Notes from RSNA 1997 Robert E. Zimmerman zimmer@bwh.harvard.edu I missed last year's RSNA but I just could not stay away this year. I think I am hooked. I am amazed at how it just keeps growing and growing. And the degree of organization is just remarkable. The big news was that there is now a North and a South building adjacent to each other and very easy to move from one to the other, at least at the level of the technical exhibit floor. The old East building was used for some events but I never even made it over there. Long waits for busses at the end of the day continue to be the worst part of the whole RSNA experience but free Metra (subway) passes encouraged many to try to travel to selected hotels using public transportation. Still not ideal but there is hope there. Never have liked being dependent on the busses provided. Weather was benign this year except on the Friday-Saturday at the beginning of the meeting. Serious fog problems at O'Hare and Midway airports caused lots of problems for many attendees who tried to arrive early. The RSNA-AAPM can be all-consuming if you allow it to happen. There is no trouble filling up all six days, and your evenings too, with worthwhile activities. I only had three days this year to accomplish two very specific tasks: 1. Clarify the distinctions between the various PET imagers available 2. Finalize an archiving solution for a Siemens ICON network I also felt obliged to take advantage of the educational aspects of infoRAD, catch a couple of scientific sessions, review some scientific exhibits and attend at least one refresher courses. Well it proved too hectic and not all goals were met and this will prove to be a very limited report. Specific Tasks PET machines There are currently only two companies providing PET imagers, GE and Siemens-CTI. The Positron Corp. seems to have gone away. I am not sure what happened to them. They were to have entered some big deal with GE some months back but that fell through and it seems the company has collapsed. Some people were sorry to see this happen as their equipment had its admirers. That leaves us to pick the best supplier/imager between GE and Siemens-CTI. Each company offers two models that need to be considered for our situation: A busy nuclear oncology practice with no cyclotron and no prospects for getting one. Hence all applications are expected to be with FDG. We want to perform rapid whole body acquisitions with attenuation correction. Hope to have a throughput of at least 8 patients/10 hour day. Will need to do the occasional brain scan and cardiac scan but theses are expected to amount to less than 10% of the workload. GEMS -PET http://www.ge.com/medical/nuclear/advance.htm There are two models from GE. At the low or mid price range there is Quest, the innovative NaI based imager from GE partner, UGM. This is the direct descendent of Penn-PET. The other model is the Advance which is a very high priced unit that has some great specs and some good possibilities for 3d (septa- less) PET. The big question with GE is their commitment to PET. After a developing a world class machine in the early 90s and picking up distribution rights to the mid-price Quest machine GE looked poised to become a major player in the PET marketplace. For some inexplicable reason they have fumbled big time by trying to handoff the PET business to a much weaker Positron Corp. Potential customers have to be very worried about GE's corporate commitment to nuclear medicine and PET in particular. Siemens-CTI http://www.sms.siemens.com/nmg/indexfull.html Between them CTI and Siemens seem to have a top flight team for designing, producing and marketing PET instruments. With three ECAT PET machines they cover the gamut from the (cheap?) partial ring ART system to the midrange EXACT and the top of the line, very expensive EXACT HR+ instrument. Their commitment to PET and nuclear medicine is unquestioned. Their track record for continued development in technology, software and refinements has been well demonstrated. The decision gets difficult because one has to weigh what can be done now with the promise of fully 3d (septaless) imaging in the (near) future. We are not PET researchers and we need a clinical product NOW. And we need continuing support in the future. And the dollars are limited and have to be invested wisely. Archiving solutions The basic problem I am faced with is the following: What is the best archiving solution or mini-pacs for a nuclear medicine department that presently and for the foreseeable future has a single vendor (Siemens e.cams with ICON computers) equipment? In my experience this situation is not common. It is much more common to have a department filled with equipment from many different vendors with a wide range of ages. In this common situation it is real tough to come up with an economical solution (<$100k). When