Converging slat collimators for hybrid PET

Hybrid PET cameras have a number of hardware compromises that degrade image quality relative to dedicated PET. Such degradation is primarily caused by poor detection efficiency given by Nal scintillation crystals, limited count-rate culpabilities, multi-head as opposed to full-ring camera geometry, and so on. The count-rate limitation is exacerbated by the relatively large axial field-of-view (FOV) offered by most hybrid gamma cameras. In order to fully utilize the available count-rate, It is necessary to shield its many singles and scattered events as possible so that the camera can process the maximum number of true coincidence events. A common approach is to use axial collimators with parallel slats to shield out-of-FOV photons. Such collimators also greatly reduce the number of scattered events. We are investigating converging slat collimators, which may be better able to shield out-of-FOV photons while greatly increasing sensitivity to the central FOV. Such collimators create a pseudo-3D imaging geometry, maintain good shielding of scattered events, and may greatly improve volume-of-interest imaging performance. The GEANT Monte Carlo program (CERN) was used to simulate a variety of parallel and convergent slat collimators. Both lead and tungsten septa were studied. Axially-focused slat collimators were found to provide greatly increased sensitivity to the central FOV while also providing better shielding to out-of-FOV photons. Whole-body NEC performance for fan slat collimators was nearly identical to that for conventional parallel slat collimators, and volume-of-interest NEC was nearly doubled using fan slat collimators.