The applicants proposed to develop, implement, and evaluate algorithms that will significantly improve image quality for clinical PET oncology imaging, an din particular for a dual PET/CT scanner under development at our institution. The motivation for this work arises from the unique sensitivity of positron emission tomography (PET) to detect increased tracer uptake associated with abnormal tumor metabolism before structural abnormalities demonstrated by CT or MRI become apparent. This ability is being increasingly used in the identification of disease remote from the primary tumor site by performing whole body scanning, where the patient bed is stepped through the scanner. The diagnostic utility of PET oncology imaging, however, is often limited in practice by low tracer uptake and low data collection rates, resulting in images with high levels of statistical noise. Whole body scanning, in particular, is constrained to short imaging times at each bed position in order to maintain a total scan duration that is acceptable to patients suffering from serious disease, leading to increased statistical noise and further degradation in diagnostic utility. The applicants suggest that these limitations on image quality can be overcome by taking advantage of two factors: the use of higher sensitivity volume-imaging (3-D imaging) to increase intrinsic scanner sensitivity, and the use of true 3-D statistical reconstruction methods that reduce noise propagation and include a priori information on image smoothness. In addition, the applicants have the opportunity to include accurately registered CT information from the new PET/CT scanner to control the local smoothing information to further improve image quality. There are challenging problems in the development of 3-D statistical reconstruction methods and in the incorporation of CT data. The solutions proposed in this work will initially be evaluated with simulation and phantom studies, and subsequently with clinical data from the current PET oncology program, using observer studies and biopsy results. This overall approach of reducing image noise to improve the discrimination of benign and malignant lesions within the body is needed to realize the full potential of PET oncology imaging and maximized its impact on patient management.
