Basic imaging performance characteristics of a variable field of view PET camera using quadrant sharing detector design

The basic imaging performance of the prototype PET camera (MDAPET) developed in our laboratory was measured. The MDAPET was developed to test the engineering feasibility of two design concepts: (a) the photomultiplier-quadrant-sharing detector design and (b) a lower cost variable detector ring which forms a densely-packed smaller ring for imaging breast/brain/animals, and a less-packed ring for whole-body imaging. The basic imaging performance was measured for the 32 and 55 cm FOV with tests similar to those described in the SNM/NEMA performance standard [12]. For the 32 cm FOV, the transaxial image resolutions at 0, 5, 10 cm were found to be 2.8, 3.1, 4.0 mm, respectively, while for the 55 cm FOV, the image resolutions were 2.9, 3.1, 3.6 mm. The axial resolution ranged from 2.5 to 4.4 mm as a function of detector-ring position. For a 21.5-cm uniform phantom, the coincidence-detection sensitivity was 82.3 kcps/mu Ci/cc at the smallest 32 cm field of view and 16.3 kcps/mu Ci/cc for the whole body configuration. The scatter fractions were 19% and 30% for the whole-body mode and the brain/breast mode, respectively. The smaller mode has 2.4 times higher noise-equivalent-sensitivity than the larger mode at 1 mu Ci/cc. High quality, artifact-free brain-phantom images were obtained in both modes.