was better in the BIDA group. The demonstration of biliary permeability (by visualization of the gut) was of similar reliability with both imaging agents.
Technetium-99m-labelled iminodiacetic acid substitutes, because of their marked hepatotropism and consequent poor excretion through the kidneys, have recently been widely used for sequential hepatobiliary scintigraphy. The first complex described in the literature [1 ] — "T cra-HIDA - was developed as a kit preparation by different groups, and has been tested on a great number of patients. The kit developed by the authors’ group proved to be suitable for cholescintigraphic studies in non-jaundiced patients as well as in patients with 459
460
W E I N I N G E R et al. I03R ~ T BIDA
HIDA
Blood Liver Gallbladder Gut
J_l__ 1_____ I_____ I___ 5 10 20
40
60
90
120
Time (min)
F IG .l.
Blood, liver, gallbladder and gut uptake o f
99
T c m-B ID A compared with " T c m -H ID A .
bilirubinaemia below 6—7 mg/dl. In icteric patients, at bilirubin levels higher than 10 mg/dl, because of the progressive reduction in the liver uptake accompanied by a corresponding reduction in the biliary tract excretion rate, the quality of liver and biliary tract visualization obtained is poor [2]. Seeking an agent suitable for imaging of the liver and the biliary tract in icteric patients, we developed the "T cm-BIDA complex as suggested by Subramanian and co-workers [3]. The first step of our work was the synthesis of high purity N-(p-butylcarbamoylmethyl) iminodiacetic acid (BIDA) [4]. The Sn-BIDA complex we developed thereafter yields, after subsequent labelling with "T cm-pertechnetate, a high labelling efficiency "T cm-complex which shows the same marked hepatotropism as "T cm-HIDA, but is accompanied by lower kidney excretion and lower bilirubin dependence. Adequate chromatographic and electrophoretic systems helped us to evaluate the labelling yield of the complex and its stability. Extensive in vivo studies in mice show that less than 1% of the injected activity is found in the stomach during a 180-min period, the uptake of the kidneys averaging 2.7% 5 min after injection and 0.94% 120 min later. The "T cmBIDA complex shows low urinary excretion: only 1—3% of the injected dose is eliminated in the urine during 150 min after injection, whereas more than 90% is found during the same period in the liver and the biliary tract. The relative slow
IAEA-SM-247/152
461
Time (mir)
F IG .2 .
Hepatobiliary clearance o f 9 9 T c m -B ID A expressed as activity per gram organ.
hepatobiliary clearance of the "T cm-BIDA complex compared with "T cm-HIDA is shown in Fig. 1. The calculated blood clearance and liver excretion half-times are: Blood clearance Liver excretion T ,/2 (min) T1/2 (min) " T cm-BIDA "T cm-HIDA
13 2
18 3
The hepatobiliary uptake evolution over a 2-h period expressed as percentage "T cm -BIDA injected activity/g organ is shown in Fig.2. A comparison of the excretion rate of both agents expressed as ratios of gallbladder uptake to blood, liver and gut is shown in Fig.3. In the authors’ kit one human dose is not more than 10 mg of the complex/70 kg body weight. The LDS0 of the Sn-BIDA mixture contained in the kit preparation is: 121.9 mg per kg in mice and 158.5 mg in rats. The margin of safety is large enough to avoid adverse reactions or toxic effects. After determination of the suitability of the radiopharmaceutical for cholescintigraphic studies in normal subjects as well as in non-jaundiced patients, a clinical trial in 30 adult jaundiced patients was performed. The results of these examinations can be summarized as follows: In 9 patients there was no visualization of the colon at 24 hours: in 8 of them the complete
462
W E I N I N G E R et al.
Time (min)
F IG .3 .
Gallbladder to blood, liver and gut ratio evolution with time o f 9 9 Tcm-B ID A compared
with 9 9 Tcm-H ID A .
TABLE I. ADULT PATIENTS (AVERAGE AGE 60 YEARS): VISUALIZATION OF COLON Medical jaundice Hepato-cellular
Surgical jaundice Chole-lithiasis
Tumour obstruction
No visualization of the colon
1
0
8
Colon visualization
9
12
0
obstruction of the bile duct was later proved to have been due to malignancies and one patient had very severe hepato-cellular jaundice. In 21 patients the radio pharmaceutical appeared in the colon 24 hours after its administration: 9 patients had hepatobiliary jaundice, 12 had chole-lithiasis or carcinoma of the gallbladder (Table I). No visualization of the gallbladder was obtained in all cases of complete obstruction o f the bile duct, acute cholecystitis and carcinoma of the gallbladder.
IA E A -SM-247/152
463
TABLE II. ADULT PATIENTS (AVER AGE AGE 60 YE ARS): VISUALIZATION OF GALLBLADDER Medical jaundice
Surgical jaundice
Hepato-cellular
Acute cholecystitis
Chronic cholecystitis
Tumour of gallbladder
Choledoco -lithiasis
No visualization of gallbladder
1
4
0
2
2
Gallbladder visualization
9
0
3
0
1
о TABLE III. BILIARY AGENTS IN OBSTRUCTIVE JAUNDICE OF THE NEW BORN
Patient
Age (weeks)
Bilirubin (mg/%)
99 Tcm-biliary
Patent
agent
Non-patent
Final diagnosis2
1
16
2.9
+
NH
2
16
2 .6
+
NH
3
3
29.0
+
NH
4
3
17.5
+
CC Kasai
5
8
5.2
+
NH
6
3
18.0
+
NH
7
8
7.2
+
BA Kasai
8
4
1 0 .0
+
BA Kasai
9
2
6 .0
+
Bile plug BA Kasai
10
4
1 0 .0
+
11
7
1 0 .8
+
BA Kasai
12
8
8 .2
+
BA Kasai
6
8.4
13
a NH - Neonatal hepatitis. BA —Biliary atresis. CC —Choledocal cyst.
+
Observation
464
W E I N I N G E R et al.
Visualization of the gallbladder was seen in all cases of hepato-cellular jaundice (Table II). From the clinical examinations it can be concluded that the lack of colon visualization proved to be highly specific in the diagnosis of complete obstruction. However, the appearance of the radiopharmaceutical in the colon is not by itself sufficient for the differentiation between hepato-cellular jaundice and incomplete obstruction. Technetium-99m-BIDA seems to have the advantage over "T cm-HIDA of a better liver concentration even in patients with high bilirubin levels. This advantage may be somewhat lost by its slower excretion into the gut. Among 13 new boms studied in the last 24 months for the detection of biliary atresia, the first 7 were studied with HIDA and the last 6 with BIDA (Table III). On the whole, the detection of biliary atresia in this population was very acceptable, and although there was a difference between both groups as far as the visualization of the liver (better in the BIDA group) was concerned, the end result —demonstration of biliary permeability through the visualization of the gut - was of similar reliability with both biliary agents, о REFERENCES [1]
[2] [3]
[4]
RYAN, J., COOPER, M„ LOBERG, M„ HARVEY, E„ SIKORSKY, E„ Technetium 99m labeled N-(2,6 dimethylphenylcarbamoyl)iminodiacetic acid - " mTc-HIDA, a new radiopharmaceutical for hepatobiliary imaging studies, J. Nucl. Med. 18 (1977) 995. WEININGER, J., TRUMPER, J., LEVI, A., LUBIN, E., Laboratory and clinical testing of 99 Tcm-HIDA for use as a cholescintigraphic agent, I.A.E.C. Annual Report 1978-79. SUBRAMANIAN, G., McAFEE, J., HENDERSON, R.W., et al., The influence of structural changes on biodistribution of Tc-99m labeled N-substituted IDA derivatives, J. Nucl. Med. 18 (1977) 624. SADEH, T., JUSZINSKY, M., Synthesis of N-(p-butylphenylcarbamoylmethyl) iminodiacetic acid —BIDA, I.A.E.C. Annual Report 1978—79.
IAEA-SM-247/153
CHOLESCINTIGRAPHY IN JAUNDICED PATIENTS C o m p a r is o n o f d i e t h y l im in o d ia c e tic a c id ( I D A ) w ith p -b u ty l-I D A R. DUDCZAK First Medical Clinic, Division of Nuclear Medicine, University of Vienna, Vienna P. ANGELBERGER, M. WAGNER-LÖFFLER Research Centre Seibersdorf, Seibersdorf K. KLETTER, P. FERENCI, H. FRISCHAUF First Medical Clinic, Division of Nuclear Medicine, University of Vienna, Vienna, Austria
Abstract CHOLESCINTIGRAPHY IN JAUNDICED PATIENTS: COMPARISON OF DIETHYL IMINODIACETIC ACID (IDA) WITH p-BUTYL-IDA. Synthesis of the inactive compounds was carried out with a modification of the method of Callery and Lofgren. After 99 Tcm labelling, radiochemical purity control was performed with ITLC, TLC, PC and HPLC. The kinetic behaviour of diethyl-IDA (de-IDA) and p-butylIDA (pb-IDA) and also their ability for sequential imaging was compared in 8 jaundiced patients (serum bilirubin: 3—25 mg/dl). Blood clearance curves were biexponential with half-times for the fast component of 4.1 and 4.9 min for de-IDA and pb-IDA respectively. Cumulative urinary excretion was higher for de-IDA. pb-IDA demonstrated a slower liver uptake and elimination. Thus on sequential images pb-IDA proved to exhibit relatively prominent hepatic uptake compared with de-IDA, whereas gall bladder filling and intestinal excretion.were delayed, and depiction of the biliary system failed in 5 patients using pb-IDA. The delayed imaging charac teristics using pb-IDA offer no diagnostic advantage over de-IDA. Misinterpretation of the scintigrams may occur more often with pb-IDA especially if the observation time is not ade quately prolonged. The data of the authors indicate that de-IDA seems to be superior to pb-IDA in its diagnostic usefulness, also in patients with severe hyperbilirubinaemia.
INTRODUCTION Advances in radioisotope imaging and in radiopharmacology have increased the potential for visualization of the biliary system by non-invasive techniques. Synthesis of IDA derivatives labelled with " T cm, which are taken up by the liver 465
466
D U D C Z A K et al.
and excreted into the bile, have improved the accuracy for the differential diagnosis of jaundice [1,2]. Hepatic uptake and excretion take place even with high serum levels of alkaline phosphatase and bilirubin when radiological contrast methods cannot be applied. However, when the degree of liver dysfunction and hyperbilirubinaemia increases, a diminished uptake will lead to inconclusive results. From previously reported animal studies [3 ] and from limited clinical experi ence in patients with jaundice [4], it was suggested that among the IDA derivatives p-butyl-IDA (pb-IDA) best maintains biliary exretion despite hepatic dysfunction and rising bilirubin levels, and offers some advantage over diethyl-IDA (de-IDA). But as there are considerable interspecies differences [5], and no direct comparison was made between de-IDA and pb-IDA, it might be important to obtain data in man and to compare them in patients with varying degrees of liver dysfunction. PREPARATION OF RADIOPHARMACEUTICALS Synthesis o f the inactive compound (in a modification of the procedure of Lofgren [6] and Callery [7], and labelling with "T cm was carried out as previously described [8]. Radiochemical purity control was performed in four independent chromatographic systems [8]: (a) thin-layer chromatography (TLC) on Merck cellulose plates developed in acetone; (b) TLC on cellulose using the solvent n-butanol : acetic acid : water (20 : 5 : 5); (c) paper chromatography on Whatman N o.l developed in acetonitrile : water (3 : 1); (d) high-pressure liquid chromatography (HPLC) on a reversed phase column (Merck LiChrosorb RP18, 7 pm , 4.6 X 250 mm) using an eluent composed o f 0 .0 1M phosphate buffer pH 6 and 20—100% (vol./vol.) methanol in a linear 20-min gradient, flow rate 1 ml/min. Technetium-99m activity in the effluent was measured by a flow through the scintillation detector and 254 nm UV absorption was also monitored. Radioactive TLCs and PCs were scanned on a thin-layer scanner. Free pertechnetate can be quantified by system (a) (less than 2% of the total activity); an independent radiochemical purity control is provided by systems (b) and (c) which separate hydrolysed reduced TC from "T cm-IDA plus TcO. System (d) shows that the "T cm-IDA complex is a molecule with a different chemical identity from the inactive IDA derivative. The chemical structure of "T cm-IDA is assumed to be an octahedral complex with 2 IDA ligands co-ordinating one technetium ion. This complex is therefore larger and more lipophilic than the unlabelled IDA derivative. PATIENTS AND METHODS Eight patients who had given informed consent were investigated, the six males and two females ranged in age from 38—66 years. Serum bilirubin levels were 3 —4 mg/dl
IAEA-SM-247/153
467
TABLE I. COMPARISON OF DIETHYL-IDA WITH p-BUTYL-IDA IN JAUNDICED PATIENTS (n = 8; x ± s)
Uptake index
Retention index
Liver peak time
Diethyl-IDA
p-Butyl-IDA
5 min
1.25 ±0.58
1.04 ± 0.58a
60 min
2.07 ± 1.32
3.26 ± 2.26a
30 min
0.97 ± 0.03
Not calculable
60 min
0.81 ± 0.07
Accumulation curve not calculable
(min)
Cumulative urinary excretion (% dose) t = 180 min
20.60 ±5.6
>60a
19.60 ±9,2
7.40 ± 4.1a
a p < 0.01, significantly different from diethyl-IDA (paired data).
in three patients, 10.6 — 14 mg/dl in another three, and 19 and 25 mg/dl in the remaining two. There was no significant change in biochemical parameters which were evaluated before each investigation. Diagnosis was established by liver biopsy, sonography, PTC, autopsy in one, and operation in two of the patients. Two patients had primary biliary cirrhosis, three cholostatic alcoholic liver cirrhosis, one carcinoma o f the liver. In the remaining two there was obstructive jaundice due to stones associated in one patient with cystic duct obstruction. All diagnostic procedures were carried out on fasting patients at a three-day interval. The activity administered was 0.06 —0.07 mCi/kg bw.1 Sequential samples for blood clearance were obtained by means of an indwelling intravenous catheter up to 180 min following injection. Urine was collected for 3 h after initiation o f the study. Scintifotos were made at 15-min intervals for 1 h and after 3 h. Time activity curves were obtained over regions o f interest (ROI) during a 60-min period after injection using a DEC PDP 11. The hepatic uptake of de-IDA and pb-IDA was assessed by means of an ‘uptake index’ [9], obtained by taking a ratio of the activity in the periphery of the right liver lobe over that from the heart both at 5 and 60 min post injection. The liver’s excretory capacity was measured by two values o f a ‘retention index’ [9], defined as the ratio of the activity in the right lobe at 30 (or 60) min over that at its maximum. 1 1 Ci = 3.70 X 1010Bq.
468
TABLE II. TIME (min) FOR VISUALIZATION OF THE GALL BLADDER (GB), COMMON BILE DUCT (CBD) AND INTESTINE (I)
Diagnosis
Serum bilirubin (mg/dl)
Alkaline phosphatase (U/l)
Diethyl-IDA GB
CBD
p-butyl-IDA I
GB
CBD
I
254
Prim, biliary cirrhosis
19
250
-
-
180
-
180
Cholostatic cirrhosis
14
570
45
-
-
-
-
Cholostatic cirrhosis
1 0 .8
255
30
30
30
60
1 0 .6
330
30
60
30
60
60
4.2
263
30
60
60
180
180
Cholelithiasis
3.4
1320
-
30
30
Choledocholithiasis
3.2
890
30
20
30
Cholostatic cirrhosis Carcinoma of the liver
Faüure to accum ulate in the liver
60
60 45
60
60 -
DUDCZAK et al.
25
Prim, biliary cirrhosis
I A E A - S M - 2 4 7 / 153
469
RESULTS
After the initial mixing phase the blood clearance curve can be adequately approximated as biexponential. The initial fast component was faster for de-IDA than for pb-IDA, with half times of 4.1 + 0.7 min and 4.95 + 0.7 min respectively. The mean cumulative urinary excretion of de-IDA was significantly higher than that of pb-IDA (Table I). Using de-IDA, the time activity curve generated over the right liver lobe reached a maximum at 20.6 ± 5.6 min after injection. In comparison, maximum liver uptake was markedly delayed for pb-IDA, where time activity curves in most patients exhibited an accumulation pattern during the first 60 min. The hepatic uptake index was higher for de-IDA at 5 min but was less pronounced at 60 min compared with pb-IDA. A faster elimination was found for de-IDA as demonstrated by the retention index, which could not be calculated for pb-IDA because of its increasing activity up to 60 min after initiation of the study. In one patient with primary biliary cirrhosis and a serum bilirubin level of 25 mg/dl, jaundice was so pronounced that no hepatic uptake or excretion could be observed. Time activity curves over ROI o f the liver were similar to background curves for both substances, thus de-IDA and pb-IDA failed to accumulate in the liver. In the remaining seven patients "T cm de-IDA imaging studies showed that after an initial increase liver activity gradually diminished. It increased on the images obtained up to 60 min using pb-IDA. Gall bladder filling was observed in five patients with de-IDA, but only in four using pb-IDA (Table II). Visualization of the biliary tract was possible in five patients between 20—60 min post injection, but delineation of the common bile duct failed in those patients with the highest bilirubin level (14 and 19 mg/dl). Activity within the intestinal tract was demonstrable in six patients, but up to 3 h after intravenous injection of de-IDA no appreciable gut entry was seen in one (bilirubin 14 mg/dl). Depiction of the common bile duct was markedly delayed for pb-IDA and could be seen only in two patients (Figs 1 and 2). Activity in the bowel was demonstrable in five, but failed to be detected in two (bilirubin 14, 3 mg/dl) until 3 h post injection (Fig.3).
DISCUSSION Although de-IDA has proved its diagnostic usefulness for cholescintigraphy [1 ,9 ], in severely jaundiced patients the degree of uptake diminishes and the results are of limited value. Therefore other compounds are being investigated
470
D U D C Z A K et al.
d e - 1D A
pb-l
D A
F I G .l. Scintifotos from a patient with stones in the bile duct associated with cystic duct obstruction (serum bilirubin 3 .4 mg/dl), illustrating the absence o f tracer in the gall bladder and pile-up o f radioactivity at the porta hepatis. In comparison, depiction o f the bile duct is fain t and delayed on images after injection o f pb-IDA.
to increase the diagnostic capabilities for investigations of patients with marked hyperbilirubinaemia. From the kinetic data one can assume a multicompartment distribution of the radiopharmaceuticals. The blood clearance, which can be approximated as biexponential, showed only slight differences between de-IDA and pb-IDA. The distribution phase was slightly prolonged for pb-IDA, and the second slope was not significantly different for both substances. This would be in agreement with an assumption of a delayed liver uptake o f pb-IDA, consistent with the findings observed.
I A E A - S M - 2 4 7 / 153
471
F IG .2. Comparison o f images obtained using de-IDA and pb-IDA in a patient with cholestatic cirrhosis (serum bilirubin 10 .6 mg/dl). Whereas there was a much greater activity in the liver at 60 min using pb-ID A , depiction o f the gall bladder and visualization o f intestinal activity was more pronounced using de-IDA.
In contrast to de-IDA, it has been shown that pb-IDA exhibits a high degree of protein binding [10]. This may possibly account for the low renal excretion of pb-IDA despite its slower hepatic uptake. The analogues o f N-substituted IDA derivatives varied in their transit time through the liver and in the degree of renal excretion. Urinary excretion was significantly higher using de-IDA. However, there was no clear relation between the degree of hyperbilirubinaemia and the amount of the tracer excreted with the urine, but in patients with severe liver disease an altered renal function has to be considered [11]. Although the qualitative pattern o f uptake of IDA derivatives is similar, considerable differences in turnover rates exist which might be due to molecular size, protein binding, type o f lipophilicity, and position of substitution [5]. The slower liver uptake and elimination of pb-IDA, which is reflected by the uptake and retention index calculated, was also demonstrable on scintifotos (Figs 1—3). In agreement with animals studies, pb-IDA proved to exhibit relatively prominent hepatic uptake compared with de-IDA. But because of its longer hepatic transit time excretion was delayed. Gall bladder filling and visualization of activity within the intestine occurred later, and delineation o f the biliary system was rarely successful. In our study there were only minor differences between the results obtained with de-IDA and pb-IDA investigations, e.g. complete obstruction o f the common bile duct was assumed using pb-IDA, and only incomplete obstruction with de-IDA,
472
D U D C Z A K et al.
de-ID A
pb-ID A 30'
180
F IG .3. Sequential images o f a patient with incom plete obstruction o f the common bile duct from a stone (serum bilirubin 3 mg/dl). Gall bladder visualization can be demonstrated with either de-IDA or pb-IDA, whereas activity in the gut, and a fain t delineation o f the upper part o f the bile duct, is only possible with de-IDA. This later finding w ould be consistent with the diagnosis o f incom plete obstruction.
which was consistent with the diagnosis. The delayed imaging of the gall bladder and the biliary tract using pb-IDA may sometimes lead to misinterpretations of the scintigrams. Our data indicate that de-IDA seems to be superior to pb-IDA in its diagnostic usefulness, also in patients with severe hyperbilirubinaemia. A virtue o f de-IDA is its rapid appearance in the biliary tract. Since the primary value of "T cm-labelled IDA derivatives is to evaluate the dynamic function, and not to visualize the liver, pb-IDA offers no diagnostic advantage for studying jaundiced patients.
I A E A - S M - 2 4 7 / 15 3
473
REFERENCES [1] RONAI, P.M., Hepatobiliary radiopharmaceuticals: defining their clinical role will be a galling experience, J. Nucl. Med. 18 (1977) 488. [2] ] WISTOW, B.W., et al., An evaluation of 99Tcm-labeled hepatobiliary agents, J. Nucl. Med. 18(1977) 455. [3] WISTOW, B.W., et al., Experimental and clinical trials of new 99 Tcm labeled hepatobiliary agents, Radiology 128 (1978) 793. [4] ROSENTHAL, L., Clinical experience with the newer hepatobiliary radiopharmaceuticals, Can. J. Surg. 21 (1978) 297. [5] СОХ, P.H., “The comparative pharmacology of technetium IDA derivatives” , Progress in Radiopharmacology 1 (СОХ, P.H., Ed.), Elsevier Press, Amsterdam-New York-Oxford (1979) 53. [6 ] LOFGREN, N.M., et al., Alkylglycinanilides, Chem. Abstr. 42 (1978) 6378. [7] CALLERY, P.S., et al., Tissue distribution of wTcm and carbon-14-labelled N (2,6 dimethyl phenylcarbomoyl methyl) iminodiacetic acid, J. Med. Chem. 19 (1976) 962. [8 ] DUDCZAK, R., et al., “Comparison of IDA derivatives in patients with liver cirrhosis” , Progress in Radiopharmacology, 1 (СОХ, P.H., Ed.), Elsevier Press, Amsterdam-New YorkOxford (1979) 207. [9] NIELSEN, S.P., et al., Hepatobiliary scintigraphy and hepatography with 99 Tcm diethyl acetanilido iminodiacetate in obstructive jaundice, J. Nucl. Med. 19 (1978) 452. [10] NICHOLSON, R.W., et al., The plasma protein binding of HIDA, Eur. J. Nucl. Med. 5 (1980)311. [11] RING-LARSEN, H., Renal blood flow in cirrhosis: relation to systemic and portal hemodynamics and liver function, Scand. J. Clin. Lab. Invest. 37 (1977) 635.
DISCUSSION on the previous two papers N.G. TROTT : Mr. Lubin, you did not refer in your presentation to the kinetics of your studies, whereas considerable reference was made to time sequence observations in Mr. Dudczak’s paper. It would be helpful in comparing the results of these papers to have your observations on the kinetics of the agents you have studied. It would also be helpful to know whether you have established a particular protocol for selecting agents to use for particular clinical problems. E. LUBIN: The oral presentation o f our paper dealt mainly with the use of the diisopropyl-IDA derivative (DIPA), which is now our biliary agent of choice, whereas the written paper summarized our work with HIDA and BIDA. HIDA shows rapid kinetics in mice but a very high bilirubin dependence, whereas BIDA has a slow excretion rate that makes for a poor visualization o f the biliary tree. The data on animal kinetics for DIPA are similar to those for HIDA but are much less dependent on bilirubin levels, as we could confirm in our severely jaundiced patients. The data on animal kinetics for HIDA and BIDA are included in the text o f the paper. All our patients are studied in fasting condition after injecting 6 —8 mCi of "Tcm DIPA in jaundiced patients and only 3 —4 mCi in acute cholecystitis patients. Patients are monitored with continuous recording for the first 30 min and thereafter studied at 60 min — 2 h —4 h — 2 4 h o r until gastro-intestinal tract visualization. Histograms are obtained during the first 30 min from four areas of interest —heart, liver, gall bladder and duodenum. The curves obtained have been analysed, but the kinetic system in the severely jaundiced has been too complex and variable to allow for a significant summary, especially in view o f the difficulty of obtaining non-contaminated data in this complex situation. A second problem we have found is that the kinetics are slow and that the 30 min that we can afford for continuous recording are frequently insufficient. As a result, we currently base our diagnostic conclusions more on the images obtained than on the kinetic data available. With DIPA the good parenchymal visualization, even with bilirubin levels o f 20 mg%, and with the biliary tract not positively visualized, has been o f additional help in differentiating complete and partial obstructions from hepatocellular jaundice. N.G. TROTT: Mr. Dudczak, have you established a protocol for selecting the best agent for a particular clinical problem? R. DUDCZAK: So far we have compared diethyl-IDA with several IDA derivatives (dimethyl-IDA and N-isopropyl-IDA) and, as reported in the paper, with N-butyl-IDA. Our data indicate that diethyl-IDA is superior to the other quoted radio pharmaceuticals. However, in the case o f severely jaundiced patients the results may still be inconclusive. 475
476
DISCUSSION
G. SUBRAMANIAN: Leonard Rosenthal o f Montreal, Canada, recently reported in Clinical Nuclear Medicine (March 1980 and August 1980) on com parative studies of "T cm-labelled 2,6 diethyl-HIDA versus p-N-butyl-HIDA and 2.6 diisopropyl-HIDA versus p-N-butyl-HIDA in patients with varying (high) bilirubin levels. His overall conclusion was that both 2,6 diisopropyl- and 2.6 diethyl-HIDA are superior to other agents in a variety of clinical conditions. K.E. BRITTON: I would like to support Mr. Lubin in that butyl-IDA is the best for severe jaundice - 123I-bromsulphthalein (BSP) is even better - and that ethyl-IDA is best for biliary disease with low levels of jaundice because of high renal excretion with ethyl-IDA. Iodine-123-BSP is better in severe hepato cellular jaundice than butyl-IDA because o f even less renal excretion. Like Rose Bengal, labelled BSP is the agent of choice in biliary atresia. The general principle is still to choose the best agent for the particular clinical problem rather than one agent for all problems. R. DUDCZAK: I would agree with Mr. Britton that that should be the general principle. Our data clearly show that the results obtained with diethylIDA are of more clinical significance. The aim of cholescintigraphy is not to investigate hepatocellular jaundice, but to evaluate the biliary system, which is served best with diethyl-IDA compared with p-butyl-IDA. The lower urinary excretion o f p-butyl-IDA may become a disadvantage as it increases background activity and the radiation burden, without increasing the diagnostic accuracy. Recently it has been reported1 that p-butyl-IDA was less conclusive than 131IRose Bengal for investigations o f patients suspected o f biliary atresia. This finding was mainly attributed to the short half-life of " T cm. Iodine-123-BSP, also mentioned, will therefore offer no advantage in this respect.
1 COLLIER, B.D., TREVES, T„ DAVIS, M.A., HEYMAN, S., SUBRAMANIAN, G., McAFEE, J.G., Radiology 134 (1980) 719.
IAEA-SM-247/39
CLINICAL APPLICATIONS OF INDIUM-111 -ACET Y LACETONE-LABELLED BLOOD CELLS P. GEORGI, H. SINN, H. WELLMAN*, J.H. CLORIUS Institute of Nuclear Medicine, German Cancer Research Center, Heidelberg W.BECKER Orthopaedic Clinic, University of Heidelberg, Heidelberg, Federal Republic o f Germany
Abstract CLINICAL APPLICATIONS OF INDIUM-111-ACETYLACETONE-LABELLED BLOOD CELLS. For evaluation of intestinal blood loss with calculation of the liver-spleen ratio and determination of the red-cell volume, the authors reported a method permitting red-cell labelling with m In-acetylacetone in 1974. White blood cells can be tagged using a similar method. In white-cell labelling, simultaneous red-cell or platelet tagging, which might produce misleading results, is avoided. Several procedures have been combined, namely dextran separation and gradient centrifugations, to develop a highly selective cell separation. In osteomyelitis it may not be as advantageous to use 67 Ga-citrate, as in inflammatory soft tissue processes, since it is also a bone-seeking agent. Differentiation between active bone localization and an inflammatory process may not always be possible. Therefore the detection of inflammatory processes with labelled leukocytes could be of great importance for the scintigraphic diagnosis of osteomyelitidies. A group of 97 patients have been examined with suspected osteomyelitis using l n In-acetylacetone-labelled leukocytes (l n In-AAL) immediately following positive routine skeletal scintigraphy. Images obtained 24 h post injection usually were the most satisfactory. In the followup group of 70 patients 21 true positives, 43 true negatives, 21 false negatives and 3 false positives were observed. These findings result in a specificity of 92%, sensitivity of 50% and accuracy of 70% with U1ln-AAL for osteomyelitis. Preliminary investigations using m In-acetylacetone-labelled thrombocytes (u l In-AAT) were carried out to detect rejection of transplanted kidneys. The platelets were separated by means of additional special density gradient centrifugations but no dextran from 15—20 ml of autologous whole blood. Scans have been obtained 15 min, 2.5 h and 24 h post injection in an initial group of 10 patients. In acute rejection, a high transplant uptake has been detected, whereas patients without acute rejection showed no or only a minimum activity accumulation. Patients with chronic rejection have intermediate uptakes.
* Alexander von Humboldt Stiftung, U.S. Senior Scientist.
477
G E O R G I et al.
478 1.
INTRODUCTION
The physical data for 51Cr and " T cm compared with that o f 111In reveal that the latter radionuclide is the most advantageous for labelling erythrocytes when the duration of a study is between 1—6 days. For evaluation of intestinal blood loss, calculation of liver-to-spleen uptake ratios and determination of red-cell volume, we have previously reported (in 1974) very satisfactory results with a method of labelling red blood cells with m In-acetylacetone [ 1]. Labelling efficiency using less than 5 ml of whole blood is greater than 95% with this method, with specific activities of 20 mCi/ml.1 Leukocytes and platelets can also be labelled by modifications o f the i n In-acetylacetone technique [2]. The methodology for such labelling and current clinical experience with the preparation is also described.
2.
INDIUM-111 -ACETYLACETONE LEUKOCYTE (INDIUM-111 AAL); SEPARATION AND LABELLING METHODOLOGY
The chief aim in labelling pure leukocytes is to avoid simultaneous labelling of erythrocytes, reticulocytes and platelets, thus reducing unnecessary high background in imaging. The technique of leukocyte separation is detailed in Fig.l. A 3% solution of 2 X 10s MW dextran is used to increase the sedimentation of erythrocytes. A metrizamide solution of sp. gr. 1.088 is used to underlay the resultant supemate containing leukocytes after sedimentation. Differential centrifugation then results in localization of leucocytes at the metrizamide inter face which are then carefully drawn off and resuspended in physiological buffer. With further centrifugation a leukocyte pellet is produced which is resuspended after pouring off the supernatant containing mainly platelets. Centrifugation is repeated, thus further purifying the leukocyte pellet, which is again suspended in physiological buffer and labelled to about 80% efficiency with up to 5 mCi m In-acetylacetone ( n i In-AAL) [2]. After labelling, a pellet is again produced which is resuspended in autologous plasma. All procedures are performed with sterile, pyrogen-free blood containers, solutions and technique. Autologouslabelled leukocytes are then ready for reinjection after about 2 h or approximately on completion of the conventional skeletal scintigraphy. Vital staining of the leukocyte preparations with fluoresceine-di-acetate (FDA) after labelling consistently demonstrates 95% in vitro viability (Fig.2).
1 1 Ci = 3.70 X 10lo Bq.
IAEA-SM-247/39
w h ite
479
c e lls
(separation,labelling)
reinjection after
F IG .l.
3.
20 '
Separation and labelling procedure fo r leukocytes using n l In-acetylacetone.
LEUKOCYTE SCINTIGRAPHY IN SKELETAL INFECTIOUS DISEASE
3.1. Rationale Bone scanning with technetium-labelled bone seekers is a highly sensitive method when seeking occult bone lesions. Since malignant and benign bone disease results in radiotracer uptake, false-positive results must be expected when
480
G E O R G I et al.
FIG.2. Comparison o f microscope photos o f a FDA-treated leukocyte pellet, made with normal conditions (upper) and with UV-light (lower) demonstrating which cells are viable.
seeking infectious bone disease [3]. Since gallium citrate is additionally a bone seeker this is also true for this radiotracer [4]. Infectious lesions, which demon strate an infiltration of polymorphcellular leukocytes, can potentially be demonstrated scintigraphically with labelled leukocytes, regardless of the organ involved [5—8]. Most experience in the use o f the m In-AAL has been in patients referred with suspect skeletal infectious processes. All such patients having positive findings on conventional skeletal scintigraphy for suspect bone infection were then studied with m In-AAL, 100 studies on 97 patients having been performed thus far. Final diagnoses have been established in 79 patients. 3.2. Methodology Immediately after drawing 20 ml blood for leukocyte labelling as described above, 15 mCi " T cm-MDP were injected for skeletal scintigraphy. Bone scinti grams were obtained approximately 2 h after injection and before reinjection of
IAEA-SM-247/39
481
the 3 mCi H1In-AAL. Leukocyte scintigrams were usually made 20—24 h after leukocyte reinjection. A gamma camera (LFOV, Searle, 400 keV collimator, window only at 240 keV ± 15%) was used in the studies. In order that optimal topographic matching would be possible between the leukocyte and skeletal scintigrams, a followup bone scan was done in the same position as that used in the leukocyte study, using a window setting of 140 keV ± 15% for this scan. Usually the hepatic and splenic localization of m In-AAL is also imaged to ensure that the preparation is satisfactory in vivo. High background, markedly pre dominant splenic localization and particularly lung localization are evidence of an unsatisfactory preparation (Fig.3). 3.3. Results Twelve of the 79 patients evaluated had a suspect acute haematogenous osteomyelitis (Table I). In only four of these patients were the leukocyte scintigrams true positive whereas the skeletal scintigrams demonstrated marked, significant radioisotope uptake due to reactive, altered bone metabolism in all. In six patients with false negative leukocyte scintigrams, the histology demon strated a sclerosing type o f osteomyelitis. With one exception, these latter results were observed in young individuals, in whom the X-ray examination included the differential diagnosis o f a bone tumour. Of the other two FN studies, one was in a 13-yr-old with chronic bland osteomyelitis and periostitis, and the second in a 60-yr-old woman with reinfection 10 weeks after removal of a soft tissue abscess with skeletal involvement. All patients with no infectious process were correctly identified with no FP studies. In the group with suspected osteomyelitis following trauma or surgical intervention, the infection site was demonstrated in 75% (Table I). In three patients with fracture and surgical repair, who demonstrated bacteriological findings of osteomyelitis, the leukocyte scintigrams were falsely negative. It was possible to rule out correctly the presence of an osteomyelitis in nearly all patients who had no infection. Only one patient 2 yr post nailing of a fractured hip had a (FP) exam, since bacteriological cultures were negative. However, radiographs and other tests in this patient also supported the presence of infection, and may indeed not be an FP result. In the small chronic osteomyelitis group (Table I) of five patients, one TP was found during acute reactivation but a FN result occurred in a patient with bacteriologically positive chronic osteomyelitis and fistula. All patients without infection were correctly identified with no false positives. Five patients with tuberculous osteomyelitis (Table I), even during reactivation demonstrated no leukocyte accumulation on scintigraphy, analogous to the classical histological picture associated with this disease process. All studies were negative in these patients.
482
G E O R G I et al.
p la te le ts s e p a ra tio n and
15 ml
labelling
blood
( H e p a t i n 0 .1/10
ml)
resuspension in autol. plasma ( r ei nj ect i on
after 2 0 ’ )
FIG.3. Separation and labelling procedure fo r platelets using 111In-acetylacetone.
Verification of infection associated with the implantation of an orthopaedic joint prosthesis (Table I) could be extremely helpful. Of the 13 examined patients with hip joint prothesis nine were removed surgically affording com parison with scintigraphic findings. Three of these patients with findings of infection had negative scintigraphic findings and are recorded as FN (Table I). In one of the latter patients a question remained as to whether the surgical infection was a sterile abscess. Again almost all absence of disease was verified with only one FP. The latter case was not histologically or bacteriologically
IAEA-SM-247/39
483
TABLE I. RESULTS OF SCINTIGRAPHY WITH INDIUM-111-LABELLED LEUKOCYTES IN PATIENTS WITH INFLAMMATORY BONE DISEASE
Scintigram versus final diagnosis Initial clinical impression groups
Patients TP
FN
TN
FP
I.
Acute haematogenous osteomyelitis
16
4
8
4
0
II.
Post-trauma or surgery osteomyelitis
20
9
3
7
1
III.
Reactivated chronic osteomyelitis
5
1
1
3
0
IV.
Reactivated tuberculous osteomyelitis
5
0
4
1
0
V.
Infection adjacent to hip prothesis
13
4
3
5
1
VI.
Infectious joint disease
20
3
2
14
1
21
34
3
Total
79
21
verified by external biopsy, but strong clinical findings, including response to therapy, suggest that the patient indeed may have had infection. In the group of patients with clinically suspected spontaneous infectious processes, most were in the hip joints. Three (TP) of five patients with bacteriologically verified disease were identified (Table I). One patient with a FP result had radiographic signs of bone destruction in the acetabulum, but infection could not otherwise be proven. 3.4. Discussion The 11‘In-labelled leukocyte examination of 79 patients revealed TP results in 21 and TN results in 34 instances, so that 70% of all examinations resulted in a correct diagnosis, that is, accuracy. Three FP and 21 FN were observed resulting in a sensitivity o f 50% and a specificity o f 92%.
4 84
G E O R G I et al.
The three false-positive results are of particular importance to the clinic. Since all clinical symptoms existed in two of these cases, but were not supported histologically and bacteriologically, the question remains whether these were really false-positive findings. The normal localization o f labelled leukocytes in the bone marrow did not appear to contribute false-positive findings in these groups of patients, although a priori this might have been expected. Problems of interpretation due to non-symmetrical radioisotope uptake in bone marrow were common, but appear to present no serious difficulty. The most significant apparent conclusion from the very high specifity (92%) is that use of 111In leukocytes can aid in the diagnostic dilemma of infectious skeletal processes. As all the 42 cases positive for infectious disease in this study had positive con ventional skeletal scintigraphy as well as the 37 patients with no infectious process, it is apparent that high sensitivity scintigraphic techniques are available for screening of skeletal infectious processes. On the other hand, the specificity of conventional skeletal scintigraphy is low and, as can be seen from the low number of FP results with m In-AAL, a positive result is very reliable in verifying the presence of infectious disease. Thus in 25% of the highly suspect group after skeletal scintigraphy an infectious process can be identified with a 92% reliability, considerably reducing the number of patients who need more extensive investi gation to verify the presence of infection. False negative results are probably to be expected with the use of labelled leukocytes in processes such as tuberculosis which are known to be associated with a minimal leukocyte reaction. Perhaps a similar explanation holds for the six cases in the haematogenous suspect group with sclerosing osteomyelitis, which is probably associated with a considerable element of chronicity. In these patients there appears to be an inordinate intensive reactive bone formation to a not very impressive leukocyte response. Together these FN cases accounted for about one half of the group of FN. The results of leukocyte scintigraphy in patients with suspect skeletal infectious disease may be summarized as follows: (a)
When a positive result is obtained, an infectious bone infection must be considered to exist. Specificity o f the method is around 90%. (b) False negative results are generally expected in sclerosing osteomyelitis and in tuberculous bone lesions. (c) Leukocyte scintigraphy does not permit differentiation between infectious and neoplastic bone lesions. This is particularly true of the sclerosing osteomyelitis. (d) Positive leukocyte scintigraphy appears to be helpful in combination with skeletal scintigraphy when endoprostheses are examined, since the method permits differentiation between mechanical loosening and infection in about one half of the cases.
485
IAEA-SM-247/39
4.
USE OF INDIUM-111 -ACETYLACETONE-LABELLED THROMBOCYTES (INDIUM-111-AAT) IN RENAL TRANSPLANT IMMUNE DISEASE
In our experience harvest and isolation of autologous thrombocytes for labelling purposes is somewhat more demanding than analogous procedures for leukocytes. The methodology for thrombocyte separation and labelling is similar to that for leukocytes (Fig.3), but differs in the details described below. Use of substances such as dextran to accelerate precipitation of red blood cells must be avoided, as this results in clumping o f thromocytes and poor recovery. Furthermore, a specialized problem has been encountered with the blood of renal transplant patients, secondary to the presence and effects of numerous medications in the patients’ sera, especially immune-suppressing pharmaceuticals. To avoid this latter difficulty, variable densities of Metrizamide solution must be used to underlay the diluted blood, proportionate to the osmolality of the sera. Four solution densities have been found satisfactory for the range of sera osmolalities as follows: Mosm.
Metrizamide density
290
1.070
295
1.075
300
1.080
310 or greater
1.088 max.
