Personal Notes from the SNM, Toronto, June 1998

Robert E. Zimmerman

Tuesday, 21 July, 1998

Toronto is a great city for the SNM. Prices are very reasonable, the city has a wealth of excellent restaurants with plenty of cultural and social outlets. It is ethnically diverse and has a very functional transportation system, downtown is small enough to walk most anywhere, and safely, what's more. I thoroughly enjoyed all aspects of the meeting this year that related to the local venue.

On the cultural side I managed to get to the Toronto Symphony Beethoven festival and catch a one-of-a-kind exhibit of Impressionists from the Courtauld Collection of London (England) at the Art Gallery of Ontario, just two blocks from my room.

We can't come here too often.

The meeting itself was subject to the usual commercial excitement (=hype?) but I came away more encouraged than ever about the state of nuclear medicine going into the next millennium.

I managed to avoid all responsibility at various council meetings (The Academic Council was summarily rescheduled with no forewarning), Categorical Seminars and User Meetings so my professional obligations did not really begin until 6:30 AM at the Siemens Physicists' breakfast meeting. Siemens put on a good show keeping us awake with a very full program.

Ben Tsui and Scott Wullenwebber of UNC reported on their PET testing of the Siemens PET/SPECT dual head coincidence ECAM+. They have only had the camera since end of February and had to quickly start competing for time with patients, unfortunately. The images shown looked very good. The performance numbers are encouraging but the industry (and potential customers) awaits some standard way to present performance measures for these not-quite-PET instruments. This is particularly true when it comes to reporting true coincidence, scatter and random count rates. The reported numbers are SO sensitive to the test conditions that almost any number one desires can be measured. Various camera companies have reported coincidence rates from 30,000 down to 1500 cps. Clearly there is not this much difference in systems all using the same basic technology. But there are widely varying conditions for making the measurement. There is a NEMA committee wrestling with this problem. (See No. 525) Let's hope they can get a set of measurement guidelines out, if only in draft form, very quickly.

Ron Nutt, of CTI, once again nearly convinced me that he would have the YSO/LSO hybrid gamma camera available for sale in a more few months. Oops, make that NaI/LSO hybrid camera. It seems that YSO is a bit trickier to get into full production and make a sandwich, or phoswich, detector. He was even able to show a bone scan taken with the current full size (?) phoswich detector.

Dave Townsend, of U Pittsburgh, showed us the very first images from his PET/CT instrument. For those of you who feel that intermodality registration is mandatory this is the machine for you. An ART PET camera on the same gantry as a Siemens Somatom spiral CT scanner. His results were really fresh having just obtained his image on the previous Friday.

Paul Shreve of U Michigan reminded us that "Roentgen can't do it all" and that physiological imaging is the key to earlier diagnosis and certainty. Applications to cellular and molecular imaging will be our future.

Simon DeBruin of Siemens introduced Profile, Siemens soon to be released attenuation correction package for the ECAM. Clinical testing is in the hands of the inventor, Anna Cellar of U Vancouver and another physicist Ed Ficaro of U Michigan. Ficaro's group has had a long clinical experience with a modified Picker STEP system and seems to have done real good work with it. ECAM users eagerly await their test of Profile. Profile is elegant. No moving parts and extremely efficient use of the Gd-153 sources that should minimize operating expenses compared to the moving line source approach.

Finally John Engdhal director of Siemens R&D presented his take on the newest possibility for a replacement to NaI - CdZnTe semiconductor cameras. I was not convinced that it will happen very soon but it should be interesting to watch Siemens work on this. Of course Digirad is offering a small field CdZnTe camera for sale. Can Siemens improve on this startup venture? They certainly will improve on the previous work of GE and Elscint that has been "shown" at previous SNM meetings. Their work seems to have died as the two companies have since joined forces and pared down their investment in nuclear medicine. The biggest problems seems to be making "cheap" CdZnTe in large quantity while making low noise connections to the crystals to keep the energy resolution substantially better than NaI.

I arrived at the plenary session late after the extended breakfast meeting. I caught the tail end of the Henry Wagner Lecture by Frans H. Corstens of The Netherlands and current President of the EANM Congress. Dr. Corstens presented an interesting overview of the "soon to be standard" NM imaging tests. Infection, thrombus and tumor imaging will find their rightful place in the diagnostic regimen. However the burden to make this happen is on the nuclear medicine physician who must be proactive in educating the referring physician.

H. William Strauss gave an overview of the program that, while elegant, we could have done without.

The Georg de Hevesy Award was shared by Heinrich Schelbert (not in program) and Nagara Tamaki for their work in developing tracers for myocardial viability and was presented by Strauss.

Paul Murphy presented the Paul C. Aebersold Award to Gerd Muehllener for his work in instrument development.

Richard Reba presented the Cassen Award to Henry Wagner. Wagner's acceptance speech was worthy of a stand-up comic in a nightclub. Lots of very funny jokes. Good ending to the plenary session.

