Research Project
10291892
NIH Project Summary (30 lines max) The proposed program aims to improve the diagnosis and prognosis of coronary artery disease (CAD) at clinics by making dynamic nuclear imaging accessible to a wide population. This protocol, currently available primarily at select research facilities, has long been known to be superior for quantitative cardiac perfusion imaging. Yet, dynamic nuclear imaging has almost no penetration in clinical diagnostics. The reason is a combination of several technical challenges, the most noticeable one being its unreasonably long scanning time. Our principal hypothesis is that the unique diagnostic parameters like myocardial blood flow and coronary flow reserve can be obtained by dynamic imaging acquisition over a much shorter scan time than what is being practiced today. Reducing the scan time will remove the primary impediment of translating dynamic nuclear imaging to the clinic. Our objective will be achieved with innovative algorithm development for the two nuclear imaging modalities: single photon emission computed tomography (SPECT) and positron emission tomography (PET). Previously we have developed cutting-edge algorithms based on a novel mathematical tool called non-negative matrix factorization (NMF) for both of these modalities. In this project, we will advance these algorithms significantly toward our objective, while also improving the accuracy of estimating diagnostically valuable information. For this purpose, we will develop additional high-level mathematical components like topology-based analysis of 1D time-series data and anatomical shapes of organs to guide the solution in the data analysis. We will validate our results with realistic numerical phantoms and clinical data with the radiotracer 13NH3 in PET and 99Tc-tetrofosmin in SPECT. Anonymized retrospective data will be available from our collaborators at the University of California, San Francisco (UCSF). The success of our project will bring the application of dynamic cardiac nuclear imaging closer to clinical utilization by proving its quantitative capability to be diagnostically more useful than the currently practiced static imaging, and thus, will add more value to the diagnosis of CAD, saving lives and cost. In addition to developing the new image processing approach. We will work on automating the dynamic data analysis, as currently, the need for manual intervention is a significant source of subjectivity and cost. The project team will include experienced medical physics researchers and dynamic SPECT imaging pioneers as co-investigators, a medical imaging data analyst as a paid consultant, and a nuclear medicine physician and a nuclear cardiologist as supporting personnel. The project will include a vital student training component with collaboration from the UCSF. If successful, our approach will be applied in larger scale studies for clinical validation at UCSF, using additional imaging agents for myocardial perfusion imaging. The result of our work will encourage patients and physicians to take advantage of dynamic cardiac nuclear imaging, which will add more value to the diagnosis of CAD, saving lives, and reducing cost.
