Personal Notes from the 1997 SNM meeting in San Antonio 31 May 1997 - 5 June 1997 Robert E. Zimmerman zimmer@bwh.harvard.edu San Antonio turns out to be a very pleasant city to visit for a convention. Weather warm but not too hot, at least for this meeting and the Riverwalk is delightful. It seemed much more developed than when we last visited this city, which was probably about 8 or 10 years ago. This year’s news is that the meeting was more hectic than ever. Lest you think this is the flawed perceptions of a senile old cynic I will explain why it is a fact. The scientific sessions were set up to start at 8 AM Tuesday through Thursday. The break between scientific sessions was only 15 minutes and the lunch break was from 11:15 to 12:30. Barely enough time to get hungry and find a sandwich and find the next scientific session. I do not think these scientific meetings are meant to be an endurance contest. There seemed to be very little time to meet vendors in the technical exhibits or confer with colleagues. I realize that the Program Committee has a tough job to maximize the number of papers and information transmitted but I am not sure this was the way to do it. A more relaxed pace would be appreciated. And what is a “Sitz Poster”. I am probably too old and set in my ways to handle so many changes in one year. Never did see these Sitz posters but presumably they were a variation on the “walking-posters” from previous years. Can anyone tell me why we need two programs? One with the Scientific Program and Abstracts and another with the Scientific Program with retyped paper titles(often in error) and Technical Program. Why can’t we get one program with everything? Why are all the paper titles retyped and put is a secondary program? Why is there no room in the abstract book to write notes as the papers are being given? Why do so many sessions start with the label: “Instrumentation and Data Analysis…”? The one saving grace of the new schedule was that Henry Wagner’s highlights talk was early on Thursday afternoon and many of us who rarely get to hear it were able to attend. It was as masterful as I remember. I never saw an attendance figure. I suspect it was less than other recent years. Either that or the convention was impeccably run. There was no overcrowding. Everything worked from my perspective. Saturday Arrived about noon and rushed to the Academic Council Board of Directors meeting only to learn that it was to beheld at the non-traditional hour of 2 PM and no lunch! May have to reconsider this activity in the future. Academic Council is having tough times. There are fewer and fewer nuclear medicine residents - to go along with fewer jobs in nuclear medicine, presumably. But the future of nuclear medicine looks, scientifically at least, very bright. But where will the docs come from who are going to be involved in the advances the scientists and current leaders of the field think are coming. Shall it be left to the radiologists and the American College of Radiology (ACR) bypassing the American Board of Nuclear Medicine (ABNM)? Or, by some means still not determined, will the ABNM win the hearts and minds of the ACR to maintain its role in the certification process? The answers are not clear but it just may be that the radiologist will control the bulk of NM in the future with the cardiologist and neurologist and oncologists peeling off bits for themselves. May not be pretty but does anyone have see a different scenario for the future? I hope so. That’s the picture I see in the states. I wonder what it is like in Europe and Asia? I just do not have the in depth perspective on the situation in other parts of the world to make definitive comparisons. I suspect that it is actually a much rosier picture but then managed care is said to be making strong inroads on the continent also. Perhaps, if I can get to Glasgow in August, I can pick up some insight on the European situation. Sunday A great day to stay away from the meeting. Although I understand that there was an interesting choice of Categorical Seminars I continued my self-imposed tradition of doing something unconnected with the meeting for at least part of a day. Ended up walking down the Riverwalk to the end and finding the San Antonio Art Museum. Located in an old brewery in an industrial section of town, it still had charm and some interesting collections. The major focus of the day was a concert of South American music by the trio Chaski using traditional (and modern) instruments. Lunch was at the famous Texas restaurant The Pig Stand, a throwback to the 1950’s drive ins. The reward for the walk back was beers at Dick’s Last Resort, where you can get insulted by the barmaids at no extra cost. Well worth the walk and playing hooky for the day. On Sunday evening the Board of Directors of the American Board of Science in Nuclear Medicine (ABSNM) met. Focus for the year has been revamping the exams. Look for an announcement of a Web Page soon. The incoming President is Lionel Lieberman. Best wishes Lionel. Keep up the good work of your predecessors. Commercial things One thing that makes for a hectic meeting is the increased attempt of the vendors to attract you to early morning or early evening meals to tell you about their latest ideas. In fact I have found that, sometimes, these sessions can be quite informative. Certainly more informative than the Sunday afternoon user meetings which I long ago gave up attending. The most interesting invitation was one from Siemens for a 6:30 AM breakfast get together for physicists. I was not disappointed. Ron Nutt gave a very convincing presentation that the future of SPECT/PET cameras was going to be with a sandwich of YSO/LSO detector material. A 2 headed system’s performance for PET would be just under the specifications for the incomplete ring system that CTI/Siemens Currently markets - ART. The SPECT performance would approach closely current SPECT instruments. There does not seem to be any way that NaI based systems could compete if Nutt is correct. Now for the downside (and this is largely based on hearsay but I think there is considerable truth in it). CTI has the world’s supply of raw material for LSO locked up. And they are gearing up to make real PET machines with it. Will they divert enough into Siemens to allow timely development of the SPECT/PET machine? Can they even gear up to produce enough of it for both “customers”. Can the Siemens/CTI alliance stand the strain? We will have to wait and see. Informed estimates indicate that salable SPECT/PET instruments based on this new technology will be 2-5 years away. In any case the future of nuclear medicine instrumentation looks very bright. About time we gave the chemists something new to work with. Also discussed by Dr Anna Cellar of U British Columbia was the attenuation correction scheme adopted by Siemens that uses multiple line sources. This seems to be a very economical, sensible approach to attenuation correction for heart and brain imaging. We need to see results, though. See papers No. 917 and 918. Mark Groch, David Cook and Brad Pratt presented some information on Siemens finally released product: ICON-IDL. This is Siemens answer to ADACs use of AVS as a “standard” high level language programming environment. Siemens has enhanced IDL to make it easy to use on the Macintosh