Characterization and Optimization of a Quasi-Monolithic Detector Module with Depth-Encoding for Small Animal PET

Abstract

Characterization and Optimization of a Quasi-Monolithic Detector Module with Depth-Encoding for Small Animal PET Yong Hyun Chung, Seung-Jae Lee and Cheol-Ha Baek Department of Radiological Science and Institute of Health Science, Yonsei University, Wonju 220-710 Kang-Joo Lee Department of Radiological Science, Yonsei University, Wonju 220-710 Yong Choi Department of Nuclear Medicine, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul 135-710 (Received 9 September 2008) A new design of a detector module with depth of interaction (DOI) based on a quasimonolithic cerium doped lutetium oxyorthosilicate (LSO) crystal, a multi-channel photo-sensor and a maximum-likelihood position-estimation (MLPE) algorithm has been developed to improve the resolution performance for small animal positron emission tomography (PET). By ning down the lookup table (LUT) sampling points, uniform resolution over the eld of view can be achieved; however, the sensitivity decreases with increasing computation time for the MLPE calculation. In this study, simulations were performed to optimize the LUT sampling points by characterizing how the number and the thickness of each crystal layer in the DOI direction a ects the image quality to simultaneously improve the resolution uniformity and sensitivity. The simulation tools GATE and DETECT2000 were used to estimate the resolution of a small PET scanner and to model the light transport for LUT generation, respectively. The number and the thickness of each layer of the quasi-monolithic crystal varied while the total thickness was xed at 10 mm. For a multi-layer detector, the ratio between each layer's thicknesses, which can minimize the radial resolution variations, was optimized. The average radial resolution and the execution time were measured as 2.73 mm, 2.17 mm, 2.13 mm and 2.09 mm and 4.14 s, 7.22 s, 10.55 s and 13.57 s per event for LUTs with optimized single, dual, triple and quadruplet layers, respectively. The results demonstrate that the resolution uniformity, as well as the system sensitivity, can be simultaneously and signi cantly improved by optimizing the ratio between each layer's thicknesses and the number of the LUT sampling points in the small animal PET detector based on quasi-monolithic crystals coupled to a multi-channel sensor with a MLPE algorithm. PACS numbers: 87.62.+n