PT - JOURNAL ARTICLE AU - Abdalah, Mahmoud AU - Boutchko, Rostyslav AU - Mitra, Debasis AU - Shrestha, Uttam AU - Seo, Youngho AU - Botvinick, Elias AU - Gullberg, Grant TI - Liver and Background Removal in Dynamic Cardiac SPECT DP - 2015 May 01 TA - Journal of Nuclear Medicine PG - 1527--1527 VI - 56 IP - supplement 3 4099 - http://jnm.snmjournals.org/content/56/supplement_3/1527.short 4100 - http://jnm.snmjournals.org/content/56/supplement_3/1527.full SO - J Nucl Med2015 May 01; 56 AB - 1527 Objectives We aim to improve the reconstructed image quality in dynamic cardiac SPECT data by generating images of different tissue types based on their tracer uptake behavior. Different tracer dynamic in the myocardium and in the liver can enable us to create an image of myocardial and surrounding tissues along with their corresponding activity from other organs and the background activity removed. This technique is important for studying and quantitating the tracer uptake in the region of the myocardium nearest to the liver, which can be much brighter than the heart.Methods A dedicated reconstruction algorithm, spline initialized factor analysis of dynamic structures (SIFADS), was applied to SPECT projections. The algorithm represented the dynamically changing volume from projections as a superposition of several coefficient arrays and corresponding temporal factors. Initially, temporal spline functions were used as factors, the results were sampled in the selected regions to produce three unique curves characterizing for that region, both the coefficients and the factors were further refined from projections. Ideally, the factors describe tissue-specific time-activity curves and the coefficients describe the spatial distributions of the respective tissues. The distribution of the myocardium coefficient array was used as the image of the heart tissue without the liver component.Results The method was applied to dynamic cardiac SPECT data consisting of 10 rotations of 120 one-second projections obtained immediately after the injection of 25 mCi of 99mTc-tetrofosmin. From these data, the algorithm produced a set of three factors and coefficients. The combination of the factors and coefficients selectively removed liver and background activity from intraventricular blood and myocardium.Conclusions The SIFADS method provides the capability of removing undesired liver and background activity in dynamic images, thus providing improved capability for delineating cardiac lesions.Research Support R01HL050663