No. 22

G Kontaxakis, LG Strauss, G van Kaick

German Cancer Research Center (DFKZ), Heidelberg

No. 23

RZ Stodilka, BJ Kemp, P Msaki, FS Prato, RL Nicholson

St. Joseph's Health Centre, University of Western Ontario, London, Ontario

An experimental determination of the relative importance of scatter and attenuation correction was attempted. Four metrics were used: Contrast recovery, relative quantitation, global bias and global uniformity. They found that non-uniform attenuation correction had little to offer over uniform attenuation correction. Scatter correction was found to be important but a using the transmission map had little to offer compared to just knowing the skull boundary. Noise was not considered.

No. 24

M Schmand, M Dahlborm, L Eriksson, M Andreaco, ME Casey, K Vagneur, ME Phelps, R Nutt

Max-Planck-Institute for Neurology Research, Cologne, CTI Inc., Knoxville, UCLA School of Medicine, Los Angeles and Karolinska Institute, Stockholm

This was the prizewinner. The phoswich detector showed energy resolution of 10% for Tc-99m and <10% at 511 keV for a block detector arrangement with 4 PMTs using Anger logic to localize to a 4 mm crystal. This is obviously a close cousin of the Siemens/CTI project to build the hybrid LSO/NaI PET/SPECT camera. Exciting!

No. 25

GK Kastis, HB Barber, HH Barrett, HC Gifford, IW Pang, DD Patton, JD Sain, G Stevenson, DW Wilson

University of Arizona, Tucson

No. 26

AJ Morguet, AF Chatziioannou, SR Cherry, ME Phelps, HR Schelbert

UCLA School of Medicine, Los Angeles

It works! A PET camera based on LSO (30 detectors with 8 x 8 crystals each going to PMTs via fiber optics) were used to successfully image infarcts in a series of 18 rats. Infarcts with dimension greater than 2 mm could be detected.

No. 27

TH Farquhar, J Llacer, J Sayre, Y-C Tai, A Chatziioannou, SR Cherry, J Czernin, M Dahlbom, SS Gambhir, CK Hoh, S-C Huang, CL Levin, LR MacDonald, HR Schelbert, MA Seltzer, DHS Silverman, EJ Hoffman

UCLA School of Medicine, Los Angeles

This seemed to be a baseline study to help further evaluations. A thorax phantom scanned on HR+, 7, 14 and 28 million trues, 0.45, 1.0 and 1.9 ml lesions with attenuation correction and standard clinical reconstruction were used. 5 different contrast levels were used. The resulting ROC curves will be reexamined with various algorithms, acquisitions and analysis methods. Probably not of immediate use to the rest of us.

No. 28

L. Green, S. Gambhir, J. Barrio, E. Bauer, K. Nguyen, M. Namavari, S. Cherry, H. Herschman, M. Phelps

UCLA School of Medicine, Los Angeles

This may have been an instrumentation session but was of relatively low interest to me. UCLA has a lot of papers!

No. 78

DS Lalush

The University of North Carolina at Chapel Hill, Chapel Hill, NC

Lalush investigated dual and triple head cone-beam orbits of pinhole cameras trying to maximize the sampling to get complete data. He concludes that a dual head approach is probably optimal and can bring practical applications of cone-beam SPECT closer to practical use while improving noise/resolution tradeoffs.

No. 79

RR Raylman, RL Wahl

West Virginia University, Morgantown, WV; and University of Michigan, Ann Arbor

The target to background ratio improved from 2:1 to 5:1 by using a stacked Si detector. The first detector gets betas and photons, the second only photons and is used to correct the events from the first detector.

No. 81

TL Mulnix, MJ Schueller, RJ Nickles

University of Wisconsin, Madison

No. 82

JE Gormley, B Pi, R Conwell, W Ashburn

Digirad Corporation, San Diego

As much as one wants to see a semiconductor camera succeed we were disappointed by this presentation. The energy resolution of the full detector was reported as a disappointing 8%. Maybe they can use energy correction to make it better? The geometry is more favorable for scintimammography since a medial view can be used.

No. 83

BMW Tsui, DE Wessell, NP Peterson, XD Zhao, PT Le, WH McCartney

The University of North Carolina at Chapel Hill, Chapel Hill; and Memorial Mission Hospital, Asheville, NC

If scintimammography is ever to be effective we must have cameras specifically designed for the purpose. This group looked into pinhole planar imaging with phantoms and patient studies. When close to the patient pinholes can be very sensitive and have increased resolution when compared to parallel hole collimators. Their clinical trial "has potential".

No. 84

FH Fahey, RL Webber, BA Harkness

Wake Forest University School of Medicine, Winston-Salem, NC

This group also looked at pinhole imaging for scintimammography, but with tomographic imaging in mind. Multiple views with fiducial markers in place allow longitudinal tomography. A phantom experiment was encouraging. ETACT is emission tuned aperture computed tomography, a reference to the fact that the scan path can be optimized for each patient.

No. 134

SC Moore, SP Müller, MF Kijewski

Harvard Medical School and Brigham & Women's Hospital, Boston and Universitatsklinik Essen, Essen

The authors found in a simulation that by using the projection data from 3 energy windows attenuation correction of Ga-67 images of the torso is feasible. No measured attenuation map is required. Scatter was not simulated. Some increase in noise was noted.