the only computer in the mix is Macintosh the decision becomes clear cut: Buy a DICOM archiving solution or buy a Macintosh RAID solution. There is no comparison with regard to expense. DICOM archiving solutions start at $100k and go up from there. Way up when you add the DICOM display stations. On the other hand great Macintosh archiving solutions exist for $25k to $50K depending just how much storage you want to buy. So I came to RSNA looking for a reason why we should spend $100k, or, phrased another way, what do we lose by going with a low-cost all-Macintosh archiving solution? The conclusion I came away with is that we lose almost nothing with the all- Macintosh solution and save the many dollars to add to the PET purchase to ensure that it will integrate easily into the department networking needs. DICOM has its attractions, especially in the context of a radiology department-wide PACS solution but as a solution to nuclear medicine archiving it has too many drawbacks and it is way too expensive, in my opinion. In our case there is already a PACS solution for the Radiology Department, it only remains that we have a means to export our nuclear medicine images. This is easily accomplished with ICON software. I think it is a no brainer. Perhaps I have overlooked something. DICOM was way overplayed at RSNA this year as it has been in recent past years as well. For example TDK, the magnetic media people, were hyping a DIOCM compatible, "Medial Grade CD-R" with "FDA 510K approval number" that comes complete with a "MedCaddy". Where does that leave "ordinary" CD-R? It is a moot point because CD-R still has only 600 MB capacity. OK for the desktop and small applications but not much for a real archiving solution for images, I think. Refresher courses The RSNA continues to have comprehensive educational programs for everyone with over 200 refresher course sessions. But if you are a nuclear medicine physicist there is almost nothing to attend in your area. But it is a great chance to broaden your knowledge in areas outside your specialty. I have never been disappointed by choosing a course directed toward the practicing physicist. This time I chose the Categorical Course in Physics: The Basic Physics of MR Imaging organized by Stephen J. Riederer and Michael L. Wood. I was not disappointed. There were 16 presentations and speakers over the 5 days (I made 8 in the first 3 days). All speakers were excellent and perhaps the best part of the course is the Syllabus which was available for $45. Excellent opportunity to see what our colleagues have been up to with all those new pulse sequences. Scientific sessions It is a problem to find good nuclear medicine scientific sessions at the RSNA. Most people bring their best work to the SNM meeting in June. RSNA gets replays and inferior work for the most part. And with only 6 + 3 minutes per presentation there is little information that gets transmitted. Nevertheless I felt obliged to pop into some AAPM physics sessions. It was a good way to get some exercise and learn the new hallways of the new, improved, gargantuan McCormick Place. Monday: Physics (Nuclear Medicine) 509 Do tracer dosimetry studies predict normal organ and tumor uptake of I-131-anti- B1 (anti CD-20) antibody in patients receiving radioimmunotherapy for non- Hodgkins lymphoma. KR Zasadny, VL Gates, S Moon, DD Regan, MS Kaminski, RL Wahl Ann Arbor The answer would appear to be yes, especially for the tumor. In-vivo biodistribution was measured with both tracer doses and therapy doses on patients at approximately 4 days post injection. 510 Comparison of coincidence imaging parameters of a dual-head scintillation camera with 3/8 and 5/8 inch thick crystals JA Patton, MB Stoner, MP Sandler, Nashville Dr Patton showed that the resolution loss with the thicker crystals was about 10% but the sensitivity gain in coincidence mode was about 180%. 511 Small field of view gamma camera for scintimammography MB Williams, r Godde, F Farzanpay, S Majewski, D Steinbach, A Weisenberger Charlottesville This planar camera uses 2000 CsI crystals, 2 x 2 x 6 mm with a single position sensitive PMT. Resolution was measured to be 2.6 mm FWHM and 4.9 mm FWTM. The energy resolution of 25% will require some scatter reduction techniques 512 Development of a quantitative imaging and MIRD dosimetry software application for radionuclide therapy treatment planning using a forth generation language WD Erwin, M Groch, SM Spies Chicago Using IDL the authors developed a program to aid the processing of data from serial WB scans. Time activity curves can rapidly be calculated from the serial planar scans, fit to equations, parameters derived and passed to an integrated version of the MIRDOSE dosimetry program. An otherwise tedious chore has become highly automated and friendly. 