Underlayment of the blood with Metrizamide using a long spinal needle is essential, as mixing o f Metrimazide with the blood would damage the platelets. With differential centrifugation thrombocytes localize at the density gradient on top of the Metrizamide with the other blood elements formed passing to the bottom of the container. Dilutions of the carefully pipetted, harvested thrombocytes in physiologic buffer and recentrifugation, usually twice, results in a high-yield, high-purity thrombocyte pellet, which is rapidly labelled with the previously reported 111In-acetylacetone technique. Beginning with as little as 20 ml of patients’ whole blood, the resultant m In-acetylacetone thrombocyte (l u In-AAT) pellet can contain as much as 5 mCi, i.e. very high specific activity. The n i In-AAT are resuspended in autologous serum to ensure continued viability. Excellent in vivo viability of the platelet preparation has been demonstrated in nearly 200 animal studies with a near normal half-life of 4.5—5 d, with pre ponderant uptake in the spleen, lesser amounts in the liver and bone marrow, and no pulmonary localization. Experimental thrombi demonstrate the expected localization of the thrombocytes. Clinical studies in renal transplant patients are progressing. The presence of immune reactions in transplant kidneys of the hyperacute, acute and chronic types are difficult to differentiate from other renal abnormalities on the basis of
486
G E O R G I et al.
classical function studies only. Involvement o f thrombocytes in the transplant immune reaction is well recognized, and serves as the rationale for employing the H1In-AAT preparation in this differential diagnostic problem with renal transplants. Serial studies are attempted in as many patients as possible to develop a spectrum of results to characterize the onset of development of immune reactions as well as the behaviour of such reactions once present. Patients receive sterile pyrogen-free 1—2 mCi o f i n In-AAT, consistent with acceptable radiation doses, with high count density large field o f view gamma camera pelvic images obtained at 2, 4 and 24 h post injection. Images also are obtained o f the liver and spleen to demonstrate preferential splenic localization and absence of pulmonary uptake. The images over the pelvis to visualize the transplant kidney uniformly display bone marrow uptake, with care being taken to exclude liver and spleen from the field of view. In this respect, the images are somewhat analogous to the use of sulphur colloid in renal transplant evaluation, so that proportionate uptake on a grading scale between kidney and bone marrow is possible. Correct autologous thrombocyte sample processing has been found to be paramount in this type of patient whose serum is replete with various administered drugs, in order to achieve dependable m In-AAT preparations. Clinical experience so far ranges from patients with no immune reaction to acute rejection. The population o f patients studied and the duration o f the study is not yet sufficient to determine the full value of this adjunct to assessing renal immune reaction disease.
REFERENCES [ 1] SINN, H., et al., Die Markierung von Erythrozyten mit radioaktiven Indiumisotopen, Nucl.-Med. 13(1974) 180. [2] SINN, H., SILVESTER, D J., Simplified cell labelling with indium-111 acetylacetone, Br. J. Radiol. 52 (1979) 758. [3] McDOUGALL, I.R., Diagnosis of abscesses by radionuclide scanning, Scott. Med. J. 24(1979) 263. [4] FORGACS, P., et al., Gallium scanning for the detection of abdominal abscesses, Am. J. Nucl. Med. 65 (1978) 949. [5] COLEMAN, R.E., et al., Indium-111 labelled leukocytes in the evaluation of suspected abdominal abscesses, Am. J. Surgery 139 (1980) 99. [6 ] COLEMAN, R.E., WELCH, D., Possible pitfalls with clinical imaging in indium-111 leukocytes, J. Nucl. Med. 21 (1980) 122 (Concise communication). [7] McAFEE, J.G., THAKUR, M.L., Survey of radioactive agents for in vitro labelling of phagocytic leukocytes: 1. Soluble agents, J. Nucl. Med. 17 (1976) 480. [8 ] McDOUGALL, I.R., Evaluation of 111-indium leukocyte whole body scanning, Am. J. Roentgenol. 133 (1979) 849.
I A E A - S M - 247/95
KINETICS AND MIGRATION OF INDIUM-111-LABELLED HUMAN LYMPHOCYTES D.A. GOODWIN, J.R. HECKMAN, L.F. FAJARDO, A. CALIN, S.J. PROPST, C.I. DIAMANTI Department of Nuclear Medicine, Veterans Administration Medical Center, Palo Alto, California and Stanford University School o f Medicine, Stanford, California, United States of America
Abstract KINETICS AND MIGRATION OF INDIUM-111-LABELLED HUMAN LYMPHOCYTES. The indium-111-oxine method was used to label autologous lymphocytes and study their kinetics and migratory patterns in 16 patients with chronic inflammatory disease. Nearly pure lymphocyte preparations (~10 8 cells) were obtained from the interface of a ficoll-hypaque gradient, re-suspended in 5 ml saline and incubated with 500 ß Ci indium-111-oxine for 30 minutes. The labelling yield was 40-70%. High-resolution electron microscope radioautographs with 0.3-0.5 pm grain size showed predominantly nuclear labelling of lymphocytes. Whole-body images were made 18-24 hours after IV injections of 100—500 f i d of labelled autologous lymphocytes. Recovery at 15 min was 25% and blood disappearance Tx for the slow phase was 53 hours. Activity is normally seen in the spleen, liver and bone marrow (in descending order of concentration) as well as cervical and inguinal lymph nodes. Any concentration outside these areas, as well as asymmetrical uptake in lymph nodes, was considered abnormal. Four patients also had indium-111-WBC (mixed leukocyte) scans within 7 -1 0 days for comparison. Eight patients had abnormal scans and one had a large spleen and one expanded bone marrow. Three patients with chronic osteomyelitis (one histologically proven) had positive lymphocyte scans with normal or only faintly positive WBC scans, and a fourth with acute osteomyelitis and a positive WBC scan had abnormal regional nodes. All large joints were intensely positive in a patient with a plasma cell disorder and arthritis. Lymphocyte viability is shown by the ability of the labelled cells to circulate and concentrate in regional lymph nodes and chronically inflamed tissues.
INTRODUCTION The ability to trace the migration of cells in vivo by means o f radioisotopic labels was chiefly responsible for the discovery of the normal recirculation of lymphocytes in mammals [1 ]. Until recently, however, because of the lack of a suitable radionuclide labelling method, there have been relatively few human 487
488
G O O D W I N et al.
studies. On the other hand, the literature concerning in vitro lymphocyte function is immense. Indium-111-oxine has recently been shown to be an ideal labelling reagent for rat lymphocytes [2]. The indium-111-label remained associated with the lymphocyte both in vitro and in vivo, and the cells migrated normally into lymphoid tissue. Results similar to those of Rannie using indium-111-labelled lymphocytes in Balb/C mice as well as in rats, have been reported by Frost [3]. Recent clinical studies have shown that indium-111-oxine is an ideal labelling reagent for neutrophils and platelets in abscess and clot localization [4-6]. This method has also been used successfully to study lymphocyte kinetics in two normal subjects and two patients with Hodgkins disease [7]. The clinical use of indium-111-labelled autologous lymphocytes in various chronic inflammatory diseases is described.
METHOD Nearly pure lymphocyte preparations (~ 1 0 8 cells) were obtained from the interface of a ficoll-hypaque gradient, re-suspended in 5 ml saline and incubated with 500 /xCi indium-111-oxine for 30 minutes [8].1 The details of the cell separation and labelling are given in the Appendix. Whole-body images were made 18—24 hours after IV injection of 100—500 ß d of labelled lymphocytes. Sixteen patients were studied; four of these patients also had indium-111-WBC (mixed leukocyte) scans within 7 -1 0 days for comparison. In five patients the 15-min recovery, blood disappearance half-time and percentage activity in circulating lymphocytes was calculated. High resolution electron microscope radioautographs were obtained on the labelled lymphocytes before injection.
RESULTS The blood kinetic data are shown in Table I. Almost pure lymphocyte preparations were obtained with little contamination from platelets, granulocytes or RBCs. The recovery at 15 min ranged from 6 to 50% (x = 21%) and the Ti. was 3 0 -6 7 h (x = 53 h). The EM radioautographs showed predominantly nuclear labelling of the lymphocytes (Fig.l). The results of the lymphocyte scans and diagnosis are shown in Table II. The normal whole-body scan showed spleen, liver and bone marrow activity (in descending order of concentration) as well as cervical and inguinal lymph 1 1 Ci = 3.70 X 1010 Bq.
IAEA-SM-247/9S
489
TABLE I. INDIUM-111-LYMPHOCYTE KINETIC DATA Percentage recovery injected cells ( 15 min)
Half-life (hours)
Percentage activity in circulating cells
48.3
67
30
76
Patient
No. of cells injection
R. McB.
0.71 X 108
J.S.
0.7 X 108
14.3
Z.H.
1.2 X 10®
20.4
—
75
E.F.
1.6 X 108
21.5
65.5
96
F.J.
4.0 X 108
42.6
67.2
92
F I G .l .
6 .6
E M radioautograph o f a human lym phocyte labelled with indium -111-oxine. Note
that at the section plane most areas o f activity are in the nucleus. Uranyl acetate and lead citrate
X 20 000.
490
G O O D W I N et al.
TABLE II. LYMPHOCYTE SCAN RESULTS Diagnosis
No. of patients
Scan +
Scan -
Chronic osteomyelitis
5
4
1
Rheumatoid arthritis
3
2
1
Chronic cystitis Reiter’s syndrome
2
2
0
Low back pain
2
0
2
Ankylosing spondylitis
1
0
1
Chronic lymphatic leukaemia
1
0
1
Mycosis fungoides
1
0
1
Chronic dermatitis
1
0
1
Total patients
16
nodes. Any concentration of lymphocytes outside these areas was considered abnormal as illustrated by the following cases. Four out of five patients with chronic osteomyelitis had abnormal scans, one of which is shown in Fig.2. Mr. R.B., a 33-year-old white male paraplegic with a high fever, and redness and swelling of the right thigh over an old area of chronic osteomyelitis around the mid femur, had a grade II positive WBC scan in this area (Fig.3). The bone scan showed marked increase in activity in the same area. The lymphocyte scan performed one week later showed practically no concentration of lymphocytes over the mid femur, but the right inguinal nodes were markedly positive. Note the lack of activity on the lymphocyte scan over the joints of the extremities with activity in spleen, liver and bone marrow predominating. Another patient, a paraplegic with chronic osteomyelitis of the left tibia, had in addition to uptake in the lesion, intense lymphocyte concentration in the left popliteal and inguinal nodes (Fig.4). A WBC scan for acute inflammation performed within one week was negative. Figure 5(a) shows a whole-body scan and spot views of the knees and ankles of a patient with chronic inflammatory joint disease. This patient, a 47-year-old white male, had a previous seven-year history o f hepatomegaly, Bence Jones proteinuria and increased IgA in the plasma with severe peripheral neuritis. He presented on this admission with a recurrence of bilateral ankle and foot pain with oedema. A bone marrow biopsy showed panhyperplasia. A tentative diagnosis was plasma cell disorder, with adenopathy, neuropathy, hepatomegaly and arthritis. The lymphocyte scan showed intense uptake in all joints of the extremities that seemed to be located in the synovium and not in the joint space (Fig.5(b)).
I A E A - S M - 247/95
491
(a)
R ig h t
F IG .2 .
L e ft
(a) A n terio r views o f knee and upper tibia in a patient with chronic osteomyelitis
o f the right tibia showing low grade uptake, (b ) Biopsy shows lym phocytic infiltration.
492
G O O D W I N et al.
"Сь*
¿..щ/*ftf-tc w / 6
¡В Ш Ш Й Й Ш Й ! F IG .3 .
Patient R. В. ( A ) Anterior view W BC scan; (В ) Anterior view lym phocyte scan
performed one week later. Both studies done 24 h follow ing injection.
DISCUSSION Chromium-51 is the label most frequently used for labelling human lymphocytes. This method has been used to study human lymphocyte kinetics in vivo [9]. In this study scintillation probes were used to study uptake of various organs by external counting. An initial rapid drop in blood activity was noted, followed by uptake in spleen, liver and bone marrow. A rather short blood Ti. of 1.7 days was thought to be due to an inability to detect the longer-lived cells in the blood because of the poor counting efficiency o f chromium-51. Decline of activity in the organs appeared to follow the death rate of the cells in the body as a whole. Studies of chronic lymphatic leukaemia patients showed a relative inability of leukaemia leukocytes to leave the circulation and enter some sites in the body. In another study utilizing chromium-51-labelled lymphocytes [10], there was a rapid equilibration following IV injection between the recirculating lymphocyte pool (RLP) and the blood, spleen and bone marrow. The size of the RLP was estimated to be approximately twice the blood lymphocyte pool, and it turned over about 12 times per day. These data were consistent with the RLP defined as a population of non-dividing small lymphocytes with
493
IAEA·SM· 24 7/9 S
.........
l
(a )
A
Ji... --------'
,...,(b)
A
8
FIG.4. (a) Anterior whole-body scans: (A) WBC; (B) Lymphocyte with uptake in L popliteal and inguinal noces. (b) Spot views of knee: (A) Lateral; (B) Anterior.
494
G O O D W I N et al.
(a )
f (b )
F IG . 5.
(a) Anterior whole-body scan shows intense uptake o f lym phocytes in the joints.
(b ) Spot views o f ( A ) anterior knees, (В & С) Lateral ankles show little activity in the jo in t space, and intense uptake around the jo in t in what appears to be synovium.
I A E A - S M - 247/95
495
an average life span of several weeks that recirculate from blood to lymphoid tissue and back to blood in a cycle measured in hours [11]. These studies have shown that much basic physiological information can be obtained with chromium-51-tagged lymphocytes in vivo. The use of indium-111-labelled autologous lymphocytes will permit the study of recirculation kinetics and migration of lymphocytes by the additional technique of clinical scintillation imaging. This method has the potential of providing diagnostic information in certain diseases in which lymphocytic infiltration is a characteristic feature, as well as showing the extent, severity and response to therapy. In addition, the technique may provide insight into the role of the circulating lymphocytes in the immunopathology of the lesions. The kinetics o f the recirculating lymphocyte in lymphoproliferative diseases and chronic lymphatic leukaemia may help to define the type of functional abnormality present, and in other neoplastic diseases might help to determine the extent of involvement of the circulating lymphocyte in host resistance.
ACKNOWLEDGEMENTS This work was supported by a grant from the Veterans Administration. We thank Carol Whiteleather, Carol I. Diamanti, and Kay L. Frank for their technical assistance.
REFERENCES [1] GOWANS, J.L., McGREGOR, D.D., The immunological activities of lymphocytes, Prog. Allergy 9 (1965) 1. [2] RANNIE, G.H., THAKIJR, J.L., FORD, W.L., An experimental comparison of radioactive labels with potential application to lymphocyte migration studies in patients, Clin. Exp. Immunol. 29(1977) 509. [3] FROST, P., SMITH, J., FROST, H., The radiolabeling of lymphocytes and tumor cells with n l indium, Proc. Soc. Exp. Biol. Med. 157 (1978) 61. [4] THAKUR, M.L., LAVENDER, J.P., ARNOT, R.M., et al., Indium-111 labeled autologous leukocytes in man, J. Nucl. Med. 18(1977) 1012. [5] GOODWIN, D.A., BUSHBERG, J.T., DOHERTY, P.W., et al., 111-In labeled autologous platelets for localization of vascular thrombi in humans, J. Nucl. Med. 19 (1978) 626. [ 6 ] DOHERTY, P.W., BUSHBERG, J.T., LIPTON, M., et al., The use of indium-111 labeled leukocytes for abscess detection, Clin. Nucl. Med. 3 (1978) 108. [7] LAVENDER, J.P., GOLDMAN, J.M., ARNOTT, R.N., et al., Kinetics of Indium-111 labelled lymphocytes in normal subjects and patients with Hodgkin’s disease, Br. Med. J. ii (1977) 797. [8 ] BOYUM, A., Separation of lymphocytes, lymphocyte subgroups and monocytes: A review, Lymphology (Stuttgart) 10 (1977) 71.
496
G O O D W I N et al.
[9] HERSEY, P., The separation and 51 chromium labeling of human lymphocytes with in vivo studies of survival and migration, Blood 38 (1971) 360. [10] SCOTT, J.L., DAVIDSON, J.G., MARIND, J.V., McMILLAN, R., Leukocyte labeling with 51 chromium: III. The kinetics of normal lymphocytes, Blood 40 (1972) 276. [11] FORD, W.L., GOWANS, J.L., The traffic of lymphocytes, Semin. Hematol. 6 (1969) 67.
Appendix LYMPHOCYTE SEPARATION AND LABELLING WITH INDIUM-111-OXINE (all procedures are performed in a laminar flow hood) 1.
Collect 2 X 49 ml venous blood in 50-ml sterile syringes, each containing 1 ml Heparin2 (approximately 1 0 0 0 units). 2. Transfer 45 ml of blood to each of two sterile 50-ml polypropylene centrifuge tubes, and add 5 ml Hetastarch3 to each. 3. Mix by gently pouring back and forth into a second 50-ml tube four to five times, and gently transfer to a third 50-ml tube, avoiding bubbles by pouring down the side. 4. Allow red cells to sediment in tubes at approximately 30° from the vertical for one hour (or less if sedimentation rate is faster). 5. Transfer platelet- and WBC-rich supernatants to new polypropylene tubes, and centrifuge at 450 G X 5 min to obtain the WBC buttons. Remove the platelet-rich supernatant and spin at 1350 G X 20 min to obtain PPP which is then set aside for the later step. 6 . Re-suspend the WBC buttons in each tube in Hanks’s Balanced Salt Solution 4 (HBSS) to a volume of 25 ml, making sure cells are well dispersed. ' 7. Draw up 10 ml of Ficoll-Paque5 into a 10-ml sterile syringe, then attach a 3-in 20-gauge ‘spinal’ needle to the syringe. 8 . Place tip of spinal needle in bottom of tube containing re-suspended WBCs and gently inject the 10 ml of Ficoll-Paque. (If volumes of cell suspension other than 25 ml are used the ratio of cell suspension to F-P is 8 ml : 3 ml.) 9. Centrifuge each tube at 800 G X 20 min, then gently remove tubes from centrifuge. The lymphocytes should appear as a hazy thin layer between the HBSS above and F-P below (RBCs and neutrophils are seen together in the cell button at the bottom of tube). 10. Using another spinal needle attached to a 3-ml sterile syringe, invert the bevel of the needle into the lymphocyte layer and carefully aspirate the lymphocytes into the syringe, gently moving the needle tip back and forth accross the lymphocyte layer. Continue aspirating until the interface between the HBSS and F-P appears as a sharp line with few cells.
Invinex (Division of Mogul Com). ‘Volex’ — 6 % Hetastarch (hydroxyethyl starch) in 0.9% NaCl from a sterile 10 -ml vial. 4 Hanks’s Balanced Salt Solution (lx ) without Phenol Red; Grand Island Biological Co. 5 Ficoll-Paque; Pharmacia Fine Chemicals; Division of Pharmacia Inc. 2
3
I A E A - S M - 247/95
497
11. Combine the aspirated cells into a single new sterile 50-ml polypropylene tube, and re-suspend the cells for washing in 3 volumes of HBSS. Centrifuge at 450 G X 5 min. The resulting lymphocyte button should appear white. 12. Discard the supernatant, and re-suspend the lymphocytes in 6 -1 0 ml of normal saline (depending on number of lymphocytes). Add the saline in 2 -3 ml increments and pipette up and down gently to break up any clumping. 13. Add l n In-oxine complex dropwise to the suspension, and incubate 30 min at room temperature, gently agitating every 5 minutes during incubation. 14. Bring total volume up to 15 ml with PPP from step 5 above, mixing as it is added, then spin at 450 G X 5 min. Remove and properly discard the radioactive supernatant after first counting and recording the activity. 15. Gently layer 3 ml PPP over the lymphocyte button and carefully remove and discard properly. Next, add 3 or 4 ml PPP to lymphocyte button and gently pipette up and down to break up any clumps. Then in increments, bring the volume up to 10-15 ml with PPP to re-suspend for injection. 16. Drea re-suspended lymphocytes into a sterile syringe with a spinal needle and measure dose in dose calibrator. Dose should be no more than 0.5 mCi. 17. Record weight of syringe before and after injections of labelled cells, and keep 0.1 ml of the injectate for a standard, if kinetic studies are to be done.
IAEA - S M - 2 4 7 / 129
USE OF INDIUM-11 1-OXINATE-LABELLED GRANULOCYTES AND THROMBOCYTES IN KIDNEY TRANSPLANTATION E.A. van ROYEN, J.B. van der SCHOOT, M.R. HARDEMAN, S. SURACHNO, J.H. ten VEEN, J. VREEKEN, J.M. WILMINK University Hospital, Wilhelmina Gasthuis, Amsterdam, The Netherlands
Abstract USE OF INDIUM-111-OXINATE-LABELLED GRANULOCYTES AND THROMBOCYTES IN KIDNEY TRANSPLANTATION. The diagnostic use of m In-oxinate-labelled granulocytes and thrombocytes in kidney graft rejection was studied in 39 transplant patients. Normal values were established for the deposition of these cells in stable, functioning kidney grafts. Although some 111 In granulocyte accumulation occurred in the graft during rejection, the increase was too slight to render the method suitable for the early diagnosis of rejection. Significant increased 111 In thrombocyte deposition was found during rejection periods, although large differences were observed in the degree of accumulation. Severity or type of rejection may relate to these differences. Post-transplantation follow-up by 111 In thrombocyte scintigraphy did not result in a much earlier diagnosis of rejection than classic clinical signs. However, more frequent bedside activity determinations might do so.
INTRODUCTION The diagnosis of kidney transplant rejection may sometimes be difficult because of the interference of complications (e.g. infective, vascular or urologie). Signs most often encountered are a swelling and tenderness of the kidney, a rise in temperature and blood pressure, a decrease in sodium and water excretion, diminution of renal function and proteinuria. The introduc tion of radionuclides in the post-operative management of transplantation has added to the correct diagnosis of rejection, ischemic damage or urinary flow obstruction by measurement of function and perfusion of the graft. However, a need still exists for a more specific sign of rejection, especially during the early stages of transplantation. Although lymphocytes and mononuclear cells are predominantly found in the interstitial tissues of the rejection graft, accumula tion of neutrophilic granulocytes and platelets also occurs. The availability of labelling techniques for these latter cells with indium-111-oxinate [1 ] prompted 499
500
van R O Y E N et al.
us to investigate the role of granulocytes and platelets in graft rejection. We were especially interested to discover whether these methods might be useful during the early phase.
METHODS Indium-111-oxinate The u l In-oxinate complex was prepared as described by Thakur and co-workers [1]. Recently a commercial preparation of the complex was pur chased from Byk-Mallinckrodt (Petten, The Netherlands). At activity reference time the specific activity was 1 mCi/25 jug 8 hydroxyquinoline (oxine).1 Granulocyte labelling The in vitro labelling was performed essentially according to Thakur [ 1], although a much lower oxinate concentration (2—5 ßg/m 1 cell suspension) was used to obtain functional granulocytes in the chemotactic assay [2]. About 50 ml of blood is taken from the patient in a 10 ml of ACD solution, mixed with 2 ml of methyl cellulose (2% in saline), after which aggregated erythrocytes are allowed to sediment by gravity during approximately 1 h at room tempera ture. The supernatant containing primarily granulocytes, but also platelets and a few erythrocytes, is centrifuged for 5 min at 500 X G and resuspended in 2 ml of buffered saline. After the addition of 500 jnCi l u In-oxinate and incubation for another 15 min, the suspension is reinjected. Routinely performed sterility tests revealed no bacterial contamination of the preparation. Thrombocyte labelling Platelet-rich plasma was prepared from 50 ml of ACD anticoagulated blood, acidified to pH 6.6 with extra ACD and centrifuged for 10 minutes at 500 X G. After 30 min the platelets were resuspended in buffered saline and incubated with 200—300 yuCi l u In-oxinate solution. Thrombocyte aggregation was not much affected by the labelling procedure and in vivo kinetics proved to be within the normal range [ 1]. Sterility tests were also negative.
1 1 Ci = 3.70 X 10IOBq.
I A E A - S M - 2 4 7/129
501
Scintigraphy In all patients scintigrams were performed at 20 min and 24 h after the injection of the labelled cells. In the thrombocyte studies, additional daily scintigrams were made up to 5 to 8 days. The acquisition time was 10 min, using a large-field-of-view gamma camera linked on line to a computer. Irregular regions of interest were defined around the renal graft and the opposite iliac region, aorta, abdomen, liver and spleen. The integrated counts in each region were corrected for injected activity and physical decay, and expressed in units (counts/10 min per 100 juCi per pixel). Patients All patients studied underwent kidney transplantation for chronic renal failure of various basic disorders at the University Hospital. The post-operative immunosuppression drug scheme consisted of initially 200 mg prednisolone and 1.5—3.0 mg/kg azathioprine daily. Rejections were treated with 1000 mg pulse dose corticosteroids and occasionally actinomycin D, heparin and dipyridamole. A modified low prednisone scheme was used in a small series of thrombocyte studies. Informed consent was obtained beforehand.
RESULTS Granulocyte studies In total 14 patients were studied, some on several occasions. Five patients were studied who had a stable functioning graft for at least 1—2 years, and who were furthermore in good health. Scans were performed 20 min and 24 h after injection. Kidney graft activity values ranged from 21—38 units for the 20-min scan and from 21—58 units for the 24-h scan. In all but one of the patients the 24-h value was higher than the 20-min value. Values obtained in the iliac region ranged from 10—21 units (20 min) and from 14—34 units (24 h). Also at this location the 24-h value exceeded the 20-min value in each patient. Relative graft activity was calculated by the graft/iliac region ratio. Values ranged from 1.65—2.36 for the 20-min scan and 1.50—2.30 for the 24-h scan. One of the patients proved to reject the graft chronically, resulting in nefrectomy some months later. The values in this patient did not differ from those obtained in patients in a stable condition. Nine patients were studied immediately or shortly after transplantation. Two of these patients were studied on two occasions. Three o f these patients were studied during a proven rejection crisis requiring therapy. In one patient high activity was recorded which proved
502
TABLE I. LEUKOCYTE ACCUMULATION IN GRAFT REJECTION3 Graft activity
Rejection Case
20
24 h
min
A
32
50
В
53
58 60
С N o rejection
3
van ROYEN et al.
25.4 ±3.4
7 studies
43.0 ± 12.8 (S.D
Numbers given in units = counts/10 min per 100 IJtCi per pixel.
TABLE II. THROMBOCYTE ACCUMULATION IN NINE NON-REJECTING GRAFTS3 Time after injection
20
Kidney graft
22.9 ±
Opposite iliac side Kidney/iliac ratio 3
24 h
48 h
72 h
8 .6
25.1 ± 8.0
26.7 ± 8.2
25.7 ± 7.9
22.0 ± 4.1
11.7 ± 3.7
14.4 ± 6.3
14.8 + 6.9
14.4 ±
1 1 .0
min
1.98± 0.49
1.83 ± 0.38
1.95+ 0.49
Values are expressed in units (counts/100 juCi per 10 min per pixel).
96 h
6 .8
1.95 ± 0.47
96 h
±
2 .0
2.0 ±0.39
16.8 ± 7.9 7.5 ± 2.7 2.17 ± 0.45
I AEA-SM-247/129
503
tim e a fte r injection
F IG .l.
Relative 111 In-oxinate-labelled throm bocyte accumulation in kidney graft as
expressed by the ratio o f graft activity ¡opposite iliac region activity during the days follow ing injection. S.E.M . is given in the figure. Stable late graft, n = 9, stable early graft, n = 9, rejection n = l l .
to be caused by a large wound haematoma around the graft. This patient has been omitted from the results given in Table I. Table I demonstrates that, although accumulation of graft activity tended to be somewhat higher during graft rejection, this increase was only slight. Thrombocyte studies In total 25 patients were studied at various times after injection and some on several occasions. Nine patients were studied who underwent transplanta tion at least 6 months earlier and were in a stable condition. All of them were followed up to 96 h after injection and some even longer. The results obtained are given in Table II. Also in thrombocyte studies the kidney graft values increased slightly between 20 min and 24 h after injection. The activity accumulation in the opposite iliac region also increased slightly, probably indicating active trapping of platelets in the RES of the pelvic bone. After 72 h values in both regions decreased, the iliac region somewhat faster than the graft resulting in an increased ratio at that time (Fig. 1). Nine studies were performed on 8 patients in the very first weeks following transplantation at a time when no signs o f rejection were present. The values obtained were not different from those in stable functioning grafts (Fig.l).
F IG .2 .
Scintigram 24 h after injection o f i n In-oxinate-labelled throm bocytes before
rejection (left), during rejection (centre) and after successful therapy (right).
IAEA-SM-247/129
505
Eleven studies were performed in 8 patients during rejection periods, as defined by clinical symptoms requiring anti-rejection therapy. In these patients the mean kidney graft values and graft/iliac ratios exceeded those obtained in stable grafts and early post-operative uncomplicated grafts except at 20 min after injection. Large differences were observed between patients during rejec tion resulting in a wide range o f values (Fig. 1). In most patients serial scintiscans taken during the rejection crisis showed a continuously increasing activity, unlike non-rejectors, suggesting ongoing platelet deposition in the former. Figure 2 shows the scintigraphic results during and after rejection in a patient who showed a marked accumulation of activity during rejection. In five studies in 4 patients we compared serum creatinine determinations with daily scintigraphic studies following repeated injections of labelled thrombo cytes in the weeks following transplantation. In one study abnormal graft accumulation occurred a few days before serum creatinine increased, and in another scintigraphy and serum creatinine pointed towards rejection on the same day. However, in one study creatinine increased one day before abnormal scintigraphic results were obtained, whereas in two others no abnormal accumulation of thrombocytes could be found in the presence of a clear-cut rejection. In 4 patients studies were made after successful anti-rejection therapy. Kidney values and kidney iliac ratios were still above normal though to a lesser degree, whereas in one study all values had returned to the normal range. Two patients were studied in whom graft function was impaired not by rejection, but through thrombosis of the renal artery. In one of them obstruc tion was seen in one of the three renal arteries present. The diagnosis, confirmed by arteriography, could be established by scintigraphy well before clinical signs developed, on the grounds of an abnormal thrombocyte accumulation near the vessel anastomosis.
DISCUSSION The results obtained with m In-oxinate labelled leukocytes in renal trans plant rejection do not sustain the preliminary finding of Frick and co-workers [3] that this method might be a useful indicator of transplant rejection. Although in our hands the use of labelled leukocytes proved to be a reliable method in the detection of inflammatory processes [4] and infection of vascular grafts [5], no major leukocyte deposition was found during rejection of a kidney graft in 3 patients. The few positive results obtained may be explained by platelets contaminating the leukocyte preparation as suggested by Smith and co workers [6]. When labelled platelet preparations were employed the results proved to be more consistent.
506
van R O Y E N et al.
First we established normal ranges of activity accumulation at the site of the graft and the opposite iliac region during the days following injection of labelled thrombocytes. Subjects who underwent successful transplantation at last 6 months before the study showed a rather narrow range of activity values. Also, patients who were studied shortly after transplant surgery showed similar activity values. Apparently, surgery itself did not appreciably increase platelets accumulation at the graft and its vessel anastomosis. Thrombosis of the renal artery, a complication which often leads to loss of the graft, could be located successfully with 111In-labelled platelets. In other conditions thrombi have also been detected by this method [7,8]. However, the detection of transplant rejection was our main concern. Regardless of the histological type of rejection (vascular or cellular), renal graft rejection proved to be accompanied in a high proportion of cases by an increased platelet accumulation at the site of the graft. In our opinion the large differences observed in accumulation of activity may relate either to the severity of the rejection crisis or to the subtype of the transplant immune reaction. It is tempting to speculate whether pharmacologic inhibition of platelet function in those crises which are characterized by intense platelet deposition would be useful to suppress the rejection. In order to answer this question one has to know in which kind of process the thrombocyte is involved, e.g. thrombosis, immune adherence, bleeding, etc. More precise relationship has to be established between platelet accumulation and the histologic picture as found in kidney biopsy. Most studies on transplant rejec tion do not provide conclusive evidence as to a predominant role of the platelet [9]. Our findings of increased platelet deposition in the rejecting graft are in accordance with those obtained by Smith and co-workers [6]. Nevertheless, when daily scintigraphic determinations of thrombocyte accumulation were compared with plasma creatinine as to its effectiveness in the early diagnosis of kidney graft rejection, the latter method proved to be superior to the former in three of the five studies performed. Possibly a more sensitive bedside detection method may compete more favourably, especially when measurements can be carried out more frequently. Such a method is at present under study. Nevertheless, the use of labelled platelets might enable us in the future to discriminate between various types of graft rejection.
ACKNOWLEDGEMENTS We gratefully acknowledge the skilful support provided by our chemical and nuclear medicine technicians. This study was supported by a grant from the Dutch Kidney Foundation.
IAEA-SM-247/129
507
REFERENCES [1] THAKUR, M.L., LAVENDER, J.P., ARNOT, R.N., SILVESTER, D.J., SEGAL, A.W., Indium-111 labeled autologous leukocytes in man, J. Nucl. Med. 18 (1977) 1014. [2] HARDEMAN, M.R., in Disease Evaluation and Patient Assessment in Rheumatoid Arthritis, Proc. V.I.S.R.A. Symposium, 10—11 Sep. 1979, Stafleu’s Scientific Publishing Company, Alphen aan de Rijn/Brussel. [3] FRICK, M.P., HENKE, C.E., FORSTROM, L.A., SIMMONS, R.A., McCULLOUGH, J., LOKEN, M.K., Use of 111 In-labeled leukocytes in evaluation of renal transplant rejection: a preliminary report, Clin. Nucl. Med. 4 (1979) 24. [4] ROEVEKAMP, M.H., HARDEMAN, M.R., van der SCHOOT, J.B., BELFER, A .I., Indium-111 labelled leukocyte scanning in the diagnosis of inflammatory disease: First results, Br. J. Surg. (submitted). [5] ROEVEKAMP, M.H., van ROYEN, E.A., KONING, J., HARDEMAN, M.R., “Indium-111 leukocyte scintigraphy: A new method in the diagnosis of infected vascular prosthetic grafts” , Proc. 29th International Congress of the European Society of Cardiovascular Surgery, Düsseldorf, July 1980. [ 6 ] SMITH, N.. CHANDLER, S., HAWKER, R.J., HAWKER, L.M., BARNES, A.D., Indium labelled autologous platelets as diagnostic aid after renal transplantation, Lancet ii (1979) 1241. [7] DAVIS, H.H., HEATON, W.A., SIEGEL, B.A., MATHIAS, C.J., JOIST, J.H., SHERMAN, L.A., WELCH, M.J., Scintigraphic detection of atherosclerotic lesions and venous thrombi in man by indium-111 labelled autologous platelets, Lancet ii (1978) 1185. [ 8 ] GOODWIN, D.A., BUSHBERG, J.T., DOHERTY, P.W., LIPTON, M.J., CONLEY, F.K., DIAMANTI, C.I., MEARES, C.F., Indium-111 labeled autologous platelets for location of vascular thrombi in humans, J. Nucl. Med. 19 (1978) 626. [9] HAMBURGER, J., CROSNIER, J., DORMONT, J., BACH, J.F., in Renal Transplantation: Theory and Practice, Williams & Wilkins Co., Baltimore (1972).
DISCUSSION
on the previous three papers G. UCHIYAMA: Mr. Wellman, do you have any information on the possibility of differentiating malignant bone diseases from inflammatory processes, using i n In-AAL? H. WELLMAN: We have no information on the use of l u In-AAL in differentiating bone neoplasms from infections. G. UCHIYAMA: Is there any difference between m In-AAL and m In-IIIoxine leukocytes in the imaging of inflammatory processes? H. WELLMAN: We have no direct comparison between 111-AAL and i n In-oxine leukocytes in the same patients, so it is impossible to determine whether one preparation is better than the other. However, no absolute àlcohol is necessary with 111In-AAL, and fluorescein diacetate staining shows 95% viability with 11‘In-AAL. G. SUBRAMANIAN : Mr. Wellman, I would like to enquire whether any kinetic studies have been done with UIIn-AAL in man. If so, what is the halflife of the labelled cells? What percentage of injected activity is associated with white cells (not in whole blood) one hour, for instance, after injection? H. WELLMAN : Mr. Sinn has shown from some 600 animal studies that n i In-AAL has normal kinetics. Likewise, as stated in the paper, the post injection distribution of i n In-AAL is normal in patients and, in particular, shows no localization in the lungs. H. SINN: After the re-injection of the labelled leukocytes we were unable to detect at any time a transfer of activity to any other blood com ponent. One hour after injection of labelled autologous leukocytes into normal animals (healthy rats), we found an average value of 45 to 55% of the total administered activity in the circulating blood, and this is associated only with the white cells. E.A. van ROYEN: I should like to make two comments. Firstly, when labelled with m In-oxinate, neutrophilic granulocytes remain only for a short period in the lung after injection. When a scintigram is taken 24 h after injec tion, there is only slight activity in the lung. Secondly, regarding cell function, we tested our granulocyte preparation in a chemotactic ‘in vitro’ assay. It was possible to find labelling conditions which preserved 85% of the function. Thrombocyte function was tested both ‘in vivo’ and ‘in vitro’. ‘In vivo’ kinetics were comparable with standard methods. ‘In vivo’ function was demonstrated by the visualization of thrombus formation in one of the renal arteries of a graft. • B.D. BOK: There is no general agreement yet about the clinical useful ness of labelled leukocytes. As a matter of fact, it seems that some properties of the normal leukocytes, such as chemotaxis, are lacking in the labelled cells:
509
510
DISCUSSION
these cells are still alive but their function is impaired, as shown by ‘in vitro’ tests or morphological studies, e.g. using a scanning electron microscope which shows damaged cell membranes. This is true when using pyrophosphate and technetium, but is even worse if one uses Thakur’s indium oxine labelling. It is perhaps too early yet to use these labelled leukocytes in humans. J.H. KRISTIANSEN: Mr. van Royen, have you any experience in differentiating between rejection and acute tubular necrosis by means of labelled thrombocytes? E.A. van ROYEN: No, we have performed no detailed studies on this subject. However, we have never observed abnormal thrombocyte accumula tion in a graft which functioned badly shortly after transplantation. At this early stage, immunological processes are less likely to occur than acute tubular necrosis, due to the extracorporeal preservation of the graft. A.F. McLAUGHLIN: How sure are the three speakers that they are labelling neutrophils, lymphocytes or platelets, when one considers that their results depend on the specificity of the blood components? E.A. van ROYEN : Neutrophilic granulocyte preparations are always contaminated by platelets. Thus, results obtained with labelled granulocytes are theoretically influenced by thrombotic processes. However, the clear results obtained in infectious lesions render the method very useful. Platelet preparations do not contain appreciable amounts of other cells. After injec tion, l u In activity of the whole blood proved to be completely bound to the platelet fraction, even several days after injection. D. A. GOODWIN : Platelets and lymphocytes may be obtained from peripheral blood with very little contamination by other cell types. Neutro phils, however, cannot be separated easily from lymphocytes by simple clinical techniques. Nevertheless, in the clinical setting this does not prevent the very effective use of the mixture (usually 88% or more neutrophils) for abscess localization.
IAEA-SM-247/70
RADIOIMMUNODETECTION OF VARIOUS HUMAN CARCINOMATA WITH RADIOLABELLED ANTIBODIES TO TUMOUR-ASSOCIATED ANTIGENS , E.E. KIM, F.H. DELAND, J.R. SALYER, . S.J. BENNETT, D.M. GOLDENBERG University of Kentucky and Veterans Administration Medical Centers, Lexington, Kentucky, United States of America
Abstract RADIOIMMUNODETECTION OF VARIOUS HUMAN CARCINOMATA WITH RADIO LABELLED ANTIBODIES TO TUMOUR-ASSOCIATED ANTIGENS. A total of 101 patients with histologically confirmed tumours were investigated by radioimmunodetection using radiolabelled antibodies to carcinoembryonic antigen (CEA), alpha-fetoprotein (AFP) or human chorionic gonadotropin (hCG) to determine whether tumours containing these antigens could be localized by external scintigraphy. Following the intravenous infusion of 1—2.5 mCi (100-500 ng IgG protein) of 131I-labelled antibodies, images of the chest and abdomen were obtained using a gamma camera, usually at 24 and 48 h. Computer-assisted subtraction of 99 Tcm-simulated background activities was used to increase lesion detectability. Primary and secondary tumour sites could be identified, and 84, 87 and 74% sensitivities by anti-CEA IgG scans were achieved for colorectal (42 patients), ovarian (21 patients), and lung (22 patients) cancers respectively. Lesions missed were usually less than 2 cm in diameter. Except for 3 seminomas and 1 embryonal carcinoma, all germ cell tumour sites were demonstrated by anti-AFP IgG (5 patients) or anti-hCG IgG (12 patients) scans. No false-positive sites were noted. The level of circulating antigens had no significant effect on tumour radioimmunodetection. The average tumour-to-non-tumour count density ratio on images ranged from 1.5 to 2.5, and the tumour-to-non-tumour image contrast ratio was enhanced more than twofold by the subtraction technique. In conclusion, radioimmunodetection offers a relatively specific method for localizing colorectal, ovarian and lung cancers, and also has a potential diagnostic significance in the evaluation and staging of germ cell tumours.