to get at Siemens data. This seems to be all that is left of the fitful Phoenix project of the last several years. It was a most impressive breakfast presentation. The sleeping giant has awoken and is moving forward. The rest of the industry had better beware. Commercial Exposition The commercial exhibit contained something of a shock. In the middle of the Toshiba booth there was an e.cam, manufactured by Siemens. Apparently Toshiba has bought 100 e.cams, married it to their workstation, a UNIX machine, and are able to sell it worldwide. I think it makes a formidable team. The e.cam is a very attractive camera and many abhor the ICON computer. While I was in San Antonio Siemens was installing an e.cam for us. It still has not done patients (26 June), largely due to our inability to complete the acceptance tests in a timely manner. Several days were lost to little things that essentially brought down the camera. We are very high on the potential of the e.cam. Incidentally, the installation of the e.cam required the other ICON computers to be upgraded to Version 7.0 of software, also. Something about being able to read the “new” e.cam files. Indeed, our Sudbury System OpenPACS did have difficulty reading the new format, but it was quickly remedied. Wonder what the change in file format was? I love it when the companies make these changes with no notice or information. All the documentation sent with the new system was for version 6.3. Not a word in the documentation about 7.0. Isn’t the FDA supposed to be worrying about things like this? Isn’t ISO 9000 or whatever demanding that documentation be updated and correct? Yeah, right! It is all just another layer of expensive bureaucracy. Park Medical Systems showed their progress in MCAT - Modular Coded Aperture Technology. While not yet in clinical tests, they did show some recently acquired clinical heart images and some intriguing phantom images that indicate reconstructed resolution of better than 8 mm with high? sensitivity. And they have FDA approval. Unfortunately rumor mills were churning about the future of Park. Their chief scientist, Daniel Gagnon left for Picker recently and is there room for a small camera company? To reinforce the feeling of company attrition, GE and Elscint were trying hard to convince us that consolidation of their research, development and production of their variable angle camera offerings in Hafia was a step forward for both. That, combined with the Toshiba/Siemens deal make it look like we should expect more shakeout in the near term. It must be extremely expensive to bring a new camera design to market successfully. And think what it costs to bring a SPECT/PET product to a viable level of performance. Who will be left in 5 years? SMV was talking about coincidence cameras and thick crystals on their digital camera. They are taking a conservative approach with coincidence by incorporating the necessary corrections (scatter, randoms, attenuation) in their product before distributing to the public. Smart approach and to be applauded! Digirad was showing a CdZnTe camera with a full field of view and offering it for sale for about $250K. Has max count rate of 500kcps. Recommended as a camera for cardiac office or clinic and for breast imaging. It is a quite small and portable device.. More details at www.digirad.com. The JNM is on CD-ROM. At least the last 10 years is. Also the Journal of Nuclear Medicine Technology. That is the good news. The bad news is that only the abstract, author name and titles are searchable. But the full text is available as an image file. This is brought to you by the SNM and Digital Publishing, Inc. I have it. I like it. The University of Newcastle, Australia is offering Postgraduate degrees in Nuclear Medicine. “First in the world to be offered via the Internet.” Check it out at http://www.newcastle.edu.au/mrs/nm. We have a National Atomic Museum here in the USA. In Albuquerque, NM. Have a look at http://www.sandia.gov/AtomMus/AtomMus.htm. Oh yes, they are looking for donations, money, not your old cameras. UNM, Ltd. Is a new (to me) company producing educational materials. Their first offering is “Technical Challenges in Nuclear Medicine”. They were handing out sample disks for take home trials. CME/Voice credit is given for these courses. This is the brainchild of Robert Henkin Of Loyola U. Chicago. More information at http://www.unm.inter.net and http://www.nucmednet.com. The sample disk contains 3 or 4 sample questions related to QC of instrumentation. Not really enough to get the full flavor of how good or effective a learning tool it may be. University of Michigan was pedaling something at Booth #747. Not sure what. One thing seemed to be a Website that was able to filter your files for you. It is planned to offer reconstruction in the future. Is this a way to take advantage of distributed computing for the average site? May want to check out http://www.tmo.umich.edu for the other offerings at the show. Also see Poster No. 916. Scientific Papers Monday Plenary Session The Henry Wagner Lectureship was presented by Michael Maisey MD of Guys & St Thomas Hospital, London, on the subject “The Cost of Success”, Measuring the Value of Nuclear Medicine. Dr Maisey presented data to show that it is possible to measure, crudely at this time, the value of a procedure. There is much hope for improving the techniques of this measurement process, primarily using computer modeling and simulation of the diagnostic process. He feels that we are on the verge of important breakthroughs in this area. This is not really an area I understand well so I will not butcher anymore of his material. Michael Devous and Tom Budinger presented overviews of what they thought were going to be the most significant papers to be presented in the areas of the brain and instrumentation, respectively. Not sure why they had this at the plenary session. Could have saved some time and moved on to the awards. Mike Welch introduced Bill Eckelman and presented to him the Georg de Hevesy Award. Al Wolf introduced Joanna Fowler and presented her with the Paul C. Aebersold Award. The Plenary session closed and we moved to the SNM Exposition. (See comments above.) Scientific Program Session 4: Instrumentation and Data Analysis: Young Investigators Competition No. 19 MICROPET I: PERFORMANCE EVALUATION OF A VERY HIGH RESOLUTION PET SCANNER FOR IMAGING SMALL ANIMALS. A.Chatziioannou, SR Cherry, Y Shao, RW Silverman, K. Meadows, T. Farquhar, ME Phelps, UCLA, Los Angeles. This was an update of a project started in Jan 1994. It has (or is that will have?) 30 detectors each of which is an 8x8 LSO array of 2x2x10 mm LSO crystals. That comes to 1920 LSO crystals. Does it meet its goal of “inexpensive”? transverse FOV is 11.2 cm dia and axial length is 1.8 cm. Volume sensitivity is 5.1 kcps/uCi/cc for 6.35 dia cylinder. Energy resolution is between 15-25% for individual crystals and for a 350-650 kev window the scatter fraction is 11%. The nominal spatial resolution is a 2x2x2 mm cube. In fact, resolution varies throughout the FOV. Intrinsic resolution is 1.48 to 1.94 mm with an average of 1.58 mm. The volume resolution varies between 5-20 mm^3. It meets or exceeds it design goals. The pictures of rat brains were amazing for their detail. They were acquired with an injection of 1.8 