No 135

MF Kijewski, SC Moore, SP Müller

Brigham & Women's Hospital and Harvard Medical School, Boston and Universitatsklinikum Essen, Essen

How much better is nonuniform attenuation correction compared to uniform attenuation correction when the task at hand is measurement of activity in a small sphere in the center of the brain? Well the answer obtained by these authors would indicate that little is to be gained. Some advantage for the nonuniform attenuation correction was noted at low contrast ratios and at low noise levels. Take home message is that in brain scanning you may not need to go to the trouble of a nonuniform attenuation correction.

No 136

N. Rajeevan, IG Zubal, Q Ramsby, SS Zoghbi, JP Seibyl, RB Innis

Yale University and Veterans Administration Connecticut, New Haven

In a patient series these authors found uniform attenuation correction underestimated activity by 6-16% when obtaining counts from central ROIs. In uniform correction greatest error was due to positioning the ellipse around the emission image. When interested in target to background measures nonuniform, uniform or no attenuation correction methods were approximately equal in performance.

No. 137

DS Lalush, BMW Tsui

The University of North Carolina at Chapel Hill, Chapel Hill

Does OS-EM break under extreme conditions of attenuation and truncation? In a rigorous investigation involving comparison to ML-EM the authors found that OS-EM is very robust and at least as good as ML-EM as long as more that 2 views/subset were used.

No. 138

L Shao, P Nellemann, G Muehllehner, H Bertelsen, H Hines

ADAC Laboratories, Milpitas

The ADAC camera uses 2 Cs-137 sources to obtain transmission data in a sequential mode with the emission data, post-injection. The sources are collimated as a fan 13 cm axially by 51 cm transversely and scan in an axial direction as the cameras rotate. OS-EM reconstructions are used throughout. It works well, they say. Emission contamination of transmission data is reported as 3%.

No. 139

RJ Smith, JS Karp, E Gualtieri, G Muehllehner

University of Pennsylvania and UGM Medical Systems, Philadelphia

This group has had considerable experience with attenuation correction using a Cs-137 source. They have improved the performance of their method by using an energy correction technique to improve the energy resolution of the system, thereby allowing tighter energy limits to be set. This improved scatter rejection and reduced contamination of the transmission data from emission photons at 511 keV to 25%. This sounds much worse that the ADAC number reported above. Why, I wonder? There is great similarity in hardware here. Clarification needed.

No. 190

JS Karp, G Muehllehner, MJ Geagan, R Freifelder

University of Pennsylvania and UGM Medical Systems, Philadelphia

Curve-Plate PET is the outgrowth of the curved plate technology for NaI detectors pioneered by Bicron. There seems to be some slight gains in performance using these plates but it is very small. One real improvement is that the light distribution is narrower at the PMTs resulting in higher count rate performance. The electronics are completely redesigned and better performance is anticipated as compared to the Quest.

No. 191

Y Shao, SR Cherry

UCLA School of Medicine, Los Angeles

The object is to build a breast or small animal PET imager. The problem is depth-of-interaction (DOI) information is obtained by measuring the light at both ends of the crystal in such a way that the size of the ring can be kept very small. The solution is to get the light out of one end with bent optical fibers. 40% light loss was measured which is thought to be OK. The LSO crystals are 21 mm long and DOI can be determined to 4.2-6 mm. Promising but not finished yet.

No. 192

SC Moore, DJ Wagenaar, RE Zimmerman

Brigham & Women's Hospital and Harvard Medical School, Boston and Siemens Medical Systems, Hoffman Estates

This was a follow-up to a paper we presented last year. Penetration in collimated 511 keV and Ga-67 imaging is severe. Measurement of the sensitivity and penetration of these collimators is important. We found that the Monte Carlo calculated and the measured sensitivity vs. distance relationship is well fit by an exponentially decaying function of distance, d: a + b exp (-cd). The parameter, a, is related to geometric sensitivity while the remaining term describes the distance dependent relationship and allows the penetration and other contamination to be characterized. Geometric sensitivity of HE collimators can now be well-specified independent of penetration or other contamination.

No. 193

NH Clinthorne, SJ Wilderman, C Hua, JW LeBlanc, WL Rogers

University of Michigan, Ann Arbor

Electron binding and Doppler broadening can strongly influence the choice of material for a Compton camera first detector. The authors used the Cramer-Rao bound and Monte Carlo techniques to determine the best material. Contrary to the published abstract they think diamond may be the best scattering detector.

No. 194

P Shreve, V Kalf, PV Kison, RS Steventon, RL Wahl

University of Michigan, Ann Arbor

In a highly controversial presentation clinician Paul Shreve found himself presenting a simple experimental attempt to define detection thresholds to a group of sophisticated physicists. He reported on the limit of detection for their ECAT 921 and 931 and ADAC MCD cameras. MCD could see no smaller than 1.3-1.5 cm no matter what the SUV. Attenuation correction did not help. The experiment was strongly criticized by Harry Barrett for its naivete in design and lack of control for false positives, i.e. no ROC curve. Steve Moore, Chris Thompson and Dale Bailey had other comments. It was strongly defended by Henry Wagner, who seemed to make a special appearance just to defend the speaker. We were due for a bit of excitement.

