Derivation of system matrix from simulation data for an animal SPECT with slit-slat collimator

Abstract—We developed SPECT imaging capability on an an-imal PET system. Our goal was to provide animal PET users the SPECT capability at a low cost and facilitate potential PET/SPECT dual modality imaging applications. The SPECT function was en-abled with a slit-slat collimator insert and by acquiring data in sin-gles mode. The focus of this paper is to establish a method for de-riving the system matrix for the SPECT system from Monte Carlo simulation. With the Monte Carlo package GATE, we simulated a uniform cylinder source which filled the SPECT field of view (FOV). To reduce the size of the original large and sparse system matrix, the detectors that were exposed to individual emission el-ements were selectively included for system matrix derivation and storage. The axial symmetry of the system was exploited so that only the base-axial volume was used for deriving system response. The system matrix derived was validated with point source mea-surements at known positions and implemented in an iterative re-construction algorithm. The imaging performance of the system matrix was evaluated with experimental phantom studies. Recon-structed phantom images were artifact free and demonstrated ex-pected spatial resolution. The method presented in this work is gen-erally applicable to other SPECT imaging systems. Index Terms—Image reconstruction, Monte Carlo simulation, PET/SPECT dual tracer imaging, SPECT, system matrix