On-going research activities have focused on advancing PET from a qualitative to a quantitative imaging modality for the non-invasive characterization of coronary artery disease (CAD). Since CAD causes 1.5 million myocardial infarctions and 520,000 deaths per year (one-third to one-half deaths between the ages of 35 and 64) in the United States, it is critically important to define its severity for cardiologists to decide objectively among dietary control and therapeutic interventions (such as thrombolytic therapy, balloon angioplasty, and bypass surgery) and for assessment of effects of these treatments. Over the years, PET has been proven as a highly sensitive and specific diagnostic tool for the detection of CAD. At the present time, PET represents the most promising non-invasive means to quantify regional blood flow and coronary reserve, which has been shown to provide a sensitive marker for the functional significance of coronary artery stenosis. Intensive research has been devoted to developing or validating kinetic models for PET perfusion agents such as N-13 ammonia, O-15 water, and rubidium-82. However, accurate quantification of myocardial perfusion using PET is difficult to achieve because of difficulties in analyzing degraded image data (due to finite spatial resolution, counting noise, accidental coincidence, instrument deadtime, and patient motion) and limitations on myocardial boundary delineation. To overcome such difficulties and limitations, we propose an integrated approach in which a 3D measurement model characterizes the imagine degradation factors, relates perfusion, myocardial boundary parameters, blood pool concentration (input function) and spatial positions of the heart (for patient motion correction) to the projection data, and allows joint estimation of these parameters directly from the projection data based on maximum likelihood criteria. The long term goal of this research is to develop and validate more accurate methods for simultaneously estimating myocardial perfusion and boundaries from gated dynamic PET data.
The aim of this proposal is to prove a hypothesis that the model-based method will provide more accurate perfusion estimates even in unaged situations. To prove our hypothesis, we will develop a 3D measurement model and will construct a 3D digital dog-thorax phantom to refine the model and to develop methods for body movement correction. We will then conduct dog heart perfusion studies using N-13 ammonia, analyze data based on the refined model and simulation experience, and evaluate the performance of perfusion estimation against conventional methods and a gold standard. We will perform preliminary trials on N-13 ammonia patient data.
