Personal Notes

SNM 96

Denver, Colorado

1 June 1996 to 6 June 1996

by Robert E. Zimmerman

Saturday

American College of Radiology Nuclear Medicine Accreditation Committee

I'll bet that many of you did not realize that nuclear medicine clinics may go the way of mammography clinics. Yep! Accreditation may be just around the corner. ACR is planning for it, working on the criteria, etc. There is still several years before you will hear anything significant about this. But you will hear more.

Nuclear Medicine PACS

A small group of nuclear medicine PACs enthusiasts met with the encouragement of Nicole Ranger from Pittsburgh. Nicole is working hard on pulling material together on all aspects of PACS for the nuclear medicine community. The major goal is to educate the consumers (nuclear medicine professionals) and manufactures on the needs and specifics of nuclear medicine PACS. This group will be a focus group of the Computer & Instrumentation Council and a task group of the American Association of Physicists in Medicine (AAPM).

Sunday

Sunday turned out to be a perfect day to visit the Rocky Mountain National Park. We were lucky to have a very experienced native guide on our quick visit. There was still lots of snow and wildlife, right at roadside. Our hike was blessed with plenty of sun, some snow and beautiful mountain lakes and streams. If ever in the Denver area make time to visit this beautiful park.

ABSNM held a board meeting Sunday evening. The new president is Dr. Ahluwalia who takes over from Dr. George Zubal. George worked hard on the revitalization of the ABSNM during his tenure. The organization is trying to make itself more attractive to nuclear medicine scientists i.e. physicists, pharmacists, health physicists and others who work in the physical and chemical side of the nuclear medicine profession. It has become very clear to me that certification of all professionals will be increasingly necessary for many of us as we continue to work in nuclear medicine. ABSNM provides a very reasonable option for this certification.

ABSNM also held 20th anniversary reception buffet for all diplomates and friends on Monday evening. Congratulations for 20 years of service to nuclear medicine professionals.

Monday

There was several themes that ran through this meeting. Dominant in my view was the impact of coincidence detection with SPECT cameras. Not more than one unit had been sold, installed and put into regular use but it has inspired numerous imitators. ADAC introduced the modern version about 18 months ago. Most companies are talking about it and trying to come up with a marketing campaigns. At least one company is actively NOT talking about it. (Park can do it better with coded apertures, see below.) Another theme is the wide role F-18 (as FDG) seems to have in the future of nuclear medicine. How many of us can actually do this back home? Well, maybe soon. Attenuation with scatter correction and ordered subsets were much talked about throughout the physics sessions.

The move to special meetings (breakfasts, receptions, dinners) for physicists is interesting. ADAC started this about 2 years ago. This year Siemens had a breakfast, Park had an evening supper and ADAC continues with an evening buffet. There may have been others. This is in addition to the user meeting, which happened on Sunday, this year, I think. I was in the Rocky Mountain Park. Most sensible thing I did that week (see above). It is flattering for the physicists to be invited to hear all this information in more detail than you could learn from the exhibit floor. But it will obviously be impossible to do it all. Let me make a proposal to all the manufacturers. Widely advertise your website. Put in-depth information on the website. Provide for a BBS-like Q and A dialog to be also on the website, something like Lunis. Let your technical people participate in the dialog with the users. Then let us have the early mornings and evenings free at SNM and RSNA. You get your story out to the physicist and we can enjoy the meeting and not feel guilty about missing your latest information, 'cause all we have to do is look at your website.

Plenary Session

As I age I find that I spend more time at these plenary sessions. Many times I know the speakers, know the awardees and understand the work for which they are receiving the awards. It makes the session more interesting. I used to think they were something of a distraction from the main part of the meeting. Interesting how we change over the years.

Theme of the meeting was Nuclear Medicine: 100 years in the making. The theme is inspired by Henri Becquerel's discovery of radioactivity in 1896. Henry Wagner reviewed the history of nuclear medicine from that point in time to the present. Dr Louis Sokoloff , of NIMH, reviewed the history and development of functional brain imaging in the featured plenary session talk. Starting with the 1948 groundbreaking work of Kety with nitrous oxide through recent work with FDG was presented by one of the history makers.

Tom Budinger, well known to physicists all over the nuclear medicine world, was awarded the Hevesy award in honor of his work at the Donner Labs, Berkeley. Dr Tom Miller presented the award and reviewed Budinger's long, productive scientific career.

The Abersold award went to Kenneth Krohn a chemist at University of California for his contributions to F-18 and Tc-99m chemistry. The award was presented by Michael Welch of St. Louis.

The Cassen award. was presented by Abass Alavi to David Kuhl in recognition of his early work in developing tomographic instrumentation, starting in the 1960s, even before there was computed tomography.

Scientific Sessions (and other meetings)

Did you ever wish that the SNM would make a machine readable version of the abstracts available. In preparing this I certainly could have used it. Just for the titles and authors. It would have been very useful. I suggest putting it on the SNM website next year.

FDG imaging seems to be taking the world of (SPECT) oncology by storm. Everyone want to do it, and cheaply. Just how useful the collimated approach is going to be is still an open question in my mind. More below.

There was a lot at this meeting concerning the imaging of F -18 on hybrid cameras. ADAC has moved to a 5/8" crystal and faster electronics. It is out of alpha trials and into beta trials as I write. No other company is as far along in the development process. Siemens is talking about, in 2-3 years, a new scintillator block detector doing hybrid studies. It would consist of YSO sandwiched to LSO. See below.

Attenuation correction is still receiving a lot of attention but with increased emphasis on scatter corrections and resolution recovery. As has been stated in several places recently, attenuation correction is no panacea. It does not suddenly make diagnosis of CAD more accurate. In my opinion, neither will scatter correction or resolution recovery. There is still the INTERPRETATION of the images. If the physician has not been trained to interpret the new, improved images there is no way diagnosis will be improved. If incorporating the improvements into your practice, you must go through the necessary self-training to understand the new images. For that reason those correction methods which allow for the preservation of the uncorrected, familiar image should have a real edge at this early stage.

Session 4: Instrumentation and Data Analysis: Young Investigators Competition

This session included papers of young members of the profession who compete to win a prize ($500?). Hence the papers cover a range of subjects.

No 18 HIGH ENERGY SLIT APERTURE AND PINHOLE SPECT, A PROPOSED METHOD FOR CORRECTING APERTURE PENETRATION TO ENHANCE RESOLUTION AND MINIMIZE PENETRATION BACKGROUND MC Worbel, NM Clinthorne, JA Fessler, Y Zhang, WL Rogers U Michigan, Ann Arbor

SPRINT-II, with its slit collimators providing magnification, can be useful for imaging small animals. This penetration becomes a problem when the isotope is F-18. The authors describe considerable success by taking a scan with the slits blocked by U to estimate, and correct, the penetration component. They propose that it could be used in more general imaging situations, also.

No. 19 ANALYSIS OF THE INVERTIBILITY AND NOISE PROPERTIES OF MULTIPLE ENERGY WINDOW PROJECTION DATA IN SPECT DJ Kadrmas, EC Frey, BMW Tsui U North Carolina, Chapel Hill

Can scattered photons help us? This paper reported efforts at reducing noise when performing scatter correction. In their simulation they found that including counts below 119 Kev increases noise in reconstructed images.

No. 20 WHOLE BODY PET ATTENUATION CORRECTION WITH TRANSMISSION IMAGE SEGMENTATION A Chatziioannou, M Dahlbom, CK Hoh UCLA School of Medicine

These authors found that segmentation of soft tissue only in the transmission images led to lower noise in the corrected images.

