A count-rate model for PET scanners using pixelated Anger-logic detectors with different scintillators

Abstract

A high count-rate simulation (HCRSim) model has been developed so that all results are derived from fundamental physics principles. Originally developed to study the behaviour of continuous sodium iodide (NaI(Tl)) detectors, this model is now applied to PET scanners based on pixelated Anger-logic detectors using lanthanum bromide (LaBr3), gadolinium orthosilicate (GSO) and lutetium orthosilicate (LSO) scintillators. This simulation has been used to study the effect on scanner deadtime and pulse pileup at high activity levels due to the scintillator stopping power (μ), decay time (τ) and energy resolution. Simulations were performed for a uniform 20 cm diameter × 70 cm long cylinder (NEMA NU2-2001 standard) in a whole-body scanner with an 85 cm ring diameter and a 25 cm axial field-of-view. Our results for these whole-body scanners demonstrate the potential of a pixelated Anger-logic detector and the relationship of its performance with the scanner NEC rate. Faster signal decay and short coincidence timing window lead to a reduction in deadtime and randoms fraction in the LaBr3 and LSO scanners compared to GSO. The excellent energy resolution of LaBr3 leads to the lowest scatter fraction for all scanners and helps compensate for reduced sensitivity compared to the GSO and LSO scanners, leading to the highest NEC values at high activity concentrations. The LSO scanner has the highest sensitivity of all the scanner designs investigated here, therefore leading to the highest peak NEC value but at a lower activity concentration than that of LaBr3.