The primary goal of this research proposal is to design, construct and perform a limited validation of a new tomograph which has the capability of operating either in X-ray CT or positron tomography (PET) mode. The proposal is to mount an X-ray CT and two arrays of PET detectors on a single support within the same gantry, and rotate the support to acquire a full projection data set for both imaging modalities. The tomograph will therefore acquire functional and anatomical images which are accurately coregistered, without the use of external markers or internal landmarks. A secondary objective is to use the CT data to improve the correction of the PET data for attenuation and for contamination from scattered photons. By using the CT image in a novel way, low-noise attenuation correction factors for PET can be generated, and by integrating the anatomical information from the CT into recently-developed scatter correction methods, an accurate scatter correction can be obtained. The use of 18F-fluorodeoxyglucose (FDG) is a rapidly growing application area for PET, in which the increased glucose metabolism of neoplasms has been used to determine the presence of recurrent glioma versus radiation necrosis, recurrent colon carcinoma versus surgical scar and radiation changes, and radiation necrosis versus reactive mediastinal lymph node. A major difficulty in all these important applications of PET is the accurate localization of the increased glucose uptake to a specific anatomical structure. While multi-modality image alignment has been extensively developed for the brain, little work has been done in the thorax and abdomen owing to the technical difficulties of identifying common structures and to the tendency for internal organs to move within the body. In the proposed combined modality PET/CT tomograph, functional and anatomical images are intrinsically co-registered. For this proposal, a preliminary study of the capability of the new tomograph to detect and localize tumors will be performed in a limited series of ten patients with known malignant disease.
Showing the most recent 10 out of 13 publications