INTRODUCTION The l o c a l i z a t i o n o f tumors by t h e m a j o r i t y o f r a d i o p h a r m a c e u t i c a l s c u r r e n t l y us ed i n n u c l e a r m e d i c i n e t o d a y i s based p r i m a r i l y on a l t e r e d r e g i o n a l p h y s i o l o g y . Such l o c a l i z a t i o n s h o u l d be e x p e c t e d t o o c c u r i n a s s o c i a t i o n w i t h o t h e r d i s e a s e c o n d i t i o n s c h a r a c t e r i z e d by s i m i l a r a l t e r e d r e g i o n a l p h y s i o l o g y . The immunological a pp r o a c h t o t umor d e t e c t i o n 511
512
K I M et al.
utilizing anti gen-anti body reactions seems to offer a specific method. However, success has been elusive since Pressman [1] suggested that radiolabeled antibodies might facilitate localization of tumors. Of prime importance are the proper ties of the antibodies used, primarily their tumor-specificity. An enormous effort has gone into attempts to detect cancerspecific antigens by immunizing various animals with human cancer tissue, but such efforts have been disappointing. In the absence of a known human tumor-specific antigen, attention has turned to the cell-surface glycoprotein, CEA, which is one of the best characterized tumor-associated oncofetal antigens. Very sensitive radioimmunoassays for CEA are currently used in clinical practice to monitor disease activity in proven or suspected cancer patients. Goldenberg et al. [2] demonstrated selective localization of transplanted human colonic cancer (GW-39) in hamsters by external scintigraphy, using I-131-labeled antibodies to CEA and improved localization of the tumors was achieved by using affinity-purified antibodies to CEA [3]. We have developed an immunoscintigraphic technique for tumor detection, called radioimmunodetection, which employs a computer-assisted subtraction process [4]. Our first clinical trials successfully localized CEA-containing tumors using this technique [5,6,7]. We have recently extended our studies to the use of antibodies to other tumor-associated antigens, such as AFP and hCG [8,9]. The purpose of this paper is to summarize our current experience with radioimmunodetection of tumors using radiolabeled antibodies to CEA, AFP or hCG in various human carci nomata. MATERIALS AND METHODS Subjects for this investigation included 101 patients with histologically confirmed diverse primary, metastatic and/or recurrent tumors (42 colorectal, 21 ovarian, 22 lung carci nomas and 17 germ cell tumors), and were investigated by immunoscintigraphy using I-131-labeled antibodies to CEA (85 cases), AFP (5 cases) and hCG (12 cases). Tumor sites detected in our study were confirmed by other procedures, such as biopsy, surgery, autopsy, radiography, ultrasono graphy, computed-tomography, or by firm clinical evidence. Prior to the injection of radiolabeled antibodies, all patients were skin-tested for hypersensitivity to goat IgG
IAEA-SM-247/70
513
and were given 10 drops of Lugol's solution by mouth daily to block thyroidal uptake of radioactive iodine. Each patient received an intravenous infusion of 1-2.5 mCi of radiolabeled hyperimmune goat antibody which was prepared against CEA purified from a human colon carci noma or hCG from human urinary hCG. We obtained goat antiAFP antiserum from E. Alpert (Massachusetts General Hospital, Boston, Mass.). After heat-inactivation of complement, antiCEA or hCG anti serum was adsorbed with human erythrocytes and then purified by affinity chromatography. The anti-AFP antiserum IgG was isolated by DEAE-cellulose column chroma tography. The antigenic specificity of the anti serum was confirmed by gel immunodiffusion, immunoelectrophoresis and radioimmunoassay. The purified goat antibodies were labeled with 1-131 by the chloramine-T method, yielding a specific activity of 5-10 y C i / pg of protein. Sephadex G-200 column chromatography showed that 90 to 95% of the radiolabeled antibody was IgG. The titers for CEA, AFP and hCG antibody were 1-2 x 10°, 1 x 105 and 0.5 x 10° respectively, as determined by radioassay. Each lot of labeled antibody was tested for sterility, pyrogenicity and acute toxicity by an independent laboratory (Scientific Associates, Inc., St. Louis, Mo.).1 Anterior, posterior and occasionally lateral images of the chest and abdomen were made with a gamma camera (LF0V, Searle), usually 24 to 48 hours following the injection of radiolabeled antibody. In order to subtract interfering blood-pool and free iodine radioactivity in the stomach, urinary bladder and thyroid, 0.5-1.0 mCi each of Tc-99m human serum albumin and Tc-99m sodium pertechnetate were administered intravenously just before imaging. The data were processed by a dedicated digital computer, and the difference between the Tc-99m and 1-131 activity was displayed in 8 color levels representing different amounts of radioactivity. Elaborate computer processing using a standard program, as well as normal ization of the heart with adjusting upper and lower thresholds, was utilized. Tumor-to-non-tumor count density (T/NT) and image contrast (T-NT/T+NT) ratios were measured from the images by outlining comparable areas of positive or negative radiolocal ization. RESULTS Table I summarizes the radioimmunodetection findings in 101 cancer patients. These results are based on the number 1 1 Ci = 3,70 X lO10 Bq.
514
K I M et al.
TABLE I. RESULTS OF RADIOIMMUNODETECTION FINDINGS IN 101 CANCER PATIENTS
Ca. type (cases)
Antibody IgG scan
+ /-
-
+
+ /-
-
Overall sensitivity (%)
No. of primary sites® +
No. of secondary sites3
Colorectal (42)
anti-CEA
9
1
1
27
0
6
84
Ovarian (21)
anti-CEA
12
0
0
9
2
3
87
Lung (22)
anti-CEA
15
1
4
5
2
3
74
anti-AFP
1
0
0
10
1
1
85
anti-hCG
5
0
0
5
2
4
71
Germ cell (17)
'
a +, tumor identified; - , tumor missed; +/—, equivocal scan result.
of tumor sites confirmed in these patients by the other diagnostic methods. The overall sensitivities (true-positive rates) of radioimmunodetection using anti-CEA antibodies in various cancer types were as follows: colorectal cancer, 84%; ovarian cancer, 87%; and lung cancer, 74%. Lesions missed were usually less than 2 cm in diameter, below the limit of the current detecting system. Except for three seminomas and one embryonal carcinoma, all germ cell tumor sites were demonstrated by the anti-AFP or anti-hCG scans. No false-positive sites were identified. One seminoma patient with a high serum hCG level had localization of the metastatic lesion by the anti-hCG scan. Another patient with embryonal cari noma and elevated serum AFP but normal serum hCG underwent anti-AFP as well as anti-hCG scans; however, only the anti-AFP scan showed the metastatic lesions. The method had a high specificity (putative true-negative rate), which was obviously dependent on the sensitivity of the other diagnostic procedures used for correlation. The tumor-to-non-tumor count density ratio derived from 30 images ranged from 1.5 to 2.5; the tumorto-non-tumor image contrast ratio was enhanced more than two fold by the subtraction technique. The level of circulating antigens appeared to have no significant effect on the success of the tumor imaging by radioimmunodetection. An example of our imaging results in CEA radioimmunode tection is shown in Fig. 1 for a patient with an undifferen tiated large cell carcinoma of the lung. His pre-operative plasma CEA level was 10.0 ng/ml. The (I-Тс) subtraction image
IAEA-SM-247/70
C E A ( l íe )
С X R
ANT F IG .l.
515
CHEST
A n ti-C E A subtraction ( I -Т с ) image o f the anterior chest shows a large focal area o f
abnormal activity in the right lung base, corresponding to ä large tum or mass with associated malignant pleural effusion seen on chest radiograph (C X R J. Diagnosis: Undifferentiated large cell carcinoma o f the lung. H : heart; L : liver; S: stomach.
of an anti-CEA IgG scan of the anterior chest shows a large focal area of abnormal activity in the right lung base, corresponding to a large mass with associated malignant pleural effusion seen on the chest radiograph. The original display in color showed better information content. DISCUSSION Our results in the radioimmunodetection of cancer in 101 patients demonstrate that radioantibodies to tumor-associated antigens are both safe and capable of locating tumors although they are not tumor-specific agents. An earlier clinical trial by Mach et al. [10], without utilizing the computer-assisted subtraction technique, failed to localize tumors with radiolabeled antibodies to CEA,, but their recent study [11] with our subtraction method confirmed our results, although they had a lower level of sensitivity. The discrepancy might be due either to different scanning techniques or to different interpretations of the scan results, especially for tumors within or adjacent to the stomach or urinary bladder. We have
516
K I M et al.
used elaborate computer processing, including standard sub traction programs as well as normalization of the heart with adjusting upper and lower thresholds. Some basic problems in the radioimmunodetection of cancer are related to the specificity of the antibody preparation, the avidity of the anti gen-anti body reaction, and the choice of radionuclide for optimal labeling. Goldenberg et al. [12] summarized pitfalls to be considered in the development of radiolabeled anti-tumor antibodies. The ideal immunodiagnostic assay for cancer should distinguish accurately between those patients with cancer and those with benign or normal conditions. The test should be organ-specific and allow for the localization of tumor masses not easily defined by other detection measures. It is interesting that we have observed no significant effects of circulating tumor-associated antigens on our radioscanning results. This suggests that the much larger quantity of CEA in or near the tumor is able to bind suffi cient radiolabeled anti-CEA so as to yield a satisfactory target-to-background activity ratio for imaging. It has been reported that more than 20% of the circulating radioactivity exists in a large molecular form in patients receiving antiCEA antibodies, presumably as anti gen-anti body complexes [13]. Further, it was found that injection of radiolabeled anti-CEA is not accompanied by a decrease in circulating antigen [13]. Radioimmunodetection may be more sensitive for tumor detection than radioassay of plasma for the corresponding tumor-associated antigen, since positive tumor radioimmuno detection was observed in patients with normal blood levels of CEA, AFP or hCG. Our clinical studies have demonstrated a less than 2% non-specific localization of benign, non neoplastic lesions in these cancer patients; a 13% non specific tumor localization rate resulted from normal goat IgG injected to a group of cancer patients [14]. Our experience in the clinical radioimmunodetection of cancer suggests that this method will be valuable as a diag nostic adjunct in evaluating and monitoring cancer patients, although it is still being improved. With better radioimmunopharmaceuticals, as well as improved instrumentation such as emission tomography, significant progress in the clinical application of this method is to be expected. Furthermore, the use of different antibody fragments of monoclonal anti bodies may make more specific localization of tumors faster and easier [11,15].
I A EA-SM-247/70
517
ACKNOWLEDGEMENTS Excellent technical assistance was provided by S. Patton, 11. Shah and R. Mueller. We are also grateful to Drs G. Simmons, F. J. Primus and W. Pettit for their continuing help in computer programming, radiolabeling, and immunochemistry. These studies were supported in part by NIH grants CA-17742 and CA-25584, and by the Veterans Admini stration.
REFERENCES [1] PRESSMAN, D., Radiolabeled antibodies, Ann. N.Y. Acad. Sei. 69 (1957) 644. [2] PRIMUS, F.J., WANG, R.H., GOLDENBERG, D.M., HANSEN, H.J., Localization of human GW-39 tumors in hamsters by radiolabeled heterospecific antibody to carcinoembryonic antigen, Cancer Res. 33 (1973) 2977. [3] PRIMUS, F.J., MACDONALD, R., GOLDENBERG, D.M., HANSEN, H.J., Localization of GW-39 tumors in hamsters by affinity purified antibody to carcinoembryonic antigen, Cancer Res. 37(1977) 1544. [4] DELAND, F.H., KIM, E.E., SIMMONS, G., GOLDENBERG, D.M., Imaging approach in radioimmunodetection, Cancer Res. 4 0 (1980) 3046. [5] GOLDENBERG, D.M., DELAND, F.H., KIM, E„ BENNETT, S., PRIMUS, F.J., VAN NAGELL, J.R., ESTES, N., DESIMONE, P., RAYBURN, P., Use of radiolabeled antibodies to carcinoembryonic antigen for the detection and localization of diverse cancers by external photoscanning, New Engl. J. Med. 298 (1978) 1384. [6 ] KIM, E.E., DELAND, F.H., CASPER, S., CORGAN, R.L., PRIMUS, F.J., GOLDENBERG, D.M., Radioimmunodetection of colorectal cancer, Cancer 45 (1980) 1243. [7] VAN NAGELL, J.R., KIM, E„ CASPER, S., PRIMUS, F.J., BENNETT, S., DELAND, F.H., GOLDENBERG, D.M., Radioimmunodetection of primary and metastatic ovarian cancer using radiolabeled antibodies to carcinoembryonic antigen, Cancer Res. 4 0 (1980) 502. [8 ] KIM, E.E., DELAND, F.H., NELSON, M.O., BENNETT, S., PRIMUS, F.J., SIMMONS, G„ ALPERT, E., GOLDENBERG, D.M., Radioimmunodetection of cancer with radiolabeled antibodies to а-feto protein, Cancer Res. 4 0 (1980) 3008. [9] GOLDENBERG, D.M., KIM, E.E., DELAND, F.H., VAN NEGELL, J.R., Jr., JAVADPOUR, N., Clinical radioimmunodetection of cancer with radiolabeled antibodies to human chorionic gonadotropin, Science 20 8 (1980) 1284. [10] MACH, J.-P., CARREL, S., MERENDA, C., HEUMANN, D., ROENSPIES, U„ In vivo localization of anti-CEA antibody in colon carcinoma: Can the results obtained in the nude mice model be extrapolated to the patient situation? , Eur. J. Cancer (Suppl.) 1 (1978) 113. [11] MACH, J.-P., CARREL, S., FORNI, M., RITSCHARD, J., DONATH, A., ALBERTO, P., Tumor localization of radiolabeled antibodies against carcinoma, New Engl. J. Med. 303(1980) 5. [12] GOLDENBERG, D.M., PRIMUS, F.J., DELAND, F.H., “Tumor detection and localization with purified antibodies to carcinoembryonic antigen”, Immunodiagnosis of Cancer, Part I (HERBERMAN, R.B., MCINTIRE, K.R., Eds), Marcel Dekker, New York and Basel (1979) 265.
518
K I M et al.
[13] PRIMUS, F.J., BENNETT, S.J., KIM, E.E., DELAND, F.H., ZAHN, M.C., GOLDENBERG, D.M., Circulating immune complexes in cancer patients receiving goat radiolocalizing antibodies to carcinoembryonic antigen, Cancer Res. 40 (1980) 497. [14] GOLDENBERG, D.M., KIM, E.E., DELAND, F.H., BENNETT, S., PRIMUS, F.J., Radioimmunodetection of cancer with radioactive antibodies to carcinoembryonic antigen, Cancer Res. 40 (1980) 2984. [15] LEVINE, G„ BALLOU, B„ REILAND, J., SOLTER, D., GUMERMAN, L., HAKALA, T., Localization of 1-131 labeled tumor-specific monoclonal antibody in the tumor-bearing BALB/c mouse, J. Nucl. Med. 21 (1980) 570.
DISCUSSION D.M. TAYLOR: Your method is of great interest as it offers the only genuine possibility of tumour-specific, tumour-localizing radiopharmaceuticals. However, your tumour/non-tumour ratios are disappointingly low. How reproducible are your results from antibody batch to antibody batch? E.E. KIM: We have just begun evaluating more specific Fab antibodies for clinical studies, but I cannot give you any information yet. V.R. McCREADY: You image the labelled antibody at 24 h and inject the blood subtraction agents just before the images are taken. Might the small tumour/ background ratio be explained by differing rates of diffusion of the IgG and the labelled albumin from the circulation into the tumour? E.E. KIM: We were also disappointed by the low turn our/non-tumour ratios. With ideal IgG, which has exactly the same biological behaviour, we might detect tumours at an earlier stage. J.A. PETERSON: In our studies using cell-type specific antibodies against mammary epithelial cells for localization of breast tumours1, we have compared the tumour/normal tissue ratios using whole antibodies with those obtained using Fab fragments. With the whole antibody labelled with 1311, the ratios were in the order of 2—3 whereas, with the 13lI-labelled Fab fragments of antimammary epithelial antibodies, the ratios were between 4 and 10. This could possibly be explained by the more rapid clearance of the l31I-labelled Fab fragments from the animals. By 24 h after injection, 99% of the 131I-labelled Fab had been eliminated. E. TOUYA: Mr. Kim, have you studied benign lesions in the lung? E.E. KIM: No, not specifically. However, our overall clinical results have revealed a < 2% non-specific localization of benign, non-neoplastic lesions in the cancer patients we studied. 1 PETERSON, J.A., WILBANKS, T., MILJLtiR, S., KAUFMAN, L., ORDENDAHL, D., CERIANI, R.L., “Use of cell-type specific antibodies for radioimmunodetection of breast métastasés with a high-purity germanium camera” , these Proceedings, IAEA-SM-247/77 (Poster Presentation).
IAEA-SM-247/6
EVALUATION OF NEW RADIOPHARMACEUTICALS FOR TUMOUR LOCALIZATION T h e v a lu e o f t h e h u m a n t u m o u r x e n o g r a f t D.M. TAYLOR Institute for Genetics and Toxicology, Kemforschungszentrum Karlsruhe, Karlsruhe and Institute for Pharmacology (Radiotoxicology), University of Heidelberg, Heidelberg, Federal Republic of Germany
Abstract EVALUATION OF NEW RADIOPHARMACEUTICALS FOR TUMOUR LOCALIZATION: THE VALUE OF THE HUMAN TUMOUR XENOGRAFT. Human tumours growing in immune-deprived mice retain most of their human histological and biochemical characteristics throughout serial transplants. They could, therefore, provide more realistic models for screening potential new tumour-localizing radiopharmaceuticals than the rapidly growing, anaplastic, transplantable tumours of rats and mice. The uptake and deposition of 67Ga citrate, 57Co-Bleomycin and three other radiopharmaceuticals were com pared in nine human tumour lines, growing as xenografts in immune-deprived CBA/LAC mice, and in five long-established, transplantable mouse tumour lines. It is concluded that there are no major differences between the xenografts and the transplanted murine tumours as far as the pattern of radionuclide uptake and deposition is concerned. Consequently, the human tumour xenograft system appears to offer no real advantages over the transplanted rodent tumour for the general screening of potential new tumour-localizing agents. However, since the xenografts retain the essential human biochemical characteristics of the original tumour, including its response to chemo- or radiotherapy, the system may be of considerable value for the more detailed investigation of new agents which show good tumour-localizing potential in rodent tumours, especially for the evaluation of the functional information about the tumour and its response to treatment which may be obtained by scanning techniques in addition to the static image.
1. I N T R O D U C T I O N In t h e s e a r c h for n e w t u m o u r - l o c a l i s i n g a g e n t s w h i c h m a y b e e x p l o i t e d to p r o v i d e f u n c t i o n a l i n f o r m a t i o n a b o u t t h e t u m o u r a n d its r e s p o n s e to t r e a t m e n t , in a d d i t i o n to a s t a t i c i m a g e of the t u m o u r d i s t r i b u t i o n , it is i m p o r t a n t to h a v e a 519
L Tï to о
T A B L E I. C o m p a r i s o n of e7G a - c i t r a t e and 57C o - B l e o m y c i n in h u m a n tu m o u r x e n o g r a f t s and t r a n s p l a n t e d m u r i n e tumours
T u m o u r and M o u s e Strain
(24h) TBR+
5.5 4.6 15.8 7.5 3.3 3.1 4.9 7.7 5.3
± 0.6 ± 0.3 ± 3.5 ± 1.4 ± 0.3 ± 1.1 ± 0.8 ± 0.9 ± 0.8
1 .7 1 .3 5.0 2.4 1 .0 1 .4 2.1 2.4 1 .7
5.1 3.4 10.5 16.7 12.5
± 1.3 ± 0.2 ± 1.2 ± 3.3 ± 2.7
4.0 3.5 8.3 11.4 12.8
5 7C o - B l e o m y c i r % Dose/g -
0.36 ± 0.09 -
0.7 9 ± 0.08 -
0 . 9 0 ± 0.08 0.91 ± 0.31 0.27 ± 0 . 1 0 -
0.71 ± 0.04 0.11 ± 0.03
“L e t t e r s in p a r e n t h e s e s indicate the d e g r e e of d i f f e r e n t i a t i o n of the t u m o u r s M - M o d e r a t e , W - Well, P - Poor. + T u m o u r B l o o d C o n c e n t r a t i o n Ratio.
M i n i m u m of 3 m i c e per group.
(3h) TBR -
1 1 .7 -
39.6 21 .4 10.0 7.3 19.7 1 .3
TAYLOR
XENOGRAFTS Immune-deprived CBA/LAC HXHC1 A d e n o c a r c i n o m a C o l o n (M)* HXGC3 A d e n o c a r c i n o m a C o l o n (P) HXIC1 Anaplastic Carcinoma Colon HXCC1 A d e n o c a r c i n o m a C o l o n (W) HXGW3 9 C a r c i n o m a C o l o n HXKR1 A d e n o c a r c i n o m a R e c t u m (P) HXPR1 C a r c i n o m a R e c t u m (W) HXCL1 O a t C e l l C a r c i n o m a Lung HXHP1 C a r c i n o m a P a n c r e a s (P) MOUSE TUMOURS Lewis Lung Carcinoma C57 Gardner Lymphoma CBA/LAC В 16 M e l a n o m a C57 H a r d i n g - P a s s e y M e l a n o m a DBA/2 ADJ/PC 6 Plasmacytoma BALB/C
67G a - c i t r a t e % Dose/g
IAEA-SM-247/6
521
realistic model system for the eval u a t i o n of p o tential n e w r a d i o p harmaceuticals. A w i d e range of t r a n s p l a n t e d t u m o u r s . i n r a t s o r m i c e is a v a i l a b l e , and currently these represent the main test system for new agents. However, such tumours are rapidly growing and are frequently highly anaplastic, thus they are not n e c e s s a r i l y good models for the more slowly growing, and often moderately or well differ entiated, human tumor i n s i t u . The development of x e n o g r a f t i n g t e chniques d u r i n g the past decade n o w p e r m i t s m a n y t y p e s of h u m a n t u m o u r to be g r o w n in immune-deprived animals, usually mice [ 1J . Such tumour xenografts grow relatively slowly and retain the histological picture and the essential b i o chemical and karyo t y p i c characteristics of the t u m o u r o f o r i g i n f o r a t l e a s t t h e f i r s t 5 t o 10 t r a n s p l a n t g e n e r a t i o n s Г1.2]. In o r d e r t o e v a l u a t e t h e h u m a n t u m o u r x e n o g r a f t system as a mo d e l for the screening of p o t e n t i a l n e w tumour localising radiopharmaceuticals, the uptake of several c l i n i c a l l y u s e d or p o t e ntial agents has b e e n com p a r e d in up to nine d i f f e r e n t hu m a n tumour xenografts, growing in immuno-suppressed mice, and in five t r a n s p l a n t e d m u r i n e tumours.
2. 2.1.
MATERIALS AND METHODS Human Tumour Xenografts
The hum a n tumours w ere grown as h e t e r o t r a n s p l a n t s in C B A / L A C m i c e w h i c h h a d b e e n i m m u n e - d e p r i v e d , two weeks previously, us i ng the techniques described b y P i c k a r d e t al. [3] o r H o u g h t o n e t al. [4] . T u m o u r samples w e r e o b t a i n e d from p a tients at o p e r a t i o n and p l a c e d in ice-cold M e d i u m TC-199 c o ntaining penicillin, s t r e p t o m y c i n and, in some instances, g e n t a m y c i n . T u m o u r p i e c e s , 8 m m 3, w e r e i m p l a n t e d s u b c u t a n e o u s l y into b o t h flanks of m a l e or female mice wit hin 6 hours of removal of the tumour from the patient. S u b s e q u e n t transplants w e r e m ade w h e n t h e t u m o u r s r e a c h e d a d i a m e t e r o f 2 cm. A l l t h e tumours u sed in this study, e xcept HXGWC-39, were in the fi r s t to t he f i f t h t r a n s p l a n t g e n e r a t i o n . Brief histological details for each tumour are given i n T a b l e I. T h e m u r i n e t u m o u r s w e r e s t a n d a r d e x p e r i m e n t a l t u m o u r l i n e s m a i n t a i n e d in C B A / L A C , BALB/C, DBA/2 or C57 mice.
522
TAYLOR
2.2 R a d i o p h a r m a c e u t i c a l U p t a k e Stu d i e s Carrier-free 67Ga citrate (Philips-Duphar, Petten, The Netherlands) was injected intraperit o n e a l l y at a d o s e l e v e l of 65 k B q p e r m o u s e . The mice were killed between 0.5 and 2 4 hours after injection and samples of blood and tumour were taken for analysis. C o b a l t - 5 7 - l a b e ll e d Bleom y c i n (prepared by the author) - 200 kBq - was injected i n t r a venously 3 hours b efore kill i n g the mice. Indium-111-labelled Bleomycin (M.R.C.Cyclotron Unit, H a m mersmith Hospital, London, E n g l a n d ) , 9 9T c m - T e t r a c y c l i n e a n d 9 9T c m - S t r e p t o k i n a s e (prepared by the author) were injected intrap e r i t o n e a l l y - at do s e levels f r o m 2 0 0 - 1 0 0 0 k B q a n d t h e m i c e w e r e k i l l e d b e t w e e n 1 a n d 24 h o u r s later. 2.3.Assay Methods Radi o a c t i v i t y was m e a s u r e d in an automatic gamma-ray scintillation spectrometer. Subcellular fractionation, enzyme assays and the meas u re m e n t of the ra t e o f D N A s y n t h e s i s in the t u m o u r s w e r e p e r f o r m e d u s i n g m e t h o d s d e s c r i b e d p r e v i o u s l y [5,6]. 3. R E S U L T S T h e 24-h.our u p t a k e s o f 6 7 G a i n x e n o g r a f t s f r o m n i n e d i f f e r e n t h u m a n t u m o u r s a r e s h o w n i n T a b l e I, together with the corresponding tumour-to-blood concentration ratios (TBR). Table I also records similar data for the 3-hour uptake of ^ C o - B l e o m y c i n in some of the tumours. The p a t t e r n o f u p t a k e of 67Ga c i t r a t e b e t w e e n 0 . 5 a n d 24 h o u r s p o s t i n j e c t i o n i s c o m p a r e d f o r t w o x e nografts HXGWC-39 and HXCLI, and one m urine t u m o u r , A D J / P C 6 , i n F i g u r e 1. L i m i t e d s t u d i e s w i t h 111 I n - B l e o m y c i n d u r i n g t h e f i r s t 24 h o u r s a f t e r injection showed b r o a d l y similar results to those s h o w n f o r 6 7 G a i n T a b l e I a n d F i g u r e 1, e x c e p t t h a t in all t u m o u r s t h e u p t a k e of x llIn w a s b e t w e e n 2 a n d 8 t i m e s l o w e r t h a n t h a t o f 6 7 Ga. T wo other agents c o nsidered to have tumourl o c a l i s i n g p o t e n t i a l , 9 9T c m - T e t r a c y c l i n e a n d 99T c m - S t r e p t o k i n a s e , w e r e e x a m i n e d in a small
IAEA-SM-247/6
523
F IG .l. The tim e course o f uptake o f 67Ga in tw o human tum our xenografts (HXCL 1, H XGW C 3 9 ) and in the A D 3/P C 6 m urine plasm acytom a after injection o f the citrate com plex.
nu m b e r of tumours, but for both r a d i o p h a r m a c e u t i c a l s t he u p t a k e in the x e n o g r a f t s a n d the m u r i n e t u m o u r s wa s l o w a n d t h e T B R w a s l e s s t h a n one.
4. D I S C U S S I O N The d ata p r e s e n t e d in T a b l e I indicate that at 24 h o u r s p o s t i n j e c t i o n t h e a c c u m u l a t i o n o f 6 7 G a in the nine h u m a n tumour x e n o g r a f t s falls in the sa m e r a n g e as t h a t in the m u r i n e tumours. Howe v e r , t h e T B R is a b o u t 4 t i m e s l o w e r f o r t h e x e n o g r a f t s than for the m u r i n e tumours, due to a slower c l e a r a n c e of 67Ga from the b l o o d of the immunedeprived mice. The limited studies of the uptake k i n e t i c s , F i g u r e 1, i n d i c a t e t h a t i n b o t h t h e mu r i n e and h u m a n tumours the initial u ptake of 6 7 G a is r a p i d w i t h at l e a s t 80 p e r c e n t o f t h e m a x i m u m uptake o c c u r r i n g w i t h i n 6 hours. B i o c h e m i c a l studies h ave shown that in the human t u m o u r x e n o g r a f t s , as in m u r i n e a n d o t h e r tumo u r s , t h e m a j o r i n t r a c e l l u l a r s i t e o f 6 7G a d e p o s i t i o n is in t h e l y s o s o m e s [5,6 , 7 ] . S t u d i e s in b o t h m o u s e a n d r a t t u m o u r s h a v e s h o w n t h a t t h e r e is a s i g n i ficant pos i t i v e c o r r e l a t i o n b e t w e e n the u ptake of 67Ga in a s p ecific t umour and the a c t i v i t y of l y s o s o m a l e n z y m e s i n t h e t u m o u r [б}. C o m p a r i s o n o f t h e
524
TAYLOR
HUSiANTUM OURS IE
M URINETUM OURS 16
RMS
HPM
-----1-------- 1 -------- 1
nl--------------- L-----1---------------------- L
ico о too га ARYLSULPHATASECONCENTRATION(ENZYME UNITS/g)
300
FIG.2. Comparison o f the relationship betw een the 67Ga uptake and the actiyity o f the lysosom al en zym e aryl sulphatase in five human tum our xenografts (left) and in a series o f long-established transplantable murine tum ours (right). The regression line shown fo r the murine tum ours was calculated from the data in R ef. [6] and th at f o r the human tum ours from the com bin ed data from this paper and th at o f H am m ersley [S ]. The significance o f the correlation is P < 0 .0 1 fo r each series.
67Ga uptake and the act i v i t y of the lysosomal enzyme aryl sulphatase in 5 of the p r e s e n t series of human t u m o u r s , a n d i n a f u r t h e r 12 t u m o u r s r e c e n t l y s t u d i e d b y H a m m e r s l e y [8], s u g g e s t s t h a t t h e r e m a y be a similar r e l a t i o n s h i p in human tumour xenografts. T h i s r e l a t i o n s h i p is i l l u s t r a t e d in F i g u r e 2 w h i c h a l s o shows the c o r r e l a t i o n p r e v i o u s l y o b s e r v e d in a l a r g e r s e r i e s o f t r a n s p l a n t e d m u r i n e t u m o u r s [6]. C o m p a r i s o n o f t h e 6 7G a u p t a k e a n d t h e r a t e o f D N A s y n t h e s i s i n t h e s e a n d o t h e r [8] h u m a n t u m o u r x e n o g r a f t s s u g g e s t t h a t , as i n r a t a n d m o u s e t u m o u r s Г61 t h e 6 7 G a u p t a k e t e n d s t o b e g r e a t e s t i n t u m o u r s s howing h i g h rates of D N A synthesis, and hence, presumably, of cell proliferation. T h e m o r e l i m i t e d s t u d i e s w i t h 5 7C o - B l e o m y c i n also show no clear differences b e tween the human t umour xenografts and the transplantable murine t u m o u r s . In c o n t r a s t to 6 7 G a t h e r e a p p e a r s to be no s i g n ificant d i f f e r e n c e in the T B R v alues for 57C o - B L M in the two g r o u p s of tumours.
IAEA-SM-247/6
525
T h e u p t a k e o f b o t h 6 7 G a a n d 5 7C o - B l e o m y c i n h a s b e e n s t u d i e d b y Y e h e t a l . [9] i n t w o h u m a n t u m o u r x e n o g r a f t s g r o w i n g in the c o n g e n i t a l l y a t h y m i c 'nude' m o u s e . T h e r e p o r t e d u p t a k e o f b o t h r a d i o p h a r m a c e u t i c a l s in a n o n - H o d g k i n s l ymphoma, and of °7G a i n a m a l i g n a n t S c h w a n n o m a , a t 4 8 h o u r s p o s t i n j e c t i o n lie in t h e r a n g e o f v a l u e s o b s e r v e d in the pr e s e n t study. F r o m t h e s e s t u d i e s w i t h a l i m i t e d r a n g e of h u m a n x e n o g r a f t s , a n d o f r a d i o p h a r m a c e u t i c a l s , i t is c o n c l u d e d that there are no m a j o r d i f f e rences between the x enografts and the transplanted murine tumours as far as the g e n e r a l p a t t e r n of r a d i o n u c l i d e u p t a k e a n d d e p o s i t i o n is c o n c e r n e d . C o n s e q u e n t l y , th e h u m a n t u m o u r x e n o g r a f t s y s t e m a p p e a r s to o f f e r no real a d v a n t a g e s o v e r the t r a n s p l a n t e d r o d e n t t u m o u r for the g e n e r a l screening of potential n e w tumour localising agents. However, since xenografts retain the essential human biochemical characteristics of the o r i g i n a l tumour, i n c l u d i n g its r e s p o n s e to ch e m o o r r a d i o t h e r a p y [l,10], the s y s t e m m a y be of c o n s i d era b l e val u e for the m ore d etailed inve s t i g a t i o n of new agents, which have shown good tumour localising p o t e n t i a l in an i m a l t u m o u r screens; e s p e c i a l l y for the e v a l u a t i o n of the a b i l i t y of the agent to p r o vide clinically useful functional information about t he t u m o u r a n d its r e s p o n s e to t r e a t m e n t t h r o u g h quantitative and dynamic imaging techniques. A c k n o w l e d g e m e n t s - T h i s w o r k w a s c a r r i e d o u t in the Institute of Cancer Research, Sutton, Surrey, England, a n d I s h o u l d like to t h a n k Drs J.A.Houghton and R.G. P i c k a r d f o r s u p p l y i n g t h e x e n o g r a f t s . In p a r t i c u l a r I w i s h to t h a n k Dr. P . A. G . H a m m e r s l e y f o r g e n e r o u s l y a l l o w i n g me to quote from his un p u b l i s h e d data.
REFERENCES [1 ] HOUGHTON, J.A., TAYLOR, D.M., Maintenance of biological and biochemical characteristics of human colorectal tumours during serial passage in immune-deprived mice, Br. J. Cancer 37 (1978) 199. [2] REEVES, B.R., HOUGHTON, J.A., Seria cytogenetic studies of human colonic tumour xenografts, Br. J. Cancer 37 (1978) 612. [3] PICKARD, R.G., COBB, L.M., STEEL, G.G., The growth kinetics of xenografts of human colorectal tumours in immune-deprived mice, Br. J. Cancer 31 (1975) 36.
526
TAYLOR
[4] HOUGHTON, P.J., HOUGHTON, J.A., TAYLOR, D.M., Effects of cytotoxic agents on TdR incorporation and growth delay in human colonic tumour xenografts, Br. J. Cancer 36(1977) 206. [5] HAMMERSLEY, P.A.G., TAYLOR, D.M., CRONSHAW, S., The mechanism of 67Ga uptake in animal and human tumours, Eur. J. Nucl. Med. (in press). [6 ] HAMMERSLEY, P.A.G., TAYLOR, D.M., The role of lysosomal enzyme activity in the localization of 67Ga citrate, Eur. J. Nucl. Med. 4 (1979) 261. [7] SWARTZENDRUBER, D.C., NELSON, B., HAYES, R.L., Gallium-67 localization in lysosome-like granules of leukaemic and non-leukaemic murine tissues, J. Natl. Cancer Inst. 46(1971) 941. [8 ] HAMMERSLEY, P.A.G., Private communication 1980. [9] YEH, S.D.J., HELSON, L., GRANDO, R., Tumour-localizing radionuclides in hetero transplanted human tumours in nude mice, Int. J. Nucl. Med. Biol. 6 (1979) 169. [10] HOUGHTON, P.J., HOUGHTON, J.A., Evaluation of single-agent therapy in human colorectal tumour xenografts, Br. J. Cancer 37 (1978) 833.
IAEA-SM-247/206
Invited Review Paper RADIATION DOSE TO THE PATIENT IN RADIONUCLIDE STUDIES H.D. ROEDLER Institut für Strahlenhygiene, Bundesgesundheitsamt, Neuherberg, Federal Republic of Germany
Abstract RADIATION DOSE TO THE PATIENT IN RADIONUCLIDE STUDIES. In medical radionuclide studies, the radiation risk has to be considered in addition to the general risk of administering a pharmaceutical. As radiation exposure is an essential factor in radiation risk estimation, some aspects of internal dose calculation, including radiation risk assessments, are treated. The formalism of current internal dose calculation is presented. The input data, especially the residence time and the absorbed dose per transformation, their origin and accuracy are discussed. Results of internal dose calculations for the ten most frequently used radionuclide studies are presented as somatically effective dose equivalents. The accuracy of internal dose calculation is treated in detail by considering the biokinetics of the radio pharmaceutical, the phantoms used for dose calculations, the absorbed dose per transformation, the administered activity, and the transfer of the dose, calculated for a phantom, to the patient. The internal dose calculated for a reference phantom may be assumed to be in accordance with the actual patient dose within a range described by a factor of about two to three. Finally, risk estimates for nuclear medicine procedures are quantified, being generally of sixth order. The radiation risk from the radioiodine test is comparably higher, but probably lower than calculated according to the UNSCEAR risk coefficients. However, further studies are needed to confirm these preliminary results and to improve the quantification of the radiation risk from the medical use of radionuclides.
1.
INTRODUCTION
Before every diagnostic procedure, the predictable benefit and risk should be assessed. The benefit may be seen in a diagnosis with consequences relating to the therapy and health outcome of a patient, or as progress in scientific research. The risk refers to events adverse to the patient’s health which may occur with any diagnostic procedure. In medical radionuclide studies, the radiation risk has to be considered in addition to the general risk of administering a pharmaceutical. An overestimation of the risk resulting in the denial of necessary diagnostic procedures is inappropriate, as is an underestimation resulting in unnecessary exposure of patients or volunteers. 527
528
ROEDLER
As radiation exposure is an essential factor in radiation risk estimation, a certain amount of knowledge in this field is desirable for the physician not only for decision making, but also for answering questions coming from patients who are often more preoccupied with the danger than the benefit of radiation. Thus, some aspects of internal dose calculation including radiation risk assessments will be presented in this review.
2.
FORMALISM
The basic equation for internal dose calculation may be expressed as follows: The mean dose D is equal to the total number A of radioactive transformations of a radionuclide multiplied by S, the dose absorbed per transformation: D = A 'S
(1)
The activity may be distributed non-uniformly throughout the body, but a uniform distribution is assumed within the source organs or tissues rh. Taking into account_the contribution o f the transformations Ah in all source organs rh to the dose D(rk ) in the target organ rk, Eq.( 1) becomes [ 1] D(rk) = 2 V S ( r k^ r h)
(2)
j In traditional or SI units, D(rk) is the mean dose (rad or Gy) in a target organ rk , Ah is the cumulated activity (jt/Ci • h or Bq ■s = 1) in a source organ rh and S(rk «-rh) is the mean absorbed dose in rk per unit cumulated activity in rh (rad/juCi'h or Gy/Bq-s = G y).1 Thus, in SI units, Ah is equal to the number o f transformations in rh and S(rk «- rh) the mean absorbed dose in rk per transformation in rh . ^ In principle, internal dose calculation is easily performed according to Eq.(2). Ah is determined by: the administered activity, the pharmacokinetical behaviour and the radioactive half-time of the radiopharmaceutical, whereas S depends on the radiation characteristics of the radionuclide, the absorption of various radiation types in different tissues and the morphologic-geometrical properties of the patient or his idealized image, the phantom. 1 100 rad = 1 Gy. 1 Ci = 3.70 X 1010 Bq.
529
IAEA-SM-247/206 2.1. Cumulated activity, A
For most radiopharmaceuticals, an appropriate mathematical description of their pharmacokinetical behaviour can be given by exponential functions. The function of activity versus time in an organ % is then approximated by a single exponential, or a sum, or the difference of several (generally not more than 3) exponentials:.
Ah(t) = A o S “hj - exp (-In 2/Thj -1) j
(3a)
where Ah(t) = activity of a radiopharmaceutical in a source organ or region r^ at timet after administration A0 = administered activity = value of the jth exponential component at time t = 0 Thj = effective half-time of the jth exponential component; 1/Т - 1/T(jj0j0gjc + 1/Tracji0active An equivalent representation of Eq. (3 a) has been used in earlier publications [2-6]:
Ah(t) = A0FhXQhj ' exp (-In 2/Tjjj •t)
(3b)
which, by the distribution factor Fh, gives a clearer picture of the distribution of the radiopharmaceutical in the source organ rh, with
“hj = Fh ' Qhj
and
Fh = 2 ahj for positive ahj j
(4)
The total number of transformations in a source organ r^ from administration o f the activity A0 is given by the cumulated activity in rh
530
ROEDLER
The graphical equivalent of Ah is the area under the curve representing activity versus time in rh. Using Eqs (3a) and (3b), the cumulated activity is Ah = 1.443 A0 2 a hjThj j
(6ä)
Ah = 1.443 A0FhS Q hjThj
(6b)
The second part of these equations may be interpreted as the residence time rh of the activity in rh, i.e. the average time that the administered activity spends in rh ; consequently Ah = A„ ' Th
(6c)
Equations (6a), (6b) or (6c) may be used for determining cumulated activities in absorbed dose calculations. 2.2. Absorbed dose per unit cumulated activity, S The dose S(rk ^ r h) absorbed in a target organ r^ from a unit cumulated activity (in SI units: from a transformation) in a source organ r^ depends on: the mean energy per particle or photon, the mean number of particles or photons per nuclear transformation, the fraction of energy emitted per transformation in r^, which is absorbed in rk (absorbed fraction), and the mass of r^. 2.3. Effective dose equivalent By introducing the concept of the effective dose equivalent [7,], non-uniform and uniform dose distributions in the body have been made comparable based on the different radiosensitivity of various tissues and organs. By this concept, which was developed for radiation protection of occupationally exposed persons, weighting factors are being attributed to organs or tissues. The effective dose equivalent is the sum of the weighted organ doses. Because the age distribution of nuclear medicine patients is essentially different from that of occupationally exposed persons, somatic radiation effects are of primary concern in nuclear medicine. Thus, a somatically effective dose equivalent HSE may be defined which takes into account only somatic, but not genetic, effects.
IAEA-SM-247/206
3.
