As the 21st century approaches, cancer imaging is undergoing an image shift of its own — from diagnostic tool to one in which imaging is used increasingly to guide treatment outcomes. As this shift occurs, however, many state-of-the-art imaging devices that were perfected in the late 1980s and early 1990s are not yet giving researchers the quality they want or need. Will combining various machines and technologies for optimal composite images make a difference, or are even newer and more inventive devices required? The answers are often site-specific. At a recent meeting on nuclear imaging at Memorial Sloan-Kettering Cancer Center in New York, nearly 200 experts tried to discern how advancing technology might impact both the diagnosis and treatment of numerous cancers. Henry Wagner, Jr., M.D., director of radiation health sciences at the Johns Hopkins University School of Hygiene and Public Health, Baltimore, noted that nuclear oncology is not just about detecting disease, but also about avoiding fruitless surgery and other invasive procedures. “In my opinion, 95% of nuclear oncology is no longer done just for diagnosis,” he said, “but rather to guide treatment outcome.” “Given the time and expense that technologies such as Positron Emission Tomography scanning now engenders,” said Jonathan Links, Ph.D., professor of environmental health sciences at Johns Hopkins, “there is a need for scans to have a maximum impact on treatment outcomes.” And toward this end, Links added, complex mathematical methodologies — only recently made possible by the use of high powered computers — may help sharpen images by maximizing the resolution and sensitivity of PET and its simpler and cheaper cousin, Single Photon Emission Computerized Tomography. Greater precision in PET scans might also help ameliorate concerns about cost if the scans could be utilized to pinpoint treatments. Ward Digby of CTI Inc., in Knoxville, Tenn., noted that as of January there were 100 PET devices installed in the United States and about 300 worldwide. At a cost of $1 million to $2 million a machine and a cost to the patient of $1,500 a scan, costs would have to come down much further, Digby and others agreed, before widespread use of PET occurs. Dual Devices David Townsend, Ph.D., associate professor of radiology at the University of Pittsburgh, said he is developing new technologies with industry that may heighten imaging capabilities. “ Our goal is to combine functional and anatomical imaging by merging a computerized tomography scanner with a PET scanner,” Townsend said. A dual modality device such as this, which recently became operational, could be particularly useful in cancers such as pancreatic cancer where images are often difficult to interpret. Cost of such a device at this stage in its development was not discussed. One seminal question at the New York meeting was how good is all this technology for diagnosing and treating individual cancers. One answer came from Ronald Blasberg, M.D., professor of neurology at Weill Medical College of Cornell University, New York, who said that for brain tumors, a combination approach is effective. “We combined an MRI [ magnetic resonance imaging] scan as an anatomical reference point with a PET scan, and slice-by-slice, co-registered them to yield an image of increased diagnostic accuracy,” Blasberg said. But as advanced and advantageous as many of these imaging systems are, they are not a perfect approach for detecting and diagnosing all cancers, according to Peter Scardino, M.D., head of the Prostate Cancer Program at Memorial Sloan-Kettering. “In 1999, a CT scan is no longer of much value in diagnosing early stage prostate cancer and ultrasound is not local enough,” he said. Scardino also said that MRI scans have better accuracy but are not very good at determining the volume of a prostate tumor. He added that bone scans have not been useful for men with prostate specific antigen levels lower than 8 and were only good for men with PSA levels over 20. Scardino hopes “that a device could be developed that would provide a good prostate cancer image so that margins for surgery or radiation could be narrowed and side effects such as erectile dysfunction and urinary incontinence reduced.” For breast cancer, Steven Larson, M.D., chief of the Nuclear Medicine Service at Cornell, said that the most important development has been sentinel node imaging. Larson believes that PET imaging really has only been useful for breast cancer mainly in later stages when looking at bone or chest wall abnormalities. Also for breast cancer, mammography is recommended by many organizations for women over 40 on an every 1- to 2-year basis. Ette Pisano, M.D., of the University of North Carolina, Chapel Hill, said in February at another seminar that a number of digital mammography clinical trials were nearing their conclusion and awaiting U.S. Food and Drug Administration approval. Pisano said that x-ray film processing is the procedure's biggest problem, but with digital detectors, photons are absorbed directly by a computerized device, reducing error rates. One manufacturer of digital mammography machines, General Electric (Trex and Fisher being two others), has technology that uses existing mammogram units with new digital screens that replace the old x-ray film screens. It is the least expensive technology, but the resolution is not quite as good as the others, according to Pisano. In a soon-to-conclude multisite pilot study of 1,250 women, 18 radiologists compared digital versus film images and the digital image was always preferred. Pisano noted however, that radiologists need time to adapt to reading computer screens with limited viewing areas. “Another developing technology that complements digital mammography is computer-assisted diagnosis, where the computer interprets images and acts as a second read to enhance human performance,” said Pisano. She speculated that digital mammography could replace current methods within a decade, but that cost issues were an important concern. Because GE's digital device only requires a retrofitting of existing machines with new screens as opposed to replacing the entire mammogram unit, the cost of switching to a digital machine could be nominal, Pisano stated. Variations in test accuracy due to reader, subject, and mechanical differences is an issue that encompasses all of these new technologies. Craig Beam, Ph.D., of Northwestern University Medical School in Evanston, Ill., said at a third meeting earlier this year that, “we need to be able to better measure the accuracy of human diagnosticians, how to analyze reader variability, and how to insure substantial sample size in clinical trials to reduce sensitivity and specificity issues.” Open in new tabDownload slide Dr. Henry Wagner, Jr. Photo courtesy Alan Mason Chesney Medical Archives, Johns Hopkins Medical Institutions. Open in new tabDownload slide Illustration shows a patient undergoing whole body positron imaging. Oxford University Press Oxford University Press