Event Processing for Modular Gamma Cameras with Tiled Multi-Anode Photomultiplier Tubes.

Abstract

Multi-anode photomultiplier tubes (MAPMTs) are good candidates as light sensors for a new generation of modular scintillation cameras for Single-photon emission computed tomography (SPECT) and Positron emission tomography (PET) applications. MAPMTs can provide improved intrinsic spatial resolution (<1mm) compared to arrays of larger individual PMTs due to their small anode sizes, and the increased number of channels also allows accurate estimation of depth-of-interaction (DOI). However, the area of a single MAPMT module is small for a modular gamma camera, so we are designing read-out electronics that will allow multiple individual MAPMT modules to be optically coupled to a single monolithic scintillator crystal. In order to allow such flexibility, the read-out electronics, which we refer to as the event processor, must be compact and adaptable. In combining arrays of MAPMTs, which may each have 64 to 1024 anodes per unit, issues need to be overcome with amplifying, digitizing, and recording potentially very large numbers of channels per gamma-ray event. In this study, we have investigated different event-processor strategies for gamma cameras with multiple MAPMTs that will employ maximum-likelihood (ML) methods for estimation of 3D spatial location, deposited energy and time of occurrence of events. We simulated anode signals for hypothetical gamma-camera geometries based on models of the stochastic processes inherent in scintillation cameras. The comparison between different triggering and read-out schemes was carried out by quantifying the information content in the anode signals via the Fisher Information Matrix (FIM). We observed that a decline in spatial resolution at the edges of the individual MAPMTs could be improved by the inclusion of neighboring MAPMT anode signals for events near the tiling boundaries. Thus in order to maintain spatial resolution uniformity throughout the modular camera face, we propose dividing an MAPMT's array of anode signals into regions such to help determine when triggers from one MAPMT need to be passed to a neighboring MAPMT so that it can contribute anode information for events between them.