531
INPUT D A T A F O R D O S E C A L C U L A T I O N S
3.1. Cumulated activity, A Most retention and distribution measurements for radiopharmaceuticals have been performed on small animals. An extrapolation of the residence time o f the total body activity from animal to man via a power function of the body weight has been proposed. It was experimentally confirmed for 6sZn-chloride, 54Mnchloride, noAgm-nitrate and 75Se-selenite in mouse, rat, dog or monkey and man [8—11 ], but could not be confirmed by the same authors for 192Ir-hexachloroiridate, 9sNb-oxalate or 106Ru-chloride [12-14]. The residence times of the activity in source organs may be transferred directly from animal to man, or corrected by a factor that takes into account the different proportions of organ and total body masses in animal and man. For extrapolation from rat to man, this correction factor ranges from 0.5 to 2 except for the testicles, the fractional mass proportion of which is only 5% in man compared with the rat [2]. Among the great number of pharmaceuticals there are few for which wellestablished values of residence times are available both in animals and man. Two o f these are 75Se-selenomethionine and 67Ga-citrate. For these radiopharma ceuticals, the following conclusions may be drawn [15]: An extrapolation o f the residence time via body weight from animal to man is not possible. In some instances, the residence times in total body and organs of animals and man agree within 50%. In others, they differ by a factor of two to seven, in one case (75Se, pancreas) by a factor of 50. More than two thirds of the residence times differ by a factor of less than three in animal and man. - These conclusions characterize the quality of the pharmacokinetic data at present available. This should be kept in mind when referring to retention and distribution data from literature or animal experiments. The problem of obtaining reliable pharmacokinetic data and extrapolating them from animal to man will be treated in depth elsewhere [16]. A survey of retention and distribution data for radiopharmaceuticals was already compiled in 1974 [17]; for an updated summary of these data for clinically used radiopharmaceuticals, see Refs [4, 5 ]. 3.2. Absorbed dose per unit cumulated activity, S Values o f •<-rj1), the absorbed dose in a target organ rk per unit cumulated activity (or per transformation, respectively) in a source organ rh,
532
ROEDLER
TABLE I. WEIGHTING FACTORS FOR CALCULATING THE SOMATICALLY EFFECTIVE DOSE EQUIVALENT, BASED ON THOSE GIVEN IN Refs [7, 31 ]
W eighting factors O rgan o r tissue
R ef. [31]
Ref. [7] . Average
Male
Fem ale
Average
Male
Fem ale
Breast
0.20
0
0.33
0.23
0
0.45
R ed bone m arrow
0.16
0.20
0.13
0.21
0.31
0.11
Lungs
0.16
0.20
0.13
0.15
0.20
0.11
T hyroid
0.04
0.05
0.03
0.10
0.09
0.10
Bone (surfaces)
0.04
0.05
0.03
R em ainder (up to 5)
0.40
0.50
0.35
0.31
0.40
0.23
1.00
1.00
1.00
1.00
1.00
1.00
have been calculated for a number of radionuclides and a 70-kg anthropomorphic phantom; the results are available in matrix form for pairs of source and target organs [18, 19]. The coefficient of variation of these values depends on the distance between source and target organs, the size of the target organ and the photon energy, and ranges from 1 to 30%. . An adaptation of S to age-dependent body and organ sizes has been attempted by three different approaches: similarity transformation of the 70-kg phantom to the dimensions o f a newborn and of a child at the age of 1, 5, 10 or 15 years, and recalculation of the absorbed fractions; data for selected organ pairs and energies above 200 keV are available [20—23]; construction of pediatric phantoms for the above-mentioned ages, starting directly from anatomic data of infants and children; for three ages, absorbed fractions are available for selected organ pairs and the photon spectrum of "T cm [24-27]; modification of the absorbed fractions for adults, based on the dépendance of S from target mass, source-target distance and photon energy. A tabulation of age-dependent S values for a number of radionuclides was published recently [28], based partly on the correcting method described in Refs [29, 30].
IAE A-SM-247/206
533
For two distribution patterns of "T cm, representing a uniform (such as " T cm-hSA) and a non-uniform (such as "T cm-colloid) distribution, the age dépendance was estimated [6, 15] using the second method described above. For children and infants at the age of 15, 5, and 1 years, S, compared with adult values, will be larger by factors of approximately 1.25, 3 and 5 respectively; for the nonuniform distribution and the gonads as target organs, these factors are 2, 5 and more than 10. Because of the smaller activity administered generally to infants and children, the absorbed dose will vary to a lesser degree with age than S. 3.3. Effective dose equivalent Modifying the weighting factors recommended in the ICRP publication 26 [7] by neglecting the genetic factors, there result relative somatic effective weighting factors as presented in Table I. This table contains values for males and females, as well as average values, based on those given by the ICRP [7] and by other authors [31 ]. The values in the first column will be used in this review.
4.
RESULTS OF INTERNAL DOSE CALCULATIONS
During the last decade, a number of internal dose calculations were published in the form of dose estimate reports [32—39] for individual radiopharmaceuticals or as comprehensive tabulations [4, 5, 17, 40], in most cases including the biokinetic data used, which is a prerequisite forjudging the results and modifying them to suit individual circumstances, if necessary. The tabulations of absorbed doses in conventional and SI units published as an annex to the proceedings of the Symposium on Medical Radionuclide Imaging in Los Angeles in 1976 [4], and [5, 40] should be consulted, because no additional substantial information concerning the biokinetics of most radiopharmaceuticals has been made available during the past few years. Somatically effective dose equivalents were calculated according to the method described above using weighting factors from Ref. [7], modified so that only somatic effects are taken into consideration. Table II contains results for the most frequently performed studies and radiopharmaceuticals according to a survey conducted in Berlin and Munich for the year 1978. The biokinetic data, mainly according to Refs [4, 5 ], are listed as residence times. The bladder activity has been taken into account by assuming that for radiopharmaceuticals, mainly excreted by the kidneys, half of the remaining body activity decays within the bladder. The somatically effective dose equivalents are listed per /uCi and per test whereby the activity A0 is administered. The dose to the breast and to the remainder was calculated for ‘other tissues’ (48 kg) according to Ref. [19]. The
534
ROEDLER TABLE II. MEAN SOMATICALLY EFFECTIVE DOSE EQUIVALENTS HgE (mrem = 10“5Sv) PER juCi AND HSE (mrem = 10"5Sv) PER ADMINISTRATION FOR THE MOST FREQUENTLY USED [40] NUCLEAR MEDICINE IN VIVO STUDIES The residence times r(h) are based on Refs [4, J],- those fo r radiopharmaceuticals mainly excreted by the kidneys are modified fo r r o f bladder contents assuming half o f the remaining body activity decaying in the bladder: RB = Remaining body, RBM = Red bone marrow
Radio pharmaceutical
Source organ
7 (h)
Co-57-Vit. B-12
Total body
634
Tc-99mPertechnetate
Thyroid Stomach RB
0.19 0.35 7.17
0.037
1 000
37
Tc-99mPertechnetate (+block. agent)
Thyroid, RB
0.019 7.69
0.013
10 000
130
Tc-99mDiphosphonate
Bone Bladder RB
4.33 1.09 1.09
0.026
10 000
260
Tc-99mColloid
Liver Spleen RBM RB
7.36 0.606 0.433 0.26
0.056
3 000
170
Tc-99mMicro spheres
Lung Liver RB
4.02 0.564 2.26
0.055
2 000
110
Tc-99mDTPA
Kidneys Bladder RB
0.433 1.40 1.40
0.029
5 000
150
1-131Iodide
Thyroid(eu) RB Thyroid(hyper) RB Thyroid(hypo) RB
84
50
4200
150
50
7400
34
50
1700
1-131Hippuran
Kidneys Bladder RB
97.2 16.8 166 25 38.1 20.1 0.157 0.372 0.372
HSE (mrem//iCi)
1.3
0.060
1% free iodide
84
1-131-Hippuran 0.3% free iodide
0.60 84
Au-198Colloid
Liver Spleen RBM
83.8 2.79 6.52
4.8
Ao (MCi)
HSE (mrem)
0.5
30
0.65
1.8
0.3
+25 =27
500 1.5
30 + 126 = 160
150
720
,
IAEA-SM-247/206
535
results may be regarded as preliminary, because the inclusion o f additional organs with higher organ doses in the ‘remainder’, instead of taking an average dose to the remainder, will influence the results. Moreover, the biokinetic data used will have to be revised as soon as more accurate data is available. The age dependence of the effective dose equivalent was recently investigated [41]: in imaging procedures, if the administered activity is adapted to the body mass m (kg) o f the child according to (m /70 kg)2A3, the effective dose equiva lent for the child at the age of 15, 10, 5, and the newborn will be higher by factors of about 1.2, 1.6, 1.7, 2.7 respectively, compared with the adult dose; in bolus and uptake studies administering equal activities to adults and children, these factors are 1.3 ± 0.2, 2.9 ± 1.2, 4.0 ± 1.6, 20.2 ±6.1; in dilution procedures with age-dependent corrections of the administered activity according to (m /70 kg), the doses in adults and children are nearly the same.
5.
ACCURACY OF INTERNAL DOSE CALCULATIONS The approximations and errors in internal dose calculation refer essentially to the biokinetics of the pharmaceutical, the phantom used, the dose per transformation, calculated for the phantom, the administered activity and the transfer o f the internal dose value, calculated for a special model, to an individual patient.
In principle, the steps to be performed for dose calculations, error consider ations and some quantifications of the errors involved, may be described as follows [15].
5.1. Biokinetics of the radiopharmaceutical (a) Collect data from measurements in humans, which may vary largely between different individuals. The standard deviation of the total body residence times may be as low as 15% for 75Se-selenomethionine and 67Ga-citrate (75Se: 24 patients measured up to 923 days [42]; 67Ga: 112 patients measured up to 112 days [43]). On the other hand, lowest and highest activity concentrations in human tissue
536
ROEDLER
specimens from autopsies may differ by a factor of 20 for the same organ (67Ga: 23 autopsies from 3 h to 227 days after administration [44]). (b) If data from humans are not available, collect data from animal experiments. Expect large intra- and interspecies variations. These will be higher again for organ than for total-body residence times. (c) Define reference values from available data. If animal data are used, be aware of errors transferring these data to humans. For 75Se-selenomethionine and 67Ga-citrate, more than two thirds o f reported mean residence time values differ from the corresponding human data by a factor of less than three. When using patient data, neglect differences in biokinetics caused by different clinical states. (d) Define reference values characteristic for selected clinical situations, e.g. for colloids for different states of parenchymal liver disease. This will result in a better approximation to individual patients. 5.2. Phantom used for dose calculations (a) Collect data concerning the geometry and the elemental composition of the human body and its organs. (b) From available information define data for a reference patient neglecting individual variations. In Ref. [45], a reference man of 70 kg and 174.5 cm has been defined, based on a large amount of anatomical and physiological data collected from literature. This reference man was created for the purpose of radiation protection mainly for occupational exposure, and thus resembles more a healthy young male than a nuclear medicine male or female patient generally more advanced in age, and o f smaller size and lower weight. (c) Develop a mathematical phantom of the reference patient, which will for practical reasons include considerable simplifications in geometry and composition. Such a phantom is currently available for reference man only (see above), but not for a not-as-yet-defined reference patient. (d) Take into account the size or age of patients, respectively. Define corre sponding phantoms, as has been done for the newborn, infant and child at the age of 0, 1 ,5, 10, and 15 years, in the absence o f other methods suitable for age and size correction. 5.3. Absorbed dose per transformation (a) Perform Monte Carlo calculations for photons and the mathematical phantoms defined above. Estimate the dose in a target organ per transformation in a source organ. The coefficient o f statistical variation will range from 1 to 50%. If it is higher, other methods (build up factor method, reciprocity theorem) should be used.
lAEA-SM-247/206
537
(b) Correct the absorbed dose per transformation for different body and organ sizes. For a phantom of 163 cm, corresponding in size to an average female (and a child phantom at the age of 14), S is higher by approximately 35% than it is for reference man. For the ovaries and a high activity concentration in the liver it is even higher by a factor of more than two; the corresponding factors for children and infants at the age of 15, 5 and 1 years with values of 1.25, 2, and 5 or 2, 5, and more than 10, respectively, have been mentioned above (see Subsection 3.2). (c) It would be desirable to confirm calculated values by those measured in phantoms or patients. For phantoms, several investigations have been performed using higher photon energies in most cases; calculations and measurements were compatible on the average within ±20% [29, 46, 47]. For a special radiopharma ceutical (99Tcm-S-colloid) and a special geometry (source organ: liver; target: LiF plate on patient’s skin), calculation and measurement agreed within ±30% for average-sized patients [48].
5.4. Administered activity (a) Collect values of prescribed activities. There will be considerable differences between various hospitals where a factor of two is not unusual. (b) Measure the actually administered activity. It may vary with a standard deviation o f about ± 10% [ 15 ]. (c) Define typical prescribed activities as reference values. For children, the following corrections have been proposed: Imaging studies Bolus and uptake studies Dilution studies
Achild = Aaduit • (mass of child/70 kg)2/3 Achad = Aadult
Achfld = Aadult •
(mass of child/70 kg)
5.5. Dose calculation (a) Perform internal dose calculations for specified data and model resulting in a dose value for this model. Using different apparently reliable sets of biokinetic data, the dose to the mathematical phantom of reference man will vary around a probable value in a range described by a factor of two, though larger variations may occur [6, 15].
538
ROEDLER
(b) Apply the result to the individual patient, neglecting the simplifications of the model. If the patient closely resembles the reference man, the dose values calculated for the reference phantom may be in accordance with the patient dose within ±30% (see Subsection 5.3). On the other hand, the doses depend on the size of the phantom, enlarging the doses to an average size female by 30%, in extreme cases (ovaries, activity mainly in the liver) up to 100% (see Subsection 5.3). In conclusion, it may be assumed that whole-body and organ doses calculated for a reference phantom, using reference input data for biokinetics and activity, will be in accordance with the actual patient doses within a range described by a factor of about two to three, the lower value referring to radiopharmaceuticals labelled with short-lived radionuclides such as " T cm, for which different biologic half-times will not influence the resulting internal radiation dose in the same proportion.
6.
RADIATION RISK
The knowledge of the internal dose and of the somatically effective dose equivalent is one of the prerequisites for estimating the possible radiation risk from radionuclide studies. Additional knowledge is needed on the frequency of radiopharmaceutical applications, the age and sex distribution of patients, and the probability of radiation effects per million per unit dose or effective dose equivalent. Using data for Berlin 1978 and risk coefficients from the ICRP [7] or from UNSCEAR [49], an attempt was made to estimate the radiation risk for nuclear medicine patients. This subject is treated in detail elsewhere [50]. The conclusions may be summarized as follows: The calculated mean somatic risk (induction of fatal cancer or leukaemia) from administration o f usual radiopharmaceuticals is of the sixth order; only for 131I-iodide it may be two or three orders of magniture higher. If this were true, the incidence of thyroid cancers should be significantly higher in patients having received a radioiodine test than in control patients. A very detailed study for Sweden [51] based on 10 000 radioiodine patients with a mean follow-up period of 17 years yielded the following results: expected number of spontaneous thyroid carcinomas: 8.3 expected number of radiation-induced thyroid carcinomas: 47—127 total number of thyroid carcinomas detected: 9
IAE A-SM-247/206
539
Thus, it might be justifiable - probably because of the dose-rate dependence of the effect —to use a risk factor for radioiodine-induced thyroid carcinomas, which is smaller by one or two orders than the value recommended by UNSCEAR [49]. This is in accordance with the results of other authors [52] showing that the effect of 131I-induced thyroid carcinoma for the same absorbed doses is only 1/70 compared with external irradiation. Further studies are needed to confirm these preliminary results and to improve the quantification of the radiation risk from the medical use of radionuclides.
REFERENCES [1] LO EV IN G ER, R., BERM AN, M., A Revised Schem a for Calculating the A bsorbed Dose from Biologically D istributed R adionuclides, MIRD pam phlet N o .l, revised, Society of N uclear M edicine, New Y ork (1 976). [2] R O ED LER , H.D ., Strahlenbelastung du rch R adiopharm aka — E ntw icklung eines m ath e m atischen D osiskonzepts und Ergebnisse von N euberechnungen der E nergiedosis, Thesis, Freie U niversität Berlin (1974). [3] RO ED LER, H.D., K A U L, A., “ D ose to target organs from rem aining body activity: R esults o f the form ally exact and approxim ate so lu tio n ” , Proc. Sym p. on R adiopharm a ceutical D osim etry, Oak Ridge (C L O U T IER , R .J., et al., Eds), HEW Publ. (FD A ) 76-8044 (1 9 7 6 ) 155. [4] KAUL, A., R O E D L E R , H.D., HIÑ E, G .J., “ In tern al absorbed dose from adm inistered radiopharm aceuticals” , M edical R adionuclide Im aging (Proc. Sym p. Los Angeles, 1976) 2, IAEA, V ienna (1 9 7 7 ) 423. [5] KA UL, A., R O E D L E R , H.D ., „D o sim etrie” , N uklearm edizin — F unktionsdiagnostik und T herapie (EM RICH , D., E d.), 2nd E dn, Georg T hiem e, S tu ttg a rt (1979). [6] R O E D L E R , H .D ., «B iokinetik u n d D osiserm ittlung” , Proc. Sym p. M edizinische P hysik '19, A lfred Hüthig, H eidelberg (1 9 7 9 ) 271. [7 ]
[8]
[9]
[10]
[11]
[12]
I N T E R N A T I O N A L C O M M IS S IO N O N R A D I O L O G I C A L P R O T E C T I O N , R e c o m m e n
datio n s o f th e In tern a tio n a l Com m ission o n Radiological P ro te ctio n , ICRP Publication No. 26, Pergam on Press, O xford (1 977). RICHM OND, C .R ., FU R C H N E R , J.E ., T R A FT O N , G.A ., LANGHAM , W.H., C om parative m etabolism o f radionuclides in m am m als: I. U ptake and re te n tio n o f orally adm inistered 65Zn by fo u r m am m alian species, H ealth Phys. 8 (19 6 2 ) 481. FU R C H N E R , J.E ., RICHM OND, C .R ., D RA KE, G.A., C om parative m etabolism of radionuclides in m am m als: III. R e te n tio n o f m anganese-54 in th e m ouse, ra t, m onkey and dog, H ealth Phys. 1 2 (1 9 6 6 ) 1415. FU R C H N E R , J.E ., RICHM OND, C .R ., D R A K E , G .R ., C om parative m etabolism of radionuclides in m am m als, IV. R e te n tio n o f silver-110m in the m ouse, ra t, m onkey and dog, H ealth Phys. 15 (1 9 6 8 ) 505. FU R C H N E R , J.E ., LONDON, J.E ., W ILSON, J.S., C om parative m etabolism o f ra d io nuclides in m am m als: IX. R e te n tio n o f Se-75 in the m ouse, ra t, m o n k ey and dog, H ealth Phys. 2 9 (1 9 7 5 ) 641. FU R C H N E R , J.E ., RICHM OND, C .R ., D RA K E, G .A ., C om parative m etabolism of radionuclides in m am m als: V. R e te n tio n o f Ir-192 in th e m ouse, ra t, m onkey and dog, H ealth Phys. 2 0 (1 9 7 1 ) 375.
540
ROEDLER
[13] FU RCH N ER, J.E ., D RA K E, G .A ., C om parative m etabolism of radionuclides in m am m als: VI. R e te n tio n of Nb-95 in the m ouse, rat, m onkey and dog, H ealth Phys. 21 (1 9 7 1 ) 173. [14] FU RCH N ER, J.E ., RICHM OND, C.R., D R A K E , G .A ., C om parative m etabolism of radionuclides in m am m als: VII. R e te n tio n of R u-106 in the m ouse, ra t, m onkey and dog, H ealth Phys. 21 (1 9 7 1 ) 355. [15] R O ED LER , H.D ., S trahlenexposition des P atien ten durch R adiopharm aka — G renzen der G enauigkeit von D osisberechnungen, Thesis, Freie U niversität Berlin (1977). [16] R O ED LER , H.D., “ A ccuracy o f radiopharm aceutical internal dose calculations” , In tern a tio n a l R adiopharm aceutical D osim etry Sym posium , O ak R idge 1980 (in prep aratio n ). [17] KAUL, A., O E FF , K., R O E D L E R , H.D., VOGELSANG, T., R adiopharm aceuticals B iokinetic D ata and R esults of R ecalculations o f In tern a l D ose, Info rm atio n sd ien st für N uklearm edizin, Berlin (1973). [18] SN YDER, W.S., FO R D , M .R., W ARNER, G.G., W ATSON, S.B., A T ab u latio n o f Dose E quivalent per M icrocurie-D ay fo r Source- and T arget Organs o f an A dult fo r V arious R adionuclides, Part 1 (1 9 7 4 ), P art 2 (1 975), O ak Ridge N ational L ab o rato ry O R N L-5000. [19] SN YDER, W.S., FO R D , M .R., W ARNER, G.G., W ATSON, S.B., “ S” , A bsorbed Dose per U nit C um ulated A ctivity fo r Selected R adionuclides and Organs, MIRD pam phlet N o .l 1, Society o f N uclear M edicine, New Y o rk (1975). [20] H IL Y E R , M.J.C., SN YDER, W.S., W ARNER, G.G., “ E stim ates of dose to infants and children from a p h o to n e m itte r in th e lungs” , H ealth Phys. Div. A nnual Progr. R ep. O R N L -4811 (1 9 7 2 ) 91. [21] H IL Y E R , M .J.C., HILL, G.S., W ARNER, G.G ., “ Dose from p h o to n em itters distrib u ted uniform ly in the to ta l body as a fu n c tio n o f age” , H ealth Phys. Div. A nnual Progr. Rep. O R N L-4903 (1973) 119. [22] SN YDER, W.S., FO R D , M .R., “ E stim ates of dose ra te to gonads o f in fan ts and children from a p h o to n e m itte r in various organs o f the b o d y ” , H ealth Phys. Div. A nnual Progr. Rep. O R N L-4903 (1973) 125. [23] SN YDER, W.S., “ D osim etry o f internal em itters for p o p u latio n e xposure” , Proc. Sym p. Population E xposures, H ealth Physics Society (1 9 7 4 ) 235. [24] HWANG, J.M .L., SHOUP, R.L., POSTON, J.W., M athem atical D escription o f a N ew born H um an fo r Use in D osim etry Calculations, Oak Ridge N ational L ab o ra to ry O RN L/TM 5453 (1976). [25] HWANG, J.M .L., SHOUP, R .L., W ARNER, G.G ., PO STON, J.W ., M athem atical D escription of a One- and Five-Y ear Old Child for Use in D osim etry Calculations, Oak Ridge N ational L ab o rato ry O R N L/TM -5293 (1976). [26] DEUS, F.S., PO STON, J.W., “ T he developm ent o f a m ath em atical p h an to m representing a ten-year-old for use in in tern al dosim etry calculation” , Proc. R adiopharm aceutical D osim etry Sym posium , O ak R idge, 1976, HEW -Publ. (FD A ) 76-8044 (1 9 7 6 ) 110. [27] JO N ES, R.M., PO STON, J.W., HW ANG, J.L ., JO N ES, T.D ., W ARNER, G.G., The D evelopm ent and Use o f a Fifteen-Y ear-O ld E quivalent M athem atical P hantom for Internal Dose Calculations, O ak Ridge N ational L ab o ra to ry O R N L/TM -5258 (1976). [28] HENRICH S, K., KAUL, A., K RA U SE, M., A ge-D ependent Values o f the Specific A bsorbed Dose, In tern a l R e p o rt, M edizinische Physik, K linikum Steglitz, 1000 Berlin 45 (1980). [29] YAMAGUCHI, H., KATO, Y., SH IRA G A I, A., T he tran sfo rm a tio n m ethod for the MIRD absorbed fractions as applied to various physiques, Phys. Med. Biol. 20 (19 7 5 ) 593. [30] YAM AGUCHI, H., E stim ation o f internal rad iatio n dose for various physiques using MIRD a d u lt absorbed fractions, A cta R adiol., O ncol., R adiat. Phys., Biol. 17 (1 978).
IAEA-SM-247/206
541
LAW, P.S., RO SEN STEIN , M., A som atic dose index for diagnostic radiology, H ealth Phys. (1978). M IRD dose estim ate re p o rt N o .l, Sum m ary o f cu rren t radiation dose estim ates to hum ans from 75Se-L -selenom ethionine, J. Nucl. Med. 14 (19 7 3 ) 49. MIRD dose estim ate re p o rt N o .2, S um m ary o f cu rren t ra diation dose estim ates to hum ans from 66Ga-, 67Ga-, 68Ga-, and ^ G a -c itra te , J. Nucl. Med. 14 (1 9 7 3 ) 755. M IRD dose estim ate re p o rt N o.3, Sum m ary o f cu rren t ra diation dose estim ates to hum ans w ith various liver c o nditions from 99T cm -sulfur colloid, J. N ucl. M ed. 16 (1 9 7 5 ) 108A. M IRD dose estim ate re p o rt N o.4, Sum m ary o f cu rren t ra diation dose estim ates to hum ans w ith various liver c o nditions from 198A u-colloidal gold, J. Nucl. Med. 16 (19 7 5 ) 173. MIRD dose estim ate re p o rt N o.5, Sum m ary o f cu rren t ra diation dose estim ates to hum ans from 123I, 124I, 12SI, 126I, I30I, 131I, and 132I as sodium iodide, J. Nucl. M ed. 16 (19 7 5 ) 857. MIRD dose estim ate re p o rt N o.6, Sum m ary o f cu rren t ra diation dose estim ates to hum ans from 197Hg- and 203Hg-labeled chlorm erodrin, J. Nucl. Med. 16 (1 9 7 5 ) 1095. MIRD dose estim ate re p o rt N o .7, Sum m ary o f cu rren t ra d ia tio n dose estim ates to hum ans from 123I, 124I, 126I, 130I, and 131I as sodium rose bengal, J. Nucl. Med. 16 (1 9 7 5 ) 1214. MIRD dose estim ate re p o rt N o.8, Sum m ary o f cu rren t ra diation dose estim ates to norm al hum ans from 99T cm as sodium p e rte ch n e tate, J. Nucl. Med. 17 (1 9 7 6 ) 74. R O ED LER , H.D., KA UL, A., H IÑ E, G .J., In tern al R adiation Dose in D iagnostic N uclear M edicine, Hildegard H offm ann, Berlin (1 978). HENRICHS, K., KA UL, A., S trah len ex p o sitio n von K indern u n d Jugendlichen in der nuklearm edizinischen D iagnostik, Nucl.-M ed. (in press). LA TH RO P, K.A., JO H N STO N , R .E ., BLAU, M., RO TH SH ILD , E .O ., R ad iatio n dose to hum ans from 75Se-L -selenom ethionine, M RD p am phlet N o.9, J. Nucl. Med. 13 Suppl. 6 (1972). W ATSON, E .E ., C L O U T IE R , R .J., GIBBS, W.D., W hole-body re te n tio n o f 67G a-citrate, J. Nucl. Med. 1 4 (1 9 7 3 ) 840. N ELSON , B., HA YES, R .L., EDW ARDS, C.L., KN ISELEY , R.M ., ANDREW S, G.A., D istribution o f gallium in hum an tissues a fte r intravenous adm inistration, J. N ucl. Med. 13 (1 9 7 2 ) 92. IN T E R N A T IO N A L COMMISSION ON RA D IO LO G ICA L PR O TE C T IO N , R e p o rt o f the T ask G roup on R eference Man, ICRP P ub licatio n N o.23, Pergam on Press, O xford (1 975). G A R R Y , S.M., STANSBURY, P.S., PO STO N , J.W ., M easurem ent of absorbed fractions for p h o to n sources distrib u ted unifo rm ly in various organs o f a heterogeneous ph an to m , H ealth Phys. 28 (1 9 7 5 ) 591. F E H E R , I., K O BLIN G ER, L ., SZABO, P.P., In tern al dose calculations and m easurem ents, H ealth Phys. 2 9 ( 1 9 7 5 ) 107. JO N ES, J.P., W AGNER, J., B R IL L , A.B., “ A n in-vivo evaluation of standard m an m odel absorbed fractio n s using 99T cm -sulfur colloid” , Proc. R adiopharm aceutical D osim etry Sym posium , O ak Ridge (1 9 7 6 ), HEW -Publ. (FD A ) 76-8044 (1 9 7 6 ) 283. U N ITED NA TIO NS SCIEN TIFIC COM M ITTEE ON TH E E FFE C T S O F ATOMIC R A D IA TIO N , Sources and E ffects o f Ionizing R adiation, UN SCEAR 1977 R e p o rt to the G eneral A ssem bly, w ith annexes, U nited N ations, New Y ork (1977). R O E D L E R , H.D., KA UL, A., „Strahlenrisiko für den P a tie n ten d urch nuklearm edizinische D iagnostik: Das Strahlenrisiko im V ergleich zu chem ischen R isiken” , Proc. Sym p. H om burg, 1980 (G LÖ BEL, B., E d.), G eorg T hiem e, S tu ttg a rt (in press). HOLM , L .E ., Incidence o f M alignant T h y ro id T um ors in Man a fte r D iagnostic and T herapeutic Doses of Iodine-131: E pidem iologic and H istopathologic S tu d y , Thesis,
542
ROEDLER D epartm ent of G eneral O ncology, R adium hem m et, K arolinska H ospital, S -10401
Stockholm , Sweden. [52] MAXON, H .R ., THOM AS, S.R., SA EN G ER, E.L., BUNCHER, C .R ., K ER EIA K ES, J.G ., Ionizing irradiation and the in d u ctio n of clinically significant disease in the hum an thy ro id gland, A m .J.M ed. 63 (19 7 7 ) 967.
DISCUSSION N.G. TROTT: You have given an excellent, comprehensive and up-to-date account of the problems of radiopharmaceutical dosimetry. However, I was not quite sure from your discussion of residence times whether you have also considered relative values of uptake. I would also like to mention that the International Commission on Radiation Units and Measurements (ICRU) has published a new report (No.32) giving a comprehensive survey of radionuclide dosimetry procedures, including a discussion of the problems of obtaining data directly from measurements on human subjects, and examples of methods of calculation of absorbed dose from biological and physical data. H.D. ROEDLER: The relative values of uptake are included in the residence times, t , and have thus been taken into account. I. SCHMITZ-FEUERHAKE: The study by Holm to which you referred (Ref. [51 ] in your paper) does not provide adequate grounds for stating that the radiation risk to the thyroid is overestimated by UNSCEAR. The author used a mean follow-up period of only 17 years, whereas the mean latency period for thyroid carcinomas is 2 5 -3 0 years. One fundamental problem which still remains is that there is not much point in determining the dose to the patient to within a factor of two or three if the uncertainties in risk estimation are ignored. Risk evaluations by various authors differ by orders of magnitude. H.D. ROEDLER: The study by Holm is an excellent investigation, charac terized by a large number of patients ( 10 000) and a relatively long observation period. The author states that, though the UNSCEAR risk coefficient is given for a follow-up time of 25 years and their average observation time was 17 years, it does not seem probable that the number of detected, possibly radiation-induced, thyroid carcinomas will rise by a factor of 100 during the next eight years. As I have stated in the paper, studies will have to be continued and should be performed in more countries. Even if, in a chain of calculations, one factor is known only by the order of magnitude, this is no argument in favour of not investigating the other factors in this calculation with reasonable accuracy. Furthermore, dose calculations are needed when comparing different methods resulting in equivalent diagnostic information, or for legal aspects, for example, documenting absorbed doses in patients’ records or in the use of radionuclides in healthy volunteers or in patients for research purposes.
IAE A-SM-247/7 7
Poster Presentation USE OF CELL-TYPE SPECIFIC ANTIBODIES FOR RADIOIMMUNODETECTION OF BREAST METASTASES WITH A HIGH-PURITY GERMANIUM CAMERA* J.A. PETERSON, T. WILBANKS Bruce Lyon Memorial Research Laboratory, Children’s Hospital Medical Center, Oakland, California S. MILLER Department o f Radiology, University of San Francisco, San Francisco, California L. KAUFMAN, D. ORDENDAHL Radiologic Imaging Laboratory, South San Francisco, California R.L. CERIANI Bruce Lyon Memorial Research Laboratory, Children’s Hospital Medical Center, Oakland, California, United States of America
A new concept of radionuclide imaging of breast tumours has been developed by using radioiodine-labelled antibodies against cell-type specific antigens of mammary epithelial cells. These antisera, that bind specifically to the surface o f mammary epithelial (ME) cells, were prepared in rabbits using milk fat globule membranes (MFGM) as immunogen. An anti-mouse mammary epithelial (anti-MME) antiserum was prepared against mouse MFGM that after absorptions with other mouse tissues [ 1] bound specifically to mouse mammary epithelial cells, but not to mammary fibroblast, nor to cells of other mouse tissues [ 1]. The anti-MME also bound to mouse mammary epithelial tumours; however, to a lesser degree than to normal mammary cells, and did not bind to tumour cells from other mouse origins [1 ]. * W ork sup p o rted b y NCI G rants Nos C A 19455 and C A 20286 and BRS G rant No. R R 054S 7 o f th e N ational In stitu te s o f H ealth.
543
544
POSTER PRESENTATION
To reduce non-specific-binding in the radioimmunoimaging, Fab fragments were prepared from the anti-MME immunoglobulins and these were then iodinated with 131I. Seventy-five pCi of this radioiodinated preparation were injected into mice carrying simulated mammary tumour métastasés.1 Twenty-three hours later they were administered 500 pCi of " T cm-pertechnetate and then an hour later the mice were imaged on a high-purity germanium camera. The transplanted mammary tumours were clearly localized in the images while non-mammary tumours were not. The specificity of the 131I-image could be increased by normalizing it to that of the " T cm-pertechnetate which compensated for the difference in intravascular and extracellular spaces in the individual mice. There was a four- to sixfold greater accumulation of 131I in the mammary tumours compared with a similar area on the opposite side of the mouse. The non-mammary tumours showed no specific accumulation of 131I-labelled antibody. In a comparable human system with antisera (anti-HME) against human mammary epithelial cell surface components [2], we have localized human breast tumours transplanted in athymic nude mice.
REFERENCES [1] [2]
C E R IA N I, R.L., PETERSO N , J.A ., ABRAHAM , S., J. N atl. Cancer Inst. 61 (1 9 7 8 ) 747. CERIA N I, R.L., THOM PSON, K.E., PETERSO N , J.A ., ABRAHAM , S., Proc. Natl. Acad. Sei. USA 7 4 (1 9 7 7 ) 583.
1 1 Ci = 3.70 X 1010 Bq.
IAEA-SM-247/38
A NEW GERMANIUM-68/GALLIUM-68 RADIOISOTOPE GENERATOR SYSTEM FOR PRODUCTION OF GALLIUM-68 IN DILUTE HC1 J. SCHUHMACHER, W. MAIER-BORST Institute o f Nuclear Medicine, German Cancer Research Center, Heidelberg, Federal Republic of Germany
Abstract A NEW GERM ANIUM - 6 8 /G A L L IU M - 6 8 RAD IO ISO TO PE G E N E R A T O R SYSTEM FO R PRO DUCTION O F GALLIUM - 6 8 IN D ILU TE HC1. A sy n th etic ion-exchange resin condensed from pyrogallol and fo rm aldehyde was tested for its su itability to serve as 68Ge su p p o rt in a 68 G e /6SGa generator. A g enerator prepared from 4 g o f this resin and 10 mCi 6 sGe, and op erated w ith 10 m l o f 4.5N HC1 as eluant, show ed a co nsistently high yield o f 6sGa over a long period o f tim e. A sm all colum n filled w ith 0.6 g o f a com m ercially available an io n exchanger (Bio Rad AG 1 X 8 ) and coupled in series w ith th e g enerator reduces th e volum e and HC1 m olarity o f th e gen erato r’s eluate to 4 ml and 0.5N HC1. The m ean recovery o f 68Ga during m ore th a n 250 days was 75%, w ith a c o n tam in atio n o f th e eluate by 68Ge o f less th an 0.5 ppm . T he eluate is free o f radiolytic p ro d u c ts and suitable for hum an use.
The availability of ionic 68Ga from a 68Ge/68Ga generator would greatly facilitate the use of this short-lived positron emitter (Ti. 68Ga, 68.3 min; Ti. 68Ge, 287d) for radiopharmaceutical preparations. For this reason several attempts have been made to replace the only commercially available 68Ga generator, which delivers 68Ga very tightly complexed by EDTA. Table I presents a survey of chromatographic procedures investigated for separation of 68Ga from 68Ge by non-complexing eluants. In our studies we investigated the separation characteristics for gallium and germanium on a specially synthesized pyrogallol/formaldehyde ion-exchange resin, because pyrogallol very selectively precipitates germanium even from 2N HC1 [9]. For preparing the resin a solution of pyrogallol and formaldehyde was condensed as a suspension in an inert, hydrophobic medium at elevated temperature. The resulting exchange resin had a binding capacity for germanium of about 40 mg/g and a reversibly bound water content of 45%. Particles of 50—100 and 100-200 U.S. mesh were taken for column experiments. 545
S C H U H M A C H E R and MAIER-BORST
546
TABLE I. CHROMATIC PROCEDURES INVESTIGATED FOR GERMANIUM-68/GALLIUM-68 GENERATORS Exchanger
E luant
68Ga yield
68Ge im purity
(%)
(%)
1. AI2 O 3
EDTA 0.005 m
70
2. A12 0
3 X 10~4
HC1 0.2 m
49
20 X
10“4
3. Sb 2 O s
Na-Oxalat 2%
80
500 X
10‘ 4
4. Z rO j, SÍO 2
HNO3
3
N ot given
5. Pyrogallol
HC1 0.3 m
60
. A nion exchanger AG 1 X 8 (Bio Rad)
H F 0.01 m
90
HC1 1.0 m
80
Na 3 P 0 4 0.1 m
60
6
7. SnO -2 8
. A12 0
3
Ref.
[1] [2 ] [3] [ 4 ,5 ]
10 X
1 0 -4
100 X IO '
[4 ,5 ]
4
[6 ]
2 X
10“ 4
П]
10 X
10~4
[8 ]
N.B. The first exchanger listed is com m ercially available.
Small test generators were built as follows: 10 to 100 ¡jlC í 68Ge in 2N HCl were shaken with 250 mg of resin for several hours.1 Distribution coefficients up to 12 000 were obtained and they were not affected by 300 mg of gallium or other impurities of the dissolved but otherwise unprocessed target. This 68Ge-laden resin was packed into a small column of 5 mm dia., which had already been filled with 750 mg of the resin as a holdback layer. The final generators had a resin bed of about 6 cm. Figure 1 shows the 68Ge leakage of small test generators during a continuous elution as affected by particle size of the exchanger and the HC1 molarity of the eluant. One fraction amounts to 0.5 1. The lowest germanium leakage could be achieved with the particle size 100—200 U.S. mesh and 4.5N HC1 as eluant. Figure 2 depicts the 68Ga recovery from small test generators during a 50-day period as affected by the HC1 molarity of the eluant. The decrease in 68Ga yield in the course of time is probably due to 68Ge migrating into the resin particles. This decrease is less pronounced at higher HC1 concentrations of the eluant. 1 1 Ci= 3.70 X 1010 Bq.
IAEA-SM-247/38 0.1 p a r t ic le
0.1
s iz e : 5 0 -1 0 0 m e sh
547
p a r t ic le
s i z e : 1 0 0 -2 0 0 m esh
H C t : 2.2 n o r m a l
HC 1 : 2.2 no r m a l
flo w r a t e : 0.26 m l/ m in . 6 8 G e in
f lo w r a t e : 0.26 m l/m m .
5.5I e lu a t e : 0.16 %
* 0D1.
6 8 G e in 5 .5 1 e lu a t e : 0 .0 9 %
aoi
îr iT iïîi
0.001
0.001 1
3
5
7
9
11
7
9
11
t r a c t io n
0.01
p a r t ic le
s iz e : 1 0 0 - 2 0 0 m e s h НС I : 4.5 n o r m a l
f lo w r a t e : 0.6 m l/ m in .
68G e
in 7.51 e lu a t e : 0 0 1 6 %
0.001
0.0001
1
3
5
7
9
11
ID 13
15
f r a c t io n
F IG .l.
Germ anium -68 leakage fro m small №G e /6SGa test generators.
For routine use in hospitals the generator should initially contain 10 mCi or more 68Ge and should be usable over a long period of time. Therefore the results from the small test generators regarding the yield of 68Ga and contamination o f the eluate with 68Ge had to be checked with a generator of enlarged size and for a longer time period. Figure 3 shows the elution characteristics of a generator prepared with 4 mCi 68Ge and 4 g of pyrogallol exchange resin with a particle size of 1 0 0 -2 0 0 mesh. This amount of resin was estimated to be suitable for increasing the amount of 68Ge up to 20 mCi. Over a 200-day period the amount, as well as the HCI molarity, was varied to achieve a 75—80% recovery of 68Ga. The best results, 83% 68Ga recovery, were achieved with 10 ml of 4.5N HCI as eluant, and 68Ge contamination remained at 2 ppm. After a resting period of 150 days this generator was again eluted 20 times within a month. The 68Ga recovery in 4.5N HCI was now 76±2% and the 68Ge content had increased to 4 ppm.
548
S C H U H M A C H E R and MAIER-BORST
a f t e r b u ild u p of g e n e r a t o r
FIG. 2.
Gallium-68 recovery from m G e /6SGa test generators.
100
%6®Ga p e r e lu t io n
70
a
50
a i с о IN О oi
_
о I с p
Ë O d
.
*
.,4..» ‘••’v* . . . t •* * ’ o I c m ÍÑ
6 o ci
Ô i c m E o Ö
a f t e r b u ild up o f g e n e r a t o r
FIG.3. Elution characteristics o f a 4 m G 68Ge generator, consisting o f 4 g pyrogallol exchange resin.
IAEA-SM-247/38
549
88/ G a -68 G E N E R A T O R (10mCi) conditions:age > 250d,about 300 elutions
Gepresent
10ml 45N
HCI eluant
10
- . c m -» •
30c m
pyrogatlol exchange resin ♦ mCi G e
10
-68
pyrogallol
exchange resin
total weight of exchange resin
40g
55
• . cm
¿.0ml
Ш Ш v
H,p eluant
ь О .б с г г и
strongly adsorbed radiolytic products from pyrogallol exchange resin
10 45 -68
waste : ml . N HCi containing % o f Ge impurity eluted from the pyrogallol exchange resin
75
3.0-cm
BioRad A G
1*8
anionexchangeri
0.5g )
.0ml 0.5N HCI ,<0.5*10 % G e -68 i 474 .6i 3.3% G o -68 recovery
37determinations
(
during
250d)
FIG.4. Schem atic diagram o f a 10 mCi m G e /m Ga generator system , consisting o f a pyrogallol exchange colum n and an anion exchange column.
