Potential for a fifth generation PET scanner for oncology

Four generations of PET scanners have emerged since the first clinical PET system. In this paper we predict the achievable performance of a fifth generation PET scanner, defined as a fully 3D system with no septa, an axial FOV approaching 30 cm with a spatial resolution mainly dictated by the positron physics. The predictions are based on the HIREZ (CPS/Siemens) performance plus current insights in detector technology, electronics and image reconstruction. The HIREZ system uses LSO scintillators, providing fast timing, good energy resolution, high spatial resolution and high system live time. LSO is an excellent candidate for a TOF detector with a proven 400 ps time resolution for detector blocks in coincidence. We have considered and integrated improvements that can be made to the key components necessary to the design of a fifth generation PET scanner. The spatial resolution can be improved by using a finer detector grid. The time window can be reduced. Packing fraction can be increased. The axial extension can be increased, giving more solid angle coverage and therefore better NEC. Based on the predicted gains in reduced randoms, dead time and TOF advantages, we have made initial simulations of the NEC for different axial field of views. An axial extension of /spl sim/32 cm with ECAT EXACT ring geometry gives at least four times higher NEC count rate than for a 16 cm axial FOV system. This will be combined with the improved image quality due to higher spatial sampling and TOF.