Performance of a PET/MR Insert for Mice Based on Monolithic Crystals

3123 Objectives: Multimodal PET/MR imaging is a powerful molecular imaging tool widely used in research and clinical practice for a broad variety of applications. Small animal PET inserts are an instrumentation tool that can be used in combination with already installed MRI systems. This work shows the initial performance and MRI compatibility of a small animal PET insert for mouse applications, based on monolithic LYSO scintillation crystals. Methods: The PET system is based on LYSO cuboids of 25x33x10 mm, laterally black painted and include a retroreflector layer at the entrance face. This layer improves the detectability of the gamma ray position interaction. The crystals are coupled to custom made 10x8 SiPM arrays with 3x3 mm active area each. 24 crystals are distributed in 3 rings defining an axial and transaxial FOV of 100 mm and 40 mm, respectively. The SiPM photosensor and readout electronics are mounted on high performance boards that avoid the generation of eddy current by the switching gradient fields.This work shows the results carried out within a Bruker BioSpec 94/20 MRI imaging instrument using a 35 mm volume RF coil. The PET insert includes an RF shielding based on carbon fiber structures. We report here the average performance of the detector blocks, and some of the most important parameters regarding PET imagingand MR compatibility. Results: After calibration of each detector block using a novel approach based on Voronoi diagrams [1], the average energy resolution of the whole system is about 13.5%, together with an impact precision determination in 2D below 1.5 mm, and about 2 mm depth of interaction (DOI) resolution.Regarding MR compatibility, the PET insert has been run under several MRI sequences including EPI with high duty cycles. No significant change in MRI performance was observed when compared to the case without the PET. The PET detector performance is not affected by interferences that could arise from different MRI imaging sequences, concerning impact determination, energy or any other parameter [2,3]. Therefore, the reconstructed images did not exhibit any degradation.The initial performance evaluation resulted in a peak sensitivity of nearly 12% at CFOV, and a NECR mouse peak of 481 kcps at 25.7 MBq . The PSF-corrected spatial resolution has been determined to be about 0.7 mm FWHM for all three space components. Using DOI allows to reach similar spatial resolution performance for the entire FOV. The good spatial resolution of the device allows to distinguish 0.8mm rods of a micro-Derenzo phantom when using 0.25 mm voxels, 1 mm virtual pixels and 25 iterations (MLEM). Conclusions: This works provides design principles needed for the realization of small PET/MR inserts for mice. The initial study has shown that the requirements for simultaneous PET/MR mouse imaging applications can be reached, thus high spatial PET imaging resolution and the absence of any interference effects for PET and MRI.The PET insert reaches an unprecedented image performance resolving well 0.8 mm Derenzo rods while simultaneously imaging MRI with EPI sequences. The high spatial resolution is obtained across the entire PET FOV, thus helping researchers for a more reliable quantification during their investigations. Bibliography: [1] M. Freire, IEEE TRPMS, 2019, doi: 0.1109/TRPMS.2019.2947716. [2] A.J. Gonzalez, IEEE TRPMS 3, 343, 2019. [3] W. Gsell, to be submitted to EJNMMI, 2020.