Development of a high-resolution four-layer DOI detector using MPPCs for brain PET

Abstract

A new high-resolution four-layer DOl detector using MPPCs for brain PET scanner has been developed. The new depth of interaction (DOl) detector was designed to compose of four layers of detector units, which were lined up five axially. Each of the detector units consists of a LYSO scintillator array finely segmented of 1.2 mm and an 8 x 8 array of multi-pixel photon counters (MPPCs), which are one of the products of silicon photomultiplier family. The MPPC is so compact and insensitive to gamma-ray that the detector units can be piled up with a small gap between each scintillator array in the depth direction. In order to have the detector in every layer equally sensitive to gamma-ray, the scintillator thickness was designed at 3 mm, 4 mm, 5 mm and 8 mm toward the bottom respectively, and the total thickness was 20 mm. We adopted an internal focused laser processing technique to a monolithic LYSO scintillator and fabricated a 2D segmented array of 32 x 32 with 1.2 mm pitch in 38.4 mm square cross-section. Each detector layer has independently front end circuits including ASICs for MPPCs and signal processing circuits for crystal identification, energy and timing detection.Each data set of four layers are fed into data interface circuits placed behind detector layers and transferred to a data acquisition unit as formatted list-mode data. The performance of the four-layer DOl detector has been evaluated. The coincidence timing resolution of the detector, with a reference BaF2 detector, was obtained 850 ps FWHM. The average energy resolution value was 24.5% at 511 keY. The crystal separation with finely segmented L YSO scintillator was also good enough at each layer.

Fabrication of finely pitched LYSO arrays using sub-surface laser engraving technique with picosecond and nanosecond pulse lasers

“…If the scintillation light is detected from one surface of a laser-engraved LYSO array with a 1.2 mm pitch, the maximum crystal thickness would be about 8 mm [4]. To address this limitation, one approach would be to detect scintillation photons from all six surfaces of a 3D crystal array using solid-state photodetectors such as MPPCs [7] because most of the scintillation photons could be detected by nearby photodetectors before they spread through the microcrack walls.…”

“…This method would reduce the production cost of scintillator arrays, especially finely pitched arrays, because fabrication can be performed automatically. Recently, a highresolution depth-of-interaction detector using multi-pixel photon counters (MPPCs) and finely segmented LYSO scintillator arrays fabricated by the SSLE technique has been developed [4]. The detector was adopted for a new human brain PET scanner [5].…”

Simulation study of a high‐performance brain PET system with dodecahedral geometry

“…The ring detector PET might be a good choice for whole body imaging while not optimized for specific organs such as the brain, due to the low sensitivity. In order to increase image quality of the PET brain imaging, dedicated brain PET systems were designed with small diameter detector rings, such as the High‐Resolution Research Tomograph (HRRT) system and the one with four‐layer MPPC DOI detectors . However, the sensitivity of these ring detector systems degrades heavily with the increase of the distance from axis in the transaxial plane .…”

“…In order to increase image quality of the PET brain imaging, dedicated brain PET systems were designed with small diameter detector rings, such as the High-Resolution Research Tomograph (HRRT) system 5,6 and the one with four-layer MPPC DOI detectors. 7,8 However, the sensitivity of these ring detector systems degrades heavily with the increase of the distance from axis in the transaxial plane. 9 The low sensitivity of the peripheral area in FOV affects quantitative accuracy for some brain regions, such as parts of cerebellum region, 10 while the cerebellum region is vital for brain imaging because this region is usually used as the reference area.…”

Development of a data acquisition architecture with distributed synchronization for a Positron Emission Tomography system with integrated front-end.

“…In timestamp-based systems, the simplest case corresponds to a single coincidence module where all timestamping nodes are directly connected [330]. This may not be possible in large scanners with a high number of timing modules, and intermediate hierarchy levels may be needed where data from lower level nodes are concentrated [343] and partial coincidence detection for their child nodes can be implemented in order to reduce their upstream bandwidth requirements.…”