mCi and imaged for 60 min. No. 20 AN INEXPENSIVE, GEOMETRICALLY PERFECT, SLIT/SLAT (SS) COLLIMATOR FOR SPECT IMAGING WITH A STATIONARY ARRAY OF PLANAR DETECTORS. JE Koss, DL Kirch, TK Johnson, PP Steele, PW Levitt, ACR Inc. and U of Colorado Health Sciences Center, Denver. This is a takeoff on the SPRINT II system but with a planar detector. It uses a one dimensional collimator made of lead strips or slats separated by 3 mm pieces of foam and a slit (or slits) perpendicular to the direction of the slats placed in front of the slats. SPECT is performed by placing several of these detectors around the patient. For example 3 are used for heart imaging. Patient or detector wobble is used to fill in the sampled volume. Sensitivity is claimed to be 35% higher than traditional SPECT detector with 3 slits. Intriguing concept. No. 21 A NOVEL DUAL DETECTOR PROBE FOR GAMMA-RAY GUIDED INTRA-OPERATIVE BETA IMAGING MP Tornai, CS Levin, EJ Hoffman, LR MacDonald, EC Glass, R Essner, KL Black, UCLA, Los Angeles This was an update of an ongoing project. The device now includes directional gamma counting, coincidence photon imaging as well as beta imaging. The detectors now have LSO crystal arrays added (2x2x10 mm) to the CaF2 beta scintillators. The LSO is used as a light guide (degrading beta spatial resolution somewhat) and as coincidence detectors and for background reduction. Intraoperative probes are becoming accepted by surgeons and we have some at our hospital using them in melanoma cases, but they are much simpler devices. This sounds so comprehensive and sophisticated can it be left to a surgical team or will it need a special trained operator, if and when it is deemed a success and ready for the operating room? No. 22 A NOVEL KINETIC PHANTOM FOR DYNAMIC IMAGING HR Tang, JK Brown, BH Hasegawa UCSF, San Francisco and Berkeley The authors used an ion exchange resin in a closed circulating system. Rates are controlled by resin type, pH, flow rates and acid type. They found it to be stable and reproducible. The phantom could be used to characterize dynamic performance of gamma cameras and kinetic system analysis. No. 23 EXTRACTION OF THE INPUT FUNCTION IN CARDIAC DYNAMIC SPECT USING FACTOR ANALYSIS EVR Di Bella, GT Gullberg, U Utah, Salt Lake City FADS was studied as a potential solution to the need for getting the input function in SPECT studies, noninvasively. Phantom simulations and dog studies with teborxine were performed. In 80% of the trials adequate input functions could be obtained. It may prove to be useful. No. 24 STRUCTURED BIAS ORIGINATING FROM INTERACTION BETWEEN A PENALIZED OBJECTIVE FUNCTION AND THE SYSTEM MATRIX CY Ng, NH Clinthorne, JA Fessler, AO Hero, WL Rogers, U Michigan, Ann Arbor The physical situation is the a brain SPECT system using a ring tomograph (SPRINT II) supplemented by a vertex view. The problem is that there is too much smearing of activity between planes when using a 2d penalized weighted least square algorithm. The authors investigated the reasons for this and concluded that a 3d penalty function may help the problem. They learned a lot about mathematical description of imaging processes. The judges had a very hard time selecting a winner from the above 6 presentations. They eventually chose Chatziioannou as the winner. The cash prize was to be presented at the business meeting on Tuesday. Session 11: Instrumentation and Data Analysis: SPECT I - Cardiac Imaging No 61 and 62 I missed No. 63 CHARACTERIZING THE LOCAL IMPACT OF ATTENUATION IN MYOCARDIAL SPECT USING INTRINSIC DUAL ENERGY PROCESSING (IDEP) JK Brown, HR Tang, EH Botvinick, RS Hattner, M Bocher, NM Ratzlaff, PP Kadkade, BH Hasegawa, UCSF, San Francisco and Berkeley These investigators took an old idea: make use of the two photons emitted and do a depth correction. In this case they used the two photons (X-rays at 70 kev and gamma rays at 170 kev) of Tl-201 to get information on the attenuation length in heart SPECT images. Their results have been encouraging, so far, but there are no claims made, yet. No. 64 ANATOMICAL VARIATIONS AFFECTING CARDIAC IMAGING IN FEMALE PATIENTS. S Jane, RJ Jaszczak, H Wang, RE Coleman, Duke U Med Ctr, Durham Missed this paper for a reason I no longer recall. No. 65 LEFT VENTRICULAR MASS FROM UNGATED PERFUSION IMAGES: COMPARISON TO MRI. TL Faber, RD Folks, CD Cooke, JP Vansant, RI Pettigrew, EV Garcia, Emory U, Atlanta Perfused cardiac mass is important to know. Knowing it can help with diagnosis and treatment monitoring. The authors assert it can be determined from gated SPECT. This paper tells us how they obtained it from ungated SPECT. Using a correction factor determined from gated SPECT and with correlation to MRI cardiac mass measurements they concluded that they could indeed determine the cardiac mass from ungated SPECT images. Not clear if this would translate to anyone’s camera without substantial effort. No. 66 MYOCARDIAL SPECT RECONSTRUCTION USING EM AND OSEM WITH COMPRESSION FOR DISTANCE DEPENDENT RESOLUTION AND POST-RECONSTRUCTION FILTERING. YH Lau, BF Hutton, Westmead Hospital, Sydney Authors investigated the gains to be obtained with compensation for detector resolution response. The parameter studied was the recovery coefficient and its standard deviation. The conclusion was that correction for detector resolution with no filtering and a small number of iterations (OSEM for speed) would suffice for reconstruction of clinical myocardial studies. Session 19: Instrumentation and Data Analysis: SPECT II - Instrumentation No. 110 DESIGN STUDIES FOR A CESIUM IODIDE SILICON PHOTODIODE GAMMA CAMERA J Strobel, NH Clinthorne, WL Rogers,Tech U of Munich and U Michigan, Ann Arbor The goal is to come up with a small gamma camera with very good spatial resolution. The crystal is to be CsI and the light detectors are to be pixelated Si photodiodes. In simulation studies the use of retroreflectors on the entrance surface of the CsI crystal improved the spatial resolution by 33%. No. 111 A THICK CRYSTAL GAMMA CAMERA DESIGN FOR PET AND SPECT USING RETROREFLECTORS. J Strobel, NH Clinthorne, WL Rogers, Tech U Munich, U Michigan, Ann Arbor. In a further study of the usefulness of retroreflectors these authors simulated their use with 25 mm thick NaI detectors for a PET/SPECT system. Resolution of 1.7 mm with 25 mm square PMTs and 3.9 mm with 60 mm square PMTs resulted from the simulations. It remains to devise a method to produce retroreflectors. No. 112 SCINTILLATION CAMERA FOR PROXIMATE SPEC IMAGING OF AN ISOLATED BREAST S Genna, V Gayshan, AP Smith, Digital Scintigraphics, Inc., Waltham If scintimammography is to reach its potential a custom SPECT instrument is necessary. The people that brought us the Ceraspect dedicated brain imager presented a design for a cylindrical SPECT system for SPECT of the pendant breast. An innovative liquid light guide and slanted collimator permits very close approach to the breast. Energy resolution of 8.5% and reconstructed spatial resolution of 5 mm is expected. No. 113 HIGH RESOLUTION BRAIN SPECT IMAGING USING HALF FAN-BEAM COLLIMATORS WITH A RIGHT-ANGLE DUAL-CAMERA SPECT SYSTEM. BMW Tsui, DS LaLush, DP Lewis, XD Zhao, EC Frey, J Li, JG Colsher, U North Carolina, Chapel Hill and GE Medical Systems, Milwaukee This paper addressed the use of one dimensional fan-beam collimators on the GE L-shaped dual crystal SPECT system for purposes of improved brain imaging. They were successful in coming up with necessary corrections and achieved 4x the sensitivity of the LEHR parallel hole collimator with equivalent resolution. Will this become a GE product? Tuesday Session 27: Instrumentation and Data Analysis: PET No. 157 A HIGH RESOLUTION SCANNER FOR INVESTIGATING TECHNICAL REQUIREMENTS AND SELECTED BIOLOGICAL APPLICATIONS OF CONTINUOS 3D DETECTORS IN SMALL ANIMAL PET. J Siedel, C. Johnson, E Jagoda, L Sha, WC Eckelamn, MV Green. NIH. Bethesda Authors are determined to get a PET animal system with high resolution. They used opposed NaI crystals with position sensitive PMT on each. EM- ML algorithm that could incorporate corrections for spatial imperfection was used. Spatial resolution of 1.5 mm in slice and 2.3 mm slice thickness was achieved. Sensitivity was very low. Use of 3 pairs of LSO detectors will permit 25 fold increase in sensitivity. N0. 