Investigation of deadtime characteristics for simultaneous emission-transmission data acquisition in PET

Modern PET systems generally utilize rotating rod sources and sinogram windowing for transmission scan and are also usually capable of simultaneous emission-transmission data acquisition. The count rate capability of a PET system using BGO is primarily limited by the deadtime of the block detectors. This deadtime is directly related to the singles event rate of the system. In a simultaneous emission-transmission acquisition, the moving rod sources generate a spatially varying singles distribution which rotates synchronously with the rods. The emission activities tend to create a nearly uniform singles distribution which is static or varies slowly depending on the half life of the radioisotope. In this study, the system deadtime is characterized for different sources of activity in a simultaneous acquisition. Typically, there are 60 to 150 MBq Ge-68 in each of the three rods causing a 26% peak-to-peak variation in the instantaneous detector deadtime. The time-averaged deadtime in a transmission sinogram has a variation of 30% peak-to-peak. Rod activities create a 4 to 10% peak-to-peak variation in the time-averaged deadtime in the emission sinogram of a simultaneous scan. The current deadtime correction method results in a 10% error in a simultaneously acquired emission sinogram with 60 MBq in each rod and 74 MBq in a plane source. A spatially variant deadtime correction technique was developed for simultaneous scans. The error was reduced to 4% for the same study.