For radiopharmaceutical preparations the eluate of the pyrogallol exchanger is unsuitable. The volume and the HCI molarity of the eluate, however, can be easily reduced by means o f a small column, filled with 0.6 g of a commercially available anion exchanger with quaternary ammonium groups and coupled in series with the pyrogallol exchange column. Figure 4 shows the schematic diagram o f a 10 mCi 68Ge/68Ga generator system. Gallium-68 is eluted from the pyrogallol exchanger with 10 ml 4.5N HCI and adsorbed on to the top of an anion exchange column. This takes about five minutes. Radiolytic products, resulting from the heavy radiation dose within the pyrogallol exchanger (10 mCi 68Ge/68Ga deliver about 0.5 Mrad during 80 h to the organic material [7]), are also strongly adsorbed from the acid eluate, whereas 68Ge, about 2 ppm compared with 68Ga, passes through the anion exchanger with only slight retention [10]. With 4 ml of water, about 95% of 68Ga activity adsorbed on the anion exchanger can be eluted within an additional two minutes, and the radiolytic products remain strongly adsorbed.
550
S C H U H M A C H E R and MAIER-BORST
The final volume o f the 68Ga-containing eluate is now 4 ml of 0.5N HCI. No radiolytic products or metal impurities could be detected. The mean recovery of 68Ga over a 250-day period was 74.6±3.3% relative to the 68Ga activity in equilibrium with the 68Ge adsorbed on the pyrogallol exchanger. By using pharmaceutical quality HCI and H20 , the eluate of a closed generator can be sterile and free of pyrogens, and suitable for human use.
REFERENCES [1] G R E E N E , M.W., TUCKER, W.D., Int. J. Appl. R adiat. Isot. 12 (1 9 6 1 ) 62. [2] KOPECKŸ, P., M UDROVÁ, B., In t. J. Appl. Radiat. Isot. 25 (1974) 263. [3] A RIN O , H., SKRABA, W.J., KRA M ER, H.H., Int. J. Appl. Radiat. Isot. 2 9 (1 9 7 8 ) 120. [4] NEIRIN CK X , R.D., DAVIS, M.A., J. Nucl. Med. 20 (1 9 7 9 ) 1075. [5] NEIRIN CK X , R.D ., DAVIS, M.A., R adiopharm aceuticals II (Proc. 2nd Int. Sym p. on R adiopharm aceuticals, March 1979, Seattle, W ashington), Soc. Nucl. M ed., New Y ork (1 9 7 9 ) 791. [6] NEIRIN CK X , R.D ., DAVIS, M.A., J. Nucl. Med. 21 (1 9 8 0 ) 81. [7] LOC’H, C., M A ZIÈRE, В., COM AR, D., J. Nucl. Med. 21 (1 9 8 0 ) 171. [8] LEWIS, R .E ., CAMIN, L.L., 3rd In t. Sym p. on R adiopharm aceutical C hem istry, 16—20 Ju n . 1980, St. L ouis, M issouri (A bstract). [9] J A N DER, G., BALSIUS, E., L ehrbuch analytischen und präparativen Chem ie, Verlag S. Hirzel, S tu ttg a rt (1962). [10] KRA US, K.A., NELSON, F., ASTM Special T echnical Publication, N o.195 (1 958).
DISCUSSION D.M. TAYLOR: Could you give an estimate of the cost o f your generator, compared with that of the ‘conventional’ 68Ge-68Ga generator? J. SCHUHMACHER: The cost should be lower than that of the currently available commercial generator, because radiochemically pure 68Ge is not required for the preparation of our model. Adsorption o f 68Ge on to the pyrogallol/formaldehyde resin is highly selective, and so a dissolved but unprocessed target (mostly gallium sulphate or gallium oxide) can be used.
IAEA-SM-247/131
INDIUM-113m-PH YTATE LIVER SCANNING AGENT E. -LACHNIK, W. ZULCZYK, J. WIZA, Institute of Nuclear Research, èwierk I. LICIÑSKA Institute of Drugs, Warsaw W. JAKUBOWSKI, W. GRABAN Radioisotope Department, Medical School, Warsaw, Poland
Abstract IN D IU M -1 13m -PHYTATE L IV E R SCANNING AG EN T. T he m eth o d o f labelling sodium p h y ta te by indium -113m is elaborated. E xam ination o f radiochem ical p u rity by thin-layer chrom atography on Sephadex G - l 5 w ith ascending 15% Na-azide indicates th a t in the final p ro d u c t the c o n te n t of labelled p h y ta te com plex was above 90% and th e c o n te n t o f u n b o u n d indium w as less th an 5%. Biological investigations included determ ining the organ distrib u tio n , th e liver, lung and spleen clearances, th e wholebody re te n tio n and urine ex cretio n o f the radiopharm aceutical a fte r IV injection in norm al Swiss m ice and W istar rats. T he biological stu d y indicates a selective u p ta k e o f n 3 Inm -p h y tate in liver - a bout 80% o f the initial dose 15 m in post adm inistration. Clinical investigation of the usefulness of this labelled com pound was p erform ed o n 120 p a tie n ts w ith various liver diseases. T he results show ed th a t U3Inm -p h y tate is very useful for liver imaging, especially for children and w hen exam inations m ust be rep eated several tim es.
INTRODUCTION In 1973 Subramanian and co-workers [ 1] first proposed " T cm-labelled stannous phytate as a liver scintigraphic agent. Since then many publications [2—6] connected with obtaining and using "T cm-Sn-phytate have appeared. In this report the method of labelling 113Inm-phytate and the results of radiochemical, biological and clinical investigations are presented. Indium-113m-phytate is a new liver scanning agent produced by the Radio isotope Production and Distribution Centre in Poland [7]. 551
552
-tACHNIK et al. TABLE I. TESTS ON PATIENTS WITH LIVER DISEASES Liver disease H epatom a
N o. o f p atien ts 5
Liver m étastasés
53
Liver cirrhosis
22
Virus h epatitis
13
Cholelithiasis
16
O bstructive jaundice
11 120
MATERIALS AND METHODS Preparation of the kit The stock solution was prepared by dissolving an adequate amount of sodium phytate in pyrogen-free water. The pH was brought to about 11 by 1M NaOH solution and then 0.5M phosphate buffer was added. The final solution was sterilized by filtration through a cellulose membrane; an amount of 1 cm3 was dispensed into a glass vial and freeze dried. Each vial contained 15 mg sodium phytate. Labelling was carried out by reconstituting the lyophilized phytate by adding 5 cm3 sterile 113Inm-chloride from the generator. Radiochemical purity The radiochemical purity of the labelled phytate was evaluated by thinlayer chromatography on Sephadex G-15 with ascending 15% sodium azide having a pH of 7.7. The chromatograms were developed in 30 minutes. The front with U3Inm-phytate moves about 12 cm during this period, unbound (ionic and colloidal) indium remaining at the origin. Animal investigations Biological studies were carried out using Swiss mice and Wistar rats of 20 and 150 g average weight respectively. Clearances of liver, lung and spleen, urine excretion and whole-body retention were carried out at 15, 30, 60, 90, 120 and 180 min post IV injection of 0.1 cm3 containing about 23 MBq activity.
IAEA-SM-247/131
553
Qinical investigations The radiopharmaceutical was clinically tested in 20 normal volunteers and 120 patients suffering from various liver diseases (Table I). The age of the patients was 3—81 years and they included 64 females and 56 males. Indium-113m-phytate was administered by IV injection using 37—55 MBq for children and 100—180 MBq for adults. All investigations were performed in three projections (AP, right lateral, PA) using the whole-body Picker Magna-Scanner 1000. The kinetic studies on 15 patients were carried out using the Jumbo Toshiba scintillation camera and Informatek SIMIS-3 processing system.
RESULTS Chromatographic analysis of radiopharmaceuticals showed that the content of the 113Inm-phytate complex was above 90%, and the content of unbound (ionic and colloidal) indium was less than 5%. The organ distribution data indicated that liver uptake reached a maximum value o f 15 post injection —about 80% of administered dose per 1 g tissue in mice. Clearances of liver and spleen in animals showed mean values at 180 min of less than 35% and 10% o f the injected dose respectively. The activity in the lung within 15—180 min was almost at the same level (Fig.l).
554
•fcACHNIK et al. % ID.
FIG .2.
Whole-body retention and urine excretion o f n3In m -ph ytate in mice.
FIG.3. D ynam ic organ distribution o f n3Inm -ph ytate w ithin 0 - 9 0 min p o st injection in norm al patients: A - heart; В, С - liver; D - spleen; E - background; F - spine.
IAEA-SM-247/131
FIG.4. N orm al liver scintigram o f n3I n m -phytate: (a) A P projection; (b) right lateral projection.
555
556
Í.ACHNIK et al.
FIG.5.
Scintigram o f liver m étastasés using n3I n m -phytate.
The whole-body retention showed that the average percentage of the injected dose at 180 min post injection was 40%, and at that time urine excretion reached 60% of the initial dose (Fig.2). In clinical evaluations analysis of curves obtained from regions of interest (liver, heart, spleen, spine, kidney, background) in 0 —30 min post injection showed that after 10 min more than 85% administered dose is accumulated in the liver, 1—3% in the spleen and 1-2% in the spine. Activity in the kidney was undetectable. T 1¡2 from blood is 2.8 ± 0.6 min. The optimal time span for scintigraphy is 1 5 -2 0 min post injection (Fig.3). In all normal and pathological cases very good quality scintigrams of the liver were obtained. There was no difference between the diagnostic quality of the liver obtained in comparative studies with "T cm- sulphur colloid and 113Inmphytate (Figs 4 and 5). In all scintigrams there was generally less accumulation of 113Inm-phytate in the spleen compared with "T cm-sulphur colloid.
CONCLUSIONS The complex of 113Inm-phytate is prepared by a one-step operation kit which is a very useful feature for clinical work.
IAEA-SM-247/131
557
Chromatographic analysis of 113Inm-phytate showed a high degree of complex formation and a low content of unbound indium. Biological investigations of this radiopharmaceutical indicated a selective uptake in liver with a maximum value within 15—30 min post injection. The authors found u3Inm-phytate to be a very effective agent for static liver scintigraphy. The quality of liver scintigrams was practically the same as scintigrams with " T cm-phytate and "T cm-sulphur colloid. There was generally less accumulation of 113Inm-phytate in the spleen compared with "T cm-sulphur colloid. Indium-113m-phytate is very useful for investigations on children and in cases when examinations must be repeated several times on the same patients.
REFERENCES [1]
[2] [3]
[4] [5]
[6]
[7]
SUBRAM ANIAN, G „ M cA FEE, J.G ., M E H T E R , A., B L A IR , R .J., THOM AS, F .D ., 99mT c-stannous p h y ta te : A new in vivo colloid for imaging th e reticuloendothelial system , J. N ucl. Med. 14 (1 9 7 3 ) 459. SEW ATKAR, A.B., N O RO N H A , P.O .D ., G A N A T R A , R .D ., G L E N N , H .J., Som e aspects o f th e radiopharm aceutics o f 99mT c-p h y tate, N ucl. Med. (S tu ttg a rt) 14 (19 7 5 ) 46. DENM AN, A .R ., G A LL A N T, M .J., C om parison o f liver scans using u 3 Inm -colloid and 99Tcm-p h y tate in a p a tie n t w ith hep atic m étastasés and enlarged spleen, Br. J. R adiol. 5 0 (1 9 7 7 ) 287. D A VIS, M .A., KA PLA N , M .L., A H N BERG , D .S., COLE, C.N ., A m odified Tc-99m p h y ta te colloid fo r liver-spleen imaging, In t. J. A ppl. R adiat. Isot. 28 (1 9 7 7 ) 123. AZOU M AN IAN, A., RO SEN TH A L L, L., SETO , H., Clinical com parison o f 99mTclabelled p refo rm ed p h y ta te colloid and sulfur colloid, J. N ucl. Med. 18 (1 9 7 7 ) 118 (C oncise com m u n icatio n ). Y A M AG ISHI, Y ., H O N D A , K .,.WATAN A BE, H ., SHIIBA, S., Y U K U T A K E, J., KARASAW A, M., N ISH IO , T ., Liver scanning w ith 99mT c-p h y tate, R adioisotopes 24 (1 9 7 5 ) 354 (in Japanese). L A C H N IK , E ., ZU LCZY K , W „ NOW AK, K ., W IZA, J., LICIN SK A , I., JAKUBOW SKI, W„ G RA BA N , W., C om parison o f u 3 m In -p h y tate, 99mT c-p hytate and 113mIn-colloid as liver scanning agents, In t. J. N ucl. M ed. B iol. 6 2 (1 9 7 9 ) 113.
DISCUSSION H.N. WAGNER, Jr.: How does the n3Inm-phytate compare with indium hydroxide particles that are also useful for liver imaging? One difference would seem to be in splenic activity. Is it not an advantage to be able to image the spleen as well as the liver? E. LACHNIK: The results are probably influenced by the chemical purity of the indium chloride from the generator. When we used indium chloride of
558
tACHNIK et al.
very high purity, we obtained poor quality scintigrams of the liver. The quality improved when we added a carrier. When the level of chemical impurities was higher, we observed a decrease in activity in the lung. Indium-113m-phytate provides good scintigrams of the liver.
IAEA-SM-247/106
SCINTIGRAPHIC EVALUATION OF CHRONIC CHOLECYSTITIS WITH CHOLELITHIASIS USING TECHNETIUM-99m-LABELLED PY RIDOX YLIDENE ISOLEUCINE E. ALP, C.F. BEKDIK, Y. LALELI, M.T. ERCAN Department of Nuclear Medicine, Hacettepe University Medical Center, Ankara, Turkey
Abstract SCIN TIG RA PH IC EV A LU A TIO N O F CHRO NIC CH O LECY STITIS WITH CHO LELITHIA SIS USING TECH NETIUM -99m -LABELLED PY R ID O X Y L ID E N E ISO LEUCIN E. T echnetium -99m -labelled p yridoxylidene isoleucine was evaluated in norm al subjects as a hepato b iliar scanning agent and in p a tie n ts w ith chronic cholecystitis w ith cholelithiasis. In all th e norm als and p a tie n ts the liver fun ctio n tests (SG O T, SG PT, AP and serum bilirubin) were norm al. Scintigraphic studies w ere perform ed at 5, 10, 15, 2 0 , 2 5 , 30 and 60 m in a fte r IV adm inistration o f the agent. In th e no rm al group (seven subjects) liver was visualized at 5 m in, the co m m on bile duct and the gall bladder at 5 —15 m in and the bow el activity w ithin 20 m in. In th e p a tie n t group (15 subjects) th e liver and the com m on bile duct were visualized at th e same tim e as in the no rm al group b u t the gall bladder w as eith er visualized later or n o t at all. T he results show ed th a t chronic cholecystitis w ith cholelithiasis m ay be diagnosed w ith 87% accuracy using 99Tcm-pyridoxylidene isoleucine cholescintigraphy. Scintigraphy w ith th is agent has also superiority over roentgenologic ex am in atio n in 20% of th e cases.
1.
INTRODUCTION
Iodine-131 Rose Bengal was the first agent introduced for hepatobiliary scanning [ 1] and has found wide use for many years. Because of its optimum physical characteristics and the wide availability of "T cm, it is preferable to iodine radioisotopes. In recent years radiopharmaceuticals labelled with "T cm have been developed for hepatobiliar scintigraphy, such as penicillamine [2], dihydrothioctic acid [3], tetracycline [4], mercaptoisobutyric acid [5], pyridoxylidene glutamate [6, 7] and 2,6-dimethylacetanilideiminodiacetic acid (HIDA) [8]. Kato and Hazue prepared several "T cm (Sn)-pyridoxylidene aminates with the tin reduction method and evaluated them in animals and one man [9]. They found that "T cm (Sn)-pyridoxylidene isoleucine was superior to the others. After intravenous injection it was rapidly cleared from the blood by the liver and 559
560
ALP et al.
was then rapidly excreted into the small intestine. The urinary excretion was minimal. In this investigation we have evaluated this radiopharmaceutical in normal subjects and in patients with chronic cholecystitis with cholelithiasis.
2.
MATERIALS AND METHODS
2.1. Radiopharmaceutical Teehnetium-99m-pyridoxal isoleucine was prepared according to the method described by Kato and Hazue [9] and sterilized by Millipore filtration (0.22 /im) into a sterile vial. Chromatographic quality control was performed by using thinlayer chromatography (ITLC-SG, Gelman) on every batch before administration. The solvent used was 75% chloroform + 25% methanol [9]. ITLC strips were scanned on a chromatogram scanner (Actigraph III, Nuclear Chicago). 2.2. Scintigraphy Seven normal subjects were chosen among volunteers who were found normal after abdominal physical examination and liver function tests such as serum bilirubin, SGOT, SGPT and AP. Fifteen patients were chosen from those who applied to the Department of General Surgery with complaints in abdomen and upper right abdomen. After physical and roentgenologic examinations, chronic cholecystitis with cholelithiasis indication was diagnosed and the patients were prepared for operation. Serum bilirubin, SGOT, SGPT and AP levels were determined in these patients. Only those with normal results were considered for this study. Scintigraphic studies were performed before operation. The operation results and pathology reports were obtained. Subjects were fasted at least for 2 h before scintigraphy. Two to three mCi "T cm-pyridoxylidene isoleucine was intravenously injected and gamma camera (Pho—Gamma IV, Searle Diagnostics, Inc.) pictures were obtained at 0, 5, 10, 15, 20, 25, 30 and 60 min.1 The time of visualization of the liver, the common bile duct and the gall bladder was recorded for each study. In some patients where the kidney activity was doubted the right lateral view was also taken to differentiate the kidney activity from that of the gall bladder [10].
1 1 Ci= 3.70 X 1010Bq.
561
IAEA-SM-247/106
TABLE I. VISUALIZATION TIME OF LIVER, COMMON BILE DUCT AND INTESTINES (time in minutes) N o. o f subjects
G roup
Liver
C om m on bile duct
Intestines
N orm al
7
0 -5
5 -1 5
1 5 -2 0
P atients
15
0 -5
5 -1 5
1 0 -3 0
TABLE II. TIME OF GALL BLADDER VISUALIZATION Tim e (m in)
No. o f G roup 0 N orm al
7
P atients
15
3.
5
10
15
2
5 2
20
1
25
1
30
1
60
4
N o t vis.
6
RESULTS
The thin-layer chromatography of the prepared complex indicated more than 95% labelling (Rf = 0.80). The amount of hydrolysed-reduced technetium (Rf = 0) and free pertechnetate (Rf = 1.0) was less than 5%. In both the normal and patient groups the liver activity was visualized at 0—5 min and the intestine activity at 10—20 min (Table I). In one patient the common bile duct was visualized at 10 min and the intestine activity was observed at 30 min. The common bile duct showed enlargement. The operation report was gall stones at the choleduct lower end. The time of gall bladder visualization is different in the patient group from that of the normal group. In the normal group the gall bladder was visualized at 10—15 min, but in the patient group only in two cases was the gall bladder visualized at 15 min, the rest being visualized later or not at all (Table II). In Fig. 1 a normal hepatobiliar scintigraphy is presented. The gall bladder was visualized at 10 min. Figure 2 is the scintigraphy of a patient with chronic cholecystitis with cholelithiasis. The gall bladder was visualized at 60 min. Figure 3 shows the scintigraphy of the patient with obstruction at the choleduct lower end.
ALP et al.
IAEA-SM-247/106
563
F IG .l. N orm al scintigraphy using 2 m Ci o f " T c ”1-pyridoxyliden e isoleucine, taken a t 5 min (a), 10 m in (b), 1 5 m in (c) and 2 0 m in (d).
In the patient group, the gall bladder visualization occurred at the normal range in 13% of the cases (false-negative). Delayed visualization was observed in 47% of the patients (20% at 2 0 -3 0 min and 27% at 60 min). In 40% of the patients no visualization occurred. In 87% of the patients either delayed or no visualization was observed. In those patients where the gall bladder could not be visualized the gall bladder was either completely filled with stones or the ductus cysticus was obstructed with them according to the operation reports. In three patients (20%) of our patient group the gall bladder could not be visualized radiologically.
4.
DISCUSSION
Technetium-99m-pyridoxylidene isoleucine was labelled with an efficiency of 95% according to the method of Kato and Hazue [9]. It is easy to prepare and shows ideal biological distribution, as was previously demonstrated. More than 90—95% activity was extracted by the liver and excreted in the bile into the small intestines in an hour. The kidney excretion was less than 10% [9]. Our results in normal subjects are in accordance with the previous study.
FIG.2. Scintigraphy o f a pa tien t with chronic cholecystitis with cholelithiasis a t 3 0 min (a) and 60 min (b).
IAEA-SM-247/106
565
In the present study the visualization time of the gall bladder was found to be significant. Delayed or no visualization indicated pathology. Delayed visualization of the gall bladder may be attributed to the size, localization and the number of stones present. No visualization indicated that the gall bladder was completely filled with stones or the ductus cysticus was obstructed with them. Ronai and co-workers [7] studied patients with cholelithiasis using "T cmpyridoxylidene glutamate, and reported that the gall bladder was visualized, but they did not indicate the time dependence. Stadalnik and co-workers [11] and Matolo and co-workers [12] also reported that the gall bladder was visualized in patients with cholelithiasis and as such cholescintigraphy was insignificant. They also did not determine the time of visualization, which is significant as demonstrated by the present study.
REFERENCES [1] T A PL IN , G .V ., M E R E D IT H , O.M ., K A D E, H ., T he radioactive (1-131 tagged) rose bengal u p tak e ex cretio n test fo r liver fu n c tio n using ex te rn al gam m a ray scintillation counting tech n iq u e, J. L ab. Clin. M ed. 45 (1 9 5 5 ) 665. [2]; K RIN SH N A M U RTH Y , G .T ., TUBIS, M „ ENDOW , J.S ., e t al., 99m Tc-penicillam ine, a new rad iopharm aceutical fo r cholescintigraphy, J. N ucl. M ed. 13 (1 9 7 2 ) 447. [3] TO N K IN , A .K ., D E L A N D , F ü . , D ih y d ro th io c tic acid: A new polygonal cell imaging agent, J. N ucl. M ed. 15 (1 9 7 4 ) 539. [4] F L IE G E L , C „ DEW A N JE E , M .K., HO LM AN, B.L., et al., 99mT c-tetracycline as a kidney and gallbladder imaging agent, R adiology 110 (1 9 7 4 ) 407. [5] L IN , T .H ., K H EN TIG A N , A., W INCHELL, H .S., 99mTc-labelled replacem ent fo r 131I-rose bengal in liver and biliary tra c t studies, J. N ucl. M ed. 15 (1 9 7 4 ) 613. [6] BA K ER, R .J., BELLEN , J.C ., R O N A I, P.M ., T echnetium -99m -pyridoxylidene g lutam ate: A new h ep ato b iliary radiopharm aceutical: I. E x p erim en tal aspects, J . N ucl. M ed. 16 (1 9 7 5 ) 720. [7] R O N A I, P.M ., BA K ER, R .J., BE LL E N , J.C ., e t al., T echnetium -99m -pyridoxylidene glutam ate: A new h ep ato b iliary radiopharm aceutical: II. C linical aspects, J. N ucl. Med. 1 6 (1 9 7 5 ) 728. [8] H A R V E Y , E ., L O B E R G , M., CO O PER, M ., " mTc-HIDA: A new radiopharm aceutical fo r h e p ato b iliary imaging, J. N ucl. M ed. 16 (1 9 7 5 ) 533. [9] K A TO , M ., H A ZU E, M ., 99mT c (S n) pyridoxylidene-am inates: P rep aratio n and biological evaluation, J. N ucl. M ed. 19 (1 9 7 8 ) 397. [10] B A K ER , R .J., M A NON , M .A., Biliary scanning w ith 99mT c-pyridoxylidene g lutam ate — the effect o f food in no rm al subjects, J. N ucl. Med. 18 (1 9 7 7 ) 793 (C oncise com m unication). [11] STAD ALNIK, R .A ., M A TO LO , N .M ., JO N SH O L T, A .L ., e t al., 99mT c-pyridoxylidene glutam ate (P.G .) cholescintigraphy, R adiology 121 (1 9 7 6 ) 657. [12] M A TO LO , N.M., R O B E R T , C „ STA D A LN IK , R .C ., e t al., Biliary tra c t scanning w ith 99mT c-pyridoxylidene g lutam ate: A new gallbladder agent, Surgery 80 (1 9 7 6 ) 317.
566
ALP et al. DISCUSSION
E. PETURSSON: H o w does the information obtained with this material differ from that obtained with HIDA or other such compounds? M.T. ERCAN: We have not made a comparative study, but "T cm-pyridoxal isoleucine was shown to be better than other pyridoxal aminates by Kato and Hazue (Ref.[9] in the paper) in that it was more rapidly excreted from the liver (>90% in 1 h), and its kidney excretion was minimal (<5%). M. ITTURALDE: Have you evaluated the behaviour of the labelled pyridoxylidene isoleucine at different serum bilirubin levels? M. T. ERCAN: No. We chose both our normal subjects and patients from among those who had normal serum bilirubin levels. N
IAEÀ-SM-247/89
E f f e c t s o f s t r u c t u r a l m o d i f i c a t i o n s o n t h e a d r e n a l u p t a k e o f s t e r o i d s l a b e l e d i n t h e s i d e c h a i n w i t h tin-117m
TIN-117m-LABELLED RADIOPHARMACEUTICALS
F.F. KNAPP, Jr. Nuclear Medicine Technology Group, Health and Safety Research Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee, United States of America
Abstract TIN-117m-LABELLED RADIOPHARMACEUTICALS: EFFECTS OF STRUCTURAL MODIFICATIONS ON THE ADRENAL UPTAKE OF STEROIDS LABELLED IN THE SIDECHAIN WITH TIN-117m. The U7Snm nuclide has a 14-day physical half-life and decays by isomeric transition with emission of a 159-keV gamma photon in 87% abundance. Because of these attractive properties, the preparation and testing of n 7Snm-labelled tissue-specific radiopharmaceuticals have been explored. A series of steroids labelled in the sidechain with the trimethyl [U7Snm] tin moiety have been prepared and tested in rats to determine the structural features required for maximal adrenal uptake of these potential adrenal imaging agents. The structural modifications that were investigated include alterations of both the steroid nucleus and sidechain. The following u 7Snm-labelled steroids were prepared in tetrahydrofuran at room temperature by reaction of the Me3 117Snm-Li reagent with steroid substrates containing a primary bromine in the sidechain: 23-(trimethylstanna)-24-nor-5a-cholan-3(?-ol (I, saturated nucleus), 24-{trimethylstanna)-chol-5en-3/3-ol (II, analogue of cholesterol), 3/3-methoxy-24-(trimethylstanna)-chol-5-ene (III, hydrophobic C-3 substituent), 23-(trimethylstanna)-24-nor-5/3-cholan-3a-ol (IV, cis-A/B ring juncture) and 17ß-[(trimethylstanna) methyl]-androst-5-en-3(3-ol (V, short sidechain). The results of tissue distribution studies in rats indicate that the important structural features required for significant adrenal uptake include an all trans ring juncture, an equatorial C-3 hydroxyl group and a side chain of moderate length. One day after administration the following levels of radioactivity (mean % injected dose/gram) were detected in rat adrenals: I, 47.13; II, 17.37; III, 12.17; IV, 6.85; V, 3.05. Steroid I (23-TSC) showed the highest adrenal uptake and only one day after injection the adrenal : blood and adrenal : liver ratios were 33 :1 and 9:1 respectively. While the adrenal : liver ratio increased steadily over a seven-day period after administration of I, this ratio decreased over this time period with steroids II-V . Since 117Snm-23-TSC showed the greatest adrenal specificity of the compounds investigated, more extensive tissue distribution and excretion studies were performed with this agent over a 2 1-day period. The adrenal uptake (mean % injected dose/gram) of 117Snm-23-TSC peaked at 3 days (50.88%) and decreased to 2.33% after 21 days. Using the rat tissue distribution and excretion data, the absorbed radiation dose values from 117Snm-23-TSC to human organs have been estimated. The calculated radiation dose values for n?Snm-23-TSC are as follows: adrenals, 83 rad/mCi; total body, 0.77 rad/mCi;
567
568
KNAPP
ovaries, 4.4 rad/mCi. These values are similar to estimates determined for a variety of other radiolabelled steroid adrenal imaging agents at present used for the clinical evaluation of adrenal disease, and suggest that 117Snm-23-TSC may be an attractive new agent for adrenal visualization in humans.
INTRODUCTION Steroids labeled with gamma-emitting radionuclides are attractive agents for the detection and diagnosis of various pathologies of the adrenal cortex. The clinical use of 6ß([131I]iodomethyl)-19-nor-cholest-5(10)-en-33-ol (NP-59) has recently been reviewed [1,2]. In addition to 131I-labeled NP-59, a 75Se-labeled steroid, 6ß-[(methyl[75Se]seleno)methyl ]19-nor-cholest-5(10)en-3ß-ol (Scintidren), has been used clinically for the diagnosis of adrenal disease [3,4]. In addition to a longer shelf-life, the 75Se-labeled agent would eliminate the high radiation dose to the thyroid glands encountered with 131I-NP-59. A variety of structurally modified steroids labeled with other radionuclides have also been studied as potential adrenal imaging agents [2]. We have prepared several 123mTe-labeled steroids as potential alternatives to NP-59 and 23-(isopropyl[123mTe]tel1uro)-24-nor-5a-cholan-3ßol (123mTe-23-ITC) and 24-(isopropyl[123mTe]telluro)-chol5-en-3ß-ol ( 123,77Te-24-ITC) show pronounced adrenal uptake in rats [5,6]. The estimated absorbed radiation dose values to human organs from 123mTe-23-ITC and 123mTe-24-ITC are also within the same range as values calculated for 131I-NP-59 [7]. The 117mSn nuclide has attractive radionuclidic properties and 1l7mSn-tartrate has been studied as a potential bone imaging agent [8]. In addition to a favorable 14-day physical half-life, 117™5п decays primarily by isomeric transition with emission of a single gamma photon with an energy of 159 keV in 87% abundance (Fig. 1). The 159-keV photon is within the optimal range for detection with Nal crystal detectors used in nuclear medicine instrumentation. These attractive radionuclidic properties coupled with the established versatility of organotin chemistry [9] have prompted us to explore the preparation of tissue-specific 117mSn radiopharmaceuticals. Recently, we reported the pronounced adrenal uptake in rats of 24-(trimethyl[117mSn]stanna)-5a-cholan-3ß-ol, a unique steroid prepared by reaction of trimethyl [*17mSn]tinlithium (Мез117mSn-Li) with 3ß-acetoxy-24-nor-5a-cholane [10,11]. The goals of the present studies were to prepare a series of structurally modified steroids by coupling M e 3117mSn-Li with halogenated steroid substrates and to determine the effect of steroid structure on the adrenal uptake of these agents in rats.
569
IAEA-SM-247/89 159 keV ; "7 m Sn -
392 keV И3т_
3
2 5 5 keV
О
о
" 3 Sn
i
о
о
ENERGY (keV) F IG .l. The Ge (Li) crystal gamma spectrum o f n7mSn produ ced in the Oak R idge High Flux R eactor (HFIR) by neutron irradiation o f 94. 74% enriched 116Sn.
REACTOR
PRODUCTION
OF TIN-117m
с
со
с г >
о <
о ч—
E \
с
СЛ Е
О CD
Q. СО
О
Irradiation Period, Days FIG. 2. The specific a c tiv ity o f lllm Sn as a fu n ction o f the irradiation p erio d o f 94. 74% enriched ll6Sn in the Oak Ridge High Flux R eactor (HFIR).
570
KNAPP
V FIG.3. Structures o f trim eth yltin -su bstitu ted steroids prepared by reaction o f brom inated steroid substrates with trim ethyltinlith iu m in tetrahydrofuran.
METHODS Radiopharmaceutical s. The 117mSn was produced in the Oak Ridge High Flux Isotope Reactor (HFIR) using the 116Sn(n,Y)117mSn nuclear reaction by irradiation of isotopically enriched 116Sn (95.74%) with a neutron flux of 2.5 x 10i5 neutrons/cm2 -s. The enriched 116Sn ($0.35/mg) was obtained from the Oak Ridge National Laboratory Isotope Sales Office. An assessment of the factors influencing production of 117mSn were based on the results of five experiments in which targets were irradiated in the same physical location in the hydraulic tube facility of the HFIR. The targets varied in weight from 10 to 73 ma, and the irradiation periods varied from 2 to 14 days. The ïi7mSn production yields (Fig. 2) are consistent with an effective cross-section (o) of 13.1 millibarns which follows the radio isotope production equation A = N ф a (1 - e -x^)[12]. This equation holds for cases in which the number of target atoms does not decrease significantly and the radionuclide product
IAEA-SM-247/89
571
TABLE I. TRIMETHYL[ mSn]TIN-SUBSTITUTEDSTEROIDS PREPAREDBYREACTIONOF Me mSn-Li W ITHSIDECHAINBROMINATEDSTEROIDSUBSTRATES 1 1 7
3 1 1 7
Substrate 3ß-Acetoxy-23-bromo24-nor-5d-cholane 3ß-Acetoxy-24-bromochol-5-ene 3ß-Methoxy-24-bromochol-5-ene 3a-Acetoxy-23-bromo24-nor-5ß-cholane 3ß-Acetoxy-17ß(bromomethyl)androst-5-ene
Yield“ (%) Saturated nucleus 49 23-(Trimethyl stanna)-24nor-5a-cholan-3ß-ol (I) trans ring structure Nuclear double 69 24-(Trimethyl stanna)bond chol-5-en-3ß-ol (II) 3ß-Methoxy-24-(trimethyl Hydrophobic C-3 34 stanna)-chol-5-ene (III) substituent Cis A /В ring 49 23-(Trimethyl stanna)-24nor-5ß-cholan-3a-ol (IV) juncture 17ß-[(Trimethyl stannaImethyl]- Short side chain 35 androst-5-en-3ß-ol (V) Product
Unique structural features
“Chemical yield values are based upon the brominated steroid starting materials since the Me n 'mSn-Li reagent was used in excess. 3
disappears primarily by radioactive decay processes. These results suggest that ll7mSn does not have a significant neutron capture cross-section and that target self-shielding is not a significant factor. The maximum specific activity of 117mSn which we can achieve is calculated to be ^ 4.4 mCi/mg. In excess of 140 days of irradiation would be required to reach the maximum value [where (1 - e -^t) ->1 as t -> ”]. Five structurally modified steroids (Fig. 3) substituted in the sidechain with the M e 3Sn- moiety were prepared by reacting the M e 3Sn-Li reagent with steroid substrates brominated in the sidechain (Step 4). The substrates that were chosen had major structural modifications of both the nucleus and sidechain (Table I). The chemical yields of the brominated steroid sub strates and the M e 3Sn-substituted products I-V were not optimized. The brominated steroid substrates used in these reactions were prepared by well-established procedures. The 3ß-acetoxy-23bromo-24-nor-5a-cholane and 3a-acetoxy-23-bromo-24-nor-5ß-cholane were prepared by modified Hunsdiecker degradation of allisolithocholic acid and lithocholic acid acetates, respectively, as described earlier [13]. The 3ß-acetoxy-24-bromo-chol-5-ene was prepared by the following route: methyl-3ß-hydroxy-chol-5-en-24-oate -*methyl3B-methoxy-chol-5-en-24-oate -+■3e-methoxy-24-hydroxy-chol-5-ene -+ 3ß-methoxy-24-bromo-chol -5-ene -»■3ß-acetoxy-24-bromo-chol -5-ene.
KNAPP
572
In a similar manner, 3ß-acetoxy-17g-(bromomethyl)-androst-5-ene was prepared as follows: methyl-androst-5-en-l 73-carboxyl ate -> methyl 3ß-methoxy-l7ß-carboxy-androst-5-en-20-oate -> 3ß-methoxy-17ß(hydroxymethyl )-androst-5-ene -* 33-methoxy-17ß-(bromomethyl )-androst5-ene -> 3ß-acetoxy-17ß-(bromomethyl)-androst-5-ene. The synthetic steroids were homogeneous by both silicic acid column and thin layer chromatography. The infrared, low- and high-resolution mass spectral, and proton nuclear magnetic resonance spectral properties were consistent with the proposed stuctures. Details of the synthesis and physical properties of the brominated steroid substrates and unlabeled trimethyltin-substituted steroids (I-V) will be described in a subsequent report. The 117mSn-labeled steroids were prepared by reaction of the substrates with M e 3 *17mSn-Li. Reactor-produced 117mSn (1 mmole) was converted to the highly reactive 117mSnCl4 intermediate by chlorination with Cl z at 150°C using the special 'flow through1 apparatus that has been described earlier [14]. The 117т5пС1ц, obtained in greater than 90% yield, was converted to M e 3 117mSn-Cl by a 'comproportionation' reaction with three mmoles of Me^Sn at 39-45°C [9]. Following room temperature reaction with excess lithium metal in 10 ml of dry tetrahydrofuran (THF), the resulting solution of M e 3 117mSn-Li (15 mCi/mmole) was decanted and 1 ml aliquots (^ 0.4 mmole) added to vials containing 50 pinoles of the steroid substrates. All of these reactions were performed in an argon atmosphere. After one hour the excess reagent was destroyed by H 20 and the products obtained by solvent extraction. The reaction scheme is outlined below. Cl о 150°C
U7msn --- 1
Step 1
Li
THF, 27 С Step 3
^ 117mSnCl 4 + 3 Me^Sn
M e 3117^Sn-Li + Steroid-Br
дс°г - M e 3 117mSn-Cl + Step 2
Step 4
Steroid-
117mSn-Me3 The 117mSn-labeled steroids were purified by silicic acid column chromatography. Elution with petroleum ether removed a radioactive non-polar component and in each case the homogeneous radioactive steroid was eluted with 25 I ether in petroleum ether. The radiolabeled steroids I-V each showed a single radioactive component by thin-layer radiochromatographic analysis in two solvent systems (CHC13 and 2% MeOH/CHCl3 ) that co-chromatographed with the unlabeled standards. Animal Studies. Female Fischer 344 rats (160-180 gm) were used for the tissue distribution and excretion studies. Both food and water were allowed ad lib itu m during the experiments.
IAEA-SM-247/89
573
TABLE II. DISTRIBUTIONOF RADIOACTIVITY INSELECTEDTISSUES OF FEMALE FISCHER 344 RATS AT 1, 3, AND7 DAYSAFTER INTRAVENOUS ADMINISTRATIONOF mSnLABELEDSTEROIDS I-V * 117
mSnLabeled Steroid; Days After Injection 1 17
%Injected Dose/gram (Range) Tissue Blood Liver
Adrenals ' !? „
n( 7
III \( 13 (7 "
I)
»
!i
47.13(37.13-53.47) 50.88(45.80-59.54) 19.37(17.54-21.26) 17.37(13.65-23.23) 18.90(17.36-21.40) 8.20(6.96-9.75) 12.17(11.73-12.46) 12.31(9.68-15.55) 4.34(3.90-4.89) 6.85(6.70-7.00) 4.48(4.38-4.54) 2.05(1.71-2.28) 3.05(2.59-3.61 ) 0.98(0.96-1.13) 0.47(0.45-0.49)
1.40(1.33-1.54) 0.80(0.77-0.85) 0.19(0.15-0.29) 1.71(1.64-1.82) 1.59(1.54-1.63) 1.26(1 .17-1.35) 0.60(0.52-0.65) 0.81(0.74-0.86) 0.89(0.60-1 .10) 0.86(0.72-0.95) 0.82(0.77-0.86) 0.73(0.70-0.75) 0.48(0.47-0.48) 0.39(0.36-0.43) 0.34(0.28-0.38)
5.18(4.74-5.71) 2.13(1 .81-2.49) 0.44(0.33-0.61 ) 5.76(5.51-6.17) 3.17(3.12-3.25) 1.63(1.51-1.73) 6.56(5.64-7.52) 4.29(3.92-4.74) 3.42(2.35-4.26) 3.71(3.51-3.90) 2.31(2.07-2.75) 0.97(0.87-1.06) 1.93(1 .64-2.13) 0.50(0.45-0.56) 0.32(0.28-0.34)
Ovaries 18.22(11 .88-21 .73) 10.99(10.24-11 .70) 4.90(3.82-6.66) 8.90(6.43-12.48) 11.91(11 .30-13.01) 5.25(4.33-5.94) 4.26(3.48-5.41) 4.53(4.32-4.22) 3.62(3.38-4.04) 5.77(5.24-6.59) 3.81(3.57-4.06) 1.95(1 .70-2.13) 1.76(1 .64-1 .91) 0.89(0.74-1.05) 0.32(0.30-0.35)
*Percent injected dose/gram values are the mean and range for three rats. Other tissues that were analyzed include the brain, heart, kidneys, lungs, pancreas, small and large- intestines, stomach and thyroid. The crystalline steroids were dissolved in ethanol and formulated in a Tween 80-ethanol-saline mixture as described elsewhere [1]. The rats were lightly anesthetized with ether and 1 ml of the steroid solution containing 5-10 pCi was injected in a lateral tail vein. The rats were killed at various time intervals (Table II) by decapitation and the organs removed, rinsed with saline, blotted dry, weighed and counted in a gamma spectrometer. For the excretion studies, three rats were individually housed in metabolism cages,and urine and feces were collected daily over the 21-day period. RESULTS AND DISCUSSION The synthesis of 117mSn-labeled organic radiopharmaceuticals requires the initial conversion of 117mSn into a suitable inter mediate for conversion to the compound of interest. Tin tetra chloride (SnClit) fills the requirements of a versatile and useful
574
KNAPP
OAYS
AFTER
INJECTION
FIG.4. Comparison o f the adrenal: b lood ratios determ ined a t 1, 3 and 7 days after intravenous adm inistration o f trim eth yl-[lllm Sn]-tin labeled steroids I - V. The adrenal: b lood ratios are calculated from the percentage injected dose/gram values sum m arized in Table II.
intermediate for the synthesis of a wide variety of organotin compounds. The high yield method that has recently been developed for the conversion of 117mSn to the pivotal 117mSnCl4 intermediate [14] has now been used for the preparation of a variety of 117mSnlabeled steroids (Table I). The M e 3Sn-moiety was introduced into the steroid sidechain since it seemed probable that such a bulky substituent would be more easily accommodated in this region of the molecule. The tissue distribution of 11?mSn-labeled steroids I-V was studied in rats to determine the structural features required for maximal adrenal uptake. Tin-117m-labeled 23-(trimethylstanna)-24-nor-5a-cholan-3ß-ol (I) showed the most pronounced adrenal uptake of the steroids investigated (Table II). The adrenal uptake was high one day after administration and decreased over the seven-day period. After one day, 2.5% of the administered radioactivity was detected in the adrenal glands^and after seven days the concen tration of radioactivity had been reduced to 1 .9% of the injected dose. These data parallel the results observed with 23-(isopropyl [123mTe]telluro)-24-nor-cholan-3ß-ol in which the adrenal accumulation of radioactivity in female rats decreased steadily from 4.5% of the injected dose after one day to 2.4% after seven days. A convenient index of the adrenal selectivity of the radiolabeled steroids is a comparison of either the adrenal:blood or adrenal:1iver ratios. The adrenal:blood ratios for the 117mSnlabeled steroids were calculated from the percent dose per gram of tissue values and are compared in Fig. 4. These data dramatically illustrate the effects of steroid structure on the adrenal specificity of the trimethyl tin-substituted steroids and indicate that well-defined structural features are required for maximal adrenal uptake. From an analysis of these data it is apparent
IAEA-SM-247/89
DAYS T IM E
AFTER
575
W EEKS IN J E C T IO N
FIG.5. The distribution o f radioactivity in tissues o f Fischer 344 fem ale rats determ ined at intervals over a three-week period after intravenous adm inistration o f 2 4 -(trim e th y /-[n7mS/i]stanna)-5a-cholan-3ß-ol (I) ( n lm Sn-23-TSCj.
that an all trans ring juncture, a sidechain of moderate length, and an equatorial hydroxyl group are required for adrenal specificity. We have found that similar structural features are required for adrenal uptake of steroids labeled in the sidechain with 123mTe [6]. Since 117mSn-23-TSC showed the highest adrenal uptake of the steroids studied, the tissue distribution and excretion properties of this agent were studied in more detail. More extensive tissue distribution studies performed over a 21-day period indicated that the radioactive contents of the adrenal glands decreased steadily after the maximal uptake was observed at three days (Fig. 5). Maximal uptake in non-target tissues such as blood, liver and kidneys was observed after one day and then decreased slowly. Although the absolute adrenal uptake of radioactivity (% injected dose) decreased rapidly after maximal values were observed between 1 (2.51%) and 3 (2.86%) days, peak adrenal:blood (102:1) and adrenal :1iver (44:1) ratios were not reached until after 7 days. Analysis of the radioactive contents of the urine and feces of rats over the 21-day period following administration of u ? m Snlabeled 23-(trimethylstanna)-24-nor-5a-cholan-3e-ol (I) demonstrated that the majority of the radioactivity was excreted in the feces. Approximately 50% of the injected radioactivity was excreted within five days. Although the identity of the radioactive contents in the feces has not yet been determined, these data suggest that the trimethyltin moiety remains attached to the steroid molecule since it is well established that neutral steroids are excreted by the fecal route.