158 SMALL ANIMAL PET IMAGING DEVICE - PRELIMINARY DESIGN STUDY JA Correia, CA Burnham, D Kaufman, DA Chesler, AJ Fischman, MGH Boston The investigators studied experimentally the use of LSO, GSO and BGO in a prototype tomograph for small animal imaging. LSO is the most promising detector. When used in 1x1x5 mm crystals in a ring configuration 1.0 mm FWHM should be realizable in a 8 cm aperture, including range and non-colinearity effects. It is interesting to note that this group had LSO delivery problems. Can CTI meet the demand? Perhaps they should loosen control. No. 159 A HIGH RESOLUTION LSO PET DETECTOR BASED ON THE 16 CHANNEL HAMAMATSU R5900-M16 PMT. S Majewski, A Boutefnouchet, Y Shao, A Weisenberger, T and R Wojcik, Thomas Jefferson National Accelerator, Newport News, UCLA, Los Angeles, U of Regina, Regina. This team has taken the new PMT and studied it application to the previously described UCLA microPET system (See No. 19). Using new fiber optic design they achieve a pixel to pixel connection of the crystal array to the PMT. Results are promising to get the cost down and maintain performance, they say. No. 160 PROTOTYPE ANIMAL PET SCANNER WITH AVALANCHE PHOTODIODE ARRAYS AND LSO CRYSTALS SI Ziegler, O. Fries, B Pichler, W Pimpl, F Roder, M Schwaiger, E Lorenz, Munich These investigators think that block designs will not work for animal PET and have looked at the feasibility of using one avalanche diode per LSO crystal. Temperature stability was found to be adequate. Energy resolution was 13.4%. Magnetic fields were no problem. They were encouraged to expand the ring detector. No. 161 Sorry, this one did not hold my interest. No. 162 RESOLUTION IMPROVEMENT IN PET USING DETECTOR DEPTH OF INTERACTION INFORMATION M Dahlbom, LR MacDonald, SR Cherry, L Eriksson, ME Casey, UCLA, Los Angeles, Karolinska, Stockholm, CPS Knoxville Layered detectors were studied. Factors of two improvement were noted for 2 layers of LSO or GSO-LSO combination. Details of implementation were not discussed. DOI (depth-of-interaction) is important for future systems. It remains to be seen how to accomplish this in real systems. It should make a big difference. Session 39: Cardiovascular Basic: Imaging Technology No. 229 FOURIER PHASE TOMOGRAPHY FOR DEPICTION OF PATIENT BODY CONTOURS IN GATED SPECT STUDIES. K Nichols, MI Friedman, EG DePuey, H Salensky, Columbia U. NY I was not able to attend this session but this paper looked so interesting I remark on it here in hopes I can follow up on it. The idea seems to be that 3rd order phase analysis of gated tomograms bears close relationship to the body contour. Investigations of the accuracy with phantom images and MRI body contours of humans indicates that there is accurate body contour information in the phase tomograms Session 41: Instrumentation and Data Analysis: Scatter/Attenuation No. 241 IMPACT OF SCATTER CORRECTION ON TUMOR DETECTION IN PLANAR SCINTIMAMMOGRAPHY: A PHANTOM STUDY I Buvat, MC De Sousa, M Di Paola, J Lumbroso, M Rcard, B Aubert, INSERM, Paris, IGR, Villejuif This was an ROC analysis on phantom images and tested two scatter correction methods, single window subtraction and factor analysis. Scatter correction did not help much or only a little at best. Another disappointment for scatter correction enthusiasts. No. 242 OPTIMIZATION OF ACQUISITION AND PROCESSING PARAMETERS IN THE COMPTON FREE IMAGING (CFI) SCATTER CORRECTION TECHNIQUE. M Monville, MK O’Conner, Mayo Clinic, Rochester. CFI is the Elscint scatter correction method. There has been very little published information on it however. This paper looked at the method with some care. While they found it could correct well for scatter it also was susceptible to making artifacts due to rigid requirements on energy corrections. Results seemed to be sensitive to the exact energy range used and on the Compton orders used in the scatter correction. No. 243 SIMULTANEOUS DUAL ISOTOPE F-18 AND TC99M IMAGING CROSS-TALK CORRECTION. JM Baron, P DeMan, IG Zubal, F. Th Wackers, AJ Sinusas, B Zaret, Tel Aviv Medical Center and Yale Medical School, New Haven When Tc-99m and F-18 images are acquired simultaneously the 511 keV photons can downscatter into the Tc-99m window and cause considerable distortion of the true count distribution of Tc-99m. This group used a Park Camera to test a correction method that used a window between 160-174 keV with a correction factor of 1.82. Cross-talk of up to 25% was found and corrected. They recommend this method. No. 244 A NEW THREE-WINDOW TRANSFORMATION CROSS TALK CORRECTION TECHNIQUE FOR SIMULTANEOUS DUAL-ISOTOPE IMAGING. K Knesaurek, J Machac, The Mount Siani Medical Center, NY Like the previous paper this groups corrects for the downscatter into the Tc-99m window. They do a convolution on the extra energy window to take into account the different spatial distribution between the correction window and the Tc-99m window. Restoration filtering is also used to improve the images. No comment on how the noise is degraded. They claim it works good. No. 245 MEAN ATTENUATION COEFFICIENTS FOR ATTENUATION CORRECTION IN EMISSION TOMOGRAPHY. MT Madsen, U of Iowa, Iowa City This paper looked at the errors caused by using the mean mu values for lung and body in SPECT and PET attenuation correction. It was found, from PET attenuation scans, that PET corrections are more critical because of the variation in lung mu values from patient to patient. SPECT was not so critical. In the cardiac region using mean values produced error in <10% for both PET and SPECT. No 246 A GENERAL APPROACH TO NON-UNIFORM ATTENUATION CORRECTION USING EMISSION DATA ALONE SC Moore, MF Kijewski, SP Mueller, Harvard Medical School, Boston and U Essen By using an ML-EM algorithm together with consistency conditions on the projections, the authors were able to successful determine a boundary for use in attenuation correction. All studies were done for a cross- section of the chest. Results are definitely better than the assumption of a uniform elliptical boundary but more work is needed to determine just how good the boundary so determined really is. Noise appears to be well controlled. Session 48: Instrumentation and data analysis: PET II - Lesion Detection No. 283 TUMOR DETECTION USING PLANAR WHOLE-BODY AND TOMOGRAPHY COINCIDENCE IMAGING ON A DUAL-HEAD GAMMA CAMERA Miller, F. DiFilippo, P Maniawski, K. Wilburn, Picker International, Highland Heights The protocol studied was this: Do a 30 min. whole body planar survey of FDG in coincidence mode and target the area for a second 45 min. tomographic acquisition. Reconstruct longitudinal planar slices on 5 mm intervals and as cross-sectional slices. Authors claim it is an efficient way to look for tumors. No data was presented to convince us though. Related: No. 863 and 844 in poster session. Picker was hawking this as PLATO (Planar Longitudinal Tomography) a.k.a. Pho-Con. N0. 