Reconstruction and filtering methods for quantitative cardiac SPECT imaging

Abstract

Cardiac SPECT has become an important imaging tool for the diagnosis of cardiac diseases. In order to obtain the best image quality for clinical diagnosis, it is necessary to understand factors which affect reconstructed images. The understanding allows us to devise image reconstruction and filtering methods which provide quantitative cardiac SPECT images. In this paper, we demonstrate individual and collective effects of major factors which degrade cardiac SPECT images obtained using conventional reconstruction and filtering methods with 180° and 360° prajection data. We also present reconstruction methods which provide accurate compensation for the image degrading factors. The improved reconstructed images are compared with those obtained from conventional methods in terms of artifacts, distortions, spatial resolution, contrast, noise and quantitative accuracy. Data from a simulation study using a realistic cardiac-chest phantom and fram a clinical Tl-201 study are used in the investigation.