No. 195

KS Pentlow, O Mawlawi, F Daghighian, RD Finn, JL Humm, SM Larson

Memorial Sloan-Kettering Cancer Center, New York

The secret: raise the window baseline to 400 keV. Some noise degradation remains but better overall performance. Resolution not affected.

No. 196

M Dahlbom, LR MacDonald, M Schmand, L Eriksson, M Andreaco, C Williams

UCLA School of Medicine, Los Angeles, Max Plank Institute, Cologne, Karolinska Institute, Stockholm, CTI Inc., Knoxville

The crystals were 4 x 4 x 15 mm for YSO and 10 mm deep for LSO. Energy resolution for 140 keV was about 14.5%. There was difficulties getting the PET mode to work. Now we see why Siemens CTI is talking about NaI/LSO phoswich.

Also on Tuesday was the Instrumentation Council' s business meeting during the midday break. You know the schedule this year was intense with 5 scientific sessions per day running from 8:00 - 17:30, only 15 minutes between them with a 1:15 break at lunch. There was little time to meet with the vendors. The "benefit" for this hectic schedule is tat you get to go home after henry Wagner wrap up talk at 12:30 on Thursday. I do not like this schedule. There is a new Program Chair coming in, Peter S. Conti. Those who would like to return to a saner schedule should contact him.

I missed some of the next session when I went to an American Society of Nuclear Cardiology (ASNC) Quality Assurance Committee meeting. This group is very interested in raising the level of competency in the Nuclear Cardiology community. A worth goal! Does it siphon off talent and energy from SNM? Probably, but there are an awful lot of cardiologists out there and wouldn't it behoove us all if they DO IT RIGHT. They have their own Journal of Nuclear Cardiology, which seems to be some science and a lot of presentation of practical technique. I have been drafted for a bit role in that effort.

I did not get to many of the papers in the next time slot but I call attention to the two immediately below:

IRCCS-Ospedale Maggiore di Milano, Milan

I did not attend this presentation but the abstract caught my eye.

Can you believe the orbit for perfusion studies has still not been settled after all these years. The common sense approach of getting close, staying close and orbiting to get a full data set (at least 180) is not good enough for some people. In this phantom study 180 vs. 360 with body contouring orbit and circular orbit was studied. The conclusion was that 180 orbits ought to have circular orbits and 360 scans should have body-contouring orbit. In my experience with 180 orbits there is little difference physically (in 90 configuration) between circular and body-contouring orbits as both are quite close to the body surface. What I never believed in is the heart-centered orbit. Too counter intuitive for me.

Vancouver Hospital and Health Sciences Centre, University of British Columbia, Vancouver and Universite Paul Sabatier, Toulouse

This presentation was less than convincing that a practical method that could be applied to dynamic SPECT. But perhaps I was confused from all the running about.

Late to this session because I tried to squeeze in a few moments in the Exhibit Hall.

No. 303

C Vanhove, M Defrise, PR Franken, A Bossuyt

Free University of Brussels, Brussels

The authors found that an OS-EM algorithm greatly improved the reconstruction of dual pinhole SPECT reconstructions compared to a Feldkamp filtered backprojection algorithm. This was a most promising paper and I expect to hear much more in this area in the near future.

No. 304

EG Hawman, JW Wallis

Siemens Medical Systems, Inc., Hoffman Estates, IL; and Mallinkrodt Institute of Radiology, Washington University, St. Louis

A fast algorithm and its details were described. I wonder if this is what Siemens is using in its latest package?

No. 305

FJ Beekman, ETP Slijpen, WJ Niessen

University Hospital Utrecht, Utrecht

In a simulation study to optimize reconstructions for the task of accurate ratios of activity in the brain it was found that using a high iteration number and optimal filtering could significantly reduce errors. High resolution is obtained and the optimized filter controls the otherwise high noise.

No. 306

G Wells, P Simkin, P Judy, M King, PH Pretorius, HC Gifford

University of Massachusetts Medical Center, Worcester and Brigham and Women's Hospital, Boston

2D or 3D, and what should that cutoff frequency be? For the simulated Ga-67 thorax images with 1 cm lesions studied this group found that 3D filters are better than 2D and cutoffs of about 0.32 were near optimum. These results based on a ROC study. This seems to be different than the published abstract.

No. 307

JE Bowsher, MF Smith, J Peter, RJ Jaszczak

Duke University Medical Center, Durham

Which is better for brain reconstructions: Larkin and Hudson's ordered-subsets expectation-maximization (OSEM), Bouman and Sauer's iterative coordinate descent (ICD), Byrne's rescaled block iterative (RBI) or Fessler and Ghia's SAGE? This group found better high frequency performance for ICD over filtered backprojection but has not finished all their investigations.

