Improved data acquisition system for brain PET using GAPD arrays

Abstract

We have previously reported that a brain PET using GAPD arrays was successfully developed. The brain PET consisted of 72 4 × 4 GAPD arrays combined with LYSO crystals (single pixel size: 3 mm × 3 mm). Each 4 GAPD arrays' output signals were sent to a 64:1 position decoder circuit (PDC) which detects the fastest gamma signal of 64 input channels. To further improve the PET system performance, several modifications were performed on the DAQ system: PET data from 3 DAQ cards were transferred and saved on one SDRAM module by rapid channel communication; parallel processing and multiplexing based FPGA algorithm was developed to detect true PET signals by real time; a more user-friendly GUI DAQ control program was developed to control 3 DAQ cards simultaneously; an accurate and fast coincidence sorting method containing 3 discrimination approaches (time, energy and line of response discriminations) was developed to improve image quality. To evaluate the improved DAQ system, several experiments were performed such as sensitivity measurement using a 25 μά Na-22 point source, spatial resolution measurement using ten F-18 line sources with different source-to-center distances (−8 cm, −6 cm, −4 cm, −2 cm, 0, 2 cm, 4 cm, 6 cm, 8 cm and 10 cm), PET images acquisition of hot rod phantom and Hoffman brain phantom. Experimental results showed that PET sensitivity of 2594 cps/ MBq at 30% energy window (350–650 kev) was achieved. Spatial resolution from 2.9 mm (center) to 5 mm (25 cm off-center) was acquired for ten different source-to-center distances. PET images of hot rod phantom and Hoffman brain phantom were successfully acquired with improved image quality. The DAQ system developed in this study allows to acquiring high quality PET images using GAPD arrays.