No. 21 FULLY AUTOMATED DETECTION OF THE MID-SAGITTAL PLANE FOR Tc-99m-ECD BRAIN SPECT IMAGES IN 6 DEGREES OF FREEDOM J Tsao, A Stundzia, M Ichise Mount Sinai Hospital and U Toronto, Toronto

This was the clear winner in the competition. Extremely well presented, addressed a significant problem with impressive results. In order to put a 3D brain image into the Talairach Proportional Coordinate System it is necessary to find the mid-sagittal plane. The automated method here can find the plane in SPECT, PET and MRI scans rapidly and with high accuracy and precision.

No. 22 QUANTIFICATION OF MEDIAN AND PHRENIC NERVE BRAIN ACTIVATION WITH Tc-99m ECD SPECT AND MRI USING IMAGE REGISTRATION PJ Slomka, UA Zifko, R Reid, Victoria Hospital, U Western Ontario, London, CANADA

The authors extended their previously reported technique for hearts to brain imaging. They found that electrical stimulation increased ECD amounts by about 20% for local regions of the brain. Registration with MRI provided anatomical reference.

No 23 THE EFFECT OF INPUT FUNCTION SHAPE AND IMAGE ACQUISITION RATE ON THE ESTIMATION OF KINETIC PARAMETERS FOR DYNAMIC CARDIAC SPECT SG Ross, A Welch, GT Gullberg, RH Huesman U Utah, Salt Lake City and U California, Berkeley

Using simulation studies, these authors found that the acquisition parameters strongly affected the bias in the estimate of kinetic parameters. High acquisition rates caused bias in the kinetic parameters because of increased noise. In certain cases longer acquisition rates led to reduced bias along with reduced storage requirements.

Session 11: Instrumentation and Data Analysis: SPECT - Attenuation

It is generally accepted (believed, hoped) that when attenuation, scatter and resolution corrections are applied to cardiac SPECT more accurate, and hence more useful images will result. One of the main problems is to get all these corrections done accurately, with no noise or artifacts and do it all quickly. It has been tough. At this meeting quite a few papers addressed this issue. Many of them were in this session.

No. 60 PERFORMANCE OF THE METZ INTRINSIC ATTENUATION CORRECTION METHOD IN PARAMETER ESTIMATION TASKS MF Kijewski, SC Moore, SP Mueller Brigham & Women's Hospital, Harvard Medical School, Boston, VA Medical School, West Roxbury and U Essen

Using the Cramer-Rao bound for several parameter estimation tasks, the authors found that an exact attenuation correction method recently reported by Metz is significantly better than previously reported techniques (Bellini, Gullberg, Treitak & Metz). Furthermore they confirm that this new method is nearly optimum and there is little expectation for significantly better methods (better in the sense of increased estimation performance) to be developed.

No. 61 COMPENSATION FOR ATTENUATION AND 2D DETECTOR RESPONSE IN CARDIAC SPECT IMAGING SJ Glick, MA King U Mass, Worcester

The method described in this paper is valid for only the heart, not surrounding tissues. This is characteristic of several current methods. Scatter corrections are not addressed. Bellini attenuation correction was performed and detector response is compensated for using the frequency distance principle in an iterative restoration technique in projection space. Reconstruction is by Fourier backprojection. It was found that more accurate bulls-eye plots and better noise properties resulted form the studied algorithm

No. 62 ANALYTICAL NON-UNIFORM ATTENUATION COMPENSATION IN TRUNCATED FAN-BEAM CARDIAC SPECT IMAGING W Xia, SJ Glick, T-S Pan U Mass, Worcester

Can an analytic attenuation correction method approach the ML-EM iterative reconstruction results? These authors found, in a simulation study, that the Bellini attenuation correction method and Fourier backprojection reconstruction with a 3-head, fan beam collimated system using a line source, can produce bulls-eye polar maps comparable to the ML-EM algorithm. The truncation in the attenuation map only slightly degraded performance.

No. 63 TOWARDS ACCURATE ATTENUATION CORRECTION WITHOUT TRANSMISSION MEASUREMENTS A Welch, R Clack, PE Christian, GT Gullberg U Utah, Salt Lake City

Can the emission data alone provide enough information for attenuation correction? The authors found, with real experimental data, that in brain SPECT corrections worked pretty well but that for heart SPECT corrections were a problem.

No. 64 COMPARISON OF TRANSMISSION MEASUREMENTS IN PET AND SPECT JR Saffer, RJ Smith, JS Karp U Penn, Philadelphia

The authors compared F-18 attenuation correction on a real PET system with attenuation correction on a Picker STEP system. They found that the PET system could determine the attenuation map with an accuracy of about 5% while the accuracy of STEP was only 20%. This may be due to scatter. They found accuracy improved with reduced pixel size and indicated 128 x 128 matrices were preferred for STEP.

No. 65 CORRECTING FOR EMISSION CONTAMINATION IN SPECT SCANNING LINE SOURCE TRANSMISSION DATA TG Turkington, M Wilk, N Wainer, RE Coleman Duke U, Durham and Elscint, Haifa

The authors discussed a means of scatter correction for the transmission data from scanning line source measurements performed according to the method of Bailey and Hutton. They found that the correction for Tc-99m scatter into the Gd-153 window substantially improved the attenuation correction results.

Tuesday

Session 19: Instrumentation and Data Analysis: SPECT Scatter

No. 108 EVALUATION OF SPECT SCATTER COMPENSATION METHODS BY THEIR EFFECTS ON PARAMETER ESTIMATION TASKS SC Moore, MF Kijewski, SP Muller, TR McIlvain, RE Zimmerman VA Medical Center, West Roxbury, Harvard Medical School, Boston and U Essen

Three optimized scatter correction methods were evaluated, using a maximum-likelihood procedure, with regard to performance in activity quantitation and lesion detection tasks. The optimum k subtraction factor for the Compton window subtraction method was determined. The widths of the energy windows were optimized in the TEW method. Optimized weights for a generalized spectral fitting method from 92-160 keV were obtained. It was found that overall the best performance was obtained with the spectral fitting method having bias < 4% and the lowest variance. TEW had bias of <7% but higher noise than the Compton window method, which had bias ranging from -2 to 22%. The author concluded by questioning the value of scatter correction for these tasks.

No. 109 EVALUATION OF THE EFFECT OF SCATTER SUBTRACTION ON LESION DETECTION IN HEPATIC SPECT IMAGING DJ de Vries, MA King, EJ Soares, BMW Tsui, CE Metz U Mass, Worcester; UNC, Chapel Hill and U Chicago

The authors investigated the effect of scatter correction on signal detection. Real observers evaluated simulated lesions in a simulated liver background. The dual photopeak window (DPW) scatter correction method was used. ROC curves were derived from the observer scores. No significant difference in the ROC curves for the primary photon images and the scatter corrected images was found. The authors concluded that improved quantitation does not lead to improved detection.

No. 110 A PRACTICAL METHOD FOR SCATTER AND ATTENUATION CORRECTION IN TL-201 MYOCARDIAL SPECT USING A DUAL-HEADED SPECT SYSTEM J Hashimoto, A Kubo, K Ogawa, N Motomura, H Hasegawa, T Ichihara School of Medicine, Keio U and College of Engineering, Hosei U, Tokyo and Toshiba Medical Engineering Lab., Tochigi

Six energy windows were used, three for Tl-201 and three for the Tc-99m transmission source. The TEW correction method was used for scatter and spilldown correction. The transmission source was an "area" source composed Tc-99m filled tubing. The method was reported to be practical and effective.

