The effect of camera geometry on singles flux, scatter fraction and trues and randoms sensitivity for cylindrical 3D PET-a simulation study

Preliminary results of an assessment of the effects of changing the axial field of view (AFOV) and detector ring diameter (DRD) of a cylindrical PET tomograph on count-rate performance are presented. The assessment was made using Monte Carlo simulations of an anthropomorphic phantom based on the Zubal phantom. This phantom was modified to include cylinders approximating arms and legs, and was sequentially stepped through the AFOV to simulate a whole-body scan covering an axial region of interest of 1 m. DRD was varied from ~60 cm to ~108 cm, and AFOV was varied from 10 cm to 60 cm. A simple activity distribution and dead-time model was assumed to allow the calculation of noise-equivalent count (NEC) rates for a situation similar to that of a typical 18F-FDG study. Both the scatter fraction and singles flux were found to be strongly dependent on DRD, but only weakly dependent on AFOV when the latter was greater than ~25 cm. Trues and randoms sensitivity were strongly dependent on AFOV, and randoms sensitivity was also strongly dependent on DRD. Scatter and singles flux do not appear to be limiting factors for extended AFOV configurations, and randoms rates, while high, appear to be manageable with existing detector technology. This initial assessment suggests that for whole-body applications, substantial gains in NEC may be possible by extending the AFOV