284 TARGET-TO-BACKGROUND RATIO THRESHOLD FOR DETECTION OF TUMOR-INFILTRATED LYMPH NODES WITH 2 AND 3-D PET: A DYNAMIC PHANTOM STUDY. RR Raylman, P Kison, RL Wahl. West Va U Morgantown, U Michigan, Ann Arbor By combining C-11 in small spheres and background organs with F-18 the authors studies the detectability of small lesions as the contrast changed over time. Spheres of 3, 6 and 12 mm dia. were filled with C- 11. One set of spheres was place in the axilla region and one in the mediastinum region of a torso phantom. A sequence of images were made in 2d mode and repeated in 3d mode using an Exact scanner. 12 mm sphere always visible, 6 mm visible above 3.5:1 contrast and 3 mm visible above 30:1 contrast. 3d mode helped a bit. Now if someone could repeat this, in a meaningful way for gamma camera PET… N0. 285 BREAST LESION DETECTION WITH PET: A DEDICATED VERSUS A WHOLE-BODY SCANNER R Freifelder, JS Karp, U Penn, Philadelphia The authors showed that a dedicated mammo PET system is better than a general whole body PET system. But is a dedicated mammo ring system better than a dedicated planar, limited angle tomo system? Not so clear except for large breasts the ring seems better. Dedicated planar system can get closer to chest wall, though. Studies continue. No. 286 COMPARISON OF LESION DETECTION PERFORMANCE FOR PET IMAGE RECONSTRUCTION USING CHANNELIZED HOTELLING OBSERVERS RM Leahy, MT Chan, SR Cherry, EU Mumcuoglu, J Czernin, A Chatziloannou, USC and UCLA, Los Angeles The authors compared standard filtered backprojection vs. a maximum a posteriori method using a mathematical observer. Data was real, lesion were real, both acquired on a ECAT 932 scanner. Lesions were placed in 90 locations. ROC curves (area) was the measure of difference. FBP got 0.934 but MAP was better at 0.974. (Numbers in abstract are WRONG.) Seems a small difference for a lot of work. No. 287 LESION DETECTION IN WHOLE BODY FDG PET USING SIMULTANEOUS EMISSION TRANSMISSION ACQUISITION AND OSEM RECONSTRUCTION (SET-OSEM)VERSUS EMISSION-ONLY (EO) AND FBP RECONSTRUCTION: AN ROC STUDY. MJ Fulham, SR Meikle, PK Hooper, KM Silver, Royal Prince Alfred H, Sydney This study involved real PET data on 80 patients. “Is attenuation correction helpful?” seems to be the main question. Results were very clear that attenuation correction is very helpful and the iterative reconstruction helps a lot when the lesion is near a hot object. THIS IS A CRITICAL PAPER, I THINK. No. 288 LIMITS OF DETECTABILITY IN THYROID CARCINOMA: A COMPARATIVE STUDY OF TC- 99M, I-31, AND F-18. YE Erdi, JL Humm, M Imbriaco, SM Larson, Memorial Sloan Kettering Cancer Center, NY Planar Tc-99m and I-131 were compared to F-18 PET and F-18 MCD. We know which is better but the question is how much? The answer: A lot. Tc- 99m and I-131 are equivalent in lesion detectability in this phantom experiment with spheres ranging from 12 to 0.4 ml, with T/NT ratio from infinity to 3:1. Tc-99m and I-131 have trouble below 5.5 ml spheres at 3:1. PET saw spheres to 0.55 ml at 3:1 contrast. MCD saw down to 2 ml spheres. There was a lot of data presented. Hope it makes it into a paper. No. 301 RAPID (<5 MIN) HIGH-RESOLUTION SPECT OF RENAL BLOOD FLOW BY CONTINUOUS INFUSION OF IRIDIUM-191m (Ir-191m) ST Treves, L Fung, AB Packard, RT Davis, N Keane, P Day, RE Zimmerman Children’s Hospital, Harvard Medical School, Boston Continuous infusion with short half-life isotopes is a most promising method to perform SPECT studies. With high count rate capability cameras and proper attention to collection parameters and data processing dramatic advances could be expected. This paper demonstrated such a technique on rabbit models using 5 sec half-life Ir91m. Radiation exposures could be less than with current isotopes. Wednesday Session 55: Instrumentation and Data Analysis: SPECT IV - Scatter Compensation No. 326 THE EFFECTS OF OUT OF FIELD-OF-VIEW ACTIVITY ON ITERATIVE RECONSTRUCTION-BASED SCATTER COMPENSATION IN CARDIAC SPECT. EC Frey, BMW Tsui, U North Carolina, Chapel Hill Authors found, in simulations, that if the length of the axial FOV is too short scatter correction using iterative reconstructions sacrificed accuracy when hot objects were present. The accuracy rapidly improved as the margins above and below the heart reached 2-3 cm. No. 327 ITERATIVE RECONSTRUCTION OF SIMULTANEOUSLY ACQUIRED TC-99M SESTAMIBI / TL-201 CARDIAC SPECT DATA WITH COMPENSATION FOR CROSS-CONTAMINATING EFFECTS. EC Frey, JA Ziffer, BMW Tsui, U north Carolina, Chapel Hill Three corrections methods for correction of the downscatter and lead x- ray contamination problem were studied using both simulation and experiments. All reconstructions were done with fast rescaled block iterative EM algorithm. In two methods a scatter model was incorporated into the algorithm and the Tl and Tc images were reconstructed either simultaneously or sequentially. The third method used a three energy window method with appropriate scaling and blurring of the correction windows. All methods reduced the bias and all methods increased the noise. The simultaneous method had very long reconstruction time. Lead x-ray corrections are important. They are still working on making this ready for regular use. No. 328 EFFECTS OF IMPROVING ENERGY RESOLUTION UPON THE PERFORMANCE OF SCATTER COMPENSATION TECHNIQUES IN SPECT. DJ Kadrmas, EC Frey, BMW Tsui. UNC, Chapel Hill How does better energy resolution translate into improved images? To investigate this dual-window scatter corrections were compared to a model based iterative reconstruction approach to scatter correction with energy resolution varying from 0%to 12%. MCAT simulations were used and contrast recovery and noise were the measures used. The dual-window method profited most from better energy resolution in contrast recovery but had higher noise. The iterative method had lower noise independent of energy resolution and better contrast recovery. No. 329 EFFECTS OF SCATTER CONTAMINATION IN TC-99M/GD-153 SIMULTANEOUS EMISSION TRANSMISSION CARDIAC STUDIES. C Riddell, WC Barker, SL Bacharach, NIH, Bethesda A 90 degree camera geometry with sliding line sources was studied with phantoms and by “artificially” contaminating the heart data with Tc scatter in the Gd window. The conclusion was important for anyone contemplating using this method: the effect of Tc downscatter is very important and the transmission image is seriously deficient, negating the efficacy of the attenuation correction method. (My words, not the authors.) Could scatter correction be avoided by segmenting the Gd derived map and using assigned mu values rather than the biased ones? In an earlier paper (Madsen No. 245) showed that lung mu values were quite variables so this has definite drawbacks. No. 330 CONVOLUTION BASED TRIPLE ENERGY WINDOW CROSSTALK CORRECTION FOR DISA