Abstract
PURPOSE/AIM: The increasing population age highlights the critical need for early brain disease diagnosis, especially in disorders such as dementia. Consequently, a notable focus has been on developing dedicated brain positron emission tomography (PET) scanners, which offer higher resolution and sensitivity than whole-body PET scanners. This study aims to design and performance evaluation of an LYSO-based dedicated brain PET scanner. MATERIALS AND METHODS: We developed a dedicated brain PET using Monte Carlo simulation based on cylindrical geometry. Each detector block consisted of a 23 × 23 array of 2 mm × 2 mm × 15 mm LYSO crystals coupled with SiPM. The performance of this scanner was evaluated based on the NEMA NU-2-2018 standard, focusing on analyzing various energy windows and coincidence time windows (CTWs). RESULTS: The results demonstrated that the noise equivalent count rate (NECR) peaked at each CTW in the 408-680 keV energy window. In addition, increasing the CTWs from 3 ns to 10 ns resulted in a decrease of 9% in sensitivity and an increase of 63% in NECR. Furthermore, the study findings highlight that using a time-of-flight (TOF) resolution of 250 ps can substantially improve image contrast relative to non-TOF reconstruction. CONCLUSIONS: We conclude that employing a broader energy window and a narrower CTW can significantly enhance the scanner's performance regarding sensitivity and NECR. Furthermore, incorporating LYSO pixelated crystals with TOF information will facilitate the generation of high-resolution and high-contrast images.