The effect of collimation on single rates in multi-slice PET

The ratio (S:C) of singles to true coincident events in a positron emission tomography (PET) depends on the collimator size and shape. As well as causing random counts when they fall within the energy acceptance window, single events are the main cause of deadtime. Monte Carlo techniques were used to compared S:C ratios in well-collimated single-slice, 15-slice, and open-collimator 3-D configurations. All systems used cylindrical arrays of BGO (bismuth germanate) or NaI crystals 51 cm in diameter and 10-cm deep. A 12-cm-deep 20-cm-diameter source gives rise to a S:C ratio of 53:1 with a low-energy threshold of 350 keV in a well-collimated single slice tomograph. When all slices are combined, the S:C ratio for both 15-slice and 3-D is 24:1 under the same conditions; however, the 3-D system detects five times more true coincident events, and is much more sensitive to events outside the scan field