576
KNAPP
Using the rat tissue distribution and excretion data, the radiation dose values to human organs from 117mSn-23-TSC have been estimated [11]. The adrenal glands receive the highest radiation dose (83 rad/mCi), which is considerably less than the 150 rad/mCi value estimated from rat tissue distribution for 1311-NP-59 [I].1 Although the maximum specific activity (1-2 mCi/mg) of reactor-produced 117mSn is limited by a rather low production cross-section for neutron capture by 115Sn, the tissue specificity of several classes of radiopharmaceuticals, such as long-chain fatty acids for myocardial imaging, should not be affected by specific activities in the 10-30 mCi/mmole range. The high plasma levels of free cholesterol in humans (150-200 mg/100 ml) significantly dilute adrenal-specific radiolabeled steroids. Under such conditions low specific activity is not a problem and radiolabeled steroids with low specific activities have been used successfully for adrenal imaging [1,5,6]. These encouraging results with 117wSn-1abeled 23-TSC suggest that the preparation of other 117mSn-labeled agents should be explored. Although preparation of the M e 3117mSn-Li reagent proceeds in good yields by the methods that have been described in this paper, the significant 75% decrease in specific activity during the 117mSnClu + 3 Me4Sn -* M e 3117mSn-Cl conversion presently limits the specific activity of 117mSn-labeled agents prepared by this method to about 60 mCi/mmole. The development of micro-scale methods to increase the specific activity to about 250 mCi/mmole should be possible.
ACKNOWLEDGEMENTS Research was sponsored by the Office of Health and envi ronmental Research, U.S. Department of Energy under Contract No. W-7405-eng-26 with the Union Carbide Corporation. The author thanks T. A. Butler, A. P. Callahan, R. A. Grigsby and L. A. Ferren for technical assistance and L. S. Ailey for typing the manuscript.
REFERENCES [1] TH R A L L , J.H ., FR E IT A S, J.E ., BEIERW ALTES, W.H., A drenal scintigraphy, Semin. Nucl. Med. 8 (1978) 23. [2] BEIERW ALTES, W.H., W IELAND, D.M., YU, T., SWANSON, D.P., M OSLEY, S.T., A drenal imaging agents: R ationale, synthesis, fo rm ulation and m etabolism , Sem in. Nucl. Med. 8 (1 9 7 8 ) 5.
1 1 rad = 1.00 X 10'2 Gy ; 1 Ci = 3.70 X 1010Bq.
IAEA-SM-247/89
577
[3] REILEY, A.L.M., The development of selenium-75 cholesterol analogues, J. Labelled Compd. Radiopharm. 16 (1979) 28. [4] CHATAL, J.F., CHARDONNEL, B., GUIHARD, D., Radionuclide imaging of the adrenal glands, Clin. Nucl. Med. 3(1978) 71. [5] KNAPP, F.F., Jr., AMBROSE, K.R., CALLAHAN, A.P., Tellurium-123m-labeled 23-(isopropyl telluro)-24-nor-5a-cholan-3)3-ol: A new potential adrenal imaging agent, J. Nucl. Med. 21 3 (1980) 251. [ 6 ] KNAPP, F.F., Jr., AMBROSE, K.R., CALLAHAN, A.P., The effect of structural modifi cations on the adrenal uptake of steroids labeled in the side chain with tellurium-123m, J. Nucl. Med. 21 3 (1980) 258. [7] WOO, D.V., KNAPP, F.F., Jr., AMBROSE, K.R., CALLAHAN, A.P., COFFEY, J.L., Radiation dosimetry of two new tellurium-123m-labeled adrenal-imaging agents, J. Nucl. Med. 21 5 (1980) 454 (Concise communication). [ 8 ] YANO, Y., CHU, P., ANGER, H.O., Tin-117m: Production, chemistry and evaluation as a bone-scanning agent, Int. J. Appl. Radiat. Isot. 24 (1973) 319. [9] NEUMAN, W.P., The Organic Chemistry of Tin, John Wiley & Sons, New York (1970). [10] KNAPP, F.F., Jr., BUTLER, T.A., CALLAHAN, A.P., AMBROSE, K.R., Tin-1 П т labeled 23-(trimethyl stanna)-24-nor-5a-cholan-3/3-ol (*Sn-23-TSC) shows significant adrenal uptake in rats, J. Nucl. Med. 21 6 (1980) 80. [11] COFFEY, J.L., KNAPP, F.F., Jr., Radiation dosimetry of Sn-117m-labeled 23-(trimethyl stanna)-24-nor-5a-cholan-3(3-ol (Sn-24-TSC): A potential adrenal imaging agent, J. Nucl. Med. 21 6 (1980) 8 6 . [12] MUGHAGHAB, S.F., GARBER, D.I., “Neutron cross sections”, Resonance Parameters 1, Brookhaven National Laboratory, Rep. BNL 325 (1973) 50. [13] KNAPP, F.F.,-Jr., Steroid synthesis: The modified Hunsdiecker degradation of bile acids and related compounds, Steroids 33 (1979) 245. [14] WOO, D.V., KNAPP, F.F., Jr., BUTLER, T.A., CALLAHAN, A.P., “An efficient micro scale preparation of tin-117m-tin tetrachloride — A pivotal intermediate for the synthesis of tin-117m-labeled radiopharmaceuticals”, Radiopharmaceuticals II (Proc. Int. Conf. Seattle, 1979), Society of Nuclear Medicine, New York (1979) 147.
DISCUSSION D.A. GOODWIN: How much 113Sn is present as a contaminant, and how much would this contribute to the radiation dose? F.F. KNAPP, Jr.: The level o f 113Sn in the reactor-produced U7Snm is very low, and arises from neutron capture of the small amount of 112Sn present in the enriched U6Sn target material. Although I do not have the exact figures available, the 113Sn amounts to only a fraction of one per cent, and our calculations indicate that its contribution to the absorbed radiation dose value would be insignificant. H.N. WAGNER, Jr.: I would like to congratulate you on another important new contribution. Have you considered the possibility that this nuclide might be useful as a red blood cell label, either for heart studies or red blood cell studies?
578
KNAPP
As you know, tin is used to sensitize red blood cells for subsequent phosphate labelling with technetium. Perhaps 117Snm might prove to be a useful red blood cell label. F.F. KNAPP, Jr.: We have not considered the possibility of labelling red blood cells with 117Snm, but it is a very interesting suggestion.
IAEA-SM-247/94
YTTERBIUM-169 CITRATE FOR DETECTION OF INFLAMMATORY LESIONS* R.L. HAYES, B.L. BYRD, J.J. RAFTER, J E. CARLTON, J.L. COFFEY Oak Ridge Associated Universities, Oak Ridge, Tennessee, United States of America
Abstract YTTERBIUM-169 CITRATE FOR DETECTION OF INFLAMMATORY LESIONS. Like 67Ga, 169Yb shows an affinity for inflammatory processes. Unlike 67Ga, 169Yb can be reactor-produced and at a much lower cost. It has photon emissions that are also quite suitable for scanning. The authors have made a detailed comparison of the tissue distribution of 67Ga and 169Yb citrates at 4 and 24 h after intravenous administration in rats bearing an experimental abscess to determine the relative worth of the two radionuclides for abscess detection. At 4 h the abscess-to-normal tissue ratios for lung, muscle, bone marrow, blood and skin were significantly higher with 169Yb, whereas with 67Ga the ratios were significantly better only with kidney and bone, suggesting that 169Yb may be of use for the more rapid detection of inflammatory lesions. At 24 h the two agents were approximately equal in utility with better ratios being obtained again for kidney and bone with 67Ga but only for muscle and blood with 169Yb. A drawback to the possible clinical use of 169Yb for detection of inflammatory processes is its long Ту 2 (32 d) which limits the 169Yb dose that could be clinically administered. On the other hand, 169Yb’s long T i/2 would result in a cost-effective shelf-life for this radiopharmaceutical, especially for facilities where 67Ga is not used because of its cost or where it is frequently unavailable.
INTRODUCTION T he r a d io p h a r m a c e u tic a l ^^G a c i t r a t e i s now w id e ly u s e d f o r th e d e t e c t i o n o f in fla m m a to ry p r o c e s s e s [ 1 ] . We h a v e p r e v i o u s l y o b s e r v e d t h a t 2 4 h a f t e r a1 d¿LQ m i n i s t r a t i o n o f ^-®®Yb c i t r a t e t h e r e w as a r e l a t i v e u p ta k e o f D Yb i n a n i n f l a m m a t o r y l e s i o n i n t h e r a t t h a t w as s i m il a r to t h a t o b s e rv e d w ith ^ G a [ 2 ] . S in c e th e b i n d in g o f 6 7 g a b y p la s m a p r o t e i n d e la y s th e c l e a r a n c e o f ^ G a fro m b lo o d , c o m p a red i n
w h e r e a s t h e c l e a r a n c e o f -*-” "Y b i s r a p i d [ 3 ] , w e h a v e d e t a i l t h e t i s s u e d i s t r i b u t i o n o f 1 6 y Yb w i t h t h a t o f
* Based on work supported by contract number DE-AC05-760R00033 between the U.S. Department of Energy and Oak Ridge Associated Universities. 579
HAYES et al.
580 67
Ga a t 4 a n d 2 4 h i n r a t s b e a r i n g a n e x p e r i m e n t a l b a c t e r i a l a b s c e s s t o d e t e r m i n e w h e t h e r ^® ^Y b m i g h t b e a s u i t a b l e s u b s t i t u t e f o r 6?G a f o r d e t e c t i n g in f la m m a to r y p r o c e s s e s , p a r t i c u la rly a t e a r lie r s ta g e s a f te r ag en t a d m in is tra tio n .
M ATERIALS AND METHODS G ro u p s
of
fiv e
m a le F i s c h e r
344
ra ts
(H a rla n
In d u s trie s ,
I n d i a n a p o l i s , IN ) w e r e u s e d . A s a lin e s lu rry o f r a t liv e r th a t h a d b e e n p a s s e d t h r o u g h a 1-m m t i s s u e p r e s s ( H a r v a r d A p p a r a t u s , M i l i s , MA) w a s i n n o c u l a t e d w i t h S^. a u r e u s 6 3 3 9 ( A m e r i c a n T y p e C u l t u r e C o l l e c t i o n , R o c k v i l l e , MD) s u c h t h a t t h e m i x t u r e c o n t a i n e d 10® b a c t e r i a / m l . To p ro d u c e a b s c e s s e s 1 -m l p o r t io n s o f t h i s s l u r r y w e re th e n i n je c te d s u b c u ta n e o u s ly on th e b a c k n e a r th e h e a d . A b s c e s s e s an d o th e r t i s s u e s w e re th e n h a r v e s te d on th e 5 th day a f t e r in je c tio n o f th e l i v e r p r e p a r a tio n . 169 The Yb c h l o r i d e u s e d i n o u r a b s c e s s s t u d i e s w a s g e n e ro u s ly s u p p l i e d b y U n io n C a r b i d e C o r p o r a t i o n , T u x e d o , NY. Y t t e r b i u m - 1 6 9 c h l o r i d e w a s a l s o p u r c h a s e d f r o m N ew E n g l a n d N u c l e a r , N . B i l l e r i c a , MA, a s w a s ^ G a c i t r a t e . The c h lo r i d e w as c o n v e r te d to th e c i t r a t e fo rm by a d d in g c i t r i c a c i d a n d a d j u s t i n g t o pH 7 w i t h s o d iu m h y d r o x i d e . B o th l ^ Y b a n d 6^G a ( 2 0 - 5 0 y C i e a c h ) w e re a d m i n i s t e r e d i n t r a v e n o u s l y b y t a i l v e i n a t a c i t r a t e l e v e l o f 1 m g /k g . T he s ta b le y tte rb iu m p r e s e n t i n th e l ^ Y b p r e p a r a t i o n s d id n o t e x c e e d 0 .1 y g /k g w hen a d m in is te re d .1 A n im a ls w e re k i l l e d b y e x s a n g u in a ti o n f o l l o w i n g a b r i e f e x p o su re to d ie th y l e th e r . W e ig h e d s a m p le s o f t i s s u e w e r e c o u n te d i n a w e l l - s c i n t i l l a t i o n c o u n te r a g a i n s t s ta n d a r d s a n d th e ^^^Y b an d ^ G a c o n c e n tr a t io n s c a l c u l a t e d a s p e r c e n t o f a d m i n i s t e r e d d o s e p e r g ra m o f t i s s u e n o r m a liz e d t o a b o d y w e ig h t o f 250 g . In a d d itio n to th e t is s u e s l i s t e d in th e t a b l e s , we a l s o s a m p le d t h e c o n t e n t s o f t h e g a s t r o i n t e s t i n a l tra c t. T he s m a ll i n t e s t i n e w as d iv id e d i n t o f o u r s e c t i o n s . W h o le -b o d y r e t e n t i o n a n d t o t a l a b s c e s s u p t a k e o f Yb a n d Ga w e re d e te r m i n e d b y t h e g e o m e tr y - in d e p e n d e n t t e c h n i q u e o f G ib b s e t a l . [4 ]. S t a t i s t i c a l a n a l y s e s w e re p e r f o r m e d u s i n g S t u d e n t 's t-te s t.
RESULTS A t 4 h th e a b s o lu te u p ta k e s v ia b le a b sc e ss tis s u e ( v ia b ility
1 1 Ci = 3.70 X 1010 Bq.
(% /g ) o f ^ ^ Y b a n d ^ G a i n d e te rm in e d b y v i s u a l in s p e c t i o n )
IAEA-SM-247/94
581
TABLE I. COMPARISON OF THE 4-h TISSUE DISTRIBUTIONS OF YTTERBIUM-169 AND GALLIUM-67 IN RATS BEARING AN S. aureus ABSCESS
Sample
Ga-67
Yb-169 % Administered Dose/g
Viable abscess Necrotic abscess Liver Spleen Kidney Lung Skeletal muscle Hèart Bone Bone marrow Blood Stomach Small intestine Cecum Colon Skin Fat Testis
1.40 0.16 0.75 0.68 1.10 0.45 0.07 0.29 2.20 0.50 0.56 0.68 0.53 0.46 0.27 0.13 0.05 0.21
+ + + + + + + + + + + + + + + + + +
Weight of abscess (g) Activity in abscess (%)
2.6 2.8
+ 0.3 + 0.4
0.10* 0.09 0.04 0.04 0.04 0.02 0.002 0.01 0.10 0.03 0.03 0.03 0.04 0.05 0.01 0.02 0.01 0.01
2.10 0.31 1.20 1.10 0.89 1.00 0.21 0.54 1.10 1.20 1.60 1.10 1.00 0.62 0.62 0.41 0.21 0.48
+ + + + + + + + + + + + + + 4+ + +
2.5 3.5
+ 0.2 + 0.4
0.05 0.11 0.08 0.03 0.04 0.54 0.02 0.05 0.07 0.14 0.09 0.05 0.09 0.04 0.02 0.05 0.09 0.01
* Standard error.
f\ 7
were nearly the same, but at 24 h the uptake of Ga in the abscess was much higher than that of i6” Yb (Tables I and II). At 4 h, except for the blood, the tissue distribution of -^*Yb was nearly complete, while with ^ G a the distribution process was still actively proceeding. At both 4 and 24 h kidney and bone uptakes were higher for 169уъ, but with all other tissues the uptakes of ^ G a were considerably higher than those of -^^Yb. Because of the basic differences observed in the tissue distributions of these two agents, a more appropriate method of assessing their relative worth in abscess detection is through a comparison of abscess-to-normal tissue concentration ratios (Table III). At 4 h the lung, muscle, bone marrow, blood and skin ratios were significantly higher with ^^®Yb, whereas with
582
HAYES et al.
TABLE II. COMPARISON OF THE 24-h TISSUE DISTRIBUTIONS OF YTTERBIUM-169 AND GALLIUM-67 IN RATS BEARING AN S. aureus ABSCESS
Sample_____________________ Yb-169____________________ Ga-67 % Administered Dose/g Viable abscess Necrotic abscess Liver Spleen Kidney Lung Skeletal muscle Heart Bone Bone marrow Blood Stomach Small intestine Cecum Colon Skin Fat Testis
1.50 0.23 0.86 0.81 0.91 0.18 0.03 0.09 3.30 0.51 0.02 0.61 0.40 0.39 0.27 0.05 0.01 0.06
+ + + + + + + + + + + + + + + + + +
Abscess weight (g) Activity in abscess (%)
2.5 3.6
+ 0.1 + 0.3
0.15* 0.08 0.06 0.05 0.04 0.01 0.002 0.001 0.09 0.02 0.001 0.03 0.03 0.02 0.08 0.01 0.001 0.003
4.20 0.20 2.30 2.30 0.95 0.47 0.18 0.17 1.40 1.50 0.28 1.30 1.20 0.75 0.68 0.20 0.04 0.32
+ + + + + + + + + + + + + + + + + +
2.1 5.4
+ 0.2 + 0.8
0.43 0.04 0.05 0.10 0.03 0.01 0.02 0.03 0.04 0.09 0.01 0.08 0.11 0.04 0.05 0.04 0.003 0.05
* Standard error.
6?Ga the ratios were higher for only kidney and bone. These results suggest that may be better than ^Ga for detecting inflammatory processes at early times after administration. At 24 h the two agents were approximately equal, with better ratios again being obtained for kidney and bone with ^ G a , while only the ratios for muscle and blood were superior for 16°Yb. A comparison of the tissue concentrations of ^ % b and ^ G a found in the gastrointestinal tract (both walls and contents) at 4 and 24 h indicated that no great differences existed between the two agents with respect to their relative uptakes in the experi mental abscess and the GI tract.
583
IAEA-SM-2 4 7 /9 4
TABLE III. ABSCESS-TO-NORMAL TISSUE CONCENTRATION RATIOS IN THE RAT AT 4 AND 24 h AFTER ADMINISTRATION OF YTTERBIUM-169 AND GALLI UM-67
R a tio o f A b scess C one, t o :
Y b -1 6 9 4 h
G a -6 7 4 h
Y b -1 6 9 24 h
G a -6 7 24 h
N e c ro tic a b s c e s s L iv e r S p le e n K id n e y Lung M u s c le Bone B o n e m a r ro w B lo o d S k in Fat
2 6 .0
11.0
10.0
2 7 .0
1 .7
1.8 1.8
1.8 2.0 1.2 3 .0 + 1 8 .0 +
0.6 2 .8 + 2 .5 * 1 1 . 0§ 3 0 .0
1 .7 1 .9 2 .4 *
2.1 10.0 2.0 * 1 .9 1 .3 5 .3 1 7 .0
1.8 1 .7 8 .4 5 8 .0 § 0 .5 3 .0 7 0 .0 * 3 3 .0 1 3 0 .0
4 .4 * 8 .9 2 6 .0 3 .1 * 2 .9 1 5 .0 2 3 .0
120.0
S t a t i s t i c a l l y s i g n i f i c a n t d if f e r e n c e w ith P e q u a l to : * , < 0 .0 0 1 ; t , 0 .0 0 1 - 0 .0 1 ; + , 0 .0 1 - 0 .0 2 ; §, 0 .0 2 - 0 .0 5 .
169 O ne d r a w b a c k t o t h e p o s s i b l e c l i n i c a l u s e o f Yb c i t r a t e i s t h e l o n g T ^/2 o f t h i s r a d i o n u c l i d e ( 3 2 d ) . E x tra p o la te d r a d i a t i o n d o s e s i n hum ans f o r anYb r a d i a t i o n d o s e s w i t h t h o s e f o r t h e n o r m a l m axim um d o s e o f ^Ga c i t r a t e ( 5 m C i ) , t h e ^ ^ Y b d o s e s f o r v a r i o u s t i s s u e s a r e b e tw e e n 1 a n d 2 tim e s th o s e o f ^ 7 g a e x c e p t f o r th e k id n e y an d b o n e ( 2 .6 an d 5 .5 tim e s t h a t o f ^ G a , re s p e c tiv e ly ). The b o n e d o s e e s t i m a t e i n t h i s s tu d y c o m p a re s w e l l w i t h t h e o n e m ad e b y A ndo e t a l [ 6 ] . A lth o u g h t h e d o s e i s r a t h e r h ig h f o r b o n e , t h e e s t i m a t e d ^°^Y b d o s e t o b o n e m a rro w w o u ld b e o n l y o n e - h a l f t h a t o f t h e d o s e f o r ^ G a .
D ISCUSSIO N 169 O ur r e s u l t s i n d i c a t e t h a t , a lth o u g h Yb i n t h e c i t r a t e f o r m d i d n o t h a v e a n a b s o l u t e u p t a k e ( % /g ) i n o u r e x p e r i m e n t a l
584
HAYES et al.
TABLE IV. COMPARISON OF RADIATION DOSES FROM YTTERBIUM-169 AND GALLIUM-67 CITRATES
Y b -1 6 9 * ( 0 . 5 m C i)
T is s u e
G a -6 7 + ( 5 . 0 m C i) ra d s
T o ta l body K id n e y Bone B one m a rro w L iv e r Lung S p le e n S to m a c h S m a ll i n t e s t i n e U pper la r g e i n t e s t i n e L ow er l a r g e i n t e s t i n e T e s tis O v a ry f H e a rt S k e l e t a l m u s c le
2.6
6.0 5 .5 1 .9 4 .5 4 .8 4 .0 4 .0
5 .5 0 .5
0.8 0.8 0 .9
1.6
3 .8 1 .4 1 .9 5 .5
3 .2
1.0
0.6
1.8
6.0
2.6 1.0
22% w i t h
+ G a -6 7 T o t a l b o d y r e t e n t i o n : 17% w i t h 83% w i t h 2 6 d b i o l o g i c T j y 2 S e p a ra te
1 .9
2.6
1 .3 1 .5 1 .5 1 .4 1 .7
Y b -1 6 9 T o t a l b o d y r e t e n t i o n : 78% w i t h 8 7 d b i o l o g i c T ^ /2 *
Í
ra tio 1 .3 1 .9 2 .5
5 .0 1 3 .5 3 .0 4 .5 1 .5 3 .9 7 .5 5 .0
Y b -1 6 9 G a -6 7
6 h b io lo g ic
T - ./? ;
30 h b i o l o g i c
T - i/2 ;
stu d y .
a b s c e s s a s h ig h a s t h a t o f ^ G a , n e v e r t h e l e s s w as c o m p e ti t i v e w i t h 6?G a o n a r e l a t i v e b a s i s ( r a t i o o f a b s c e s s c o n c e n t r a t i o n to t h a t in n o rm a l t i s s u e s ) . Y tte r b iu m - 1 6 9 h a s p h o to n e m is s io n s t h a t a r e q u ite s u i ta b l e f o r s c a n n in g . F u rth e rm o re , 1 6 9 Yb m ay w e l l b e s u p e r i o r t o ^ G a f o r d e t e c t i n g i n f l a m m a t o r y p r o c e s s e s a t e a r l y tim e p e r i o d s a f t e r a g e n t a d m i n i s t r a t i o n . 169 A p a rtic u la rly a ttr a c tiv e fe a tu re o f Yb i s t h a t i t c a n b e p ro d u c e d i n a r e a c t o r fro m e n r ic h e d 168 y b c o n s e q u e n tly w o u ld b e c o n s i d e r a b l y l e s s e x p e n s i v e t h a n a c c e l e r a t o r - p r o d u c e d 6 ^G a. F u r t h e r m o r e , t h e n ,Y c r o s s - s e c t i o n f o r ^ ® Y b i s 3 5 0 0 b a r n s w h i c h w o u ld r e s u l t i n s p e c i f i c a c t i v i t i e s i n t h e C i/m g Yb r a n g e . H ig h s p e c i f i c a c t i v i t y r a r e e a r t h r a d i o n u c l i d e p r e p a r a ti o n s a r e d e s i r a b l e , s i n c e o n ly m o d e st l e v e l s o f th e
585
IAEA-SM-2 4 7 /9 4
s t a b l e e le m e n ts a l t e r t h e i r t i s s u e d i s t r i b u t i o n s [ 2 ] . In th e p r e s e n t s t u d y we f o u n d t h a t w hen -^ ^ Y b w as a d m i n i s t e r e d a lo n g w ith 100 yg o f s t a b l e y tte r b iu m p e r k g o f b o d y w e ig h t th e u p t a k e o f -*-69уЪ i n l i v e r w a s t r i p l e d a n d t h e u p t a k e i n t h e s p l e e n , k i d n e y , lu n g an d b o n e m a rro w w e re a l s o i n c r e a s e d .
169
T he m a jo r d is a d v a n ta g e a s s o c i a t e d w ith th e u s e o f Yb f o r d e t e c t i n g i n f la m m a t o r y p r o c e s s e s i s i t s T ^ /2 ° f 32 d w h ic h n e c e s s a r ily l i m i t s th e d o se th a t can b e a d m in is te re d (s e e T a b le I V ) . T h i s i n t u r n w o u ld i n c r e a s e t h e s c a n n i n g t i m e r e q u i r e d f o r 1 6 9 у ь s t u d i e s o v e r t h a t n e e d e d w i t h 6 7 g a# Qn t h e o t h e r h a n d , H is a d a a n d h i s c o - w o r k e r s [7 ] i n a s tu d y o f t h e u s e o f 1 6 9 Yb c i t r a t e f o r tu m o r d e t e c t i o n i n m an h a v e e s t i m a t e d t h e 169уь r a d i a t i o n d o se to b e 1 .7 ra d to th e t o t a l b o d y an d 5 .8 r a d t o t h e b o n e , v a l u e s t h a t a r e d e c id e d ly lo w e r th a n th o s e we h a v e e s tim a te d fro m r o d e n t d a t a . On t h e p o s i t i v e s i d e , l ^ Y b ' s r e l a t i v e l y l o n g T ^ / 2 w o u ld n e c e s s a r i l y p r o v i d e f o r a l o n g s h e l f - l i f e f o r t h i s ra d io p h a rm a c e u tic a l.
REFERENCES [1]
S T A A B , E .V ., M c C A R T N E Y , W.?!., Role of gallium 67 in inflammatory disease, Semin. Nucl. Med. 8(1978) 219.
[2]
H A Y E S , R .L ., “ Factors affecting uptake of radioactive agents by tumour and other tissues” , Tumour Localization with Radioactive Agents (Proc. Advisory Group
[3]
Meeting Vienna, 1974), IA E A , Vienna (1976) 29. P O G G E N B U R G , J.K ., H A Y E S , R .L ., “ Radiopharmaceuticals for function studies, present and future” , Dynamic Studies with Radioisotopes in Medicine 1974 (Proc. Symp. Knoxville, 1974) 1, IA E A , Vienna (1975) 85.
[4]
G IB B S , W .D., H O D G E S, H.D., L U S H B A U G H , C.C., Precise geometry-independent radioassay of large biological samples, J. Nucl. Med. 9 (1968) 264.
[5]
L O E V IN G E R , R., B E R M A N , M., A revised schema for calculating the absorbed
dose
from biologically distributed radionuclides, M IR D No. 1, revised, New York, Soc. Nucl. Med. (March 1976). [6 ]
A N D O , A., H IR O F U M I, M., A N D O , I., H IR A K I, T., H IS A D A , K., Whole-body retention studies of
[7]
169
Yb-citrate, Radioisotopes (T o k yo ) 26 (1977) 602.
H IS A D A , K., T O N A M I, N., H IR A K I, T., A N D O , A., Tumor scanning with
169
Yb-citrate,
J. Nucl. Med. 15(1974) 210.
DISCUSSION D.A. GOODWIN: Was the gut excretion o f 169Yb less than 67Ga? It might be an advantage in scanning abdominal abscesses. R.L. HAYES: Yes, but so also was the abscess uptake so that, as a result, we saw no difference in the target : non-target ratios for either the walls of
586
HAYES et al.
the gasto-intestinal tract or for the tract content. We had, of course, hoped that we would find 169Yb to be superior to 67Ga in that respect. H.R. WAGNER, Jr.: Have you investigated the use of 169Yb for detecting neoplasms? R.L. HAYES: This question has been rather extensively studied by the Japanese1. Ytterbium-169 for that purpose was found wanting. They found that in general 67Ga was superior to 169Yb.
1
H IS A D A , K., A N D O , A., S U Z U K I, Y ., “ Radiolanthanides as tumour-localizing
agents” , Tumour Localization with Radioactive Agents (Proc. Advisory Group Meeting Vienna, 1974), IA E A , Vienna (1976) 113.
IA EA-SM -247/60
S P A L L A T IO N -P R O D U C E D T H U L IU M -1 6 7 F O R M E D IC A L A P P L IC A T IO N S G.-J. BEYER, R. MUENZE, W.D. FROMM Academy of Sciences of the German Democratic Republic, Rossendorf, Dresden W.G. FRANKE, H. HENKE Nuclear Medicine Department, Carl Gustav Carus Medical Academy, Dresden, German Democratic Republic V.A. KHALKIN, N.A. LEBEDEV Nuclear Problems Laboratory, Joint Institute for Nuclear Research, Dubna, USSR
Abstract
SPALLA TIO N-PRO DUCED T H U L IU M -167 F O R M EDICAL APPLICATIONS. The biokinetics and organ distribution of 155
Tb,
167
Tm,
172
139
Ce,
143
Pm,
144
Pm,
146
Eu,
147
Eu,
153
Tb,
Lu and 113L u were investigated in normal, tumour (virus-induced
mammary carcinoma, UV-induced U V T 14306 sarcoma, chemically induced D M B A 788 sarcoma), and aseptic inflammation-bearing male X V II/ B ln strain mice. The radiolanthanides were accumulated in all investigated tumours. A blood activity decrease of 0.006—0.01% of the injected dose was generally observed during 72 h. When the individual nuclides were compared, the biological characteristics and the decay properties were most favourable in the case of I67Tm with regard to tumour imaging. Satisfactory scintigrams were obtained already 4—6 h after administration. Thulium-167 showed tumour/blood ratios of 1.5-3.0 5 h and 20—30 24 h p.i. respectively. The tumour/muscle ratios (12.0±2.0 for the U V T 14306 sarcoma and 10.0±1.5 for the D M B A 788 tumour) remained constant between 5 and 50 h. Because of the higher skeletal uptake (about 40% of the injected dose), the tumour/bone ratios did not exceed a value of 1. Thulium-167 is not tumour specific. As in the case of 67Ga and u l In, a considerable uptake in inflammations was observed. The elimination of the
167
Tm-activity from
these lesions was slower than for tumours. Preliminary clinical studies showed satisfactory delineations o f human tumours in the cranial, thorax and abdominal region. Thulium-167 offers advantages in soft tissue tumour imaging, especially in the abdominal region.
587
588 1.
BEYER et al.
INTRODUCTION
According to previously published results [ 1] on the biodistribution and tumour uptake of radioactive lanthanides, thulium and ytterbium could succesfully compete with 67Ga in certain cases. Whereas the clinical importance of 169Yb-citrate has been evaluated recently [2—6], there is still not enough knowledge on the biokinetics and tumour affinity of the other lanthanides for a conclusive assessment. Thulium-167 is distinguished by favourable decay characteristics (half-life 9.25 days, no beta emission, 42% 208 keV gamma rays) for camera imaging. An accumulation of radioactive rare earth elements in animal tumours (Yoshida sarcoma) was reported by Ando and co-workers [7]. Yano and co-workers [8], Hisada and co-workers [9] and Hayes and co-workers [10] confirmed these results, extending the investigation to further tumours (mice hepatoma 5123 C). In addition to the cases of 67Ga and 111In, activity storage in the malignancies is strongly influenced by the presence of minor (microgram) amounts of stable nuclides of the radioisotopes investigated [10], resulting in higher liver uptake and diminished tumour accumulation. For radiolanthanides even the presence of weighable quantities of neighbouring elements leads to such unfavourable effects because of the close chemical relationship of these elements. For this reason, the production methods have to be critically evaluated in regard to stable rare earth elements (micrograms) and long-lived radioactive contaminants in the final product.
2.
PRODUCTION OF THULIUM-167
To meet these requirements, various production methods have been published. Chandra and co-workers [ 11 ] used holmium or enriched erbium oxide-167 targets irradiated with alpha particles (30 MeV) and protons (15 MeV) respectively. In the second case a yield of 2.8 MBq/microamp in one hour was achieved. The main drawback of these procedures is high 168Tm (half-life 93 days) contamination. Steinberg and co-workers [12] obtained about 10 MBq 167Tm in a one-hour bombardment using enriched erbium targets irradiated with deuterons. Yano and co-workers bombarded an HOCl3-target with 32 MeV alpha particles but gained only small activities (0 .5 -0 .7 MBq/microamp in an hour). An entirely different method was reported by Ando and co-workers [13] using a photonuclear reaction. Via the process I68Yb(7 , n) 167Yb »• 167Tm they obtained about 18 MBq 167Tm after a 10-h irradiation (60 MeV) followed by a chemical separation procedure lasting three days. The main problem in the methods discussed is to avoid contamination of the final product by stable lanthanides from the large amounts of the rare earth element targets.
589
IAEA-SM-247/60
Ce
Pm Eu
Tb
Tm Lu
58
61
65
69
63
71
FIG.l. Organ distribution o f carrier-free radiolanthanides in XVII/Bln strain mice bearing UVT14306 tumours 16 h p.i. (% injected dose/g tissue).
Scholz and Sodd [14] evaluated the possibilities of producing 167Tm by spallation o f non-lanthanide targets. The authors [15] have produced 167Tm together with the other investigated radiolanthanides via spallation of tantalum. The material (about 5 g o f tantalum metal) was irradiated at the internal proton beam o f the Dubna synchrocyclotron (Ep = 680 MeV; Ip = 2.3 microamp; 5 -1 0 h). The rare earth elements were separated from the target material by co-precipitation with extremely pure carrier lanthanum as fluorides from a hydrofluoric-nitric acid solution of the target. After washing and dissolving the precipitate in boric-hydrochloric acid mixtures, the separation of the individual radiolanthanides was carried out with small cation exchange columns filled with
590
BEYER et al.
Ce
Pm Eu
Tb
Tm Lu
FIG.2, Intercomparison o f the organ distribution o f radio lanthanides normalized for the 167Tm-accumulation rare earth element Z (% injected dose/g tissue)z /(% injected dose/g tissue)'6Tjm as a function o f the atomic number Z.
Dowex 50 X 8 resin in the ammonium form. Alpha hydroxyisobutyric acid was used as an elution agent. The fractions were purified by further chromatographic runs. The final solutions o f the radiochemically pure radiolanthanides, each in some drops o f solution, were evaporated to dryness and dissolved in two drops of 0.001N hydrochloric acid. By adding the calculated quantities of isotonic Na-citrate solution ( 2 - 5 ml), radiolanthanide solutions with radiochemical concentrations of 10—70 MBq/ml ready for injection were obtained. The 167Tm activity per batch was in the range 0 .4 -1 GBq containing merely negligible contaminations of 165Tm (half-life 30 h) and 168Tm (0.25%) two weeks after the end of the bombardment. This method therefore offered the possibility of attaining radioactivities of 167Tm sufficient for clinical application. The mixture for the multi-element biodistribution studies was reconstituted from suitable fractions of the several radiolanthanides and 67Ga, and met the requirements for satisfactory counting statistics in the samples.
591
IAEA-SM -247/60
T (h )
T ( h) 1
5 10
50
1■•■!■■■■!__I i I—1—L_
100
1
5 10
50
100
100
Г 10
z ш 5 x О ОС
z
г 0.1
> S
0.01
VT FIG.3. Gallium-67 and l61Tm uptake (% injected dose/g tissue) and relative enrichment [(% injected dose/g tissue)штТт/(% injected dose/g t i s s u e ца]for a virus-induced mammary carcinoma (XVII/Bln strain mice).
3.
ORGAN DISTRIBUTION STUDIES AND BIOKINETICS
Biodistribution and biokinetical studies were carried out with male XVII/Bln strain mice, some with experimental tumours and inflammations. The aim of the investigation was to evaluate the aptitude o f 167Tm in tumour detection compared with the other radiolanthanides and 67Ga as the currently most prominent tumour scanning agent. A total of 207 animals was investigated including 112 tumour-bearing (47 mammary carcinoma; 40 UVT sarcoma, 25 DMBA sarcoma), 45 with inflamma tions and 40 controls. The tumours used were: a virus-induced mammary tumour, histologically characterized as carcinoma medulläre partim adenosatum; an undifferentiated DMBA 788 sarcoma, chemically induced by painting the skin with dimethylbenzanthracene, and a spindle cell sarcoma UVT 14306 induced by UV irradiation. The inflammations were initiated by injection of 0.05 ml highly purified turpentine oil into the thigh. Preliminary studies indicated a high variability of the organ uptake in individual animals, probably because o f the varying size of the tumours compared with the animal weight (1 - 2 g tumour, 20 g total weight of mouse),
592
BEYER et al.
t (h) 1
5 10
T ( h) 50
100
FIG.4. Tumour/blood ratios (% injected dose/g tissue) for virus-induced mammary carcinoma.
1
5 10
50
100
FIG.5. Tumour, muscle and blood activities o f 167Tm in U VT14306 sarcoma as a function o f time.
especially for the longer time intervals. To minimize these effects and to achieve a reliable intercomparison o f the investigated nuclides in an individual organism we injected a mixture of representative radiolanthanides ( 139Ce, 143Pm, 144Pm, 146Eu, 147Eu, 153Tb, 15STb, 172Lu, 173Lu, and 167Tm) together with 67Ga intra venously. The activity uptake of the different organs and radioisotopes was analysed by means of high resolution Ge(Li)-gamma spectroscopy followed by computer processing (radioactive decay) of the results. In this way the number of animals necessary for statistically significant results could be drastically reduced. Finally, the biodistribution was studied by autoradiographic techniques. The first preliminary clinical studies were begun with clinically verified cases of non-Hodgkin lymphoma and colon cancer demonstrating the delineation of human malignancies as expected from the animal studies.
4.
RESULTS AND DISCUSSION
4.1. Animal experiments The organ distribution of the radiolanthanides approached an almost steady state during 10—20 h, and varied regularly with the atomic number in all organs investigated (Fig. 1).