No. 308

C Riddell, F Jousse, L Aloj, P Mansour, M Whatley, JA Carrasquillo, ME Daube-Witherspoon, SK Libutti, DN Danforth, SL Bacharach

The National Institutes of Health, Bethesda

OS-EM gave better noise performance in all practical cases when compared to filtered backprojection. Detection was therefore better.

West Virginia University, Morgantown

I got in late for this talk but the author seems to have simulated the detection of beta emitting isotopes in tumors. He found that the energy of the beta is an important factor.

No. 359

A Saoudi, D Rouleau, R Lecomte, M Andreaco, M Casey, R Nutt, H Dautet

Universite de Sherbrooke, Sherbrooke, CTI, Inc., Knoxville and EG&G, Vaudreuil

The authors concluded the DOI measurements in phoswich detectors using BGO will not work but LSO/YSO will be feasible. Gerd raised the question about the variable light decay time of LSO and pointed out the crystals would have to be selected.

No. 360

T. Turkington, D. Gilland, R. Coleman, C. Laymon

Duke University Medical Center, Durham

This group reported on their progress in fitting multiple Cs-137 sources onto an Elscint Varicam in order to obtain transmission scans. Transaxial septa (1.3 cm spacing, 3 mm thick, 6 cm depth) are used. The sources are placed between septa on one camera and data acquired with the other. They project that 4 min transmission scans will be possible.

No 361

D. Wilson, E. Clarkson, H. Barrett

University of Arizona, Tucson

Synthetic collimation consists of a multiple pinhole aperture, a high resolution detector, away to move the detector relative to the aperture and an algorithm.

Advantages are mm spatial resolution, high sensitivity when close to object. Artifacts must be explored. This is neat stuff. See also Poster 781.

No. 362

P. Marsden, H. Parkes, A. Cave, P. Garlick

Guy's and St. Thomas' Hospital, London, UK and UCLA School of Medicine, Los Angeles

PET and MRI scans were done simultaneously on an isolated perfused rat heart. The PET scanner is a MRI compatible small animal unit with LSO crystals. The MRI has a 9.4 T magnet.

No. 363

RH Huesman, GJ Klein, BW Reutter, JH Reed, S. Grootoonk, TF Budinger

Lawrence Berkeley National Laboratory, Berkeley, University College, London, UK and CTI, Inc., Knoxville

The group investigated retrospective gating to get cardiac motion and respiratory phase information. They think it will work without degrading quality of the data.

No. 364

WH Wong, J Uribe, H Li, H Baghaei, N Zhang, S Yokoyama, J Wang

University of Texas MD Anderson Cancer Center, Houston

Wong presented a new PET camera design that has a variable FOV. It has BGO detectors block detectors, modular construction, requires rotation of the gantry in body mode and should be a lower cost scanner.

No 421

RE Coleman, CE Laymon, TG Turkington

Duke University Medical Center, Durham

They compared dedicated PET, camera based PET and 511 keV collimated SPECT. The phantom was a realistic torso with 4 spheres of 6, 10, 13 and 22 mm added to the lung area. Attenuation correction was applied to all based on transmission map from dedicated PET. Chief finding was that the lower noise in dedicated PET allowed the 6 mm sphere to be detected. Camera based PET could only see greater than 10 mm sphere and the SPECT system needed 13 mm spheres. Attenuation correction helped correct distortion and improved detection in camera based PET. Simplistic experiment but is probably helps a bit.

No. 422

M Baehre, B Melter, I Lauer, H Luig, E Richter

University of Luebeck, and University of Gottingen

Report of results with a Picker Prism 2000 XP PCD. They found that iterative reconstruction is better than filtered backprojection. They applied to >150 patients and think it work swell.

No. 470

EC Frey, DJ Kadrmas, BMW Tsui

University of North Carolina, Chapel Hill, and The University of Utah, Salt Lake City

Authors have developed a clinically usable algorithm that is equivalent in performance to ML-EM.

No. 471

C Bai, GL Zeng, GT Gullberg

The University of Utah, Salt Lake City

Larry described the practical use of the slab-by-slab response correction combined with scatter correction. Slab correction faster than slice correction. Only about 30% increase in computation time than without scatter correction.

No. 472

S Jang, RJ Smith, JR Saffer, LI Araujo, KJ Kilroy, JS Karp

University of Pennsylvania, Philadelphia

It was concluded that attenuation corrected SPECT MIBI on Prism 3000 is equivalent to N-13 attenuation corrected PET on PENN-PET.

No. 473

MP Tornai, C McCulloch, RJ Jaszczak, KL Greer, VE Johnson, RE Coleman

Duke University, Durham

The gated cardiac dynamic image was morphed into a template of end diastole or end systole. Contrast seemed to get worse but statistics were better and motion blurring effects were minimized.

No. 474

W Kool, DJ Kadrmas, EC Frey, BMW Tsui, DA Piers

University Hospital Groningen, Groningen, University of Utah, Salt Lake City and University of North Carolina at Chapel Hill, Chapel Hill

Authors describe their simulation using MCNP code for UHE collimator imaging situation. They got good agreement to measured line spread functions but did not state if there was absolute agreement.

