DEVELOPMENT AND SIMULATION-BASED VALIDATION OF PRACTICAL METHODS FOR ESTIMATING WHOLE-HEART BLOOD-FLOW USING PERFUSION TRACERS AND A DUAL CAMERA SPECT SYSTEM
Embargo until
Date
2017-02-26
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Johns Hopkins University
Abstract
Dynamic changes in tracer uptake in the heart may be useful as early predictors of cardiovascular disease or heart failure [1]. Dynamic imaging can provide additional information to aid in diagnosis. Measuring the ratio of blood flow at stress (maximum flow) and rest (normal flow) can provide quantitative information about presence of either focal defects or microvessel disease. Knowing the state of microvasculature could provide an early indicator of cardiovascular disease before blockages of large arteries develop. Dynamic SPECT imaging has been proposed as a means to measure the kinetics of a tracer, and the resulting kinetic rate parameters represent underlying absolute physiological measures such as myocardial blood flow.
However, dynamic myocardial perfusion SPECT is complicated by a number of factors. First, SPECT acquires only small number (typically 2) of projection views of the patient at a time. Thus, dynamic SPECT must use fast rotation or complicated reconstruction methods that model the change of tracer distribution during acquisition. These reconstruction methods are complicated and computationally intensive, so the full physics is often not implemented. Thus the activity distribution estimates are degraded by a number of factors, resulting in the potential for biased kinetic parameters. In addition, the noise levels in SPECT images are high and voxel values are likely not estimable, in the sense that unbiased estimators do not exist, due to the ill-posed nature of the SPECT reconstruction problem. Thus, dynamic images are likely to be very noisy and regularization is required.
We propose a new approach to Dynamic SPECT based on the previously-developed QPlanar (Quantitative Planar) method for estimating values of activities in VOIs from a pair of projections. This allows estimating the activity in the myocardium, blood pool, and defects at each projection without requiring complicated acquisition protocols and time consuming, complex reconstructions. Kinetic analysis of these time-activity data could allow estimation of flow in a small number of VOIs in the heart from standard SPECT/CT instrumentation. In this work we performed computer simulation studies based on published literature of Thallium-201 (201Tl) and 99mTc-teboroxime kinetics to validate the proposed method and to investigate several potential limitations. The results of these studies, demonstrate the potential advantages of this approach over the previous SPECT method, thus warranting it’s in vivo implementation for further validation.
Description
Keywords
Dynamic SPECT, Perfusion imaging