Improved gamma camera performance using event positioning method based on distance dependent weighting

Event positioning algorithms for scintillation gamma cameras are typically based on centroid ("Anger") positioning algorithms that weight photomultiplier tube (PMT) responses by their PMT location. Iterative maximum-likelihood based algorithms use measured PMT responses to provide more accurate position estimates, but suffer from long computational times that limit the use in large field-of-view gamma cameras in real-time systems. An improvement to the centroid-based method is to use an iterative distance dependent weighting method to modify the individually measured PMT responses. The distances from the initial event position estimate to each PMT location is calculated. Each distance is used to index into a weighting table designed to optimize PMT signal-to-noise for better detector resolution, improve the non-linearity of the PMT response in scintillation cameras for more accurate position estimates, and correct for energy dependent mean detector light responses. The Philips EPIC-AZ detector with a single 200 MHz TMS320C6204 DSP has a throughput of 4.3 musec per event for two iterations. Intrinsic spatial resolution has improved by ~0.2 mm to typically 3.1 mm FWHM. Intrinsic integral flood uniformity of <2.5% for Ga-67 can be achieved with a Tc-99 flood correction table by fixing the PMT high voltage and calibrating PMT gain tracking errors