SPECT W Kool, ATM Wilemsen, AMJ Paans, DA Piers. U Hospital Groningen, The Netherlands DISA is Dual Isotope Simultaneous Acquisition, in this case F-18 and Tc- 99m of the heart. Crosstalk correction clearly needed. Like similar authors reported this year and previous years, good results are obtained by taking data in a third window (at 170 kev) and transforming it based on the impulse response function to represent the downscatter at 140 kev. Instrument used was the MS2 with EHE collimators. No discussion of noise, unfortunately. Session 71: Instrumentation and Data Analysis: PET IV - data Correction No. 421 A SPIRAL CT APPROACH TO RECORDING ACCURATE SINGLE PHOTON TRANSMISSION DATA IN PET. Dl Bailey, WF Jones, T Brun, J Young, R Nutt. Hammersmith Hospital, Londn, CTI, Knoxville Using uncollimated Cs-137 source and BGO detectors results in serious scatter problems leading to underestimates of mu. Authors studied use of 2 mm slit collimation on the partial ring ECAT ART system. Results were much better, giving mu values within expected range. Multiple such sources may result in enough flux for rapid attenuation scans. The NaI based PENN-PET does not have much problem with Cs-137 scatter. Is it because of the better energy resolution? No. 422 CLINICAL VALIDATION OF SINGLES TRANSMISSION FOR ATTENUATION CORRECTION IN PET. RA deKemp, WF Jones, CC Watson, MT Aung, RS Beanlands. U of ottawa heart Inst., Ottawa and CTI, Knoxville. Using the ECAT ART these authors found the Cs-137 attenuation correction efficacious and rapid in clinical use. They developed an empirical correction for correction to 511 kev mu values and for effects of scatter. They found the singles transmission scan better than a coincidence transmission scan for resolution and noise standpoint. No. 423 A COMPARISON OF SINGLES VERSUS COINCIDENCE ATTENUATION CORRECTED FDG PET BODY IMAGES. MP Bedigian, R Smith, F Benard, A Alavi, U Penn, Philadelphia These authors also reported significant gains when using a Cs-137 source over coincidence transmission measurements. It appears this is the way to go. Will coincidence gamma cameras also use this method? No. 424 EVALUATION OF A DEADTIME CORRECTION TECHNIQUE FOR SIMULTANEOUS EMISSION- TRANSMISSION DATA ACQUISITION IN PET. Y-C Tai, A Chatziioannou, M Dahlbom, EJ Hoffman, UCLA Los Angeles In the rotating rod coincidence method of attenuation correction (performed simultaneously with collection of emission data) there are significant deadtime issues to be addressed. The authors did address this for the ECAT EXACT machines and found large errors if exact nature of deadtime was not accounted for. Thursday Session 85: Instrumentation and Data Analysis: Image Registration No. 505 Withdrawn No. 506 NORMAL BRAIN METABOLISM PATTERNS COMPUTED FROM UNIFIED PET AND MR IMAGERY TL Faber, JM Hoffman, JR Votaw, ME Brummer, EV Garcia, Emory U, Atlanta The goal is automated analysis of PET brain images for e.g. finding epileptic foci. The tools include MRI brain scans and knowledge of the PET psf and 17 normal volunteers. Normal database for PET scans was compiled using MRI, PET registration, segmentation and warping for 10 of the volunteers. Seven of the volunteers were then compared to the database and found to be essentially “normal”. No. 507 DETERMINATION OF PATIENT BOUNDARIES BY USE OF COMPTON SCATTER DATA FOR REGISTRATION OF ABDOMINAL CT AND SPECT. K Sjogreen, M Ljunberg, L Floreby, K. Erlandsson, S-E Strand, Dept. Radiation Physics, Lund The patient has I-131 on board. How to best get an abdominal contour for registration with CT and to aid therapy planning? Authors simulated the situation and studied various way to get the contour out of the scatter window from 50-150 keV. Abstract and presentation seemed to differ some but the best approach involved some kind of thresholding. Liver activity caused problems. No. 508 A 3-D NON-LINEAR REGISTRATION ALGORITHM FOR FUNCTIONAL SPECT BRAIN IMAGING USING OPTICAL FLOW. C Janicki, J Meunier, J-P Soucy, B Imbert, A Guimond, Hopital Notre-Dame and U Montreal, Montreal. The gist of this paper was that optical flow (a type of non-linear warping) can be used to take already registered SPECT images and improve the registration to the Talairach coordinate system. This aids functional studies across different individuals. No. 509 USE OF VASCULAR STRUCTURE FOR CT/SPECT IN-111 MoAb CYT-356 REGISTRATION IN THE PELVIS. RJ Hamilton, BD Milliken, CA Pelizzari, S Vijayakumar, D Calvin, MJ Blend, U Chicago and U Illinois, Chicago. Tc-99m red blood cell labeling of the vascular system is used to register the SPECT and CT scans. The simultaneously acquired MoAb images are then registered also. A good trick to accomplish a difficult task. A major question: How to ensure that CT and SPECT were done in comparable positions. All the couches and positioning techniques seem to be different! No. 510 USE OF UNIMODALITY REGISTRATION AND FIXED REGIONS OF INTEREST TO IMPROVE DIAGNOSTIC ACCURACY OF BRAIN SPECT BY PROBABILISTIC NEURAL NETWORK. P Maski, FS Prato, DL Silver, J Davis, B Kemp, P Slomka, A Kertesz, RL Nicholson, St. Joseph’s Health Centre, UWO, London, Ontario There were 19 controls and 51 patients in the study. Manual methods were compared to the method based on neural network. With fixed ROIs. Specificity was improved with the automated method. Will apply to larger series. Session 90: Instrumentation and Data Analysis: General I - Advances in Instrumentation No. 535 DEDICATED BREAST IMAGING SYSTEM BASED ON A NOVEL SOLID STATE DETECTOR ARRAY. Be Patt, JS Iwanczyk, Photon Imaging, Inc, MP Tornal, EJ Hoffman UCLA, Los angeles and C Rossington, Lawrence Berkeley National Lab. The design studies for a small CsI planar camera was presented. Si photodiodes would be coupled to 1.5x1.5x6mm CsI crystal matrix. Intrinsic resolution would be 1.5 mm. So far a small portion of the total array has been built and tested. Energy resolution is about 8.7% for this array. The small detector should improve planar scintimammography compared to a standard gamma camera. No. 536 DEVELOPMENT OF MODULAR 64X64 CdZnTe IMAGING ARRAYS WITH MULTIPLEXER READOUT. Jm Woolfenden, HB Barber, HH Barrett, ES Dereniak, JD Eskin, DG Marks, KJ Matherson, ET Young, FL Augustine, U arizona Tucson and Augustine Engineering, Encinitas This was a progress report on an ongoing project to develop a solid state detectors for nuclear medicine imaging. Integrated devices consisting of detector elements, bonding and multiplexer readout incorporating high speed amplifiers and sample and holds have been fabricated. Thermoelectric reduces leakage current and noise. System tests on such elements exhibit ~1 mm resolution. Very encouraging. No. 537 PRODUCTION OF A RADIO-DIRECTED BIOPSY SYSTEM (RDBS) FOR USE IN INTERVENTIONAL NUCLEAR SCINTIGRAPHY. RH Wagner, SM Karesh, GL Dillehay, RE Henkin, Loyla U MC, Maywood, IL That should be radioactive not radio in the title. Needles are made radioactive by electroplating a radioactive metal on them. In this case Tc-99m was used with encouraging results. No. 538 A TIME-OF-FLIGHT MULTI-PROBE SYSTEM FOR QUANTIFYING POSITRON-EMITTING TRACER IN A TARGET REGION. N Satoh, M Inubushi, H Okada, E. Yoshikawa, S Nishiyama, M Futatsubashi, S Nobezawa, Y Ouchi, T Yamashita, Hamamatsu, and Kyoto U. Goal is to measure rCBF noninvasively. Technique studied uses multiple probes and TOF to detect counts in heart region. 