513 Combining X-ray and radiotracer mammography images IN Weinberg, WA Berg, R Pani, F Scopinaro, G Bakale, LP Adler Bethesda By using a crystal array miniature PET imager, compression paddles and X-ray system the authors succeeded in registering the FDG breast image with the mammogram for a patient. 514 Serial SPECT and correlative image registration to improve quantitative analysis DA Weber, M Ivanovic, DK Shelton, J Xu Sacramento Does image registration in the brain aid in the reproducibility of the evaluation of brain perfusion? The authors found that improvement in ROI placement and slice selection in the registered images improved the accuracy and reproducibility of the quantitative analysis. Physics (Image Processing: PACS, Workstations) 699 Implementation of an HL7/DICOM broker for automated entry of patient and exam information into computed radiography modalities SG Langer, BK Stewart Seattle Technologist productivity improved 44% when the RIS-HIS-DICOM path for transmission of patient and exam information was implemented to seven CR systems. Availability of accession number at the CR system aided image tracking in the distributed PACS system. The impressive part of this paper is that this was accomplished with a legacy HIS and RIS (IDXrad). Solutions as described in this paper are important. It is hoped that software like this can be marketed in all commercial PACS systems. 700 DICOM Image integration into a web-browsable electronic medical record BK Stewart, SG Langer, JI Hoath, P Tarczy-Hornoch Seattle Stewart described the approach for making an electronic medical record (with images) available using web server technology. Browsers can easily access text and images. The RIS accession number serves as the link to the image database and test record. Sounds like a straightforward approach. Are commercial companies adopting solutions like this? The Brigham and Women's Hospital in Boston has previously described a similar approach. http://www.partners.org/cgi-view/dbml.exe?template=/pw- view/optional/optional-interpage.dbml&item_id=7109&dept_id=4). There may be other hospitals that have also solved this problem. Now I know that the URL above is not something anybody would want to type, but I cannot find a shorter URL that works. 701 Reliability of a hospital-wide picture archival and communication system (PACS) using a central architecture: Analysis of the four-year experience at the Baltimore VA Medical Center EL Siegel, BI Reiner, SM Pomerantz, C Siegler, FJ Hooper Baltimore Authors found 1% downtime due to scheduled downtime and 1% due to unscheduled problems, CR contributed heavily to this downtime as did maintenance activities such as upgrades. Some degree of redundancy will be necessary to improve on these figures. 702 PACS archive architecture integrating DICOM and World Wide Web access FM Behlen, JB Weldy, E Rackus, D Ye Chicago Both a DICOM server and A WWW server for images were found to play complementary roles. Web based queries were especially useful. 703 A quality control program for paper printers used with teleradiology/PACS AD Maidment, M Albert, EJ Halpern Philadelphia Authors feel that regular testing of paper printers is necessary. A pattern is recommended (J. Digital Imaging 8:3-9, 1995). Cheap paper printers have inherent limitations in their dynamic range. Roller marks were a problem. Dye sublimation printers were more consistent. 705 Accuracy of radiologists using a computer workstation in comparison to film in the interpretation of CT examinations BI Reiner, EL Siegl, FJ Hooper, AW Dahlke, SM Pomerantz, Z Protopapas, et al It was found that not only is the use of workstations faster than viewboxes accuracy was improved by 2-8% (non-significant) for four readers. Further software development and improved presentation of data is expected to offer even more speed improvements. 706 Full-field direct digital mammography (FFDDM) display workstation (DWS) MI Jahangiri, SL Lou, DR Hoogstrate, HK Huang Existing workstation cannot load the eight 40 MB images in required 1.5 sec. The window level display needs improvement. This demanding display task cannot be performed, yet, on state-of-the-art equipment. 707 Interpretation of chest radiographs on high-resolution PACS monitor: Optimal ambient illumination in reading room. KS Lee, JH Yoon, DW Ro, TS Kim, JH Hwang, D Choi Seoul Because of the low brightness of monitors ambient light must be greatly reduced, but not all the way to 0 fL or eye fatigue can be a problem. The authors found that 0.9 fL is best for viewing monitors in their comprehensive test using ROC methodology. 