Experience with scintillators for PET: towards the fifth generation of PET scanners

Since the ECAT 2, the first commercial positron emission tomograph developed by EG&G ORTEC, four generations of scanners can be identified. The first such scanners were based on sodium iodide (NaI(Tl)) scintillators, although as early as 1978 the transition to bismuth germanate (BGO) detectors had begun. By 1981, second-generation PET scanners with up to four rings of BGO detectors were available commercially. The BGO block detector appeared in 1985, initiating the third generation of PET scanners with the potential to increase the axial coverage in a cost-effective manner. As with the second generation, the third generation of PET scanners incorporated lead septa to collimate the annihilation photons within transverse planes, thereby reducing the acquisition of scattered and random coincidences and limiting detector dead time. The fourth generation of PET scanners offered up to 15 cm axial coverage and incorporated retractable septa that permitted both 2D and 3D acquisition within the same scanner. Therefore, fifth-generation scanner should be a fully 3D system with no septa, a 25–30 cm axial field-of-view, and a spatial resolution approaching the limits set by the physics of positron emission. Rather than using septa, limitation of randoms and scatter should be achieved directly at the detector by using a scintillator with high-light output, good energy resolution, and a fast scintillation decay time. The fast decay time will ensure low deadtime. The recent development of lutetium oxyorthosilicate (LSO), a scintillator with the required properties, suggests that a fifth generation of positron emission tomographs can now be attained.