No. 111 AN INVESTIGATION OF COMBINED ATTENUATION AND SCATTER COMPENSATION FOR TL-201 CARDIAC SPECT EC Frey, BMW Tsui, UNC, Chapel Hill

The authors tackled the often repeated story that attenuation correction causes spurious increase in counts in cardiac regions. They studied scatter and attenuation with simulation methods and applied two scatter correction methods: a deconvolution method and a model-based approach. They found that both methods improved the spurious count problem with the model-based approach being better but costing more computer time.

No. 112 A METHOD FOR ESTIMATING THE SCATTER RESPONSE FUNCTION FOR FAN-BEAM SPECT IN HOMOGENOUS AND INHOMOGENEOUS MEDIA J Li, EC Frey, BMW Tsui, GT Gullberg UNC, Chapel Hill, and U Utah, Salt Lake City

The authors extend previous work in Slab Derived Scatter Estimation with parallel hole collimators to fan beam collimators. They found that the technique has some problems when used with inhomogeneous media and Tl-201. The results for Tc-99m were better. More importantly, for me, they concluded that it is not enough to just do attenuation correction. Scatter correction must also be done.

No. 113 CORRECTION FOR DOWNSCATTER FROM F-18 IN DISA (FDG/MIBI) SPECT JA Patton, JW Forrester, MP Sandler Vanderbilt U Medical Center, Nashville

DISA means dual isotope, single acquisition. There is concern that the spilldown from F-18 will cause diagnostic errors in the MIBI interpretation. A third energy window at 170 keV +/- 10% was used to estimate this spilldown. The subtraction factor was 0.67. Phantom measurements indicate the errors were reduced. In patients the images looked better but no diagnostic improvement could be detected.

Instrumentation Council Business Meeting:

Margaret Daube-Witherspoon presiding.

New officers are Fred Fahey, Bowman Gray, Winston-Salem, NC, president-elect and Jim Halama, Loyola U, Chicago, secretary-treasurer.

Steve Bachrach called our attention to the new interchange format for nuclear cardiology based on CD-ROMs. It is based on the DICOM file format and allows images from any modality to be viewed with a universal reader. He was passing out free CDs with samples. Manufactures now need to support it. Steve ran a full session on this later in the week to explain more fully what this does for nuclear medicine.

Nicole Ranger asked that a PACS Focus group be formed. It was. See above for what this group is trying to do.

Another focus group on telemedicine headed by Trevor Cradduck will be writing a final draft this summer.

The 1997 mid-winter meeting will be in Palm Springs. Fred Fahey may be running this meeting. Contact him with your ideas. Jack Correia may also be working on this. Contact him, too, if you have any topics that can attract physicians and physical scientists to Palm Springs.

The SNM program committee is entertaining ideas for new presentation methods at the annual meeting. For example some type of interactive presentation may be tried in the future. How about those walking poster sessions? Do you like them? Let them know. See below for my comments on them.

Session 34: Instrumentation and Data Analysis: General 1

No. 200 DESIGN OF A SMALL SCINTILLATION CAMERA WITH PHOTODIODE READOUT FOR IMAGING MALIGNANT BREAST TUMORS CS Levin, EJ Hoffman, MP Tornai, LR MacDonald UCLA, Los Angles

If mammoscintigraphy is ever to prove useful special purpose cameras will have to be developed specifically to image the breast with the highest possible spatial resolution consistent with useable sensitivity. The UCLA group reported their progress with a CsI scintillator with a silicon photodiode readout which they hope will provide intrinsic resolution of <2 mm. They admit that photodiode noise will be a major problem.

No. 201 MURINE NUCLEAR ANGIOGRAPHY WITH MULTIWIRE PROPORTIONAL GAMMA CAMERA AND Ta-178 JL Lacy, TA Osipova, E John Proportional Technologies, Inc., Houston

Dr Osipova reported on their Monte Carlo simulation of pinhole imaging on the multiwire proportional camera. With a magnification of more than 10, two mm in vivo resolution is predicted and they feel that useful mouse heart images will be obtained.

No. 202 EVALUATION OF A 48x48 CdZnTe ARRAY WITH MULTIPLEXER READOUT FOR HIGH-RESOLUTION IMAGING JM Woolfenden, HB Barber, DG Marks, HH Barrett, ES Dereniak, JD Eskin, KJ Matherson, ET Young U Arizona, Tucson

The Arizona group updated us on the FASTPECT - semiconductor version. By summing adjacent 125 um pixels in the 48x48 array 7.5% energy resolution was obtained with a spatial resolution of 214 um at 140 keV. Next readout will consist of 64x64 array with 380 um pixels.

No. 203 INITIAL COINCIDENCE IMAGING EXPERIENCE WITH A SPECT/PET DUAL HEAD CAMERA EC Glass, P Nelleman, H Hines, MA Mandelkern, WH Blahd West Los Angeles VA Medical Center and ADAC Labs, Milpitas

This is the institution that has the ADAC prototype SPECT/PET camera. The current prototype has crystal thickness of 5/8" and improved electronics to achieve much higher count rates than the original prototype (see below). Dr Glass reported on his comparison of the prototype camera to a real PET system. He feels that the thicker crystal and improved electronics have lead to a machine that is significantly better than the original and is useful for oncology, brain and heart imaging.

ADAC Physics Reception:

Attenuation Correction

ADAC presented technical details on their latest products at an evening buffet. The ADAC collaborators presented some details of their ongoing clinical trials. Mark Groch of Northwestern told us about his experiences with the attenuation correction - present generation. While, in general, it worked pretty well he did notice some errors due to downscatter of Tc photons into the Gd window. Jim Cullom, Emory U. told us about improvements he was working on that would provide scatter correction, downscatter correction and resolution recovery. He uses a ML reconstruction method.

SPECT/PET Camera

Gerd Muehllener reported in considerable detail regarding the latest PET coincidence camera improvements. The crystal thickness is 5/8" an the counting rate capability has been improved to 2,500,000 cps max. singles rate. The information was presented:

Resolution in coinc 4.7 mm
Sensitivity 100 kcps/uCi/cc
Singles deadtime 5% at 1 Mcps
Max singles rate >2.4 Mcps/det
Crystal Thick 10 mm 16 mm
At 511 keV:
Intrinsic Sens 28% 45%
Coinc Sens 7.8% 20.2%
Coinc Frac 0.8% 2.0%
Singles Mode:
Intrinsic 3.5 mm 3.9 mm
System Res 7.05 mm 7.25 mm

Dr Glass again told us about his promising clinical results.

The results are impressive and ADAC is clearly leading this race to get coincidence imaging working on a gamma camera. But I hope we can hear that it is routinely available soon. It is getting to be a long wait.

Wednesday

Session 41: Instrumentation and Data Analysis: SPECT Reconstruction

No. 239 ACCELERATED SPECT IMAGE RECONSTRUCTION USING A MODIFIED ORDERED SUBSETS EXPECTATION MAXIMIZATION METHOD: A FAST LINE SEARCH STRATEGY YH Lin, AJ Sinusas, JA Browne, FJ Th. Wackers Yale U., New Haven

Missed this paper. This was the first of lots that used the ordered subsets method to accelerate the EM iterative solution.

No. 240 ACCELERATED FULL 3D SPECT RECONSTRUCTION USING DUAL MATRIX ORDERED SUBSETS C Kamphuis, FJ Beekman, MA Viergever, PP van Rijk U Hospital Utrecht

Including the scatter psf in iterative fully 3D ML-EM has been shown to work better than subtraction based scatter correction. The problem is the tremendous amount of computation. By using separate projection and backprojection matrices with ordered subsets EM speed increases of 170 were realized.

