Modeling of a preclinical PET system based on monolithic scintillator crystal using the GATE/GEANT4 and Castor codes

The advance in pre-clinical research and technologies made the use of monolithic scintillator crystals a field of research. The preclinical scanners Albira Trimodal e Molecubes β-cube are the state of the art in sensitivity and spatial resolution performance. This study aims to model and analyze the performance of a pre-clinical PET system based on monolithic scintillator crystals using GATE code and CASToR image reconstruction code. A pre-clinical PET was simulated based on 10 monolithic scintillator crystals with a sensitive area of 50 x 50 mm2 and different crystal thicknesses: 2, 4, 6, 8 and 10 mm. Spatial resolution and sensitivity were evaluated with a 1mm diameter point source of ²²Na inserted into a 1 cm³ PMMA cube in different radial positions: 0, 5, 10, 15, 20, 25 mm, from -25 to 25 mm, in 5 mm steps along the axial axis, following NEMA NU4-2008 protocol. Simulation and reconstruction of phantom image for small rodents from NEMA NU4-2008 was also performed. The images were reconstructed with CASToR that use iterative method MLEM (Maximum Likelihood - Expectation Maximization). The lowest spatial resolution obtained with the 2 mm crystal thickness at the FOV center was (0,545 ± 0,003) mm, while the highest value was obtained with the 10 mm thick crystal was (1,1 ± 0,1) mm at the FOV center and (1.8 ± 0.1) mm on the radial axis offset of 25 mm. The highest sensitivity results were obtained with PET using 10 mm crystals in the FOV center, reaching 7.4% and by 25 mm in the axial offset axis reached 3.2%. The lowest sensitivity results were achieved with 2 mm thick crystals, reaching 1.5% in the FOV center and 0.7% in 25 mm axial offset axis. It is observed that for thicker crystals there is an increase in system sensitivity and a degradation in spatial resolution, while for thinner crystals, there is a gain in spatial resolution and loss of sensitivity, data that agree with the literature. The Monte Carlo simulation proved to be an effective tool in the evaluation of these new technologies. The reconstruction method was able to provide results close to those of commercial systems and is capable of reconstructing volumes like the phantom for small rodents from NEMA NU4-2008. (author)