Initial Results of the MINDView PET Insert Inside the 3T mMR

Abstract

Hybrid molecular and anatomical imaging devices, especially when simultaneously working, have shown to provide advantages over sequential acquisitions. In particular, we present in this preliminary study, the working performance of a brain positron emission tomography (PET) insert, within a 3T magnetic resonance imaging (MRI) system. To our knowledge, this is the largest PET system based on monolithic LYSO blocks. It consists of 60 scintillator blocks of <inline-formula> <tex-math notation= LaTeX >$50\\ \\text{mm} \\times 50\\ \\text{mm} \\times 20\\ \\text{mm}$ </tex-math></inline-formula> arranged in 3 rings of 20 detector modules each. An effective field of view (FOV) of 240 mm in diameter and 154 mm axially is defined. The crystals included a retroreflector layer at the entrance face, and are coupled to custom arrays of <inline-formula> <tex-math notation= LaTeX >$12 \\times 12$ </tex-math></inline-formula> silicon photomultipliers (SiPM), <inline-formula> <tex-math notation= LaTeX >$3\\ \\text{mm} \\times 3\\ \\text{mm}$ </tex-math></inline-formula> each. Frontend electronics provide <inline-formula> <tex-math notation= LaTeX >${X}$ </tex-math></inline-formula> and <inline-formula> <tex-math notation= LaTeX >${Y}$ </tex-math></inline-formula> projections of the scintillation light by sampling each row and column of the SiPM arrays. The insert is thermally stabilized by using temperature-controlled air to about 27 °C. The PET insert has been installed at the Klinikum rechts der Isar (Munich) and tested within the whole-body Siemens biograph mMR, a 3T MRI combined with a PET scanner. A system sensitivity of almost 7% for an energy window of 350–650 keV was measured using a small size source at the center of the FOV (CFOV). Current system evaluation showed a spatial resolution at the CFOV of 1.7 mm using iterative algorithms, being below 2 mm within a centered diameter of 120 mm. Rods of a small Derenzo phantom of 2.5 mm were clearly resolved, independently of all tested MRI sequences including echo-planar imaging, ultrashort time echo, MPrage, and T2-flair or arterial spin labeling. The PET insert did not show any count rate degradation also under those sequences for a variety of MR imaging acquisitions.