No. 241 INFLUENCE OF ORDERED SUBSET RECONSTRUCTION AND ELLIPTICAL ORBITS ON SPECT 3D SPATIAL RESOLUTION T-S Pan, D-S Luo, MA King U Mass, Worcester

These investigators reported a factor of 10 reduction in computation time with ordered subsets-ML vs ML-EM when doing collimator blur correction with attenuation correction for circular orbits. Elliptical orbits take two or more ML-OS iterations to reach the ML-EM solutions. This was a noise free simulation study.

No. 243 MEAN-VARIANCE ANALYSIS OF FAST ITERATIVE RECONSTRUCTION ALGORITHMS FOR ATTENUATION COMPENSATION IN CARDIAC SPECT DS Lalush, BMW Tsui, U North Carolina, Chapel Hill

These authors also confirmed the great utility of ordered subsets -EM algorithm. They found that the speed up was from 5-30 times over ML-EM.

No. 244 RAPID 3-D PROJECTION IN ITERATIVE RECONSTRUCTION USING GAUSSIAN DIFFUSION JW Wallis, TR Miller, MM Miller, J Hamill Washington U school of Medicine, St. Louis, Siemens, Hoffman Estates

These authors reported on a way to efficiently incorporate blur and attenuation correction using a Gaussian diffusion method. This is an approximate method that seems to agree well with the more complete and detailed convolution filtering method.

Session 51: Instrumentation and Data Analysis: General II

No. 299 A COMPTON SCATTER CAMERA DESIGN FOR IN VIVO MEDICAL IMAGING OF RADIOPHARMACEUTICALS: RESULTS OF SIMULATION STUDIES RC Rohe, JD Valentine U Cincinnati, Cincinnati

M. Singh has worked extensively with Compton scatter gamma camera using a Ge detector as the first stage and a NaI camera as the second stage. These authors have extensively studied Compton scatter cameras using a Monte Carlo simulation and conclude that having a Si camera as the first stage and a CdZnTe as the second stage offers potential performance advantages particularly at higher energies.

No. 300 APPLICATION OF DPW TO A DIGITAL, STATIONARY, DEDICATED BRAIN CAMERA RE Zimmerman, BB Williams, KH Chan, SC Moore, MF Kijewski Brigham & Women's Hospital, Harvard Medical School, Boston and VA Medical Center, West Roxbury

The authors attempted to implement the dual photopeak window (DPW ) scatter correction method on a dedicated brain camera (Ceraspect). A stable, unique calibration curve could not be obtained. Simulation of the calibration process confirmed the difficulties. The DPW method may not have sufficient stability to be practical, especially for a brain system.

No. 301 COLLIMATOR DESIGN AND IMAGING GEOMETRY FOR IMPROVED MAMMOSCINTIGRAPHY BMW Tsui, XD Zhao, D Wessell, K LaCroix, NP Peterson UNC, Chapel Hill and Asheville Imaging Center, Asheville

Pinhole collimator! Simulation and phantom studies confirm that resolution of about 4 mm and sensitivity of about 2x the HR collimator can be obtained at the center of the breast. Modifications to a camera could permit compression and biopsy. Encouraging!

No. 302 DEADTIME OF TWO MULTI-HEAD ANGER CAMERAS IN I-131 DUAL-ENERGY-WINDOW ACQUISITION MODE KF Koral, KR Zasadny, RJ Ackerman, EP Ficaro U Michigan Medical School, Ann Arbor

The authors were interested in quantitatively imaging therapy doses of I-131 labeled monoclonal antibodies with multiple head gamma cameras and were concerned about deadtime in the dual window mode. They performed experiments on a Picker 3-head system and on a Siemens MS-2. The paralyzable model was used. For less than 30 mCi in the field they found that accurate dead time corrections could be performed and that the Siemens was a considerably faster camera.

No. 303 CHARACTERIZATION AND REJECTION OF ANNIHILATION GAMMA-RAY BACKGROUND FOR A SURGICAL BETA RAY CAMERA THAT UTILIZES POSITRON EMITTING RADIOPHARMACEUTICALS CS Levin, MP Tornai, LR MacDonald, EJ Hoffman UCLA School of Medicine and UCLA Physics Dept.

A 1.2 cm diameter beta ray camera using CaF2(Eu) is used to detect tumors emitting positrons or electrons. There is a problem with annihilation photons when betas are used. Rejection methods using beta-gamma coincidence techniques were studied and reported. The background can be reduced using these techniques.

No. 304 BISMUTH-213 IMAGING: PRE-CLINICAL CHARACTERIZATION OF AN ALPHA-PARTICLE EMITTING RADIONUCLIDE G Sgouros, JL Humm, M McDevitt, J Kennedy, R Schumaker, A Pagliarulo, SM Larson, DA Scheinberg Memorial-Sloan Kettering Cancer Center, NY and ADAC Labs

Bi-213 emits alphas and a 440 keV gamma ray. In order to image the 440 keV gamma ray with HEHR collimators the gain of the gamma camera had to be adjusted. After adjustment the Bi-213 could be imaged successfully.

Session 53: Cardiovascular - Clinical: Attenuation Correction

No. 311 CLINICAL EVALUATION OF ATTENUATION AND SCATTER COMPENSATION IN TL-201 SPECT WH McCartney, BMW Tsui, KF Adams, DP Lewis, DS Lalush, LS Bujenovic, KJ LaCroix, RE Johnston, JR Perry, AHR Lonn, S Maniam, R Culp UNC, Chapel Hill, GE Medical Systems, Milwaukee

I missed the presentation of this paper but the abstract was so interesting I include it in this summary. Authors had 55 patients, 20 of whom had cath. The instrument was a GE L-shaped Optima with scanning Gd-153 line sources. Sequential TCT data acquisition followed the stress study and 4 hours later the redistribution study. There were 3 reconstructions done. Sensitivity and specificity are given below:

Reconstruction Method Sens Spec
FBP w/o compensation 83 58
ML-EM w/ AC 80 72
ML-EM w/ AC & SC 88 67

Clearly something is happening but is it all good? Does it need resolution recovery?

No. 312 CLINICAL EVALUATION OF SIMULTANEOUS ACQUISITION OF TRANSMISSION AND EMISSION DATA USING TECHNETIUM-99m TRANSMISSION SOURCE ON THALLIUM-201 MYOCARDIAL SCINTIGRAPHY S Tomiguchi, Y Oyama, T Kira, M Kira, R Nakashima, A Tsuji, A Kojima, M Takahashi, Y Inobe, H Yasue, Kumamoto U Medical School, Kumamoto

In a study with 21 patients using a fan beam collimator with a 3-head camera system the investigators found significant differences between corrected and uncorrected using the ML-EM algorithm. Some of these differences were attributed to scatter. They concluded that BOTH image sets are needed to make an accurate diagnosis.

