Impact of detector parameters on the imaging performance of a dedicated brain PET scanner

High performance brain scanners suitable for accurate and precise measurement of uptake in small regions of the brain require high spatial resolution (1.5-2mm) and effective sensitivity. While small cross-section crystals are needed to achieve high spatial resolution, effective sensitivity requires long crystals and good TOF resolution. In addition, depending on crystal length some level of depth-of-interaction (DOI) capability may also be needed to reduce parallax error. However, long crystals with DOI capability typically degrade TOF resolution. Hence, brain system development normally involves a trade-off between crystal cross-section, crystal length, DOI, and TOF resolution. The focus of this work is to use Monte Carlo simulations and task-based metrics to evaluate these trade-offs in detector design to better guide a new brain system development. Fixing the crystal cross section at 1.6 mm x 1.6 mm, we simulated different LSO crystal thickness (10 mm, 15 mm, and 20 mm), number of DOI layers (1, 2, and 4), and DOI layer assignment accuracy (100% and 80%). Spatial resolution was evaluated as 3-dimensional Full Width of Half Maximum (FWHM) of point sources reconstructed with a list-mode TOF OSEM algorithm. A brain-size phantom was also studied for imaging evaluation. For crystal thickness, we found that 10 mm LSO without DOI performance system performs better than a 20 mm LSO with 2-layer 100% accuracy DOI capability system. Further evaluations using smaller crystal cross-sections and varying TOF resolution will be performed to benchmark not only spatial resolution. In addition, the combined impact of all the detector design parameters on the accuracy and precision of uptake measurement in small structure within the brain and discrimination of small brain structures using a detectability metric will also be evaluated.