593
IAEA-SM -247/60
15
10 0) D ф (A z 5 CT 's. «(0 о "О 5 24 72 h
5
24 72 h
1 2 3 4 5 6
1 2 3 4 5 6
FIG.6. Thulium-167 uptake in tumour (UVT 14306 sarcoma) and aseptic inflammations and retention ratios 24 h p.i. 1: inflamm,/blood 2: tumour/blood 3: inflamm./muscle 4: tumour/muscle 5: inflamm./liver 6: tumour/liver
In good agreement with results previously published by other authors [10], a high liver uptake of the lighter lanthanides was observed. The spleenic accumulation varied with the atomic number in a similar way. Both effects seemed to be caused by the well-known lower stability of citrate and protein complexes of these elements resulting in an increased production o f hydrolysed species in the serum. The excretion pattern is scarcely dependent on the atomic number and showed a nearly constant ratio between the urinary tract and the faeces. This behaviour suggests a common mechanism of the lanthanide biodistribution without any particular interaction of a single element of the series. The behaviour seems to be mainly determined by the small chemical variations caused by the regularly varying ionic radii. The distribution pattern of 67Ga as well as its kinetics cannot be explained in the framework o f the same hypothetical approach. According to its much smaller ionic radius (0.6 Â compared with 1.0—1.2 Â for the lanthanides), a higher tumour storage, but also a much higher skeletal uptake, would be expected. The experimental results indicate a higher tumour accumula tion but a lower bone uptake than for the radiolanthanides. An intercomparison of the lanthanide biodistribution (Fig. 2) indicated 167Tm as the most promising nuclide for tumour scanning in the lanthanide series. Further discussion is therefore focused on the evaluation of 167Tm in comparison with 67Ga.
5 94
BEYER et al.
FIG.7. Macro-autoradiogram: XVII¡Bln strain mouse with UVT14306 sarcoma.
FIG.8. Macro-autoradiogram: XVII/Bln strain mouse with aseptic turpentine oil inflammation in the right thigh.
The tumour uptake o f 67Ga is higher than 167Tm, at least immediately after administration (Fig. 3). However, a delayed and smaller decrease in radiothulium in the lesion led to similar concentrations of the two elements at the normal time o f 67Ga scanning (three days after administration). In contrast to 67Ga, the biokinetics of 167Tm reveal a good tumour delineation already four - six hours p.i. because of the faster blood clearance (Figs 4 and 5) and slower decrease in activity in the tumour. The tumour/blood ratios for 167Tm increased in the first hours to reach a value of 50 within a few days. Gallium-67 showed no significant change in this parameter during this period. The hepatic and gastro-intestinal activity concentrations are much higher in the case o f 67Ga, particularly within the first hours after administration. Thulium-167 was concentrated in aseptic inflammations even more than in the investigated tumours. Figure 6 demonstrates that the activity concentration in the inflammation was about twice as high compared with UVT 14306 sarcoma. Not only the uptake but also the time course o f the activity is different in both kinds o f lesions. Macro-autoradiograms gave first impressions of the morphological distribution o f the activity inside the organs and lesions. Figures 7 and 8 show the results for UVT 14306 sarcoma and a turpentine oil induced inflammation. As expected,
595
IAEA-SM -247/60
67 Ga
®7Ga
'67Tm
167 y m -
. . - ~=iï-Г' -
-
Й Р : -Э р !!: Se -
1Г-
;JtS« -iiiTf:* - "* ' -=î=»i' -"*. ---;*=-.* 7-•I■-
_hî; --=rÈ" ::=:UppUfin
~ Élil -s
FIG.9. Normal XVII/Bln strain mice with 67Ga- and 167Tm-citrate 4 h p.i.
K ï
FIG.10. Mammary-tumour-bearing mice with 67Ga- and 167Tm-citrate 4 h p.i.
the activity is concentrated in the skeletons, the liver and in the two lesions. Considerable activity was found in the mucosa o f the stomach, the intestine, and perhaps also in the salivary glands. In both lesions the radioactivity was non-uniformly localized. In the tumour only the viable tissue showed activity uptake, the inner necrotic parts being almost free. In the inflammation the spreading paths were delineated with high contrast. The scintigrams o f normal and tumour-bearing mice (Figs 9 and 10) confirm the possibility o f early tumour scanning using 167Tm.
5 96
BEYER et al.
FIG.11. Scintigram: lymph-node metastasis o f non-Hodgkin lymphoma delineated by 161Tmcitrate. The shadowed areas are regions o f higher activity.
4.2. Preliminary clinical studies First scintigraphic studies o f patients with malignancies were carried out to obtain information on the scanning o f human tumours with 167Tm as expected from the animal experiments. Figure 11 demonstrates the delineation o f affected lymph nodes in a case of non-Hodkin lymphoma; métastasés of a colon adenoma in the upper and abdominal region are seen in Fig. 12. Surprisingly, a radiation-induced pneumonitis was not detected.
5.
CONCLUSIONS
Thulium-167 has the properties of a good ‘tumour-localizing’ agent: high tumour uptake, in vitro and in vivo stability as well as a favourable decay characteristic for camera imaging. The production method described in this study offered the possibility o f obtaining activities sufficient for clinical application. In comparison with 67Ga, 167Tm possesses some advantages by
597
IAEA-SM -247/60
FIG.12. Scintigram: métastasés in the upper and abdominal part o f an adenoma o f the colon visualized by 167Tm-citrate.
reason o f its different biokinetics permitting earlier tumour scanning and a better detection in the abdomen. The more pronounced osteotropic properties o f 167Tm can be considered a drawback but, on the other hand, bone lesions distinguished by an increased metabolic activity will be found simultaneously. The mechanism of the tumour uptake and localization is not well understood. It is not clear whether the radiolanthanides follow, at least partially, similar reactions and pathways as gallium. Further studies, particularly on possibilities of influencing the distribution and kinetics, might be useful.
REFERENCES [1]
H IS A D A , K ., S U Z U K I, Y ., et al., Clinical evaluation of tumor scanning with Radiology 116(1975) 389.
169
Yb-citrate,
598
BEYER et al.
[2]
H IG A S H I, T., F U JIM U R A , T., et al., On the accumulation of I 6 9 Yb-citrate in malignant
[3 ]
K A N O , I., O n the accumulation o f Sosho 9 6 ( 1 974) 49.
[4]
C H A T A L , J.F ., L E M E V E L , B.P., et al., 169Ytterbium citrate: interest in cancerology,
[5]
A N D O , A., H IR A K I, T., et al., Relationship between strains of tumor-bearing animals
tumors, Kaku Igaku 10(1973) 27. 169
Yb-citrate in malignant tumor, Kanazawa Irigaku
Radiol. Electrol. Med. Nucl. 56 (1975) 401. and the tumor affinity of 169Y b and [6 ]
67
G a, Radioisotopes (T o k yo ) 24 (1975) 104.
K U E H N E , A., F R A N K E , W .G., et al., 169Y b citrate for tumour scintigraphy — first experiences, Radiobiol. Radiother. 17 (1976) 111.
[7]
A N D O , A., H IS A D A , K ., et al., A ffin ity of 161Tm for malignant tumor and bone, Radioisotopes (T o k yo ) 24 (1975) 109.
[ 8 ] Y A N O , Y ., C H U , P., Cyclotron-produced thulium-167 for bone and tumor scanning, Int. J. Nucl. Med. Biol. 2 (1 9 7 5 ) 135. [9]
H IS A D A , K ., A N D O , A., Radiolanthanides as promising tumor scanning agents, J. Nucl. Med. 1 4(1973)615.
[10]
H A Y E S , R .L ., B R O W N , D.H., et al., Comparison of the subcellular distribution of the tumor-localizing agents
[11]
67
Ga, l u In,
206
Bi, and
167
Tm, J. Nucl. Med. 15 ( 1974) 501.
C H A N D R A , R., H E R N B E R G , J., et al., I 6 7 Tm: A new bone scanning agent, Radiology 100 (1971) 687.
[12]
S T E IN B E R G , M., R A S M U S S E N , J.W ., E N N O W , K ., R O Y - P O U L S E N , N.O., V O E T M A N N , V ., P O U L S E N , B „ A N D E R S E N , M.-L., “
167
Tm-citrate for bone imaging” ,
Radiopharmaceuticals and Labelled Compounds (Proc. Symp. Copenhagen, 1973) 2, IA E A , Vienna (1973) 151. [13] A N D O , A., H IS A D A , K., et al., A ffin ity o f 167Tm for malignant tumor and bone, Radioisotopes (T o k yo ) 24 (1975) 41. [ 14] SC H O L Z , K .L ., SO D D , V .J., Production of thulium-167 for medical use by irradiation o f lutetium, hafnium, taiitalum and tungsten with 590-MeV protons, Int. J. Appl. Radiat. Isot. 27 (1976) 263. [15] B E Y E R , G. J., F R A N K E , W .G., et al., Comparative kinetic studies o f simultaneously injected
167
Tm- and
67
Ga-citrate in normal and tumour bearing mice, Int. J . Appl.
Radiat. Isot. 29(1978) 673.
DISCUSSION J.P. OLIVA GONZÁLEZ: Have you carried out studies to compare the uptake for malignant and benign tumours? R. MUENZE: No, not yet. The paper gives the first preliminary results with patients. We intend to continue the investigations and will certainly look into that aspect, too.
IA EA -SM -247/113
S T U D Y O F T H E D O S E A N D D IS T R IB U T IO N O F T E C H N E T IU M -9 9 m -M D P IN V A R IO U S B O N E S A N D T IS S U E S I. OTHMAN*, N.M. SPYROU, R.A.A. KHAN+ Medical and Environmental Physics Group, Department o f Physics, University of Surrey, Guildford, United Kingdom
Abstract STUDY OF THE DOSE AND DISTRIBUTION OF TECHNETIUM-99m-MDP IN VARIOUS BONES AND TISSUES. The variation with time of the percentage uptake and retention of injected activity per gram of various bones and organs of rat were measured as part of a series of experiments designed to study the distribution and localization of the wTcm-MDP bone imaging complex. These values were used retrospectively to estimate the radiation dose absorbed in the human body. Although the results obtained were in fairly good agreement with those reported in the literature, those for bone were significantly higher, and therefore the calculation for the latter was repeated this time making use of in vivo dog data where the radiopharmaceutical was directly injected into the nutrient artery to the tibia. The absorbed bone radiation dose from the canine data was found to be in very good agreement with the value previously calculated from rat data. It was interesting to observe that percentage uptake and retention of injected activity per gram of various rat bones at two hours correlates with the inverse of the percentage of trabecular mineralization of the bone. The usual difficulties encountered in calculating blood dose values are also expressed.
INTRODUCTION The use of the " T c m-Sn complex of methylene diphosphonate, first proposed as a skeletal-imaging agent in 1973 by Subramanian and co-workers [1] and shown to offer significant advantages when compared with other technetium-labelled bone-seeking complexes, as reported by the same group in 1975 [2], has now been widely adopted in bone-scanning investigations, and has become a routine clinical procedure for screening purposes and in a number of applications as, for example, in orthopaedics. However, it is also known that the * On leave from Nuclear Medicine Centre, Damascus, Syria. Present address: Nuclear Medicine Department, Medical Branch, University of Texas, Galveston, Texas, United States of America.
599
600
TABLE I. COMPARISON OF TECHNETIUM-99m-MDP UPTAKE IN VARIOUS RAT BONES (% INJECTED ACTIVITY/g) Rat number
1
2
3
4
5
Mean ± S.D.
Tibia
2.75
2 .8 8
2.87
1.99
2.05
2.51 ±0.40
Femur
1.78
2.40
2.45
2.15
2.03
2.16 ± 0.38
3.30
2.95
2.53
2.73
2.79 ± 0.35
2.35
2.58
2.36
1.84
1.87
2 . 2 0
Pelvis
2.15
2 .2 0
2.48'
1.78
2.24
2.17 ± 0.26
Vertebra
1.63
2 .0 2
1.89
1.79
1.96
1.86 ±0.30
Rib
3.91
4.08
3.91
3.85
3.78
3.91 ± 0.29
Scapula
3.07
3.58
3.85
3.22
3.69
3.48 ±0.17
Clavicle
-4.43
5.14
5.43
4.20
5,50
4.94 ± 0.52
± 0.28
Sternum
1.31
1.72
1.58
1.76
1.51
1.58 ±0.14
Skull
1.92
2 .2 2
1.80
2.15
2.29
2.08 ± 0
. 2 0
OTHMAN et al.
. 2.46
Radio-ulna
Humerus
IAEA-SM -247/113
601
mechanism of uptake of phosphate compounds in bone is not clearly defined, and in late 1974 a collaboration between the University of Surrey, Hammersmith Hospital and the Royal Postgraduate Medical School was set up to study this. It is an on-going project which more recently has also involved the Department of Surgery, Princess Margaret Rose Orthopaedic Hospital, Edinburgh. During the course of the study experimental data became available which it was suggested may also be useful in the calculation of radiation dose. Since a quick survey of the literature indicated that little had been published with respect to dose when " T c m-MDP is administered, apart, from the pioneering studies and the figures quoted by manufacturers of the product, we decided to extract the information necessary for the calculation from the first stages of the research project, reported in a doctoral thesis [3]. Although some theses are published, they do not generally attract a large reading public, and we are fairly certain that in this case the experimental data therein have not been used before for the calculation of radiation dose which forms the basis of this paper.
EXPERIMENTAL DATA The data were obtained from a set of experiments where a group of 48 rats were used; measurements were carried out to study the distribution with time of " T cm-MDP in the organs and the bones of the animals, as well as to test the effects o f dilution, pH value and length o f storage of the preparation. All studies were carried out at two hours post injection except where the optimum period for performing a bone scan was being determined. Table I shows the percentage of the injected activity per gram in various bones o f rats two hours after administration. Each rat was anaesthetized with Nembutal (30 mg/kg) intraperitoneally and a dose o f 4 0 —60 juCi o f the "T cm labelled agent administered intravenously into a tail vein.1 The technique for preparation of " T cm-MDP has been described previously [4]. Rats were killed, as mentioned above, at 2 h post injection and also at 15 min, 1 ,2 ,4 and 6 h respectively. The following tissues were removed, cleaned, weighed and counted in a well scintillation counter: blood, liver, kidney, skeletal muscle, spleen, thyroid, femur, tibia, humerus, radius, pelvis, rib, vertebra, scapula, clavicle, sternum, skull and bone marrow. All samples were counted for " T cm activity with reference to a specially prepared 2-ml standard and some considerable effort went into ensuring that geometrical variations were taken into account. It must also be pointed out that wet bones were used in the radioassay since earlier experiments had established that no significant differences were found between wet and dried bones; this made the
o\ о го
TABLE II. TECHNETIUM-99m-MDP UPTAKE (% INJECTED ACTIVITY/g) IN RATS KILLED AT VARIOUS TIMES AFTER INTRAVENOUS INJECTION (mean o f six rats ± S.D.) Tissue
15 min
1
h
2
4h
h
6
h
1.39 ±0.30
2.02 ± 0.32
2.79 ± 0.27
2.19 ± 0.60
1.70 ± 0.31
Blood Muscle
0.46 ±0.14 0.09 ±0.01
0 .0 2
0.09 ± 0.02 ± 0 .0 0 1
0 .0 1
0.04 ± 0.01 ± 0 .0 0 1
0 .0 1
0.02 ± 0.004 ± 0 .0 0 1
0 .0 1
Liver
0.12 ± 0.04
0.05 ± 0.02
0.03 ± 0.01
0 .0 2
Spleen
0.12 ±0.03
0.04 ± 0.01
0.05 ± 0.03
0.03 ± 0.01
0.03 ± 0.01
Kidney
4.65 ±2.61
0.47 ± 0.01
0 .5 8
± 0.13
0.42 ± 0.20
0.34 ± 0.03
Thyroid
0.25 ±0.06
0.05 ± 0.01
0.06 ± 0.03
±
0.04 ± 0.001
0 .0 2
±
0 .0 1
0 .0 0 1
0 .0 2
± ±
0 .0 1 0 .0 0 1
0.03 ± 0.01
Femur/tissue concentration ratios (g/g) Femur/blood
3.0
± 1.1
22.4 ± 6.7
69.4 ± 21.2
110.0 ± 39.7
85.5 ± 32.0
Femur/muscle
15.4
± 3.7
101.0 ± 16.4
229.0 ±27.5
219.0 ±66.7
170.0 ±65.4
OTHMAN et al.
Femur
IAEA-SM-247 /1 1 3
603
10
■ femur
• spleen
о kidney
D thyroid
• blood
о
v
muscle
a
liver
1
0.5
0.1 □ 0. 05
0.01 0. 005
1
2
3
4
5
6
7
8 h
FIG.l. Distribution o f" T c m-MDP activity with time in various tissues.
604
OTHMAN et al.
process simpler and much more rapid as well as providing greater statistical counting precision. Table II shows the distribution in various tissues of " T cm-MDP expressed as percentage of injected activity per gram o f tissue with time, and Fig. 1 is a graphical representation of these values. Although it is always dangerous (i.e. inaccurate) to determine half-lives from incomplete data, since the requirement for these measurements is for data points to extend beyond several times the value of the half-life of interest, the temptation is there, and we have used both graphical and analytical methods to determine the approximate half-life of the longer-lived compartment(s) o f each of the tissues. The values in hours found for femur, kidney, blood, liver, spleen and thyroid are: 11.3, 11.0, 4.0, 7.8, 9.6 and 6.3 respectively, with fitting errors in the region of ± 20%. However it must be emphasized that these values were not used, as explained later, for determining the radiation dose.
RADIATION DOSE CALCULATIONS AND RESULTS The calculation of absorbed dose in various organs and the estimate of total body dose following the administration of a radioactively labelled bone-seeking agent becomes particularly important, as with all radiation procedures, when a test is carried out for screening purposes, in the assessment of risk/benefit factors. The absorbed dose for biologically distributed radionuclides can be calculated either by the classical equation using the exposure rate and the appropriate geo metrical factors, or by the present formalism recommended by MIRD [5] using the equilibrium dose and the specific absorbed fraction. The estimation of the absorbed dose in humans for a large number of radioactive isotopes can now be carried out with the latest of the software packages, CAMIRD/III [6]. For these calculations the input data assumed as kno wn include the fractions and energies in the radionuclide disintegration schemes, the model, build-up factors, absorbed fraction and specific absorbed fractions between source and target, the last being listed in MIRD pamphlet No. 11 [7] as an absorbed dose per unit cumulative activity for selected radionuclides and organs. In the situation where only limited biological disappearance data are available for dose calculation, as in our case, the cumulative concentration of activity can be obtained by numerical integration o f the area under the time distribution curve. The formula used to estimate the radiation dose due to self-absorption is
D (v*-v) = Cv 2 i
Ai0i(v
v) rad
(1)
IA EA-SM -247/113
605
where Cv is the cumulative concentration o f activity in volume v and has units of /uCi-h/g. This cumulative concentration was obtained in parts, first by succes sively summing the areas of trapezia between the first and last experimental data points, then adding the contribution made from time zero to the first datum point and that from the last point to infinity. The cumulative concentration between time zero and the first datum point was calculated by assuming that either the activity at time zero was zero when there is no uptake beyond the first point or an activity equivalent to the first datum point when continued uptake is exhibited by the second datum point. After the last experimental point recorded, it was assumed that elimination of the activity was due only to physical decay of the radioactive isotope and the contribution was summed to infinity. Hence the final form of the cumulative concentration is MCi—h (rj + 1 -T ¡)+ 1.44CnT1/2-
(2)
where r ^ 2 is the half-life of " T cm. In Eq.( 1) 2 is the equilibrium absorbed dose o f radiation of type i i = 1 ,2 ,..., each with a fractional frequency n¡ per disintegration and a mean energy Ej in MeV; when units are taken into consideration g-rad 4 1 = 2 1 3 "‘E‘ ^
(3)
and ф[ is the absorbed fraction of the energy in the source region. The value was estimated for both bone and soft tissue by assuming that the activity is homogeneously distributed in the organ and the unit volume of the organ is replaced by a point source in the centre of that volume. The mass energy coefficient for both bone and soft tissue was calculated using Hubble’s data [8]. For the purpose of these calculations all soft tissue was taken as water whereas the percentage composition of bone was assumed to be similar to that given by Woodward [9]: Щ 0.0339), C(0.155), N(0.0397), Na(0.0006), Mg(0.0021), P(0.102), S(0.0031) and Ca(0.222). For the radiation dose from organ to organ the general equation of average absorbed dose was used, i.e.
D (v ■«- r) = — 2 mv Î
Д;0; (v •*- r) rad
(4)
606
TABLE III. NUCLEAR DATA FOR TECHNETIUM-99m USED IN CALCULATIONS Mean No. of photons/ disintegration corrected for internal conversion
Mean photon energy Ej (MeV)
i.c. Electrons of 7 2
-
-
0.0109
0.0232
i.c. Electrons of 7 3
—
—
0.0013
0.00277
?i
0.986
0 .0 0 2
0.00197
0.00420
72
0.90
0.140
0.126
0.268
7з
Type of radiation
niEi (MeV/disintegration)
Ai g-rad mCi-h
0.142
0.00057
Ke , X-ray
0.0455
0.0183
0.00083
0.00177
Ka 2 X'ray Kß X-ray
0 .0 2 2 1
0.0182
0.0004
0.0008
0.0103
0.0206
0 .0 0 0 2 1
0.00045
OTHMAN et al.
0.004
0 .0 0 1 2 2
IA EA -SM -247/113
607
where mv is the mass of the target volume v in grams, and Ar is the cumulated activity in the source; 0¡ as before is the absorbed fraction for the ith type of radiation, calculated by a point isotropic source in the unit volume, etc. In MIRD pamphlet No. 11 [7], the quantity 2 (v «- r) is referred to as the i ‘S’ value; where Ф^= Ф^/ту is the specific absorbed fraction for the target organ of volume v, for the ith radiation emitted from source organ r. ‘S’ values for the photons have been estimated using Monte Carlo techniques as described in MIRD pamphlet No. 5 [10], and are the ones we have used whenever possible. By making use o f the nuclear data available for the decay of " T cm [11], shown in Table III, and assuming that the emitted internal conversion electrons competing with the radiative transitions from the energy levels 0.140 MeV and 0.142 MeV are considered as non-penetrating radiations as are the 2 keV photons emitted in the decay process, it was possible to calculate from the experimental rat data presented in Table II and Fig.l the absorbed dose in rad/mCi o f injected activity in man. Although extrapolation of animal data to man can be questioned, we have adopted one o f the prescribed systematic procedures and used the following conversion formula:
where mr , mh is the mass o f the corresponding organs in rat and man and Mr , Mh is the mass of the whole body o f rat and man respectively. In our calculations we have used ICRP 23 [12] for the tissue mass data o f 70 kg man. Table IV lists these results in some detail. Note that although the original experimental rat data did not contain any direct information for the bladder compartment, the radiation dose to the bladder walls was calculated by assuming that the amount o f injected activity which did not reside in the meas ured tissues would accumulate in the bladder over the first four hours, after which complete voiding would occur. Thereafter it was assumed that activity in the bladder will have a biological half-life equal to the physical half-life o f the isotope, and the additional cumulated activity was calculated from that period onwards to infinity. Since no rat data were available for the uptake of activity in the testes, the calculation o f dose is based only on the testes as targets. Note also that the bone dose results are based on the cumulated concentration obtained for the femur.
DISCUSSION AND CONCLUSION Table V summarizes the total absorbed dose for various organs and the total body from this work, and includes values found by other workers for
608
TABLE IV. CALCULATED ABSORBED RADIATION DOSE IN RAD PER juCi OF INJECTED ACTIVITY FOR 70 kg MAN Target
Source Muscle
Liver
Spleen
Kidney
Thyroid
Bone
6.8 X 10'5
2.9 X 10"7
4.7 X 10'8
4.2 X 10'9
1.6 X 10’ 7
Muscle
9.4 XI O'6
7.9 XIO'7
4.7 XI O'8
5.4 XIO'9
1.5 X IO’ 7
Liver
6.4 X 1 0 -6
7 3.2 X 10‘
2.0 X Í0"6
3.7 XIO'9
Spleen
5.6 XI O'6
4.1 XIO'7
3.9 XI O'8
1.3 X 10'6
Kidney .
7.9 X 10'6
3.8 XIO'7
1.7 XIO'7
3.5 X10'8
Thyroid
7.6 XI O'6
3.8 XIO'7
6.5 X 10'9
3.3 XIO'10
Bladder
Total body
3.8 X 10'10
2.4 X 10"6
4.5 XIO'10
4.9 XIO'10
4.7 XIO"6
3.4 X 10'10
4.5 X 10“ 7
3.5 XIO'11
4.4 XIO'7
4.0 X 10'10
9.9 XIO’ 7
4.2 XIO'11
1.7 XIO'6
4.0 XIO'10
2.2 XIO’ 6
1.3 XIO'11
7.0 XIO'7
4.0 XIO'10
5.5 XIO*9
8.8 XIO'6
5.4 X IO’ 9
2.7 XIO"10
3.2 XIO'8
7.6 XIO'13
4.2 X IO'4
4.2 XIO'10
1.9 XIO'13
2.21 X lO-6
3.1 XIO'10
4.9 XIO'6
3.6 XIO'10
Bladder
OXIO"6
5.2 XIO'7
7.3 XI O'9
4.6 XI O'10
Testes
6.2 X10‘ 6
3.2 X 10‘ 7
3.8 XI O'9
1.8 XIO’ 10
1.0 XIO'8
Total body
1.9 Х1СГ7
5.5 XIO'7
9.5 XIO'8
8.4 XIO’ 9
2.5 XIO'7 . 6.8 XIO'10
OTHMANetal.
Bone
609
IAE A-SM-24 7 / 113
TABLE V. TOTAL ABSORBED DOSE IN rad/mCia OF INJECTED TECHNETIUM-99m-MDP Other
This work
Organ
Ref.[2]
Ref. [ 13]
0.068
0.07
0.038
Muscle
0.015
0.009b 0.008
Liver
0 .0 1 0
Spleen
0 .0 1 0
Kidney
0.031
Thyroid
0.016
Bladder
0.42
Testes / Total body
0.18
О О
Bone
0.025
0.007
0.031
0.229 о
-J
0.44
(0.030)c 0.016
a M ultiply by 2.707 X 102 to convert to AiGy/MBq. k Average soft tissue. c
Ovaries.
TABLE VI. THE CONTRIBUTION TO THE ABSORBED DOSE FROM BLOOD (rad/mCi o f injected 99Tcm -MDP)
Dose .........
Skeleton
Liver
Spleen
Kidney
Thyroid
7.5 X 10"s
3 .5 X 1 0 ’ 4
9 X 1 0 "4
4X104
2 .5 X 1 0 "4
comparison. It is gratifying to see that the values quoted for kidney, bladder and liver for radiation dose estimates in both this work and that by Subramanian and co-workers [2] are in such good agreement; however it is disturbing to note the significant differences in the bone dose estimate in the two values quoted. It has been pointed out above that our calculation for absorbed dose in the bone was based on the cumulated concentration in the femur and this may have intro duced an additional error. Examination o f Table I suggests that the uptake of injected activity per gram, at two hours post-injection, is more or less equal for
610
OTHMAN et al.
the bones studied except for the rib, scapula and clavicle where a significant higher mean uptake per unit mass is recorded. It can also be seen that a somewhat lower uptake than the average occurs for the vertebra and sternum. However, when we compared the values for the mean uptake of various bones with the percentage of the total mineralized bone mass which was o f trabecular nature (e g. in the vertebral column ~ 73% of mineralized bone is trabecular, in the femur ~ 33%, in the tibia ~ 26%, in the sacrum ~ 25%, in the radius ~ 16%, in the ulna ~ 13% and in the chest cage, including clavicle, sternum and ribs, it is only 6% [12]), there seemed to be a correlation between a high uptake and a low percentage of trabecular mineralized bone with the sternum being the obvious exception. The average figure of trabecular mineralization quoted for the chest cage may, however, be an underestimate o f the degree of trabecular mineralization in the sternum, as the ratio of trabecular-to-cortical bone thicknesses in the sternum is higher than that for the ribs, and thus may bring the former more into line with this simple-minded explanation. From the experimental data in Table I a mean uptake, normalized by weight, was calculated for the various rat bones and found to be 2.40 ± 0.29% of the injected activity per gram o f bone, at two hours after injection. This compares fairly well with the value at two hours of 2.79 ± 0.27% o f injected activity per gram o f femur (see Table II) which was used in the calculation of absorbed skeleton dose, but suggests that the value of the latter is an over estimation, especially when one notes that the mean uptake at two hours in the femur, from the rat data in Table I, is 2.16 ± 0.38% o f injected activity per gram. Nevertheless, it must be noted that cumulated concentration for the bone (i.e. the area of the activity distribution per unit mass versus time curve) will not alter significantly since errors associated with these calculations are about ± 20%, and the skeletal dose remains significantly higher than that quoted in Ref.[2], although it is closer to the value given in Ref.[ 13]. Tо ascertain whether we overestimated the value o f the radiation dose to the skeleton we also made use of experimental data from a completely different set o f investigations on dogs. The results have shown that the transcapillary exchange of phosphate compounds in the canine tibia occurs by passive diffusion [14] and suggest that cortical bone consists of four definable compartments through which minerals pass from the capillaries to the bone crystal [3, 15, 16], as well as indicating that after osteotomy in the canine tibia the residue of " T cm-MDP in bone increased at two weeks [17]. In all these cases adult greyhounds were used and surgical procedure ensured that injection of " T cm-MDP was selectively into the nutrient artery o f the tibia, without loss of tracer to the surrounding tissues. In one set of experiments the change in radioactivity over a given region of bone following a bolus injection was measured externally, using a Nal(Tl) detector, as a function of time for a period up to six hours; typical results were:
611
IAEA-SM -247/113
Component
2 3 4
Half-life
Size of compartment (arbitrary units, %)
12 s
11 000(31.5)
2.1 min
15 400(43.5)
13
min
9.2 h
2 500(7) 6 350(18)
From the above, the absorbed dose in the human skeleton, this time using the canine tibia as the representative bone, was calculated and found to be 0.085 rad/mCi o f injected activity, which is in good agreement with our previous calculation using the rat experimental data. In addition to the experiments described above, investigations at the microstructure level o f the localization o f "T cm-MDP have been undertaken using microautoradiographic [18] and electron microscopic [19] techniques, but these require further study with respect to dose calculations. Finally, to conclude with the most uncertain estimation, that pertaining to the absorbed dose in blood and its contribution to the dose o f other target organs is shown in Table VI. The values quoted in the table were obtained by assuming that the blood in each organ acted as a uniformly distributed source of activity within each organ and therefore the dose was calculated by accepting the listed ‘S’ value for self-absorption in that organ. As for the dose absorbed in the blood itself, we felt unable to decide how the calculation could be performed and here we can only offer an indication o f extreme limits; in the first case where the dose is considered to be due to an isotropic point source in the centre of the total volume of blood, and in the second case where the specific activity resides in a unit spherical volume of 1 cm3 of blood and the total dose absorbed is a summation of the dose for each individual volume. The two values were found to be 9.0 X 10"4 rad/mCi and 5.0 rad/mCi respectively, and cover almost four orders of magnitude. A clearer understanding o f the mechanism of uptake o f phosphate com pounds labelled with technetium has been found necessary for diagnostic proce dures and quantitation studies; that it is also important for more accurate dose calculations is to us abundantly clear.
ACKNOWLEDGEMENTS The authors wish to thank K. Kouris o f this Department for useful discussions.
612
OTHMAN et al.
REFERENCES [1] S U B R A M A N IA N , G., B L A IR , R .J., K A L L F E L Z , F.A ., et al.,
99
T c m-MDP (methylene
diphosphonate): A superior agent for skeletal imaging, J. Nucl. Med. 14 (1973) 640. [2] S U B R A M A N IA N , G., M c A F E E , J.G ., B L A IR , R .J., K A L L F E L Z , F.A ., T H O M A S , F.D ., Technetium-99m-methylene diphosphonate, a superior agent for skeletal imaging: Comparison w ith other technetium complexes, J. Nucl. Med. 16 (1975) 744. [3] K H A N , R .A .A ., Localization o f Technetium-labelled Diphosphonate Preparations in Canine Bone, PhD thesis, University of Surrey (1978). [4]
K H A N , R .A .A ., S H E A S B Y , B ., Preparation o f a stannous methylene diphosphonate kit and its biological distribution in animals, J. Clin. Pharmacol. New Drugs 7 (1976) 163.
[5]
L O E V IN G E R , R ., B E R M A N , M., A Revised Schema for calculating the Absorbed Dose from Biologically Distributed Radionuclides, M IR D Pamphlet N o .l (Revised), Society of Nuclear Medicine, New York (1976).
[ 6 ] B E L L IN A , C.R., G A Z Z A R D I, R ., C A M IR D /Ш : A revised version of the C A M IR D / II and MIRD-S packages for internal dose calculation, J. Nucl. Med. 21 (1980) 379 (Concise communication). [7]
S N Y D E R , W.S., F O R D , M .R., W A R N E R , G.G ., W A T SO N , S.B., “ S ” Absorbed Dose per Unit Cumulated A ctivity for Selected Radionuclides and Organs, M IR D pamphlet No. 11, Society of Nuclear Medicine, New Y o rk (1975).
[ 8 ] H U B B L E , J.H ., V E I G E L E , V n.J., B R IG G S , E.A ., B R O W N , R .T., C R O M E R , D.T., H O W E R T O N , R .I., Photon cross sections attenuation coefficients and energy absorption coefficients from 10 keV to 100 keV, J. Phys. Chem. Ref. Data 41 (1975) 471. [9]
W O O D W A R D , H.Q., The elementary composition of human cortical bone, Health Phys. 8 (1962) 513.
[10] S N Y D E R , W.S., F O R D , M .R., W A R N E R , G.G., et al., Estimates of Absorbed Dose Fractions for Monoenergetic Photons Sources Uniform ly Distributed in Various Organs [11]
of Heterogeneous Phantom, M IR D pamphlet No. 5, J. Nucl. Med. 10 (1969) Suppl. No.3. L E D E R E R , C.M., H O L L A N D E R , J.M ., P E R L M A N , I., Tables of Isotopes, 6 th edn., John W iley, New Y ork (1968).
[12] IN T E R N A T IO N A L C O M M IS SIO N ON R A D IO L O G IC A L P R O T E C T IO N , Report of the Task Group on Reference Man, IC R P Publication No. 23, Pergamon Press, Oxford (1975). [13] G R A H A M , L.S., K R IS H N A M U R T H Y , G.T., B L A N D , B.H ., “ Dosimetry of skeletalseeking radiopharmaceuticals” , Proc. 21st Annual Meeting of the Society of Nuclear Medicine, J . Nucl. Med. 15 (1974) 496 (Abstract). [14] H U G H E S , S.P.F., D A V IE S , D .R., B A S S IN G T H W A IG H T E , J.B ., K N O X , F.G ., K E L L Y , P .J., Bone extraction and blood clearance of diphosphonate in the dog, Am. J. Physiol. 232(1977) H341. [15] H U G H E S , S.P.F., D A V IE S , D .R., K H A N , R .A .A ., K E L L Y , P.J., Fluid space in bone, Clin. Orthop. Relat. Res. 138 (1978) 332. [16] K H A N , R .A .A ., H U G H E S , S.J., L A V E N D E R , J.P ., S P Y R O U , N.M., “ Does the bone scan represent bone blood flow ? ” , 2nd Congress in Nuclear Medicine, the European Society of Nuclear Medicine, London (1978) (Abstract). [17] H U G H E S , S.P.F., K H A N , R .A .A ., D A V IE S , D .R., L A V E N D E R , J.P., The uptake by the canine tibia of the bone scanning agent
99
T c m-MDP before and after osteotomy,
J. Bone Jo in t Surg. 60-B (1978) 579. [18] L E O N , M., M Sc dissertation, University of Surrey, 1977. [19] M O L F E T A S , M., M Sc dissertation, University of Surrey, 1978.
IAEA-SM-2 4 7 /1 11
S T R U C T U R E -A C T IV IT Y R E L A T IO N S H IP O F T E C H N E T IU M -9 9 m - A N D I N D I U M -1 1 1 L A B E L L E D N -D I P H O S P H O N A T E S P. ANDREOU, E. CHIOTELLIS, A. VARVARIGOU Radiopharmaceuticals Laboratory, Nuclear Research Centre Demokritos, Athens C. KOUTOULIDIS 401 General Military Hospital, Athens, Greece
Abstract S T R U C T U R E - A C T IV IT Y R E L A T IO N S H IP O F TECHNETIUM -99m - A N D IN D IU M -1 11 - L A B E L L E D N -D IPH O SPH O N A T ES. The pharmaceutical preparation and analysis as well as the biodistribution in experimental animals of various N-diphosphonates, labelled with
99
T cm and
111
In, are
described. The biological distribution of these complexes was studied in mice comparing the bone scanning radiopharmaceuticals, " T c m-IDP and l u In-EDTM P. The derivatives labelled with
99
T cm showed low bone concentration compared with
99
T c m-IDP. The
indium complexes presented higher bone values of activity, among which m In-EDDMP exhibited quite satisfactory bone concentration, comparable with that of l n In-EDTM P. The molecule of E D D M P carried only two phosphonic groups, offering thus free sites for substitution and may be evaluated as basic structure for the production of indium radiopharmaceuticals.
INTRODUCTION The technetium-99m-labelled phosphorus compounds, currently used in bone scintigraphy, can be classified in three main categories [1 ], compounds o f the types: P- О —P, P—С—P and P—N—P. Studies were performed on the effects of increasing the length of the chain P—С—P [2], as well as of carrying substitution on the central carbon atom [3, 4]. The results indicated that the biodistribution and the bone localization of the resulting 99Tcm-chelates may be influenced by the various substituents. Polyfunctional phosphonates labelled with 113Inm were also prepared and evaluated as potential bone imaging agents [5—7]. 613
614
ANDREOU et al.
TABLE I. STRUCTURE OF PHOSPHONATES STUDIED (A )
H 2 0 3 PC H 2 - Îji- C H 2 P 0 3 H 2 R R = - H (I), - C H 3 ( II) , - C H ,C 6 H 5 ( I I I )
(B )
H 2 0 3 PC H 2 - N H - C H 2 C H 2 - N H - C H 2 P 0 3 H 2 E D D M P ( IV )
In the present work the complexing properties o f four N-diphosphonates (Table I) with the radionuclides " T cm and m In were studied, and their biodistribution in experimental animals was compared with that of "T cm-IDP1 and m In-EDTMP2.
EXPERIMENTAL Chemical Iminodimethylene-diphosphonic acid (Table I (A) R=H—) was prepared by the reaction of phosphorous acid diethylester with dilute ammonia and formaldehyde [8]. N-methyl and N-benzyl iminodimethylene diphosphonic acid (Table I (A) R=CH3_, C6H5CH2_) were prepared by the reaction of the corresponding amines with formaldehyde and phosphorous acid [9]. Ethylene-diamine-N, N'-dimethylene-diphosphonic acid, EDDMP (Table I (B)), was synthesized from ethylene diamine, formaldehyde and phosphorus diethylester, followed by hydrolysis [10] o f the reaction product. EDTMP and IDP were commercial products. Labelling with technetium-99m The phosphonate derivatives were labelled with technetium-99m by the stannous ion method: In 20 mg of each derivative in solution, 250 pg of SnCl2 in 5N HCI were added, and pH was adjusted slowly to 6.5. Pertechnetate was then added (0 .0 2 5 -3 .0 mCi, from CIS Elunatic III generator) up to 4 ml final volume.3 The solution was shaken and filtered in an evacuated vial 1
ID P: Imidodiphosphate.
2
E D T M P: Ethylene-diamine-tetraraethylene-phosphonate.
3
1 Ci = 3.70 X 101 0 Bq.
IAEA-SM-2 4 7 /1 11
615
(Millipore 0.22 щт). The above preparations were also formulated to instant freeze-dried kits (laboratory freeze drier, Virtis Co., 10—800). Labelling with indium-111 The labelling procedure consisted o f mixing 0.025—3.0 mCi o f n i InCl3 in 0.05N HCI (Amersham Rad. Centre) with 2 ml (20 mg) o f each derivative (pH = 1.0-2.0). The pH was then adjusted to 6.5—7.0, and the volume was brought to 4 ml, if necessary, and filtered in an evacuated vial (Millipore 0.22 pm). Technetium-99m-IDP and i n In-EDTMP were prepared by methods previously described [1, 5]. The radiochemical purity of the " T cm-complexes was determined by ITLC4, using acetone, saline and AcCN : H20 (3 : 1) as solvents. The l l l In-compounds were analysed by electrophoresis on Whatman paper N o .l, with tris-barbital buffer pH 8.8 (Gielman Chemical Co.). Electrophoresis was run in 200 V for one hour. Biodistribution studies Organ distribution of the labelled compounds was studied following intravenous administration (0.1 ml, approx. 1 juCi), in the tail vein of Swiss Albino mice. The animals were sacrificed by ether asphyxiation at various time intervals. The several organs were removed, as well as the two femurs and tibias, samples of the blood and skeletal muscles. The specimens were counted in a well-type gamma counter (ICN-500) against a standard and the percentage dose per organ was calculated. Images were performed in halflop rabbits by injecting through the ear vein 1 - 2 mCi of the labelled derivatives. The animals were imaged using a 3C Picker Dyna camera, adapted with the suitable collimator for "T cm or m In.
RESULTS AND DISCUSSION The chromatographic and electrophoretic assays for the technetium-99m and indium-111 complexes are presented in Table II. Most of the radioactivity for all the preparations was found under the required chelate form, whereas free pertechnetate was found to be 0.2—2.0%, hydrolysed technetium 1.5-4.6%, and ionic indium-111 0.1-3.0% . In Table III (A and B) the distribution data o f the labelled compounds in mice, 4 Gielman Chemical Co., S.A. type.
616
TABLE II. ANALYTICAL DATA OF THE LABELLED N-DIPHOSPHONATES Percentage o f radioactivity
" T c m-complexa
( I)
" T c m 0 4-b
Hydrol. " T c mC
n l In-complexd
m In (O H )3e
92.10
2 .0
4.60
94.30
3.00
(H )
95.50
0.90
2 . 1 0
96.30
0.50
(H I)
96.00
0 . 2 0
1.50
84.30
1.50
( IV )
94.10
2 .0 0
1 .2 0
99.00
0 . 1 0
Acetone: R f = a 0.00, b 0.90, c 0.00; A cC N :H 2 0 , R f = a 0.85, b 0.90-1.0, c 0.00. Electrophoresis: migration distance, d 5 cm approx., e 0.00.
