Validation of SimSET Monte Carlo simulations of the Siemens Biograph mCT PET scanner

Abstract

Introduction: SimSET is a popular Monte Carlo simulation toolkit that features fast simulation time, as well as variance reduction tools to further enhance computational efficiency. However, SimSET has lacked the ability to simulate block detectors until its most recent release. Our goal is to validate new features of SimSET by developing a simulation model of the Siemens Biograph mCT PET scanner. Methods: SimSET is used to simulate positron emission and annihilation, model gamma ray photon absorption and scatter in objects and the LSO block detectors. In-house code is used for front-end detector pulse processing, electronic multiplexing, and coincidence generation. We use the NEMA NU-2 2007 scatter fraction, count rates, and spatial resolution protocols to validate the SimSET simulation model and its new features. Results: The SimSET model overestimates the singles, trues+scatter, randoms, and NEC count-rate curve by 11-22% compared to the experimental results, which compares favorably to previous validation studies of other PET scanners in the literature. For the spatial resolution test, the SimSET model underestimates the experimental results by 11-30%. This is expected as the detector model does not account for blurring due to light sharing and variations in light collection efficiency of the photodetectors. Variations of resolution with position in the field of view track those seen experimentally. Conclusions: The new block detector model in SimSET generates reasonable agreement with the experimental results.