Efficient Monte-Carlo based system modelling for image reconstruction in preclinical pinhole SPECT

Minh Phuong Nguyen,Marlies C Goorden,Ruud M Ramakers,Freek J Beekman
DOI: https://doi.org/10.1088/1361-6560/ac0682
IF: 3.5
2021-06-16
Physics in Medicine and Biology
Abstract:Abstract The use of multi-pinhole collimation has enabled ultra-high-resolution imaging of SPECT and PET tracers in small animals. Key for obtaining high-quality images is the use of statistical iterative image reconstruction with accurate energy-dependent photon transport modelling through collimator and detector. This can be incorporated in a system matrix that contains the probabilities that a photon emitted from a certain voxel is detected at a specific detector pixel. Here we introduce a fast Monte-Carlo based (FMC-based) matrix generation method for pinhole imaging that is easy to apply to various radionuclides. The method is based on accelerated point source simulations combined with model-based interpolation to straightforwardly change or combine photon energies of the radionuclide of interest. The proposed method was evaluated for a VECTor PET-SPECT system with (i) a HE-UHR-M collimator and (ii) an EXIRAD-3D 3D autoradiography collimator. Both experimental scans with 99m Tc, 111 In, and 123 I, and simulated scans with 67 Ga and 90 Y were performed for evaluation. FMC was compared with two currently used approaches, one based on a set of point source measurements with 99m Tc (dubbed traditional method), and the other based on an energy-dependent ray-tracing simulation (ray-tracing method). The reconstruction results show better image quality when using FMC-based matrices than when applying the traditional or ray-tracing matrices in various cases. FMC-based matrices generalise better than the traditional matrices when imaging radionuclides with energies deviating too much from the energy used in the calibration and are computationally more efficient for very-high-resolution imaging than the ray-tracing matrices. In addition, FMC has the advantage of easily combining energies in a single matrix which is relevant when imaging radionuclides with multiple photopeak energies (e.g. 67 Ga and 111 In) or with a continuous energy spectrum (e.g. 90 Y). To conclude, FMC is an efficient, accurate, and versatile tool for creating system matrices for ultra-high-resolution pinhole SPECT.
engineering, biomedical,radiology, nuclear medicine & medical imaging
What problem does this paper attempt to address?