No. 313 DIAGNOSTIC ACCURACY AND IMAGE QUALITY OF A SCATTER, ATTENUATION AND RESOLUTION COMPENSATION METHOD FOR Tc-99m SESTAMIBI CARDIAC SPECT: PRELIMINARY RESULTS SJ Cullom, RC Hendel, L Liu, EV Garcia, ML White, H Kiat, DS Berman Emory U, Atlanta and Northwestern U, Chicago and Cedars-Sinai Medical Center, Los Angeles

This is early results from a collaborative study with ADAC Vantage attenuation correction. Only 16 low-likelihood (LL) patients and 6 patients with CAD were studied with stress MIBI. Comparison included routine FBP vs standard Vantage attenuation correction with ML reconstruction vs Cullom's novel scatter, attenuation and spatial resolution corrections program. A Mean Summed Stress Score for 20 segments and Image Quality were scored: The numbers:

MSSS IQ
LL CAD
FBP 4.3 14.6 2.3
ML 2.0 12.3 1.3
SAR 1.8 13.6 1.2

Although the spatial resolution correction was done with the frequency distance principle the noise problems seen by others were not observed. While admitting the numbers are too low, Cullom is encouraged to explore his more complete correction method.

No. 315 COMPARISON OF SUPINE, PRONE, AND ATTENUATION CORRECTED STRESS Tc-99m SESTAMIBI MYOCARDIAL PERFUSION SPECT H Kiat, S Reuter, K Van Train, M Patterson, J Areeda, X Kang, G Germano, RC Hendel, JD Friedman, DS Berman Cedars-Sinai, Medical Center, Los Angeles and Northwestern U, Chicago

Presented by Berman, before No. 314 because of scheduling problems. Some people think that prone imaging obviates the need for attenuation correction by decreasing diaphragmatic attenuation. In this study with 20 low-likelihood patients they found near perfect agreement between supine images with attenuation correction and prone imaging with no correction. Maybe we only have to roll the patient over and stop all this effort. In defense of more physics work, there was no correction for Tc spilldown to Gd window and with stress imaging there is low liver uptake. So the results are not conclusive.

No. 314 DIAGNOSTIC ACCURACY OF 180 AND 360 RECONSTRUCTIONS FOR UNCORRECTED AND ATTENUATION CORRECTED TC-99M CARDIAC SPECT EP Ficaro, SR Pitt, PE Wawrzynski, PA Rose, JR Corbett U Michigan, Ann Arbor

This Ann Arbor group has a Picker 3-head camera with STEP hardware but only use only one fan beam collimator and 2 parallel hole collimators. This permits them to acquire datasets and calculate straight uncorrected or corrected heart scans. They used this setup to test 180 vs 360 rotations. The line source was Am-241. They found that there was little difference between 180 and 360 with or without attenuation correction AS LONG AS PROPER DATABASES WERE USED. I guess it is obvious that we need to understand and have a knowledge base to make decisions.

No. 316 QUANTITATIVE ANALYSIS OF SPECT IMAGING IN COMBINATION OF FIRST-PASS RADIONUCLIDE ANGIOGRAPHY PERMITS IDENTIFICATION OF INFERIOR WALL ATTENUATION FROM SCAR MYH Shen, AJ Sinusas, Y-H Li, P DeMan, JA Mattera, AF Osman, BL Zaret, FJ Th Wackers Yale U, New Haven

Using a dedicated first pass camera, this group found a quantitative first pass study helped to distinguish scar from inferior wall attenuation in 15-30% of patients that were known to have fixed inferior wall defects from previous SPECT test. Does first pass test have a role to play in nuclear cardiology?

Thursday

Session 79: Instrumentation and Data Analysis: SPECT

No. 467 INVESTIGATION OF USING PET BLOCK DETECTOR TECHNOLOGY FOR SPECT IMAGING L MacDonald, M Dahlbom, M Paulus UCLA School of Medicine, Los Angeles and CPS, Knoxville

The abstract talks about hand waving claims for an LSO detector. The presentation was about a YSO/LSO phoswitch detector that the authors hope will prove capable of SPECT and PET imaging. The simulations were not particularly convincing.

No. 468 THEORETICAL PERFORMANCE LIMITS FOR ELECTRONICALLY-COLLIMATED SINGLE-PHOTON IMAGING SYSTEMS NH Clinthorne, SJ Wilderman, JE Gormley, GF Knoll, DK Wehe, WL Rogers U Michigan, Ann Arbor, MI

This study evaluated the Cramer-Rao bound for a Compton scatter camera. The model used was a simplified version in order to make the calculation tractable. I did not really understand the conclusion. It sounded like the authors were encouraged enough to continue the study with a more realistic and complex model.

No. 469 FEASIBILITY OF PINHOLE FLOURINE-18 SPECT OF THE BRAIN; A MONTE CARLO STUDY H Wang, RJ Jaszczak, RE Coleman Duke U, Durham

This simulation study seemed to come to a positive conclusion on the possibility of pinhole brain SPECT imaging with 511 keV photons. In poster No. 965 this same group presented experimental data on I-131 SPECT imaging of the brain and it seemed to work, too. Compensation for those penetrating high energy photons is important.

No. 470 MULTIPLE LINE SOURCES FOR SPECT TRANSMISSION IMAGING A Sitek, A Celler R Harrop Vancouver Hospital, U British Columbia and Simon Fraser U, Burnaby

This is a very clever idea. An array of 8-10 line sources of 26 cm length are mounted opposite the detector. Line spacing is optimized in simulation study. Requires 360 rotation. Can do simultaneous emission-transmission scanning. Lower cost as hottest central sources can be moved to outer positions as they decay. They will be studying the effect of scatter. I think this is the idea that Siemens has bought into for their new e.cam system.

Walking Posters

Several sessions called walking posters were held. These require the audience (smaller than in a meeting room) to walk from poster to poster and listen to the presenter. It is hard to sort of stand in one spot for 90 minutes, it is hard to hear if the crowd exceeds 20 or so. It is nice to interact with the presenter and interested participants. On balance I do not like them, however. A more standard poster session where you can circulate and speak to a select group of authors of your choosing is better, I think.

Walking Poster Session: Instrumentation and Data Analysis: Image Generation and Analysis

There were 11 papers in this session. I describe here only those that most interested me.

No. 674 NONSTATIONARY RESOLUTION RECOVERY FOR ROI ACTIVITY ESTIMATION IN SPECT SC Moore, MF Kijewski, SP Muller, VA Medical Center, W Roxbury, Harvard Medical School, Boston, U Essen

The authors extended their work for activity estimation with restoration for stationary psf to the case where the psf is depth dependent. This represents the case in SPECT. Restoration was performed using a Metz filter that was depth dependent. The recovery coefficient increased to about 1. There was a some increase in noise. They conclude that nonstationary resolution recovery is necessary for ROI activity estimation.

No. 676 AN EXPERIMENTAL PHANTOM BASED ON QUANTITATIVE SPECT ANALYSIS OF PATIENT MIBI BIODISTRIBUTION DR Gilland, RJ Jaszczak, MW Hanson, KL Greer, RE Coleman Duke U, Durham

This physical phantom has fillable organs with MIBI uptake (includes R and L myocardium and chambers, R and L Styrofoam beads for lungs, liver, gallbladder and bowel. It has removable breasts. Imaging tests were close to those of real humans.

Walking Poster session: Instrumentation and Data analysis: Technology and Correction

No. 746 INVESTIGATION OF MULTIPLE ENERGY RECONSTRUCTIONS IN SPECT USING MLEM HM Hudson, BF Hutton, R Larkin, C Walsh Macquarie U, Westmead Hospital, Sydney

Authors found little computational cost for including simultaneous reconstruction of the attenuation corrected low and high photons of Tl-201. They found significant improvement in the when using two photons. Of course, ordered subsets were used.