No. 475

J Peter, RJ Jaszczak, RE Coleman

Duke University Medical Center, Durham

Description of the object using solid geometry surfaces defined by second-degree equation in 3 variables makes for efficient specification and computation.

No. 476

DA Dougherty, IT Hsiao, W Wang, GR Gindi

State University of New York, Stony Brook

In an elegant experiment using autoradiography with C-14 and partial volume corrected FDG PET images the authors were able to quantify partial volume errors in monkey brain.

No. 519

IG Zubal, L Daley, C Ng, H Dey, F DiFilippo, JM Baron, JP Seibyl

Yale University School of Medicine, New Haven, Picker International, Highland Heights and Tel Aviv Medical Center, Tel Aviv

A Picker 2000 PCD system was equipped with low energy collimators and test performed to see if simultaneous imaging of Tc-99m and F-18 could be performed. The answer seems to be that it is possible.

No. 520

C Hua, WL Rogers, SJ Wilderman, JW LeBlanc, NH Clinthorne

University of Michigan, Ann Arbor

The conclusion is that performance increases at low energy will not be great because of Doppler broadening. (See No. 193) However at higher energies significant performance improvement can be expected.

No. 525

JS Karp, RJ Smith, G Muehllehner, ME Daube-Witherspoon, H. Hines

University of Pennsylvania, Philadelphia, UGM Medical Systems; NIH, Bethesda and ADAC Laboratories

NEMA is coming out with a new PET testing document. This abstract discusses the new image quality test than is being worked on. The new test attempts to compare scanners based on clinical performance with a focus on lesion detection. Acquisitions will be normalized to 100 cm/hr in attempt to compare widely different hardware. Phantom is based on IEC PET phantom. Out of field activity will be accounted for. Contrast and SNR will be measured with ROIs.

Can't come too soon.

The poster layout was about the worst I have ever seen. Tall, narrow panels that caused one to crouch on the floor to read the posters or to crane one's neck high in the air. It was hell for wearer of trifocals. And the posters were in artificial corners so it was impossible to walk the aisle and scan for the interesting poster. Hope the new program committee gets the message.

No. 767

PJ Slomka, D Cheng , AA Driedger

University of Western Ontario, London, Ontario

This looks like a nice approach. Of course, there are many such browsers being developed. Just take a look around InfoRAD at RSNA. How to judge the quality and utility? I think only using it for a week or two under field conditions would permit one to judge.

No. 768

D Dey, PJ Slomka, L. Hahn, R. Kloiber

University of Calgary and Foothills Hospital, Calgary and University of Western Ontario, London, Ontario

Registration techniques using a radionuclide transmission scan and a normal CT scan was developed. Error using fiducials on chest phantom was 2.8 +/- 1.7 mm for a counts based method and 3.5 +/- 2.3 mm using a uniformity index method. Reproducibility was about half the above error. The claim is that it will work for any section of the head or body.

No. 770

J Li, Z Liang, J You, F Yang, H Kuan, C Wong, G Wang

State University of New York, Stony Brook and Brookhaven National Laboratory, Upton

It took me a long time to understand why one would be interested in the combined sampling patterns of the PMTs and the collimator. Late in the poster we see that they are interested in collimated F-18 imaging and I think they are after a filter to remove the hole pattern. But it is not clear.

No. 773

HWAM de Jong, FJ Beekman PP van Rijk

University Hospital Utrecht, Utrecht

The authors call attention to the need to consider backscatter when using energy windows other than the photopeak for Tc-99m. For Tl-201 even the photopeak has appreciable backscatter.

No. 774

NK Doshi, A Boutefnouchet, SR Cherry

UCLA School of Medicine, Los Angeles

Another dedicated imager was simulated. This one is maxPET or mammary and axillary region PET system. Planar detectors of 2 mm LSO crystals could be used for planar imaging to limited angle tomography.

No. 777

F Daghighian, S Vallabhajosula, NH Bander, SJ Goldsmith, SM Larson

Sloan-Kettering Institute for Cancer Research and New York Hospital/Cornell Medical Center, New York, NY

No. 778

R Freifelder, JS Karp, S Surti, JA Wear

University of Pennsylvania, Philadelphia

The curved panel crystals of NaI are also being used to build a dedicated PET scanner for breast. Limited angle tomography would be used. 3.8 mm spatial resolution, 10% energy resolution. Curve plates will help in count rate issues since light will not spread as far. I wonder if any of these dedicated breast imagers will make it past the prototype stage. I think so.

No. 779

A Savi, C Landoni, MC Gilardi, L Gianolli, I Castiglioni, V Bettinardi, G Rizzo, L Travaini, F Fazio

University of Milano, Milan

They measured the physical parameters of MCD vs. GE Advance. Resolution comparable or better for MCD. Sensitivity, count rate and scatter worse for MCD except in Advance 3D mode. Nice table in the abstract.

