List-mode reconstruction for the Biograph mCT with probabilistic line-of-response positioning and event-by-event motion correction

In this study, we address two issues that degrade spatial resolution in PET. First, conventionally, the Siemens Biograph mCT scanner collects data in 32-bit list mode, so each LOR is defined as a sinogram bin (instead of detector pair) with axial span and transaxial mashing of adjacent angles reducing the number of LORs, thus degrading resolution. In this work, we proposed a probabilistic assignment of LOR positions (pLOR) that addresses axial and transaxial LOR grouping in 32-bit data. Secondly, respiratory motion results in image artifacts in whole-body scans. Gated reconstruction does not correct for intra-gate motion. We have previously demonstrated for brain imaging that event-by-event correction can eliminate intra-frame motion. Therefore, we developed a list-mode reconstruction algorithm for the Siemens Biograph mCT with pLOR positioning and event-by-event motion correction, based on MOLAR (Motion-compensation OSEM List-mode Algorithm for Resolution-Recovery Reconstruction). This algorithm was compared to the manufacturer´s PSF+TOF algorithm. Static phantom and animal studies demonstrated that MOLAR with pLOR gives faster contrast recovery than the PSF+TOF algorithm that uses the average LOR positioning for all sub-LORs within each sinogram bin. Moving phantom and a whole-body human study suggested that event-by-event motion correction is able to reduce image blurring introduced by the intra-gate motion. We conclude that list-mode reconstruction with pLOR positioning and event-by-event motion correction can generate high quality images for the mCT, and has great potential to optimally recover fine structures in wholebody PET scans with motion.