According to recent statistics by American Heart Association that coronary artery disease (CAD) causes approximately 42% of all deaths each year. Currently available myocardial perfusion SPECT is perhaps the most cost-effective means to diagnose the disease. However, its lack of quantitative capability compromises the evaluation of myocardial viability and, therefore, the assessment of myocardial function. The long-term objective of this project is to achieve quantitative myocardial perfusion SPECT by (I) developing efficient reconstruction methods for accurate mapping of radiopharmaceutical distribution inside the myocardium with an improved detector configuration and (II) investigating automatic segmentation techniques for accurate quantification of reconstructed regions-of-interest (ROIs).
The specific aims of this renewal are: (1) To investigate an improved detection configuration: A variable focal-length fan-beam collimated dual-head vertex SPECT system with scanning line sources will improve detection efficiency and spatial resolution and achieve optimal sampling without truncation in both transmission and emission scans. (2) To develop efficient reconstruction and segmentation strategies for the improved detection configuration: Accurate reconstruction and simultaneous compensation for object-specific photon attenuation, scatter and spatially-variant detector response, as well as effective suppresses of noise propagation will be accomplished. Automatic segmentation for quantitative ROI analysis will be achieved. (3) To evaluate the accuracy of reconstruction and segmentation by phantom experiment and patient diagnosis: Criteria of bias-variance graph and miss-segment ratio will be used to quantify the reconstructed images and segmented ROIs against their simulated and experimental anthropomorphic phantoms of the thorax. The performance of variable focal-length fan-beam collimator will be compared to the currently used parallel-hole collimator by means of both receiver operating characteristic curve (human observer) and Hotelling trace criterion (computer observer). Diagnosis of CAD, as well as pulmonary embolism and breast cancer will be assessed