No. 780

DL Kirch, JE Koss, PP Steele, TK Johnson

Western Cardiology Association and University of Colorado, Denver

The same people that started 7 pinhole imaging are multiplying by 3 using 3 Picker Dynamos. No motion required after placing the cameras orthogonal to each other. Suitable for dynamic SPECT. Wide energy range. High sensitivity. Good angular sampling. While I am favorably inclined toward pinhole imaging this is not what I would bet on.

No. 782

IW Pang, HH Barrett, G Kastis, DD Patton, JD Sain, DW Wilson

University of Arizona, Tucson

FASTSPECT, the 24 pinhole camera, was used for dynamic SPECT of an artificial heart. Parameters were obtained and spatial resolution of 7.8 mm could be achieved.

No. 783

TR Oakes T Hurtado, S Jivan, TJ Ruth

University of British Columbia, Vancouver

Want a low scatter source? Use a gas source. These authors used F-18 in 3.2 mm of Lucite to achieve a scatter fraction due to the source of 2.09%.

No. 784

V Sossi, M Krzywinski, P Cohen, K Knitzek, K Hudkins, J De Rosario, KS Morrison, S Jivan, RR Johnson, TJ Ruth

University of British Columbia and Lions Gate Hospital, Vancouver

This was one of the best sources of info on MCD type detectors that was presented at the meeting. Noise equivalent count rate (NECR) was reported for photopeak mode (3.9 kcps) and for Photopeak- Compton mode (5.1 kcps). System sensitivity was found to be 2.1 and 3.8 kcps/kBq/ml respectively. Measurements were made with a 20 cm dia. by 20 cm long cylinder.

Using a long phantom with out of field activity the 5.1 goes to 1.5 kcps, the 3.9 goes to 1.3 kcps. Very nice study.

No. 785

EM Smith, K Hopkins, V Chengazi

University of Rochester Medical Center, Rochester, NY

This looks to be a useful phantom for some imaging tests but it has 17 fill ports which may make it very difficult to set up. It is basically cylinders in a cylinder.

No. 787

DG Marks, HB Barber, HH Barrett, JD Eskin

University of Arizona, Tucson

Getting uniformly good CdZnTe crystals is very difficult. This leads to worse than expected energy resolution. By using ML estimation the energy resolution improves significantly from 15 keV to 10 keV in one example.

No. 795a

JM Baron, IG Zubal, C Ng, H Dey, J Seibyl

Tel Aviv Med Center, Tel Aviv, VA Connecticut and Yale U

Another phantom study for lesion detectability. The instruments studied were a Posicam, 2 picker PCD systems with 3/4 crystal and 3/8 crystal and 2 SPECT systems 1 Park dual head and 1 Picker 3000. The ranking was as expected. At 8:1 contrast the Posicam could clearly see 10.5 mm sphere while the 3/4 and 3/8 crystals had lower detectability. Results changed significantly as other hot lesions were put in the field and as count rate increased. Authors call for careful phantom models to characterize these systems.

No. 796

SR Meikle, S Eberl, RR Fulton

Royal Prince Alfred Hospital, Sydney

Did not see the authors of this poster from down under. They did a phantom study to determine the best way to extract parameter form dynamic ET studies. They concluded that the method of Carson and Lang (EM-PIRA) is superior .

No. 799

VY Panin, GL Zeng, GT Gullberg

University of Utah, Salt Lake City

If the computer could determine the correction for attenuation without an explicit map greatly increase throughput would follow. In No. 134 Moore shows that it can be done in Ga-67 imaging. It is tougher with Tc-99m. These authors were able to get the attenuation map of the MCAT phantom with their improvements.

No. 800

T Takayama, T Ichihara, N Motomura

Toshiba Medical Engineering Laboratory, Otawara

The Tl-201 emission data was found to be 10% contaminated with Pb X-rays from the Gd source. Using a simulated Gd spectrum this could be accounted for to get a lead free Tl spectrum.

No. 804

RD Badawi, MA Lodge, PK Marsden

Guy's and St. Thomas' Hospital, London, UK

While the authors were able to implement 3 methods for variance reduction techniques in randoms correction it proved to provide a surprisingly small gain in the over all image quality.

No. 807

ML Goris, A Cortes, WM Pace, C Fujii, S DeBruin, HW Strauss

Stanford University School of Medicine, Stanford and Siemens Medical Systems, Hoffman Estates

In a patient study the authors attempted to sort out the contribution of attenuation correction by comparing difference between prone and supine imaging. They found that attenuation correction does indeed minimize attenuation effects and most of this comes from the attenuation correction and not from the iterative reconstruction.

As usual the scientific exhibit section was rather small. It is too bad it is so small since there is a very high educational value to a properly prepared scientific exhibit. SNM could learn a lot of RSNA in this regard. Of course to be properly done it would need a proper infrastructure something the RSNA has in abundance and SNM does not. Maybe we are just too small.

No. 1305

Da Wilson, RK Halkar, JR Galt, D Schuster, KA Scheidt, WJ Fajman

Emory University Hospital, Crawford Long Hospital and Grady Memorial Hospital, Atlanta

Maybe attenuation correction should become the standard of practice.

