Positioning annihilation photon interactions in a thin LSO crystal sheet with a position-sensitive avalanche photodiode

Using scintillation crystal sheets instead of discrete crystal arrays in high-resolution PET has the advantage of reduced complexity. In order to evaluate the positioning capability of a position sensitive avalanche photodiode (PSAPD) using a sheet lutetium oxyorthosilicate (LSO) crystal scintillator, we need to understand the functional dependence of a detected event position on the known source position. We studied positioning with both collimated 57Co 122keV and coincidence-triggered 22Na 511keV sources, which were stepped across the face of an 8 mmtimes8 mm LSO sheet crystal coupled to an 8 mmtimes8mm PSAPD at 160 mum intervals using a voltage-driven mechanical stage with a LabVIEW controlled acquisition system. We analyze the energy resolution, sensitivity, photopeak position and energy gated full width at half maximum (FWHM) spread of the detected position for a particular known source position. We have observed a 10% variation in average energy from the center of the crystal to the edge with 57Co and <1% for 22Na and an average point spread function FWHM of 2.86 mm and 1.12 mm for 57Co and 22Na respectively. We investigated methods to create a 1-1 map between (1) the four positioning signals from the PSAPD and the recorded energy and (2) the true position of the annihilation photon interaction. We found the average energy change over the 1.2 mm near the edge of the continuous LSO crystal to be ~5% - insufficient to resolve with the prototype PSAPD (energy resolution 12%) with an Anger-type logic positioning algorithm. Simulation using the annihilation photon interactions from GATE and scintillation photon transport from DETECT2000 have confirmed the effects observed in experiment