708 Portable DICOM-compliant PC/NT-based diagnostic workstation with asynchronous transfer mode technology H Zhang, G Lei, AW Wong San Francisco The incorporation of ATM communication board was found to speed up a PC/NT workstation to where it could compete with a UNIX based workstation in speed at lower cost. Speed improvements of up to 10x were observed compared to Ethernet. Display software needs to have preset windowing and fast DICOM retrieval to be effective. (see also Scientific Exhibit 0417PH). Scientific Exhibits I am always impressed with the amount of work that goes into the scientific exhibits at RSNA. There is a lot to be learned by the critical participant here. While I was not able to give the time many of these exhibits deserve, a quick cruise through the scientific exhibits provided the following nuggets: SPACE 0221MS Use of sentinel node lymphoscintigraphy in malignant melanoma AP Yudd, JS Kempf, J Goydas, TJ Stahl New Brunswick, NJ This technique is much underutilized, I think. SPACE 0229MS Octreoscan scintigraphy in a variety of neoplasms P Farid, CS Firment, KL McDonald A nice tutorial about the uses of In-111 Octreotide. One of the few nuc med exhibits in the Multisystems/Special Interest section of the posters. SPACE 0235MS CT evaluation of damage, repair, and internal wood grain patterns of solist quality bowed stringed instruments SA Sirr, MB Plunkett, LB Hassan, J Waddle, M Darnton Minneapolis Fascinating CT study of old instruments. SPACE 0238MS Image registration between 18FDG-PET and MR imaging using normal internal landmarks H Uematsu, N Sadato, Y Yonekura, N Hayashi, K Yamamoto, Y Ishii Japan Registration improved when sagittal FDG images landmarks such as oropharyngeal mucosa, spinal cord and brain features were included. (See also paper 1097) SPACE 0254MS Hybrid Rendering: A new approach to 3D image reconstruction DR Stelts, CB Siege, DK Ahn, DJ Vining Winston-Salem This may bear further study. www.vec.bgsm.edu Space 0390PH Evaluation of efficiency on CRT diagnosis K Kudoh, T Nambu, S Terae, H Yoshikawa, N Fujita, K Miyasaka Sapporo These investigators think that CRTs can replace viewboxes for routine reading of CT and MR. They did prefer a viewbox for an overview and when the number of images was very large. SPACE 0392PH New image display `Virtual Volume Display" S Azemoto, M Gono, Y Hiramatsu, S Murakami Japan Using technology similar to the 3D postcards this group has produced 61 CT images on one card. May not be practical but it is cute. SPACE 0349PH Three-dimensional display of vascular image with cylindrical lens array: touchable organic image in air A Heshiki, M Amanuma, S Tukuda Japan Another way to make a 3D display. The authors think it would be useful in education and for surgical interventions. Needs more work and an in-depth evaluation. SPACE 0396PH A new CT software for quantitative in vivo measurement of thyroid iodine Y Imanishi, N Ehara, T Shinagawa, D Tsujino, Y Mitogawa, S Konno, et al Japan I was surprised to learn that this could even be done. Can it help the patient? Not answered here. SPACE 0402PH High fidelity electronic display of digital radiographs MJ Flynn, J Kanicki, A Badano, WR Eyler Detroit This was a big display with a lot of meat in it. Looks significant. Wish I had time to digest it all. SPACE 0416 Wavelet compression and its effect on medical image contrast KR Persons, P Palisson, A Manduca, BJ Erickson, V Savcenko, TJ O'Connor Rochester, MN This tutorial was an excellent review of wavelets and their use in image compression. The wavelet compression parameters should be "tuned" to the modality and content for best results. SPACE 0417PH Portable DICOM-compliant PC/NT-based diagnostic workstation with asynchronous transfer mode technology G Lei, H Zhang, AW Wong San Francisco (See paper 708) SPACE 0419PH A comparison of conventional and photothermographic laser camera technology TM Bogucki, WE Paul, PC Bunch, KL Yip Rochester NY These Kodak authors urge us to be cautious in using the new heat-processed films for laser cameras. SPACE 0460PH Testing data compression for teleradiology including JPEG, LOTS, and wavelets A Todd-Pokropek, V Chameroy London UK We are warned to test all compression algorithms on relevant tasks using ROC techniques. Nijmegen phantom can be useful to optimize the algorithms. Artifact are to be expected. SPACE 0846CECH Myocardial perfusion imaging with quantitative gated SPECT: An Interactive multimedia tutorial HD Tran, CM Oliner, AF Guirgues, JA Gubernick, B Shapiro, WB Garfinkle, et al Philadelphia This NM tutorial looked useful to those who wanted to learn more about nuclear cardiology. SPACE 0847CECH Functional cardiac imaging: Normal and pathologic anatomy W Stanford Author used MR and electron beam computed tomography (EBCT) to educate the viewer. Not enough time for this time-challenged physicist to do this one justice. infoRAD I just did not have much time to wander the infoRAD floor of the exhibit. It was still located in the Basement of the North building but was "out of the way" for me this year. Did manage to learn a few things however: SPACE 9100LIT The National Library of Medicine was much in evidence. The