A N D R E O U e ta l.
Compound
617
IAEA-SM-2 4 7 /1 11
TABLE III. BIODISTRIBUTION STUDY IN MICEa OF THE LABELLED COMPOUNDS, TWO HOURS AFTER THE INJECTION A:
99
T c m-complexes
Percentage dose per organ
Organ (I)
0.57 ± 0.19
(ID P )
( IV )
(III)
( II)
1.62 ±0.28
Blood
1 .1 1
Liver
6.64 ±0.35
0.28 ± 0.07
1 .1 1
1.51 ±0.25
1.17 ± 0.57
Muscle
1.22 ±0.03
0.46 ± 0.20
6.62 ±0.71
0.91 ± 0.12
1.33 ±0.29
Spleen
0.23 ± 0.02
0 . 1 0
0.07 ± 0.01
0.03 ± 0.00
0.03 ± 0.01 0.87 ±0.22 0.77 ± 0.29
±0
.1 2
5.51 ± 0.31
±0 .0 1
±0 .2 1
Kidneys
1.82 ±0.18
0.54 ± 0.20
0.93 ± 0.31
2.55 ±0.25
Stomach
0.50 ±0.23
0.39 ± 0.06
0.62 ±0.45
0.82 + 0.53
1.17 ±0.32
5.82 ± 2.17
2.07 ± 0.54
3.08 ±0.51
8.82 ± 1.30
2.04 ± 0.80
Bones
11.71 ± 0.81
12.68 ± 0.71
13.42 ± 1.02
1.27 ± 0.09
35.23 ± 1.53
Urine
57.59 ±6.74
76.95 ±2.89
54.81 ± 2.87
75.48 ± 6.02
56 . 1 3 ± 3 . 8 6
Intestines
B:
111
In-complexes
Percentage dose per organ (I)
(И )
(III)
( IV )
(E D T M P )
Blood
4.18 + 0.64
8.64 ± 0.37
21.53 ±2.80
2.15 ±0.31
0.17 ±0.07
Liver
1.43 ±0.20
2.48 ±0.10
8.41 ± 2.31
0.78 + 0.12
0 . 2 0
Muscle
10.25 + 2.11
1 8 .5 7 ± 2 . 1 4
22.74 ± 2.26
4 . 3 8 ± 0 .6 9
0.99 ± 0.27
Spleen
0.09 ± 0.01
0.18 ± 0.04
0.56 ±0.02
0.06 ± 0
0 . 0 2
Kidneys
1.84 ±0.44
2.29 ±0.15
2.89 ±0.52
1.02 ±0.33
0.91 ±0.54
Stomach
0.38 ± 0.01
0.83 ±0.24
0.59 ± 0.03
3.14 ±0.29
0.18 ± 0.09
Intestines
. 0 0
±0
±0
.0 2
. 0 0
2.95 ± 0.33
4.79 ±0.39
10.02 ±0.46
3.75 ± 0.76
2.22 ± 1.25
Bones
21.35 ± 2.32
17.61 ± 0.15
12.45 ±0.82
32.44 ±2.70
34.63 ± 4.22
Urine
46.17 ± 6.14
23.89 ± 3.14
11.32 ± 2.82
39.71 ± 9.36
56.03 ± 9.56
3
Each value is the average o f five animals.
TABLE IV. BIODISTRIBUTION OF THE INDIUM-11 1-ETHYLENEDIAMINE-N, N'-DIMETHYLENEDIPHOSPHONIC ACID (INDIUM-111-EDDMP) AT 0 .5 -4 8 h AFTER THE INJECTION Percentage dose per organ Organ
0.5
Blood
2.40 ± 0.55
Liver
1.78 + 0.19
24
4
2
1
1.65 ±0.30
1.52 ± 0.49
0.78 ± 0.38
48 0.36 ± 0.05
0.51 ± 0.11
0.71 ± 0.07
0.77 ±0.12
0.78 ±0.16
1.83 ± 0.64
1.21 ± 0.67
Muscle
4.25 ± 0.51
4.18 ± 1.07
4.38 ± 0.69
3.33 ± 0.58
4.56 ± 1.00
4.10 ±0.61
Spleen
0 . 0 2
0.03 ± 0.00
0.06 ± 0
0 . 0 2
±0 .0
1
. 0 0
±0
. 0 0
Kidneys
1.98 ± 1.08
1 .1 2
+ 0.18
1.02 ± 0.33
1.13 ±0.04
Stomach
3.02 ± 1.10
3.38 + 1.24
3.13 ± 2.29
1.12 ± 0.59
0.13 ± 0.03
0.15 +0.01
1.22 ± 0.27
1.18 ± 0.23
0.15 +0.04
0.12 + 0.04
4.28 ± 2.08
3.72 + 4.57
3.37 ± 1.76
3.36 ± 1.89
2.21 ± 0.83
1.57 ±0.57
Bones
23.16+ 2.66
26.05+ 1.54
32.44 ±2.70
28.79 ±3.11
28.20 ± 1.81
25.64 ± 2.67
Urine
45.89 ± 7.34
47.73 + 8.35
39.71 ± 9.36
52.73 + 5.19
Intestines
-
-
R atios a Bone/blood Bone/muscle
a
7.22
10.40
1 2 .6 6
14.40
25.64
51.20
23.10
26.00
30.40
36.00
28.20
28.40
The ratios were calculated from the values of the percentage dose per 1% body weight.
IA EA-SM -247/111
619
two hours after the administration, are comparatively presented. Technetium-99mIDP and l u In-EDTMP are the reference radiopharmaceuticals. The " T cm-complexes (Table III, A) showed significantly lower skeletal uptake compared with " T cm-IDP. All the "T cm-derivatives were mainly excreted by the urinary system, presenting negligible amounts of radioactivity in stomach, liver and spleen, sites of free pertechnetate and hydrolysed "T cm respectively. Bone localization o f the l u In-compounds (Table III, B) was found greater than that of the " T cm-complexes. Indium-111-EDDMP, two hours after administration, showed high bone concentration, similar to that of l u In-EDTMP, whereas the low liver values of all the 111In-complexes indicate the absence of the in vivo indium hydroxide formation. In Table IV the distribution data in mice of m In-EDDMP are presented 0 .5 -4 8 h after administration. Indium-111 -EDDMP exhibited a rapid blood clearance. After 30 min only 2.5% of the total radioactivity was found in the blood. The maximum bone concentration was achieved two hours after the injection and remained practically unchanged until the 24-h interval. Liver activity was negligible for all the time intervals studied. Rabbit images 2 and 24 hours after the injection (Fig.l) show a good delineation of the skeleton.
CONCLUSIONS The studies described overall indicate that -P -C H 2-N -C H 2- P - and - P - N - C H 2-C H 2- N - P - configurations (Table I, A and B) are incapable of producing " T cm radiopharmaceuticals possessing significant bone localization. The compounds based on configuration A may be considered analogous to the molecule o f IDP, containing methylenes between the nitrogen-phosphorus bridge. The comparative distribution data of the " T cm-complexes of type A point out that the loss in bone affinity is probably due to this change of the - P - N - P — configuration. This is emphasized by the fact that substitution of the iminohydrogen o f type A compounds by aliphatic or aromatic groups did not alter bone localization or change the in vivo orientation of the resulting "T cm-cheiates. Indium-111-EDDMP showed high bone concentration and exhibited biodistribution similar to that o f l u In-EDTMP. The data demonstrate that the two phosphonic groups of 11 Чп-EDDMP are sufficient to ensure high bone localization. The molecule o f EDDMP labelled with 113Inm may be considered as a possible bone indium agent. Moreover, EDDMP presents a basic structure with position on the nitrogens available for substitution by groups other than phosphonic, for the production o f new indium radiopharmaceuticals.
620
ANDREOU et al.
A
B
FIG. 1. Rabbit scintigram 2 h (Aj and 24 h (B) after the IV injection o f n iIn-EDDMP.
REFERENCES [1]
S U B R A M A N IA N , G., M c A F E E , J.G ., B L A IR , R .J., R O S E N S T R E IC H , М., COCO, M., D U X B U R Y , C.E., Technetium-99m-labeled stannous imidodiphosphate, a new radiodiagnostic agent for bone scanning: Comparison with other
99
mTc-complexes,
J. Nucl. Med. 16(1975) 1137. [2] W A N G , T.S.T., H O S A IN , P., S P E N C E R , R.P., A H L Q U IS T , K., H O S A IN , F., Synthesis, radiotechnetium labeling and comparison o f biologic behaviour of longer-chain analogs of methylene-diphosphonate, J. Nucl. Med. 19 (1978) 1151.
621
IAEA-SM-2 4 7 /1 11
[3]
U N T E R S P A N N , S., Experimental examinations on the suitability of organoaminoaminomethane-bis-phosphonic acids for bone-scintigraphy by means of Tc-99m in
animals, Eur. J. Nucl. Med. 1 (1976) 151. [4] W A N G , T.S.T., M O JD E H I, E.G ., F A W W A Z , A .R ., JO H N S O N , M.P., A study of the relationship between chemical structure and bone localisation of Tc-99m diphosphonic acids, J. Nucl. Med. 20 (1979) 1066 (Concise communication). [5]
S U B R A M A N IA N , G., M c A F E E , J.G . R O S E N S T R E IC H , М., COCO, M., Indium-113 labeled poly-functional phosphonates as bone-imaging agents, J. Nucl. Med. 16
(1975) 1080. [ 6 ] JO N E S , A.G., D A V IS , M .A., D E W A N JE E , M .K.,
113
mIn-labeled bone scanning agents,
Radiology 117 (1975) 727. [7]
S E W A T K A R , A .B., N O R O N H A , O.P.D., G A N A T R A , R.D ., Evaluation of 113mIn labelled ethylene-diamine-N,N,N',N'-tetramethylene-phosphonate, E D T M P , as a bone imaging agent, Int. J. Nucl. Med. Biol. 4 (1977)
6 6
.
[ 8 ] P E T R O V , K.A., M A K L Y A E R , F .L ., B L IS N Y U K , N .K., Synthesis of aminodiphosphates and aminotriphosphates, Zh. Obshch. Khim. 29 (1959) 591. [9]
M O E D R IT Z E R , K., IR A N I, R .R ., The direct synthesis of a-aminomethylphosphonic acids, Mannich-type reactions with orthophosphorus acid, J. Org. Chem. 31 (1966) 1063.
[10] K A B A N IK , M .I., M E D V E D , T.Ya., K O Z L O V A , G.K ., B A L A B U K H A , V.S., M IR O N O V A , E .A ., T IK H O N O V A , L .I., Synthesis and examination of complex-forming properties of some organophosphorus compounds, Izv. Akad. Nauk. S S S R , Ser. Khim. No. 4 (1 96 0) 651.
IAEA-SM -247/122
S O M E A S P E C T S O F T H E M E C H A N IS M O F U P T A K E O F T E C H N E T IU M -9 9 m -D IP H O S P H O N A T E S R.A.M.J. CLAESSENS Department of Radiotherapy and Nuclear Medicine, St. Radboud Hospital, Catholic University, Nijmegen Z. KOLAR Interuniversity Reactor Institute, Delft J.E.J. SCHMITZ, J.G.M. van den LINDEN, J.J. STEGGERDA Department o f Inorganic Chemistry, Catholic University, Nijmegen I. KAZEM Department of Radiotherapy and Nuclear Medicine, St. Radboud Hospital, Catholic University, Nijmegen, The Netherlands Abstract SO M E A S P E C T S O F T H E M E C H A N IS M O F U P T A K E O F T EC H N ET IU M -99 ra-D IPH O SPH O N A T ES. A mechanism for the uptake of technetium-99m-labelled bone seekers is proposed. In boneseeking technetium-labelled radiopharmaceuticals, stannous tin and a technetium species are kept in solution by diphosphonate. In bone, the diphosphonate is adsorbed by hydroxy-apatite; the stannous tin loses its stabilizing agent and precipitates. The technetium species seems to compete with the diphosphonate for hydroxy-apatite binding sites. Adsorption experiments with E H D P, tin-EHDP and technetium-tin-EHDP on hydroxy-apatite showed that E H D P is separately adsorbed from technetium and tin. Presumably the EHDP-tin complex has to dissociate before adsorption. A set of six diphosphonates was selected out of 26 diphosphonates after testing in rabbits. Fo r these, calcium and tin-complex stabilities and compositions were determined by pH titration using the Schwarzenbach method. Results for only four of these diphosphonates are at present available. M D P appears to be the best bone seeker when labelled with
9 9
Tcm although
it is the weakest complexing agent. To estimate technetium diphosphonate complex stabilities, pertechnetate had to be reduced electrolytically in the presence of diphosphonate. In order to choose the right potential for reduction, polarograms were made o f the respective systems. These experiments showed that a normal redox reaction between pertechnetate and tin is unlikely. Nevertheless, in the presence of diphosphonates a reaction between pertechnetate and tin takes place in a molar ratio o f 1:2. The electrolytical reduction of pertechnetate in the presence of M D P does not yield a technetium-MDP complex. Under the same conditions, however, in the
623
624
CLAESSENS et al.
presence of D CM D P a weak technetium4DCMDP ) 2 complex could be identified. It is considered that, as a first step of the reaction of pertechnetate with tin, a complex between technetium and tin is created with a molar Tc:Sn ratio o f 1:2. There is evidence that at least at higher pH this complex remains stable for hours. A t lower or neutral pH a technetium species is created which differs from electrolytically reduced pertechnetate and contains no tin. This species appeared to be relatively stable to atmospheric oxygen, even without the presence of stannous tin.
1.
INTRODUCTION
Notwithstanding its extensive use in nuclear medicine, the mechanism of uptake of technetium-99m-labelled diphosphonates in bone has not yet been fully understood. Several mechanisms have been described. On the basis of autoradiographical evidence [1, 2] we concluded that the hydroxy-apatite crystal in bone is where the uptake of technetium-labelled bone seekers takes place.
2.
MATERIALS AND METHODS
2.1. Adsorption experiments Adsorption experiments were carried out with EHDP, tin-EHDP and technetium-tin-EHDP. Carbon-14, 119Snm and " T cm were used as labels. Three ml of solution were incubated at 37°C under nitrogen with 100 mg hydroxy-apatite (spec, surface area 65 m2/g). The time needed to reach equilibrium was found to be between 48 and 72 h for EHDP concentrations not exceeding 0.01M. For each adsorption isotherm a set of seven different EHDP (or tin-EHDP or technetium-tin-EHDP) concentrations was incubated for 1 or 72 h. Table I shows an outline o f all the adsorption experiments. After incubation, the supernatant was filtered over a Millipore filter (0.22 pm). Half a millilitre of the filtrate was acidified with 0.5 ml of 0.5N HCI, and 10 ml of a scintillator was then added. 2.2. Tissue distribution experiments Twenty-six diphosphonates were labelled with " T cra up to 30 min before administration to rabbits. Paper chromatography was carried out to check the quality of labelling (Whatman N o .l, methanol 85%). To each rabbit (New Zealand white albino, weight ± 2200 g) about 500 /¿Ci of labelled disphosphonate (0.5 ml) was injected intravenously.1 1 1 C i= 3 .70 X 1010 Bq.
625
IAEA-SM -247/122
TABLE I. ADSORPTION EXPERIMENTS WITH HYDROXY-APATITE3
Exp.No.
E H D P conc.
S n :EH D P
(M )
(molar)
9»T c m
pH
H S A conc.
Time
(mg/ml)
(h)
1
0
.0
1
-0
.0 0 1
-
-
6
.3/7. 6
15
72
2
0
.0
1
-0
.0 0 1
-
-
6
.4/7. 8
30
72
3
0 .0 1 -0 .0 0 1
1:20
-
6.4/7. 8
15
72
4
0
.0
1
-0
.0 0 1
1 :2 0
-
6.4/7. 8
15
5
0
.0
1
-0
.0 0 1
1 :1
-
6
1 :2 0
+
6.4/7. 8
15
1
1 : 2 0
+
6.4/7. 8
15
1
-
+
6
.4/7. 8
15
1
.0 0 1
1 :2 0
+
6.4/7. 8
-
1
0 .0 0 1
1 : 2 0
+
6.4/7. 8
15
1
6
0.01-0.00005
7
0
.0
1
-0
8
0
.0
1
-
9
0
.0
1
-0
0
.0
1
-
1 0
.0 0 1
0 .0 0 1
.4/7. 8
-
Remarks
1
72
R IH S A
a H SA = human serum albumin. R IH S A = radio-iodinated human serum albumin.
One hour after injection the rabbits.were sacrificed. Tissue and blood samples were taken. Uptake of " T cm was calculated as percentage of dose per gram tissue. Also bone-to-soft-tissue ratios were calculated. Each labelled diphosphonate was tested in 2 - 1 0 rabbits. 2.3. Complex stability estimations The Schwarzenbach method was chosen to estimate diphosphonate complex stabilities [3, 4]. This method is based on the fact that metal ions compete with protons for ligand binding sites. Complex formation will lower the pH o f the diphosphonate solution. The extent of the pH change produces a measure for the complex stability. First, acid dissociation constants were calculated from the results of titrations of the pure ligands with base. Next, these titrations were repeated after adding different amounts o f metal ion. Data from the results of these titrations, together with the corresponding acid dissociation constants, were entered in a computer; complex stability constants as well as the equilibrium concentrations of all the species present in the diphosphonate solutions were calculated by the SCOGS program [5].
626
GLAESSENS et al.
TABLE II. ELECTROLYTICAL EXPERIMENTS
(M D P) = 3-10'3M
M D P + T c 0 4 (4:1)
pH = 6.5+ 9.1
- 0.75 V
MDP + T c 0 4'(8 :1 )
pH = 2.8 + 3.0
- 0.5
V
D C M D P + T c 0 4 (5: 1)
pH = 2.9 + 5.1
- 0.9
V
D CM D P + T c 0 4 ( 5 :l)
pH = 2.9 + 3.4
- 0.5
V
DCM D P + T c 0 4 (5:1)
pH = 6.3 -
- 0.9
V
(D C M D P) = 1.75- 10-3M
6 .6
Acid-base titrations were carried out with 3 • 10 -3 M solutions of diphosphonic acid. The starting volume was 40 ml; 0.1M tetramethylammoniumhydroxide was used as base. All titrational and polarographical experiments were carried out at a constant ionic strength of 0.1 by adding tétraméthylammonium nitrate or sulphate. The following metal-to-diphosphonic acid ratios (on a molar base) were used: 1:2, 1:4 and 1:8. Calcium was added as calcium nitrate, tin as stannous sulphate and technetium as "Tc-ammonium pertechnetate which was previously electrolytically reduced. All experiments were carried out at 25°C in duplicate under argon. 2.4. Polarography In order to estimate the most suitable potential for electrolysis of pertechnetate in the presence of ligand, polarograms were made of solutions of 3- 10"3M MDP or DCMDP with stannous sulphate or pertechnetate added at pH values between 2.5 and 8.5. Metal-to-diphosphonic acid ratios were 1:2, 1:4 and 1:8. Tetramethylammoniumsulphate (0.033M) was the supporting electrolyte. A dropping mercury and a saturated calomel reference electrode were used. Experiments were carried out under argon. Polarograms were made for the region between + 0.1V and - 2.0V. 2.5. Electrolytical reduction of pertechnetate A mercury pool was used as a cathode. Table II gives an outline of the electrolysis experiments. The mercury pool was gently stirred.
627
IAE A-SM -247/122
mm ol/l
FIG.I. Adsorption isotherms for EHDP and tin.
Before and after electrolysis, polarograms of the reaction mixtures were made. The resulting solutions were used for pH titration after adjustment o f the pH to its original value. 2.6. Titrations o f stannous diphosphonates with technetium-9 9 -Tc0 4 To obtain more information on the reaction between tin and pertechnetate, 40 ml of solutions containing MDP-tin (4:1 ; MDP concentration = 3 • 10~3M) or DCMDP-tin (2£ : 1; DCMDP concentration = 1 .8 - 10_3M) were titrated with "Tc-pertechnetate at pH values of 2.8 and 6.2. After each addition of pertechnetate, polarograms were recorded. After reaching the equivalence point for MDP-tin mixtures the reaction product was used for pH titration after adjustment o f the pH to its original value. The results of these pH titrations were processed by the SCOGS program.
3.
RESULTS AND DISCUSSION
3.1. Adsorption experiments A typical adsorption isotherm for EHDP and tin is shown in Fig. 1. EHDP adsorption results, plotted according to Langmuir (Ce versus Ce /Q, where Cg is the concentration at equilibrium and Q is the total amount of absorbed EHDP per gram hydroxy-apatite), render straight lines. This leads to
628
CLAESSENS et al.
TABLE III. BONE-TO-SOFT-TISSUE RATIOS (B:S) FOR SIX DIPHOSPHONATES3
B :S
MDP
19.50
EHDP
9.67
PDP
7.41
D CM D P
7.66
N-ethyl PD P
4.18
PC M D P
2.97
a P D P = Pyrrolidon diphosphonic acid. D C M D P = Dichloromethylene diphosphonic acid. PC M D P = Phenylchoromethylene diphosphonic acid.
the conclusion that hydroxy-apatite has a limited number of indistinguishable binding sites for EHDP. EHDP will cover the hydroxy-apatite with a monomolecular layer. The maximum adsorption level of EHDP is reduced to half its original value by the presence o f equimolar amounts of tin. Small amounts of tin do not show any influence on the EHDP adsorption. After one hour, EHDP adsorption reaches 75% of its value in the equilibrium state. The tin adsorption characteristics do not fit the Langmuir model. Tin curves never reached a constant maximum level in our experiments. After one hour, tin ‘adsorption’ values equal the 72-h values. At lower pH, the slopes of the upper part of the tin curves are much steeper (up to fourfold) than at higher pH value. The behaviour of tin in this system could be explained as precipitation. At any rate, EHDP and tin behaved independently. Technetium showed adsorption characteristics which differed from both the tin and the diphosphonate behaviour. Technetium seemed to compete with diphosphonate for hydroxy-apatite binding sites. From adsorption experiments with carrier technetium we concluded that the technetium species, interacting with hydroxy-apatite, presumably contains no tin; this species is, even in a medium without tin, relatively stable to atmospheric oxidation. After three days under normal atmosphere not more than 20% of the technetium was present in the form of pertechnetate.
629
IAEA-SM -247/122
TABLE IV. ACID DISSOCIATION AND COMPLEX STABILITY CONSTANTS'1
Carbon
Tin
pK2
pK3
pK4
logKML logKM jL
l ° g K MHL
logKML logKM¡L
logKMHL
M DP
2.75
7.10
10.75
5.97
3.84
3.00
12.43
-
9.46
EH D P
2.74
6.98
11.18
6.19
4.73
-
15.68
10.31
-
PD P
2.44
6.95
10.70
6.91
-
3.85
14.58
8.93
-
DCM DP
-
5.97
9.72
5.95
3.00
3.24
13.64
9.46
9.11
a L = Ca, Sn. M = M DP, E H D P , PD P, DCMDP.
At the experimental pH of 6.5, over 95% of tin is normally bound to EHDP in the complexes tin-EHDP and tin2-EHDP. Independent adsorption of EHDP can only occur if EHDP has been released from the complex by dissociation. Stannous ions, not being stable at pH = 6.5, might precipitate (Fig. 1). The stability of the diphosphonate complex might influence its bone-seeking qualities. The presence o f HSA has almost no influence. \ 3.2. Tissue distribution experiments Bone-to-soft-tissue ratios for six diphosphonates are given in Table III. The tissue distributions for the diphosphonates tested, which are of lesser relevance for this paper, will be published elsewhere. 3.3. Complex stability estimations Acid dissociation and complex stability constants for four diphosphonates are given in Table IV. Figures 2 and 3 show the relative quantities of the different tin species present in the MDP or DCMDP media, as a function of pH. The corresponding curves for EHDP and PDP resemble Fig. 3. For calcium, similar curves were obtained. At pH = 7.4, with MDP almost no tin complex formation is present (Fig. 2). Stannous hydroxide is present in a supersaturated solution. This is stabiUzed by the presence of MDP. Diphosphonates are known to inhibit the crystallization of slightly soluble compounds such as calcium sulphate.
630
CLAESSENS et al.
pH '
FIG.2. Relative quantities o f different tin species in the presence o f MDP (L = MDP).
pH
FIG.3. Relative quantities of different tin species in the presence of DCMDP (L = DCMDP).
When we compare the results from the tissue, distribution experiments and the complex stability estimations, we see that the weakest complexing agent is the best bone seeker when labelled with technetium. This was also suggested by the results of the adsorption experiments. 3 .4 . P o la ro g ra p h y
Table V shows the half-wave potentials for the oxidation of tin (II) and the half-wave potentials o f all technetium waves observed between + 0.1V and — 1.5 V. The metal-to-diphosphonate ratio did not influence the half-wave potentials significantly. It was our intention to perform electrolytic reduction of pertechnetate at the half-wave potential for the stannous oxidation; we hoped to obtain a technetium diphosphonate complex that would resemble the stannous reduced technetium diphosphonate as much as possible. Unfortunately, electrolytical reduction o f pertechnetate is not possible at any o f the tin half-wave potentials given in Table V. Because all the tin oxidation potentials given in Table V appear at lesser negative values than the corresponding technetium reduction potentials, a normal redox reaction will presumably not take place. With the Nemst equation [6] and the results from Table V, the ratio TcOj/TcOj was calculated for a reaction between stannous tin and pertechnetate with tin (II) concentration = 10_3M, tin (IV) concentration = 5- 10~SM and the technetium concentrations being negligible when compared with tin. The result is T c0j/T c02 = 7.6 106.
631
IAEA-SM -247/122
TABLE V. HALF-WAVE OXIDATION POTENTIALS (V) FOR STANNOUS TIN AND HALF-WAVE REDUCTION POTENTIALS (V) FOR PERTECHNETATE
M = tin
M = technetium
M DP: M = 4:1 pH = 2.5
-0 .1 8
6.5
-0 .3 2
-0 .4 2
- 1.03
-0.75
- 0.84
-0.75
-0 .84
7.4
-0 .0 4
-0 .2 4
8.5
-
-0 .38
-0 .84
0 .0 2
-0 .37
-0 .32
- 1.24
D CM DP: M = 2 : 1 pH = 2.5
-0.08
-
0 .2 2
-0 .32
- 1.03
6.5
-0 .08
-0.23
-0 .75
-0 .8 4
7.4
+ 0.09
-0.23
-0 .29
in co
+ 0.05
-0 .26
-0 .39
-0 .4 4
- 1.24
-0 .8 4 -0 .84
3.5. Electrolytical reduction of pertechnetate During electrolysis at potentials given in Table II, the colour changed from colourless to brown; no precipitate was observed. With MDP no technetium complex is formed. With DCMDP only one complex could be identified: Tc- (DCMDP)2 ; at pH = 10, 85% of technetium is bound in this complex (see Fig. 4). Again, MDP turns out to be a weaker complexing agent than DCMDP. Technetium complexes appear weaker than the corresponding stannous or calcium complexes. 3.6. Titrations of stannous diphosphonates with technetium-9 9 -Tc0 4 For MDP at pH = 2.8 and 6.2, and for DCMDP at pH = 6.0, the equivalence points show that technetium reacts with two tin units. Titrations of stannous diphosphonates in acid media showed a bright yellow colour; in neutral media, a pink colour appeared. Titrations with pertechnetate in acid media caused no pH change; at pH = 6, addition of pertechnetate caused the pH to increase. This corresponded to a consumption o f about one proton per pertechnetate ion. This pH change is much smaller than in electrolytical reductions.
632
CLAESSENS et al.
DCMDP - Te
5.1
FIG.4. Relative quantity o f technetium in technetium-(DCMDP)2 (L = DCMDP).
Analysis of the polarograms made during the titrations demonstrated that different tin species reacted with pertechnetate to a different degree. Tin waves decreased in the course of the titration and disappeared at the equivalence point, as technetium waves showed up. Only in the presence o f MDP one technetium-tin species with a molar Tc:Sn:OH ratio of 1:2:2 could be identified from the pH titrations of the products of pertechnetate titrations. This complex exists only above pH = 8.5 in appreciable amounts. Its relative abundance as function of the pH is shown in Fig. 5. No technetium diphosphonate complex could be identified in the presence of tin. This was also pointed out by Steigman [7]. Full details of all experiments described will be published elsewhere. Although a normal redox reaction between stannous tin and pertechnetate will presumably not take place, a reaction between these two reagents definitely occurs. Some evidence was found for the existence of a technetium-tin complex at higher pH. The adsorption experiments demonstrated that the product of the interaction between pertechnetate and tin contains no tin, differs from electrolytically reduced pertechnetate and is relatively stable to air oxidation. From this technetium species pertechnetate is slowly liberated. It is possible that we are dealing here with a polymerical technetium compound. This technetium species might have the same kind of interaction with diphosphonate as was postulated between MDP and tin. This technetium species has a smaller affinity to hydroxy-apatite than diphosphonates.
IA EA-SM -247/122 M D P + S n - Те
633
2:1
80 % Tc
60
FIG.5. Relative quantity o f technetium in TcSn2 (OH)2 after titration o f tin with pertechnetate in the presence o f MDP.
REFERENCES [1 ]
T IL D E N , R .L ., JA C K S O N , J., Jr., E N N E K IN G , W .F., D E L A N D , F.H ., M c V E Y , J.T., 99mTc-polyphosphate: Histological localization in human femurs by autoradiography, J. Nucl. Med. 14(1973) 576.
[2]
JO N E S , A .G ., F R A N C IS , M.D., D A V IS , M .A., Bone scanning: Radionuclide reaction
[3]
mechanisms, Semin. Nucl. Med. 6 (1 97 6) 15. R O S S O T T I, F .J.C ., R O S S O T T I, H., The Determination of Stability Constants, McGraw-Hill, New York (1961).
[4]
H E N D R IC K S O N , H.S., Computer analysis of metal complexing equilibria from pH
[5 ]
S A Y C E , I.G., Computer calculation of equilibrium constants of species present in mixtures of metal ions and complexing agents, Talanta I S (1968) 1397.
titration data, Anàl. Biochem. 24 (1968) 176.
[6 ]
SK O O G , D.A., W E S T , D.M., Fundamentals of Analytical Chemistry, Holt, Rinehart, Winston, New Yo rk (1963) 389.
[7]
S T E IG M A N , J., M E IN K E N , G., R IC H A R D S , P., The reduction of pertechnetate-99 by stannous chloride: II. The stoichiometry o f the reaction in aqueous solutions o f several phosphorus ( V ) compounds, Int. J. Appl. Radiat. Isot. 29 (1978) 653.
DISCUSSION B.F. NOSSLIN: Do you have an explanation for the common finding of individual differences in soft-tissue uptake? One and the same preparation may give good pictures in one patient but bad pictures in another. Is there an interaction with pharmaceuticals which the patient may be taking, such as cytostatic compounds?
634
CLAESSENS et al.
R.A.M.J. CLAESSENS: Owing to the weak binding between technetium and MDP, and the stronger interaction between calcium ions and MDP, low bone : soft-tissue ratios can be expected in patients with high plasma calcium levels. Moreover, the diminished availability of young bone crystal, which has the highest hydration level, can be a cause of low bone : soft-tissue ratios. This can be expected in older patients. Pharmaceuticals which show a strong complexing ability can also influence the distribution of technetium after administration of " T cm-MDP. Most non-bone uptake of technetium after injection of such technetium-labelled radiopharma ceuticals is caused by the interaction of diphosphonate with non-bony tissue, for instance, in the outer border of the necrotic zone in myocardial infarcts, or the degenerative tissue in amyloidosis. R. MUENZE: Have you checked the reversibility of the complex equilibrium, since this is the basis for interpreting the measured change o f hydronium concentra tion in terms o f stability constants? R.A.M.J. CLAESSENS: The reversibility of technetium and tin complex formation is proven by the results o f the adsorption experiments with hydroxyapatite. The reversibility of tin diphosphonate complex formation can also be derived from the results given in Fig. 4 of the paper.
C H A IR M E N O F S E S S IO N S
Session 1
K.E. SCHEER
Federal Republic of Germany
Session 2
W.J. MACINTYRE
United States of America
Session 3
C. KELLERSHOHN
France
Session 4
D. IVANCEVIC
Yugoslavia
Session 4a
D.M. TAYLOR
Federal Republic of Germany
Session 5
N.G. TROTT
United Kingdom
Session 6
R. HÖFER
Austria
Session 7
M. IIO
Japan
Session 8
E. TOUYA
ALASBIMN
Session 9
H.N. WAGNER, Jr.
United States of America \
S E C R E T A R IA T
E.H. BELCHER
Division of Life Sciences, IAEA
B. VAVREJN
Division of Life Sciences, IAEA
Administrative Secretary:
G. SEILER
Division of External Relations, IAEA
Editor:
S.M. FREEMAN
Division of Publications, IAEA
Records Officer:
R. LUTHER
Division of Languages, IAEA
Scientific Secretaries:
635
The following conversion table is provided for the convenience o f readers and to encourage the use o f SI units.
FACTORS FOR CONVERTING SOME ОТ THE MORE COMMON UNITS TO INTERNATIONAL SYSTEM OF UNITS (SI) EQUIVALENTS N O TES: (1)
S I base units are the m etre (m ), k ilo g ra m (kg), second (s), am pere (A ), kelvin (К ), candela (cd) a n d m ole (m ol).
<2)
►
indicates S I derived u n its and those accepted fo r use w ith S I;
^
indicates ad d ition al u n its accepted fo r use w ith S I fo r a lim ite d tim e.
\For further information see The International System of Units (Sll, 1977 ed., published in English by HMSO, London, and National Bureau of Standards, Washington, DC, and International Standards ISO-WOO and the several parts of ISO-3 J published by ISO, Geneva. | (3)
T h e correct abb revia tion fo r the u n it in c o lu m n 1 is given in c o lu m n 2.
(4)
*
indicates c on v e rsio n fa cto rs g iven exac tly ; o ther factors are g ive n ro u n d e d , m o stly to 4 significant figures.
h
indicates a d e fin itio n o f an S I d erived unit: [ ] in c o lu m n 3 + 4 e nclo se fa cto rs given for the sake o f com pleteness.
Column 1
Column 2
M ultiply data given in:
Column 3
Column 4
by:
to o btain data in:
Radiation units ^ becquerel
1 Bq 1 s“ 1
*
> roentgen
1 Ci 1R
(has dimensions o f s 1) = 1.00 X 10° Bq = 3.70 X 1 0 '° Bq = 2.58
X 10"“
C/kgJ
*
► gray
1 Gy
= 1.00
X 10°
J /k g |
*
> rad
1 rad
= 1.00
X 10~2
Gy
1 Sv
= 1.00
X 10°
J /k g ]
* *
1 rem
= '1 .0 0
X 10“ 2
J/kgJ
*
► unified atom ic mass u n it ( ^ o f the mass o f l2C)
1u
= 1.660 57 X 10“ 21 kg, a p p ro x .]
^ tonne (= m etric ton) pound mass (avoirdupois)
1t
= 1.00
1 Ibm
= 4.536 X 10~'
kg
ounce mass (avoirdupois) ton (long) ( - 2240 Ibm)
1 ozm 1 to n
= 2.835 X 10'
g
= 1.016 X 103
kg
to n (short! (= 2000 Ibm)
1 short ton = 9.072 X 102
kg
disintegrations per second {= dis/s) > curie
sievert (radiation protection only) rem (radiation protection only)
*
Mass X 103
kg]
*
Length statute mile
1 mile
nautical m ile (international)
1 n m ile
= 1.609 X 10° = 1.852 X 10°
km km
*
yard
1 yd
= 9.144 X 10~'
m
*
fo o t
1 ft
inch m il (= 10-3 in)
1 in 1 m il
= = =
3.048 X 1 0 '1
m
2.54 2.54
mm mm
X 10' X 10-2
*
* *
Area > hectare > barn (effective cross-section, nuclear physics) square acre square square square
mile, (statute m ile)1 yard fo o t inch
1 ha
= 1.00
X 104
m 2]
1b
= 1:00
X 10~28
m2 ]
1 m ile2 1 acre
= 2.590 X 10° = 4.047 X 103
km 2 m2
1 yd2 1 ft2 1 in2
= 8.361 X 10“ ' = 9.290 X 10“ 2 = 6.452 X 102
m2 m2
= 1.00
m 3] m3
■* *
mm2
Volume ► litre cubic yard
1 I o r 1 Itr 1 yd3
X 10“ 3
= 7.646 X 10_l
cubic fo o t
1 ft3
= 2.832 X 1 0 "J
m3
cubic inch gallon (im perial)
1 in 3
= 1.639 X 104
mm3
1 gal (U K )
= 4.546 X 10 "3
m3
gallon (US liq u id )
1 gat (US)
= 3.785 X 10~3
m3
1 ft/s 1 ft/m in
= 3.048 X 10“ ‘
fo o t per m inute
= 5.08
m/s m/s
m ile per hou r (= mph)
1 m ile/h
*
Velocity, acceleration fo o t per second (= fps)
t> k n o t (international) free fa ll, standard, g fo o t per second squared
1 knot 1 ft/s 2
X 10 "3
Í4.470 X 1 0 '' [1.609 X 10°
* ■#
m/s km /h
= 1.852 X 10°
km /h
= 9.807 X 10°
m /s2
= 3.048 X 10~‘
m /s2
* *
This table has been prepared by E.R.A. 8eck for use by the Division of Publications of the IAEA. While every effort has been made to ensure accuracy, the Agency cannot be held responsible for errors arising from the use of this table.
Column 1
Column 2
M ultiply data given in:
Column 3
Column 4
by:
to obtain data in:
Density, volumetric rate pound mass per cubic inch
1 lb m /in 3
= 2.768 X 104
k g/m 3
pound mass per cubic fo o t cubic feet per second cubic feet per m inute
1 lb m /ft3
= 1.602 X 101
k g /m 3
1 f t 3/s
= 2.832 X 1 0 '2
m 3/s
1 f t 3/m in
= 4.719 X 10~4
m 3/s
Force ► new ton
1N
dyne
[ = 1.00
X 10°
m -kg s 2] * * N
1 dyn
= 1.00
kilogram force (= k ilo p o n d (kp))
1 kgf
= 9.807 X 10°
N
poundal pound force (avoirdupois)
1 pdl
N
1 Ibf
= 1.383 X 10“ ’ = 4.448 X 10°
ounce force (avoirdupois)
1 ozf
= 2.780 X 10_1
N
X 10~s
N
Pressure, stress N /m 2 ] * *
► pascal
1 Pa
> atm osphere3, standard
1 atm
> bar
1 bar
= 1.00
X 10s
Pa
1 cmHg
= 1.333 X 103
Pa
centim etres o f m ercury (0°C) dyne per square centim etre
[ = 1.00 X 10° = 1.013 25 X 10s:
1 d y n /c m 2 = 1.00 X 10“ 1 1 f tH 20 = 2.989 X 103 1 inHg - 3.386 X 103
feet o f water (4°C) inches o f m ercury (0°C) inches o f water (4°C) kilogram force per square centim etre
1 in H 20
= 2.491 X 102
1 k g f/c m 2 = 9.807 X 104 1 lb f / f t 2 = 4.788 X 101 1 Ib f/in 2 = 6.895 X 103 1 to rr = 1.333 X 102
pound force per square fo o t pound force per square inch (= psi) ^ to rr (0°C) (= mmHg)
Pa
Pa
* *
Pa Pa Pa Pa Pa Pa Pa
Energy, work, quantity o f heat 1J 1 eV
► joule (= W-s) ► electronvo lt
X 10° N -m ] * [= 1.602 19 X 10" 19 J, approx.] [ ее 1.00
1 Btu
= 1.055 X 103
J
1 cal
= 4.184 X 10°
1
cal i t 1 erg
= 4.187 X 10° = 1.00 X 1 0 "7
J J J
fo o t-p o u n d force kilo w a tt-h o u r
1 f t Ibf
= 1.356 X 10°
J
1 kW h 1 k t yield
= 3.60 4.2
J J
♦X-
k ilo to n explosive yield (PNE) (= 1012 g-cal)
J/s]
*
B ritish therm al u n it (Internationa l Table) calorie (therm ochem ical) calorie (Inte rn a tio n a l Table) erg
X 106 X 1012
* * •
Power, radiant flux 1W
► w a tt B ritish therm al u n it (In te rn a tio n a l Table) per second
1 Btu/s
calorie (Internationa l Table) per second foot-pou nd force/second horsepower (electric) horsepower (m etric) (= ps) horsepower (550 f t lbf/s)
[ = 1.00 X 10° = 1.055 X 103
1 cal|-|-/s
= 4.187 X 10°
1 f t ■Ibf/s 1 hp
= = = =
1 ps 1 hp
1.356 7.46 7.355 7.457
X X X X
10° 102 102 102
W
w w w w w
*
Temperature ► tem perature in degrees Celsius, t where T is the therm od ynam ic tem perature in kelvin and To is defined as 273.15 К
t = T - T0
to
degree Fahrenheit
t (in degrees Celsius) *
-3 2
T (in kelvin) ДТ ( = At)
degree Rankine
ДТор (=Atop)
degrees o f tem perature d iffe re n c e c
Thermal conductivity 1 B tu- in /( ft 2 ■s ° F ) 1 B tu /( ft * s-°F>
(international Table Btu) (International Table Btu)
1 c a ln -A c n v s ^ C )
Wm
= 6.231 X 103
W -гтГ1 K “ W -m "1 K "
= 4.187 X 102
atm abs, ata: atmospheres absolute;
^
atm (g), atü: atmospheres gauge. The abbreviation fo r tem perature difference, deg
= 5.192 X 102
{-
K'
Ib f/in 2
{= psig) : gauge pressure;
Ib f/in 2 abs
(= p s ia ): absolute pressure.
degK = degC), is no longer acceptable as an SI u n it.
* *
INTERNATIONAL ATOMIC ENERGY AGENCY V I E N N A , 1981
SUBJECT GROUP: I Life Sciences/Nuclear Medicine PRICE: Austrian Schillings 980,