No. 747 COMPTON-SCATTER CORRECTION IN SPECT USING ENERGY SPECTRAL DECONVOLUTION: COMPARISON WITH MULTI-ENERGY WINDOW METHODS A Kojima, Y Zhang, C-y Ng, WL Rogers, NH Clinthorne Kumamoto U, And U Michigan, Ann Arbor

Authors compared TEW and DEW with an energy spectral deconvolution technique. Lowest relative mean square error was with TEW. Their spectral method was OK, too. No accuracy and bias reported.

No. 749 LIMITATIONS OF SCATTER CORRECTIONS METHODS FJ Beekman, C Kamphius U Hospital Utrecht

The authors emphasized the importance of high energy resolution and an accurate scatter model. Scatter correction techniques always came out second best to a good model calculation. At low count rates it is better to use the scatter than not use it.

Other posters seen:

Numerous posters on FDG imaging in oncology were up. I could not find any negative results. It seems that FDG is a real good thing and we all should begin imaging it, even if only poorly, like with a big collimator, maybe. I searched in vain for guidance on just how small the lesion could be before they could be visualized. In other words what is the threshold for visibility in PET and collimated FDG? I could not find the answer to this.

No. 651 COMPARISON OF A ROTATING, PARTIAL RING PET SCANNER (ECAT-ART) WITH A DUAL-HEADED GAMMA CAMERA FOR WHOLE BODY IMAGING WITH Fe-52 TRANSFERRIN D Glass, H Young, M Myers, D Baily, J Awotwi-Pratt, M Deenmamode, D Swirsky, AM Peters Hammersmith H. and MRC Cyclotron Unit, London, UK

Fe-52 is a beta emitter. ECAT-ART is a real PET system but only a partial ring. Sensitivity of ECAT-ART is 11400 with resolution of 6.0 to 7.7 mm. The MS-2 sensitivity was 70 and resolution was 12.7 to 17.9 mm. Both instruments were able to localize extramedullary erythropoiesis but EXACT-ART could localize activity within the bone marrow.

No. 652 SIMULTANEOUS TRANSMISSION EMISSION IMAGING OF REST/REDISTRIBUTION THALLIUM MYOCARDIAL PERFUSION IMAGES: THE EFFECT OF ATTENUATION CORRECTION ON DEFECT REVERSIBILITY IN PATIENTS WITH REGIONAL DYSSYNERGY PE Wawrzynski, EP Ficaro, PA Rose, JR Corbett

Using their own technique which uses a single fan beam collimator with two parallel hole collimators, the authors claimed that attenuation correction improve identification of defect reversibility. There were no negative comments with regard to attenuation correction.

No. 653 IMPROVED Tc-99m SESTAMIBI UNIFORMITY WITH ATTENUATION CORRECTION IN PATIENTS WITH LOW LIKELIHOOD FOR CORONARY ARTERY DISEASE J Viggiano, JM McClellan, A Alavi, LI Araujo U Penn, Philadelphia

Used commercial STEP system. All favorable comments about results with attenuation correction. Found no problem reading the scans.

No. 655 THE ADDED DIAGNOSTIC VALUE OF GATING TO EXERCISE Tc-99m SESTAMIBI FOR DETECTION OF CORONARY ARTERY DISEASE J Erel, H Kiat, JD Friedman, G Germano, DS Berman Ceaders-Sinai medical Center, Los Angeles

Gating increased normalcy rates in the 100 low likelihood patients and increased diagnostic confidence but did not affect sensitivity or specificity for detection or localization of CAD.

No. 658 THE USE OF THE ROTATING PLANAR PROJECTION VIEWS TO IMPROVE SPECIFICITY OF MYOCARDIAL SPECT PERFUSION IMAGING AND INTERACTION WITH ATTENUATION CORRECTION METHODOLOGY RC Hendel, H Kiat, WP Follansbee, GV Heller, SJ Cullom, DS Berman Northwestern U, Cedars-Sinai Medical Center, U Pitt, Hartford H and Emory U.

These authors found that in a group of 70 patients viewing the rotating projections lowered sensitivity, but raised specificity and accuracy with or without attenuation correction. The best results seemed to be with attenuation correction, but it was difficult to be sure.

No. 811 CORRECTION OF PHOTON ATTENUATION IN SPECT MYOCARDIAL PERFUSION IMAGING: PRELIMINARY RESULTS OF A MULTICENTER TRIAL RC Hendel, H Kiat, WP Follansbee, GV Heller, SJ Cullom, DS Berman Northwestern U, Cedars-Sinai Medical Center, U Pitt, Hartford H and Emory U.

Looks like the 70 patients from the above study are back again. This time it was found that ADAC's Vantage attenuation correction increases normalcy rate. Image quality was rated better for attenuation corrected images. Does not seem to be any problems and AC helps.

No. 952 PHANTOM EVALUATION OF COMBINED SCATTER AND ATTENUATION CORRECTION WITH A SIMULTANEOUS TRANSMISSION / EMISSION SPECT SYSTEM (STEP) SD Miller, PJ Maniawski, HT Morgan, MA King Picker Intl., Cleveland, U Mass, Worcester

Adding two extra windows one FWHM below and above the photopeak centerline allows scatter (and spilldown) information to be collected. This is something like a TEW correction. A solid phantom containing a Tl-201 line source and Co-57 transmission source was scanned to determine if it worked. It did.

No. 953 FEASIBILITY OF USING SHORT AsF TRANSMISSION CT FOR ATTENUATION CORRECTION OF CARDIAC SPECT IMAGING EF Hollinger, D Yu, G Huang, W Chang Rush-Presbyterian-St Luke's Medical Center, Chicago

In conclusion, a 15 min. ECT followed by a 4 min. TCT study provided uncontaminated transmission and emission data at the statistical level of the ECT.

No. 954 RECONSTRUCTION OF TRUNCATED PROJECTION DATA FROM AN L-SHAPED DUAL DETECTOR SYSTEM FOR ATTENUATION COMPENSATION IN MYOCARDIAL SPECT BMW Tsui, DP Lewis, DS Lalush, AHR Lonn, S Maniam, R Culp UNC, Chapel Hill, GE Medical System, Milwaukee

The GE detectors are small and miss part of the body on big patients. Using an ML-EM algorithm can provide satisfactory reconstructions.

No. 955 IMAGE PERFORMANCE OF A SIMULTANEOUS TRANSMISSION - EMISSION IMAGE SYSTEM INCORPORATING A SCANNING TRANSMISSION LINE SOURCE HT Morgan, RN Lexa Picker Intl., Cleveland

The dual head Picker 2000 now has the standard scanning line source. Scatter correction and crosstalk is included. No surprises. It works pretty good.

No 956 THE EFFECT OF DEFECT SIZE, LOCATION, AND CONTRAST ON THE DIAGNOSIS OF MYOCARDIAL DEFECTS IN SPECT IMAGES WITH AND WITHOUT ATTENUATION COMPENSATION KJ LaCroix, BMW Tsui UNC, Chapel Hill

This was an attempt to address the question of how defect size, location and contrast may affect the diagnosis of myocardial defects with and without attenuation correction. It was a simulated study using noise free projections of the MCAT phantom. A computer did the diagnosis. Location is the big factor and attenuation correction helped in areas subject to attenuation. Not surprising results, is it?

No. 957 A MODIFIED WIENER FILTER METHOD FOR NONSTATIONARY RESOLUTION RECOVERY WITH SCATTER AND ITERATIVE ATTENUATION CORRECTION FOR CARDIAC SPECT L Liu, SJ Cullom, ML White Emory U, Atlanta

Providing depth dependent resolution recovery takes lots of computer time. There were several papers that addressed this problem. In this work resolution recovery was performed using a modified depth-dependent Wiener filter. Authors report good convergence properties and accurate results with this method.