MEDLINE database is now searchable for free right from your very own web browser. www.nlm.nih.gov/nlmhome.html This has got to be the best bargain in medicine. Paid for with our tax dollars and available to the world. Who says Uncle Sam is ugly? There are two front ends to the search process: Internet Grateful Med or PubMed. Search away. I wonder what will happen to the vendors of search interfaces for MEDLINE. Will have to develop some sort of value added service, I guess. This free searching seems to meet my needs, however, and at my own desktop, with my own computer and browser. Can't be beat. NLM also offers International MEDLARS for distribution of retrieved documents. This seems to work in conjunction with local libraries and other agents around the world. SPACE 9106CAI Creating a multi-institutional Internet-based teaching file with a common database schema and distributed query MW Freckleton, RA Greenes, JG Smiirniotopoulos, A Boxwala, JP Ehresman, RB Evans, et al Texas MedShare was created by a consortium of teaching hospitals as a way to share teaching files in a unified way. Looks like a neat way to increase efficiency for the providers and for the learners. They want others to join them. http://dsg.harvard.edu/public/medshare or http://206.39.77.2/qry/medshare/medshare.qry?function=form SPACE 9107 CAI 3D volume rendering - Theory and method: A CAI demonstration program PS Calhoun, BS Kuszyk, DG Heath, EK Fishman Baltimore A really good program to interactively educate and demonstrate the utility of 3D rendering. SPACE 9110CAI Atlas97: An interactive atlas of radiological anatomy with presentation enhancements developed in Java J Eng, EP Tamm Baltimore I'm still looking for useful Java programs. This comes close. It can be tried at welchlink.welch.jhu.edu/~jeng/javarad SPACE 9208WB Issues in developing radiology discussion forums on the World Wide Web D Cheng, LA Fox, V David, HA Caruthers, BL Holman, GL Mammone Boston This group has pioneered in some uses of the web for education. Here they share some of their thoughts on what makes for a good discussion forum. It might be found at http://www.med.harvard.edu/BWHRad/ SPACE 9215 RSNA EJ http://ej.rsna.org/ Well, a free peer reviewed journal that can come to our desktop whenever we are ready. This web distributed radiology journal is blazing new ground. Definitely worth a look. I counted 35 articles and counting for volume 1. Publication started in summer of 1996. Technical Exhibits In searching for the useful I certainly discovered the unusual this year: Does anyone else get confused by the similar names: The Ann Arbor MedImage www.medimage.com the German MEDIMAGE www.compubase.de/medimage.htm and the Maimi Medasys Digital Systems www.medasys.com All of them are doing pretty much the same thing - DICOM archives and viewing stations. MedIamge (DELTA manager, MEDView and Galenon on Mac platforms) and Medasys (UNIX based Dx software and Wintel Based Dx Win) had their genesis in the old days of nuclear medicine, I think, but that German, MEDIMAGE, is new on my radar horizon. Their web pages are in German but AltaVista does a very good job of translating the pages to English. They do seem to have a large user base in Europe. And who would think that a company with the name "American Predator Corporation" with products like "Predator 747" and "Hunter Triton" would be the leading supplier of industrial embedded computers to the radiology OEM market. At least that is what they told me. Brit Systems (Beyond Roentgen Imaging Technologies) www.brit.com of Dallas was showing their version of a DICOM archive based on (you choose): X- windows, Windows, Linux and/or AIX. E-MED a Raytheon company with origins in E-Systems, a defense contractor, tried to entice with their version of a DICOM archive. The trouble is these system are just not cost effective in the existing nuclear medicine environment. Digital Equipment Corp. hosted a booth that housed a number of "solution providers". Included were Dartnell Enterprises Inc., JABR Technology Corp., and ImageLabs. Together this group can build you a complete PACS system at very economical price. Ed Smith of the U Rochester was featured describing his hospital's solution using these vendors. I was tempted but still too much money for a small nuclear medicine department. Apple Computer, Inc. was here with a rather large display. I used their people to get names of Macintosh RAID dealers in the Boston area. This was probably worth the trip to Chicago. Learned a lot about Mac dealerships from the booth personnel. I did not see half of what I think I would have profited from, but I had to leave to get the plane. And I missed it. Have to plan better next time. Bob Z Robert E. Zimmerman Joint Program in Nuclear Medicine Harvard Medical School Department of Radiology 25 Shattuck St. Boston MA 02115 zimmer@bwh.harvard.edu