No. 1310

DL Kirch, JE Koss, PP Steele, TK Johnson

Western Cardiology Associates, Englewood and University of Colorado, Denver

This received an honorable mention. The authors try to reduce routine QC of a SPECT system to a suite of 5 tests. Trying to keep it simple and manageable. A worthy goal. This exhibit is especially oriented to heart imaging. Especially attractive because it checks gated acquisitions.

No. 1312

C Burger, K Mikolajczyk, M Grodzki, P Rudnicki, M Szabatin, A Buck

University Hospital, Zurich and Warsaw University of Technology, Warsaw

This won 2^nd place. In order to make quantitative processing more available and useful these tools were developed. Performance is acceptable, they indicate.

No. 1313

YE Erdi, H Kalaigian, SM Larson, JL Humm

Memorial Sloan-Kettering Cancer Center, New York

The Memorial folk have been working to put their vast knowledge of thyroid dosimetry on the Web via Java programs. This will be well worth a visit when it is ready.

No. 1314

RE Wendt III, ES Delpassand, DA Podoloff

UT MD Anderson Cancer Center, Houston

No. 1315

DC Truong, SC Huang, C Hoh, D Vu, SS Gambhir, ME Phelps

UCLA School of Medicine, Los Angeles

This also won a second place. Never did see the first place tag. I think this is a good idea. Can it gain widespread acceptance? I suppose it would have to be commercialized to make it widely available. But then Linux is not really commercial and it is widely available. May it just has to be great.

No. 1316

SR Cherry, A Chatziioannou, NK Doshi, Y Shao, RW Silverman, AJ Annala, M Pedarsani, ME Phelps

UCLA School of Medicine, Los Angeles

Where do the guys at UCLA get all the time and resources to prepare the educational materials as well as develop the hardware? A correspondent has since informed me that this won the first prize for Scientific Exhibits.

Precious little time to see many vendors. Here are a few observations I was able to make:

Siemens had a big booth with lots of equipment and activity. They seem to be creating a lot of interest these days with lots of new products. Of course, trying to get specs on the new coincidence camera is impossible, they have none. They will take orders, they have 510k but they do not have specs. Just images and precious few of those. Isn't American health industry wonderful?

Siemens is once again pushing MedView software from MedImage as an alternative display package. Smart move. We have it (for 2 months) and like it. It runs on PCs too, not just Macs. What I need from MedImage is a contact that can help us customize it. So far no one has given me other than customer support names and that is not enough.

GE had a small booth crammed full with equipment to the point it was difficult to get around. Lots of people it seemed.

SMV was there in force but I am still confused by the two camera lines. I tried to get specifics about the coincidence effort but was not able to hook up with the one guy who had the answers. Why only one guy with the answers?

ADAC had a large presence with C-PET in the forefront. That is the new curved plate NaI PET camera from UGM. That seems to be the only product from UGM as the Quest/PENN-PET model seems to have gone away. I did not get specifics on this. Who supports the older UGM stuff? GE (groan) or ADAC (groan) or UGM(?). Curve plate technology will improve the NEC for UGM from 35 to 43 kcps. FORE/OS-EM reconstruction will be released. 5.0 to 5.5 mm resolution and the coincidence window will be reduced from 12 to 8 ns. Energy correction will get scatter fraction down to <30%. I was impressed that they had full data sheets for this new product.

Trionix had a LARGE booth but few people in it. I am afraid for them.

Picker was present in force but I did not explore their booth.

Digirad was present and taking orders for the CdZnTe camera. I was disappointed in the 8% energy resolution. But not surprised. It is difficult material. Hope they can weather the storm. The major feature I see of their camera is it has limited dead space at the edge and can do e.g. medial views of the breast at very close distance.

A new company for me was Summit Technologies of Mundelein IL. Seem related to old Medx. Is Medx still around? They are selling the NuQuest computer system that runs on a PC. This system had its origins in Argentina and has been at SNM for several years in a row.

ICP (Institute for Clinical PET) deserves commendation for fighting the good fight and finally seeing PET receiving acceptance. I seem to have several membership applications in my pile of stuff (Expensive!). Did you see their new quarterly journal Clinical Positron Imaging.

SNM Marketplace would indicate that the SNM is going into business for itself. Just a catchy word for the books and educational materials they sell, I think. But there was a vendor in with them, Digital Publishing, Inc. These people have the JNM on CD-ROM for high prices. They have this years abstract on CD, too. I say high prices because many societies are GIVING their members CDs with their journals on it. Why not SNM? Are we too commercial? What is the link between Digital Publishing Inc. and SNM?

Syncor was pushing tungsten shielding for pigs and syringe shields. Isn't that terrible expensive? But they would be far more durable that lead.

Robert E. Zimmerman

Medical Physicist

Joint Program in Nuclear Medicine

Harvard Medical School

Brigham & Women's Hospital

Department of Radiology

75 Francis St.

Boston MA 02115

zimmer@bwh.harvard.edu

Tel: 617 732 7196

FAX: 617 732 6336