No. 959 INFLUENCE OF IMPROPER ATTENUATION / EMISSION REGISTRATION ON THE ACCURACY OF SPECT TUMOR IMAGING H Surova-Trojanova, C Barker, JA Carrasquillo, SL Bachrach NIH, Bethesda

CT was registered with SPECT in the abdomen and thorax. Errors in x or y of 5 mm led to quantitation errors of <10% . Quantitation errors could be much higher at the edges of an organ. Errors in z of 10 mm (CT slice thickness) could lead to very large errors depending on size of object.

No. 960 IMPROVED FLOOD SOURCE UNIFORMITY USING CHELATION ON In-111 IN A LIQUID FILLED PLEXIGLAS TANK MJ Blend, M Friedman, T O'Dowd, S Tantayakom, B Rubas, H Levi, BC Penny, J Mukherjee U Illinois and U Chicago, Chicago

We often forget or ignore details, like: does the isotope stay in solution within our phantom? These authors found significant problems with uniform distributions of In-111 until they used DTPA to chelate the In-111. We need more hints like these compiled someplace.

No. 961 EVALUATION OF A SCANNING LINE SOURCE METHOD FOR ATTENUATION CORRECTION USING AN ANTHROPOMORPHIC PHANTOM MW Groch, SM Spies, H Hines, J Lieberg, RC Hendel Northwestern U, Chicago and ADAC Labs, Milpitas

Moral of this abstract is that you need to correct for Tc downscatter into Gd-153 window. It is so easy, why not do it from the start?

No. 962 RESULTS OF ATTENUATION AND SCATTER CORRECTION IN CLINICAL CARDIAC TL-201 SPECT PERFUSION IMAGING JA Case, MA King, ST Dalhberg, MSZ Rabin U Mass, Worcester

This is an attempt to improve on the STEP system that Picker supplies with the Prism 3000. STEP was the first of the commercial releases of cardiac attenuation correction. Picker is to be congratulated for making it available. But it is not the final solution. There are problems with the commercial version of STEP that we, and others have seen. Problem 1: there is no way to ease into attenuation correction. You have to stop using the parallel hole collimators, install 3 fan beam collimators and start using the iterative reconstruction program. At this point you can compare non AC images with AC images. But you have no diagnostic skills developed to use either the fan beam images with or without AC. Problem 2: there is no scatter correction in the commercial version of STEP. There is downscatter correction. Problem 3: There is no depth dependent resolution recovery in commercial STEP.

What is needed is a way to ease into attenuation correction by retaining the original parallel hole images, allowing comparison of the old with the new, improved "fully" corrected images. Several groups have been working on such "improvements" to STEP. The U Mass work is reported in this abstract. I also learned at this meeting that a group associated with EP Ficaro from U Michigan has a similar approach.

The U Mass approach uses one fan beam collimator and a line source from the original STEP hardware and uses two parallel hole collimators. The transmission study is acquired after the emission study. TEW scatter correction is provided. Reconstruction is by FBP for the old and familiar AND by ML-EM for the attenuation corrected images. No resolution recovery operation is described. Some artifacts remain in the corrected images but "physician education" can be done by using the old images to learn how to use the "new, improved" images.

No. 967 MISLEADING ESTIMATES OF SMALL-LESION DETECTABILITY IN PHANTOM STUDIES OF TL-201 MYOCARDIAL PERFUSION (TL) SPECT R Wernikoff, H Gerwirtz, Newton, MA and Mass General Hospital, Boston

The authors found that the wall thickness of the lesion inserts gave significant errors with regard to nominal lesion size. This led to more favorable object contrast than expected and overestimation of system performance.

No. 969 COMPUTER SIMULATIONS FOR THE OPTIMIZATION OF COLLIMATOR DESIGN BJ Kemp, R Stodilka, FS Prato, RL Nicholson St Joseph's Health Centre, U Western Ontario, London

A stochastic ray tracing program that explicitly models hole shape, fan beam or parallel hole parameters, resolution vs depth etc. is described. Optimum collimator for brain SPECT is arrived at by using a MRI brain scan with radiopharmaceutical distributions as a source object and comparing reconstructed image with object. Sound like a good way to design a collimator.

No. 971 LIMITED-ANGULAR-VIEW ML-EM PINHOLE SPECT FOR BREAST TUMOR DETECTION J Li, C Scarfone, RJ Jaszczak, H Wang, RE Coleman Duke U, Durham

The authors report favorable phantom trials using pinhole SPECT and a ML-EM algorithm. They indicate the collimator would have to be further optimized for mammoscintigraphy.

No. 972 COMBINED ATTENUATION AND SCATTER CORRECTION SIGNIFICANTLY REDUCES THE EFFECT OF EXTRA-CARDICAC ACTIVITY ON TL-201 MYOCARDIAL PERFUSION SPECT PJ Maniawski, SD Miller, HT Morgan Picker Intl., Cleveland

The authors incorporated two windows on either side of the Tl-201 photopeak ala TEW, but not abutting the photopeak window. Rest of the method seems like standard STEP. The images were better with scatter correction.

I have heard that Toshiba has patented TEW and is the only manufacturer that can "sell" TEW so I guess the other companies have to come up with workarounds. Isn't that like patenting the photopeak window? I do not think that it is reasonable to patent acquisition windows or what you can do with the data from those windows. Hope we are not breaking the law. I would hate for the Japanese industrial machine to come down on Harvard.

No. 973 A 3D ITERATIVE FILTERED BACKPROJECTION ALGORITHM FOR ATTENUATION COMPENSATION AND RESOLUTION RECOVERY FOR SPECT WITH FAN BEAM COLLIMATION G Yin, JR Galt, SJ Cullom Emory U, Atlanta

The authors report on their FBP algorithm that improves resolution over a re-projection scheme from 7.1 mm to 5.8 mm. There is a slight increase in noise.

No. 974, CORRECTION OF ERRORS DUE TO PARTIAL VOLUME EFFECTS INCLUDING CONTRACTILE MOVEMENT IN RESTING MYOCARDIAL MOVEMENT IN RESTING MYOCARDIAL SPECT BF Hutton, A Osiecki Westmead Hospital, Sydney

In theory one can obtain the regional myocardial uptake per gram of tissue of a tracer in the myocardium. In PET it is easier that in SPECT. The authors extend their work in PET partial volume correction to SPECT in this paper. Method involves two simultaneous emission and transmission acquisitions. One for blood pool and one for perfusion. Attenuation and scatter correction are performed. The estimate of uptake can be made independent of partial volume if emission and transmission images have same partial volume errors. Sounds futuristic. We can't even get attenuation correction to work where I come from.

No. 976 COMPENSATION OF ATTENUATION MAP ERRORS FROM Tc-99m-SESTAMIBI DOWNSCATTER WITH SIMULTANEOUS Gd-153 TRANSMISSION SCANNING SJ Cullom, L Liu, ML White Emory U, Atlanta

The platform was ADAC, I believe. Correction for downscatter is important. I suppose the next iteration of Vantage for ADAC will have something like this.

No. 978 ENHANCED CROSS-TALK CORRECTION TECHNIQUE FOR SIMULTANEOUS DUAL-ISOTOPE IMAGING: A TL-201/TC-99M MYOCARDIAL PERFUSION SPECT DOG STUDY K Knesaurek, J Machac Mount Sinai Medical Center, NY

The details must be published elsewhere. The method uses two convolution equations applied in frequency space and incorporates restoration filtering. Makes contrast better. No statement as to how noise changes.