2.4 cm region is defined with 160 ps resolution. It seems to be able to handle the activity from standard 0-15 injection. No. 539 A PILEUP PREVENTION METHOD FOR GAMMA CAMERAS WH Wong, H Li, J Uribe, H. Baghaei, N Zhang, J Wang, S Yokoyama, U Texas MD Anderson Cancer Center, Houston This circuit used a variable processing time for pulses based on the interval between gammas. Prototype performance to 2 Mcts/sec with projected performance in a real situation to 3-4 Mcts/sec. It looks promising but it was difficult to evaluate how this would be used by camera manufacturers since it uses the X+, X-, Y+ and Y- gamma camera signals. Most cameras do not produce these signals anymore if they are all-digital. At least that is my understanding. No. 540 EVALUATION OF A CURVED NaI(Tl) DETECTOR FOR SPECT AND PET APPLICATIONS. JA Wear, R Freifelder, JS Karp, U Penn, Philadelphia, and DJ Herr, PS Raby, RS Schreiner, Bicron Corp, Newbury, OH Bicon has been advertising curved detection plates of NaI(Tl) for several years now. I had been wondering if an evaluation of a system using these plates was being undertaken. Well here is the first report I have seen on such an evaluation. Thickness of these plates is 0.375”. A bit thin for any PET work. The report was modestly encouraging but not revolutionary. One advantage over flat plates may be the limited range of the light from single events leading to a reduced light background. Sitz Posters and Posters No. 611 COMPARISON OF IODONE-123 AND IODONE-131 AS A SCANNING AGENT FOR THE DETECTION OF METASTATIC THYROID CANCER. GJ Bautovich, JE Towson, S Eberl, J Turtle, E Chua, R Fawdry, I MacKinnon, D McHarg, Royal Prince Alfred Hospital, Sydney Ten ablated patients with suspected mets were studied. I-123 was scanned 1 and 2 days after administration. I-131 was scanned 3 days after administration. In favor of I-123 scans: Less dose to tissues, better scans, more foci seen. In favor of I-131: I-123 missed one foci. Others have reported similar favorable results with I-123. We need a decisive study to answer this question. No. 618 WHICH IS THE MOST USEFUL TRACER FOR THE EVALUATION OF MALIGNANCY GRADE OF GLIOMA, THALLIUM-201, C-11 METHIONINE, OR F-18 FDG? M Sasaki, Y Ichiya, Y Kuwabara, T. Yoshida, T Fukumura, T Morioka, M Fukui. Kyushu U, Fukuoka Answer: Tl-201 in these 28 patients. No. 631A PERFORMANCE OF DUAL HEAD COINCIDENCE MODE “SPECT”: COMPARISON WITH FDG PET IN ONCOLOGIC DIAGNOSIS P Shreve, RL Stevenson, E Deters, MD Gross, RL Wahl, Ann Arbor VA and U Michigan, Ann Arbor. 22 patients were in the series. ADAC Vertex MCD compared against ECAT 931 or 921 EXACT. Attenuation was used on true PET images but not on MCD. PET was considered gold standard. Breakdown by location shows MCD had great trouble with mediastinum and abdomen. Will attenuation correction help? No. 717 FEASIBILITY OF SIMULTANEOUS STRESS TECHNETIUM-99m SESTAMIBI/ REST THALLIUM-201 DUAL-ISOTOPE MYOCARDIAL PERFUSION SPECT IN THE DETECTION OF CORONARY ARTERY DISEASE. M Nakamura, K Takeda, H Shimizu, Y Saito, Y Nomura, T Ichihara, N Isaka, T Konishi, T Nakano. Mie U and Toshiba, Japan Using the correction method of Moore, et al this group determined that for their test group of 50 patients the method worked well and enhanced throughput and patient convenience. No. 842 ACCURACY OF EVENT POSITIONING IN A PIXELATED BGO ARRAY COUPLED TO A MINIATURE POSITION-SENSITIVE PHOTOTUBE. J Seidel, S Siegel, WR Gandler, MV Green, NIH Bethesda A 9x9 array of isolated BGO detectors (2x2x10 mm) was coupled to a position sensitive PMT. 86% localization to the originating pixel was noted. Promising results for a small animal imager. No. 847 WHOLE-BODY PET IN CLINICAL ONCOLOGY: COMPARISON OF ATTENUATION-CORRECTED AND NON-CORRECTED IMAGES. F Bengel, S Ziegler, N Avril, C Laubenbacher, M Schwaiger, Technische U, Munchen This group says that attenuation correction not helpful to increase detection rates. Their attenuation correction method was not described but they used an EXACT scanner. Noise was a problem. N0 848 THE KFA TIER-PET: A FLEXIBLE PET DETECTOR SYSTEM FOR LABORATORY ANIMALS H Herzog, A Terstegge, S Weber, R. Engles, R. Reinartz, P Reinhart, F Rongen, HW Muller-Gartner, H Halling, Julich A small animal scanner with resolution of 2 mm was reported. Design uses YAP (yttrium aluminum perovskit) scintillator 2x2x15 mm crystals in 20x20 array coupled to a position sensitive PMT. The 2 opposed arrays can rotate 90 degrees. Sensitivity is 1.8 kcps/uCi/ml for 5 cm dia cylinder. Count rates saturation at 2200 cps. No. 849 A NOVEL METHOD FOR THE REMOVAL OF STREAK ARTIFACT FROM PET IMAGES TH Farquhar, G Chinn, CK Hoh, S-C Huang, EJ Hoffman, UCLA, Los Angeles Streaks from high activity in FBP reconstruction are a problem in PET and SPECT. The authors reduced the severity of the artifact by separating the sinogram into high activity and low activity parts. Then they separately filtered and reconstructed the parts and recombined them. It makes a difference. Clever! No. 852 PERFORMING VOLUME IMAGING PET SCANS IN THE PRESENCE OF HIGH RESIDUAL SINGLE PHOTON ACTIVITY. RJ Smith, JS Karp, F Bernard, A Alavi. U Penn, Philadelphia The patient comes for a PET scan with 30 mCi of Tc-99m on board and you have light buildup in your NaI PET camera. How to do the PET scan? These authors succeeded by adding lead filters in front of the PET camera reducing the Tc rates much more than the 511 rates. Should be applicable to any NaI PET imager. No. 863 PLANAR WHOLEBODY COINCIDENCE IMAGING ON A DUAL-HEAD GAMMA CAMERA. F DiFilippo, M Heller, K Kellar, S Miller, R Zahn, Picker International, Highland Heights One patient was scanned on Picker’s PCD system and longitudinal tomographic images produced for different depths. Said to be like PET results. 15 mm lesion detectability claimed. Good enough? I doubt it! See also comments on Picker commercial exhibit and abstract No. 282 and No. 844. No. 865 POSTINJECTION TRANSMISSION MEASUREMENTS IN CARDIAC PET. PK Hooper, S Eberl, SR Meikle, KM Silver, AR Anayat, MJ Fulham, Royal Prince Alfred Hospital, Sydney. Careful comparison with 6 patients and a phantom study showed that the post-injection attenuation correction could work well. No. 869 DEVELOPMENT OF A REALISTIC MULTIMODALITY BREAST AND AXILLARY NODE PHANTOM. Nk Doshi, M Basic, SR Cherry, LJ Pang. UCLA and Radiology Support Devices, Long Beach Authors report on a rigid phantom to aid their research in breast imaging. One problem limiting usefulness may be that breasts are not rigid and certain imaging techniques will not be able to use a rigid phantom like this. No. 876 AN IMPROVED METHOD FOR MEASUREMENT OF GAMMA CAMERA SENSITIVITY FOR HIGH ENERGY PHOTONS. DJ Wagenaar, RE Zimmerman, TC Hill, Joint Program in Nuclear Medicine, Harvard Medical School, Boston Authors propose a method to measure system sensitivity of collimated F- 18 gamma cameras that minimizes the errors caused by penetration. Essence of the methods is to use a small area source (10x10 cm) of Ge-68 at a 40 cm distance with an ROI within the image of the source. Penetration effects are shown to be greatly reduced. No. 878 PERFORMANCE CHARACTERISTICS OF A DUAL HEAD COINCIDENCE CAMERA FOR THE DETECTION OF SMALL LESIONS. SI Ziegler, A Enterrottacher, G Boning, P Nieland, J