Comparison of measurement and simulation using GATE with optical photon tracking for a DOI PET detector

Objectives We are using GATE simulations to model a dual-ended readout LSO based PET detector with DOI capability. In this work we attempt to validate simulations of DOI resolution against previously published experimental results.

Methods Simulations including optical photon transport were performed using GATE ver. 4.0.0. Dual-ended readout detectors were simulated with LSO crystals of widths 1.0, 1.3, 1.5 and 2.0 mm and length 20 mm. Side surfaces were set as rough-teflon-wrapped and end surfaces were polished. Each end was coupled to a 'perfect APD' surface. Crystals were irradiated with a pencil beam source of 511 keV photons oriented perpendicular to the long axis in steps of 2.5 mm. DOI position was calculated as the ratio APD1/(APD1+APD2). DOI resolution was determined as a ratio of the FWHM of the distribution of the ratio signal and its slope versus irradiation depth. Simulated data were compared against published experimental results (Shao et al., IEEE Trans Nuc Sci, 49(3) 2002; Yang et al., Phys Med Biol, 51 2006).

Results Table 1 shows the average DOI resolution for the simulated and experimental results. There is agreement in the trend of better resolution with thinner crystal, but systematic differences for events with energy > 350 keV (simulation better than experiment) and for energy < 350 keV (experiment better than simulation).

Conclusions While GATE results give the general trend in DOI resolution, there are systematic differences that need to be further explored. We are examining the source of these variations through refinements of the simulated optical surfaces.

Research Support Supported by funding from NSERC Discovery Grant 360020-2008 and Manitoba Health Research Council Studentship Award.