Design, Optimization and Performance Evaluation of Bm-Pet: A Simulation Study Publisher



Sheikhzadeh P^{1, 2, 3} ; Sabet H⁴ ; Ghadiri H^{2, 3} ; Geramifar P⁵ ; Ghafarian P^{6, 7} ; Ay MR^{2, 3}
Source: Nuclear Instruments and Methods in Physics Research# Section A: Accelerators# Spectrometers# Detectors and Associated Equipment Published:2019

Abstract

Introduction: The aim of this work is to report on modeling and evaluation of potential performance characteristics of Brain Mapping PET (BM-PET) which is under development at our center. BM-PET is based on expansion of our recently developed small animal PET system (Xtrim-PET) and optimize the geometry of its detector system for brain imaging applications. Methods: We modeled an accurate representation of Xtrim-PET in GATE and after validation of the model, we extended our model to design and optimize the BM-PET scanner followed by defining a cost function and running series evaluation tests in terms of different geometrical, physical and detection parameters. Resolution recovery and attenuation correction techniques were developed and applied to the simulated data. Finally, the performance of the BM-PET system is evaluated and compared with two simulated models of commercially available brain scanners. Results: The optimized rotating cylindrical BM-PET based on LYSO-SiPM detectors with dimension of 2×2×20 mm3 resulted in system sensitivity of ̃16 cps/kBq and 2.1 mm FWHM spatial resolution when tested using the NEMA standard. Detection efficiency of BM-PET is 1.2x and 2.1x higher than our simulated NeuroPET and HRRT scanners. The calculated spatial resolution of the BM-PET in rotation mode is approximately 0.8 and 0.2 mm better compared with that of the mentioned scanners. Conclusion: Optimization of the BM-PET geometry and detector configuration, implementation of scanner rotation/wobbling and utilization of image correction and enhancement techniques, improved the performance of our modeled brain PET and direct us to manufacture of a prototype scanner. © 2019 Elsevier B.V.