No. 979 FULLY 3D ITERATIVE SCATTER COMPENSATION IN NON-UNIFORM ATTENUATING MEDIA FOR SPECT FJ Beekman, JM den Harder, MA Viergever, PP van Rijk U Hospital Utrecht

The details of the method were not easy to discern. Water equivalent depths were used to account for the non-uniform attenuation in the chest. Since it uses the psf I would think that it would perform some sort of resolution recovery also, but that was not discussed in this poster.

No. 983 USE OF ITERATIVE METHODS IN RECONSTRUCTION OF 180 DATA JW Wallis, TR Miller Washington U, St. Louis

The authors report on their difficulty in getting ML-EM, iterative Chang and their own iterative reconstruction algorithms to converge on 180 degree acquisitions.

No. 1008 COINCIDENCE MODE IMAGING USING A STANDARD DUAL-HEADED GAMMA CAMERA RS Miyaoka, WLS Costa, TK Lewellen, MS Kaplan, SK Kohlmyer, F Jansen U Washington, Seattle and GE Medical Systems, Waukesha

GE has apparently relegated the development of coincidence imaging to the academic arena. I suppose that this is some kind of holding pattern so they can determine if a first class, expensive development is needed to catch up to ADAC's position. Meanwhile it gives the students at U Washington some very interesting projects.

But I think we would all agree that a PET/SPECT camera has to be designed from the ground up to be suitable for PET. It is hard to do it good enough to be clinically useful.

Scientific Exhibits

I did not have any time to spend at the scientific exhibits bit I do want to call attention to one:

No. 1418 A WORLD WIDE WEB-BASED RESOURCE FOR THE PHYSICAL EVALUATION OF 511 KEV-COLLIMATED IMAGING DJ Wagenaar, AD Van den Abeele, GM Kolodny, TC Hill, RE Zimmerman JPNM, Harvard Medical School, Boston

The Joint Program in Nuclear Medicine is in a position to test many different vendors cameras. Even to compare (eventually) collimated FDG imaging with PET in certain circumstances. This web project: http://www.med.harvard.edu/JPNM/physics/511keV/511keV.html is a compilation of our results and notes to date. We intend to make this useful to the full nuclear medicine community. We welcome your comments.

Commercial exhibits

3M Medical Imaging Systems

The business of the dry laser printer started by Polaroid has serious competition. 3M Medical Imaging Systems, St Paul, MN, has one. Claims 1024 gray levels and 300 dpi. I wonder if anyone has ever compared these things? I have always thought that these things would be replaced by CRTs, but I seem to be wrong.

ADAC

ADAC continues to lead in coincidence counting, announcing 5/8" crystals and new electronics that push max. singles rate to 2.4 Mcps. Their Vantage attenuation correction seems to want downscatter correction. They have a website and it seems to have some good information on it. They may be incorporating net browser ideas into their imaging viewing programs. (See report above on their physicist meeting for more ADAC details.)

Digirad Corporation, San Diego CA

Bill Ashburn was showing a semiconductor camera (some kind of CdZnTe, I think). It was one quarter of the final size. A test image was displayed. GE also had a CdZnTe camera (see below). It is 3 mm thick now, working on 5 mm thick, could be up to 10 mm thick. Interesting.

Digital Scintigraphics Inc., Waltham MA

Makers of Ceraspect dedicated brain scanner are planning to develop a dedicated mammoscintigraphy camera. This a great idea. The folks who brought you the first digital gamma camera, the first (and only) cylindrical gamma camera are turning their creative minds to a superior way of imaging the breast. New innovations include a small cylindrical shell of NaI with a liquid light coupling medium. They should arrive at the optimum geometry for mammoscintigraphy. Feasibility studies were very encouraging.

General Electric

GE also had a CZT camera in the booth. Definitely a works in progress. They have an Israeli connection for the detectors. And developed with the aid of your (US) tax dollars. That's leveraging your investments!

GE was also showing a mockup of a dual head camera "Millennium". Very premature showing so it was hard to make firm judgments about it. It was pretty. Will have big crystals, do whole bodies and turn out to do sitting patients. Could not give specs on the coincidence detection option.

Picked up a nice booklet: "Gamma Technology Briefing" Pub. No. 96-4546 from GE. Is useful to understand what they are doing.

Harrison Medical Corp.

The Mountz Imaging System was shown by Harrison Medical Corp., Helena, AL. It is a head fixing device based on eyeglass frames. Supposed to guarantee reproducible head positioning across modalities. Full system includes software, conversion software for different image formats.

MEDX

MEDX, Wood Dale, IL, continues to sell its Nuclear MAX, a PC based low cost computer system. Also had a very interesting deluxe PC-based computer system from an Argentine company named Sitco - NuQuest. This was a pretty impressive machine, as demonstrated by its inventor and programming genius.

ONES Medical Services

There is a move to get independent service groups to come into a hospital and bid against the manufacturers for service contracts. While this has always been around to some extent, I have always cautioned against it, especially for new systems. ONES Medical Services, Goffstown, NH is one of these companies. They seem to be pretty successful at it. Also sell the former Strichman computer, re-manufactured cameras, etc. We have been having such poor service from GE I would be tempted.

Park Medical Systems Inc.

Park is not going to coincidence counting. They are re-evaluating coded apertures. Using a modular coded aperture (random array of uniformly redundant pinholes, modified with a blocking code), and taking advantage of <2 mm intrinsic camera resolution they hope to significantly improve sensitivity. No scientific information is available yet. No trials till autumn. Interesting but I am skeptical. They intend to do rotating tomograhy AND planar imaging with this technology. First version is in W for low energy work, second version will be in U for high energy work.

Their holospectral imaging remains unproven but interesting. No serious trials or results have been presented. Independent evaluations of their impressive technology would be desirable.

They have gone to 1/2"crystal for improved imaging at high energy. Do not know if their depth correction technology is being used yet.

This is a company with lots of promise and lots of ideas. I think we will have to just be patient to see results on these ideas.

Picker

Picker is in a slump. Can't find much new here. No viable coincidence project. No STEP improvements released. Their EPICON detector project is now part of the main company but it is not clear to the outside what is happening. Maybe the bombshell will be heard in the future.

Siemens

This sleeping giant is stirring. A new camera, e.cam, looks pretty good. All digital technology. Very flexible in positioning and clinical uses. Will be developing coincidence technology, but nothing to report yet. Will eventually have F-18 collimators. Will not be delivered until next year. This camera will be able to receive the hybrid YSO/LSO detector that CTI and Siemens are developing. That is a whole 'nother story. It is very unproven and still being conceptually designed. Too early to put a confidence level on it.

The e.cam will have a unique attenuation correction source, which will be cheaper, and maybe more effective than some of those out there now. Development seems to be in conjunction with a group in Vancouver. See No. 470 above. It uses two arrays of Gd-153 line sources of graded activity. Have to watch this.

In computers they are moving to put all new software and port existing software to run under IDL. IDL is software that runs on many different platforms. Programs developed under IDL can also then run on many different platforms. This sounds attractive. We will have to see if they can pull it off.

General comment on cameras:

There is a lot of futures being sold right now. It is probably better to focus on what will be delivered NOW. One important question that needs to be addressed on these new cameras is collimator changing. It took ADAC quite a while to get their automatic version working and now they are offering a semi-automatic version. What are all those new cameras doing for collimator changes? Collimators are getting heavier and they are multiplying.

Hope to see you at the 44th SNM meeting in San Antonio, Texas 1 June - 5 June 1997