Design considerations for a partial-ring, multi-modal compatible whole-body TOF PET scanner: Flexible PET

Along with the popularization and technological maturation of integrated PET-CT, the center of research interest has moved to integrated PET-MR. In this situation, we are developing a multi-modal compatible flexible PET scanner. In our concepts, the scanner consists of adjustable two detector units and their support apparatus, and then scans a patient lying on a bed equipped by other medical imaging/non-imaging device. Therefore, depending on the size and shape of the patient and the patient bed and other demands, the detector geometry can be changed flexibly. Although, due to the limited angular coverage, distortion- and artifact-free image reconstruction is not feasible without detector rotation, it is known that accurate time-of-flight (TOF) information helps to improve reconstruction quality in such a partial-ring scanner. In this study, we performed whole-body 2-D imaging simulations to investigate detector geometric and TOF requirements for the partial-ring scanner for obtaining comparable image quality to a conventional full-ring non-TOF scanner. In particular, required angular coverage by the detector units in relation with the timing resolution and the influence of the arrangement of the opposed detector units (i.e., missing direction is perpendicular to patient's front-to-back or left-to-right direction) to the reconstruction image quality were investigated. The simulation results on the image appearance and contrast-noise performance showed that the partial-ring scanner consisting of the two detector arcs each having the central angle of 150 degree and 400-500 ps timing resolution could provide comparable image quality to the conventional scanner regardless of the arrangement of the opposed detectors. However, in the case of the 130 degree coverage with the 400-500 ps timing resolution, the contrast-noise performance of the partial-ring geometries could not reach partially that of the conventional scanner due to the missed spatial and/or statistical information. In this point, we showed that the use of the object support constraint in the iterative reconstruction could reduce the intensity bias and blurring and then improves the contrast-noise performance.