The Department of Nuclear Medicine, in conjunction with the National Cancer Institute and the Department of Radiology, performs clinical research in the use of imaging in oncology. In particular, they are studying the use of positron emission tomographic (PET) images, in conjunction with computer tomography (CT) and magnetic resonance (MR) images, to evaluate the effects of therapy on tumors. Several therapeutic agents are being studied, among them various anti-angiogenesis therapies. The PET scanners are used to measure glucose metabolism, blood flow, and blood volume in tumors over the course of therapy. CT scans are used to determine tumor morphology, and MR imaging is used to determine both morphology and parameters related to tumor perfusion. This research is geared toward developing, implementing, and testing methods to better quantify the data obtained from the images and to determine if these methods are efficacious for the monitoring of tumor therapy. These methods involve determination of tumor morphology and optimal determination of functional parameters such as blood flow, metabolism, and blood volume. The overall goal is the development of a clinically useful methodology for determining tumor response to therapy at an earlier phase of therapy than is currently possible. Such a methodology could permit optimal adjustment of the course of therapy while the therapy was still proceeding, potentially improving both tumor response and patient morbidity. Several areas of investigation are being pursued to achieve this goal, including these: oDevelopment of new reconstruction techniques that will improve noise properties of the image sets. This noise reduction will improve the ability of the physician to visually interpret the images and improve the noise in quantitative parameters derived from the images. This work has recently been completed and a paper submitted for publication. oAssessment of the physiologic models employed for blood flow measurement, using O-15 water. Several models are being analyzed, especially with regard to their utility in producing functional flow images. The results of these PET flow models are being compared with similar data obtained from Gd-DTPA dynamic MR images. The variability and reproducibility of each of the methods is also being determined, using replicate measurements. Some of these data have now been analyzed and accepted for publication as a full article in the Journal of Nuclear Medicine. oAssessment of methods for making accurate, noninvasive measurement of the arterial input function. These methods compare LV cavity and aorta-derived arterial input functions with actual arterial sampling. Several schemes are being explored to correct for partial volume and spill in/out effects. Preliminary results of this work were accepted for presentation at the June 2000 Society of Nuclear Medicine meeting. oExploration of methods for using three-dimensional region growing to more accurately assess tumor volume, metabolic volume, and perfused tumor volume. These methods will be employed to make objective assessments of the various physiologic parameters (e.g., FDG """"""""uptake"""""""", and ROC analysis used to determine which of these quantitative indices are best for detecting disease, and to determine if quantitative measures are better than subjective visual assessment. oInvestigation of methods to reduce transmission scan time in PET imaging. Transmission scan time is now a major limiting factor in the clinical application of PET to oncology whole-body imaging. """"""""Watershed"""""""" methods are being explored to segment very short attenuation scans, thus significantly reducing transmission scan time. This work has recently been completed and has been published as an article in the Institute of Electrical and Electronic Engineers Trans. Nucl. Sci.