Kretschklo, M Schwaiger, Technischen U, Munchen and G Muehllehner, E Gualtieri, UGM Labs, Philadelphia ADAC MCD camera was studied with Tc-99m and F-18. For Tc-99m sensitivity is 1840 count/min/MBq and resolution was 5.5, 6.7, 8.2, 9.8, 11.5 mm FWHM at 0, 5 10, 15, 20 cm distance. Coincidence sensitivity is 113k cts/sec for 20 cm dia cylinder. Spheres in a body phantom with F- 18 and F-18 background could be detected if bigger than 17 mm dia. Comparison with real PET showed that with lesion size <13 mm gamma camera PET fails. Tc performance not compromised. No. 879 QUANTITATIVE EVALUATION OF SPECT FIELD UNIFORMITY MT Madsen, U Iowa, Iowa City Author shows that annular averaging about the COR reduces noise in uniformity measurement without destroying contrast of artifact. May have promise in QC because a measure of SPECT uniformity is sorely needed. No. 880 TUMOR DETECTION WITH 511 KEV SPECT IMAGING: A SIMULATION STUDY. MT Madsen, DL Bushnell, D Khan, U Iowa, Iowa City Authors produced a table of SUV needed to detect a spherical lesion against typical body backgrounds for different parts of the body. Useful. Discouraging? E.g., a SUV of 12 is needed to detect a lung lesion. No. 886 THE HARDWARE AND SOFTWARE PERFORMANCE (HASP) SPECT PHANTOM. EM Smith, K Hopkins, U Rochester Med Ctr, Rochester, NY This phantom has virtues in testing SPECT performance. Smith has all the details at emsmith@biophysics.rochester.edu No. 887 SIMULTANEOUS SPECT AND COINCIDENCE IMAGING USING A DUAL DETECTOR SCINTILLATION CAMERA - WORKS IN PROGRESS EM Smith, U Rochester and WK McCroskey, DS Vickers, P Vernon, SMV America, Twinsburg Do Tc-99m imaging and coincidence imaging through the LE collimator. Interesting idea. Was not convinced that it will be useful. No. 888 QUANTITATIVE PLANAR IMAGING METHOD FOR MEASURING RENAL ACTIVITY USING A CONJUGATE-EMISSION IMAGE AND TRANSMISSION DATA. A Kojima, Y Ohyama, S Tomiguchi, M Kira, M Matsumoto, M Takahashi, Kumamoto U, Kumamoto I was impressed with this technique and will be requesting a reprint from the authors. akojima@kaiju.medic.kumamoto.ac.jp No. 897 AN INTERNET-BASED INTERACTIVE NUCLEAR MEDICINE IMAGE DISPLAY SYSTEM IMPLEMENTED IN THE JAVA PROGRAMMING LANGUAGE. NX Phung, JW Wallis, Washington U, St. Louis Looks like a good teaching aid using the Internet. May have uses beyond teaching. Makes heavy use of compression to achieve good performance. See http://gamma.wustl.edu/abs97.html for more. Demo did not work on my laptop from home. Will have to try on my big computer with real Internet connection at work. No. 900 A COLLIMATOR-SPECIFIC TREATMENT OF DATA FOR F-18 SPECT IMAGING. J Li, Z Liang, H Kuan, C Wong, J Cheng, G Wang, D Schlyer, N Volkow, D Harrington, SUNY-SB, and BNL, NY Authors demonstrate improvement of phantom images using low pass filters derived form measured collimator response. Looks interesting. Needs more data. No. 901 THE LIMITATION OF UNIFORM ATTENUATION CORRECTION FOR BRAIN SPECT QUANTIFICATION RZ Stodilka, BJ Kemp, FS Prato, RL Nicholson, Lawson Res Inst and St Joseph’s Health Centre, London Ontario Errors of at least 10% and usually 15% will be faced when trying to quantify brain values. Non-uniform correction is necessary. No. 907 SCATTER ELIMINATION, ATTENUATION CORRECTION, AND RESOLUTION RECOVERY - THREE MAJOR STEPS FOR QUANTIFICATION IN SPECT: FIRST CLINICAL IMAGING. V Rappoport, J Riznar, WS Kim, K Armstrong, CB Lim, Trionix Research Labs, Inc The Trionix approach to these important correction techniques is described. No. 909 No. 944 EVALUATION OF NONUNIFORM ATTENUATION CORRECTION IN QUANTATIVE BRAIN SPECT IMAGING. N Rajeevan, RB Innis, Q Ramsby, SS Zoghbi, JP Seibyl, IG Zubal, Yale u. New Haven Errors were 5-12% if uniform assumption was used. Presumed to be less if nonuniform Chang correction used but numbers not given or I missed it. No. 916 FRAMEWORK OF COMMUNITY-WIDE EVALUATION OF NUCLEAR MEDICINE DATA- PROCESSING METHODS. NH Clinthorne, P Chiao, U Michigan, Ann Arbor Submit your images through a Java enabled browser to a site that will offer processing scripts. Is this the way of the future? No. 917 No. 918 DEVELOPMENT OF A MULTIPLE LINE-SOURCE ARRAY FOR SPECT TRANSMISSION SCANS A Cellar, A Sitek, E Stoub, D Lyster, C Dykstra, D Worsley, A Fung, U British Columbia and Siemens Get a preview of the Siemens approach that will be used in the e.cam. Most promising approach to SPECT attenuation correction. Implemented on a DIACAM here, authors also studied FBP and iterative reconstruction methods. No. 969 BIODISTRIBUTION AND RADIATION DOSIMETRY OF [I-123]FP-CIT. E Busemann-Sokole, J Booij, M Stabin, K deBruin, AGM Janssen, Academic Medical Center, Amsterdam and Oak Ridge Useful information on a new dopamine imaging agent. No. 991 SIZE OR HISTOLOGY? FACTORS AFFECTING UPTAKE IF TC-99M SESTAMIBI IN BREAST LESIONS IN PATIENTS WITH SUSPECTED PRIMARY BREAST CANCER JB Cwikla, RV Barlow, JR Buscombe, SM Kelleher, J Hinton, SP Parbhoo, DS Thakrar, A Deery, J Crow, AJW Hilson Royal Free Hospital, London Authors note that only 80-90% of breast cancer will take up enough MIBI to be imaged. This is because of histological type of the tumor. Ductal carcinomas have the best uptake for MIBI. Scientific Exhibits No. 1300 PINHOLE SCINTIGRAPHY OF THE PEDIATRIC SKELETON LP Connolly, ST Treves, SA Connolly, R Zimmerman, Z Bar-Sever, D Itrato, RT Davis. Children’s H, Boston A 3rd place winner for this important exhibit calling attention to the valuable role of pinhole imaging. No. 1302 A NETWORK-INTEGRATED FUNCTIONAL DATA STORAGE SOLUTION FOR NUCLEAR MEDICINE APPLICATIONS. NT Ranger, P Murtha, L Collins, M Charron, ML Brown, SUNY-SB, Stoney Brook, SMV America, Twinsburg and U Pitt, Pittsurgh. A demo of a clinical information system in nuclear medicine. Sounds like a great idea. Unfortunately the computer was not allowing the demo to function. No. 1309 KINETIC MODELING WITH A SPREADSHEET PROGRAM MM Graham, U Washington, Seattle This received an honorable mention. Nice use of a spreadsheet in nuclear mediicne. Look at ftp.u.washington.edu for more. No. 1314 THE INTERACTIVE PET WHOLE BODY AND BRAIN ATLAS: AN INTERNET BASED MULTIMEDIA LEARNING SYSTEM FOR PET. Ss Gambhir, j Strommer, C Shen, A Panchal, K Aprk, V Litman, C Ho, R Dennis, H Lo, G Lacson, S Choi, H Tiet, ME Phelps, UCLA First Place! Look at http://laxmi.nuc.ucla.edu:8000/html_docs Unfortunately this computer was down when I found it and I could not have a look. The group did Let’s Play Pet and I suppose this is very good, too. No. 1318 A PORTABLE NUCLEAR ANGIOGRAPHY CAMERA WITH ON-BOARD SHORT-LIVED RADIONUCLIDE GENERATOR. JL Lacy, D Dai, C Martin, Proportional Technologies, Inc., Houston Honorable Mention! This table was manned and it worked! The MWPC camera for first pass cardiology lives! Contact them. No. 1319 DEMONSTRATION OF A WORLD-WIDE-WEB-BASED VIEWER FOR DICOM-FORMATTED MEDICAL IMAGES. RE Wendt III, JD Hazle, DF Schomer, DA Podoloff, U Texas MD Anderson Cancer Center, Houston Honorable Mention! Java and all that: http://recon.mdacc.tmc.edu/snm97demo No. 1320 MULTIMODALITY BRAIN IMAGE REGISTRATION USING A 3-d PHOTOGRAMMETRICALLY DERIVED SURFACE OR Mawlawi, BJ Beattie, JL Humm, RG Blasberg, SM Larson, Memorial Sloan Kettering Cancer Center, NY Second Place! Use light to get the face contours to register to MRI and thence to PET. Great results posted.