Impact of Time-of-Flight Reconstruction on Cardiac PET Images of Obese Patients.

Abstract

The benefit of time-of-flight (TOF) information in PET oncology studies is well established, demonstrating improved signal-to-noise ratio and enhanced lesion detection. In cardiac PET studies, tracer distribution and study interpretation differ substantially from oncology PET, and the benefit of TOF is less established. We investigate the impact of TOF on reconstructed myocardial distribution in Rb PET perfusion studies of obese patients, for whom TOF would have most significant impact. Cardiac Rb PET data of 13 obese patients (8 male, 5 female patients; weight, mean, 139 kg [range, 109-191 kg]; body mass index, mean, 49 kg/m [range, 36-76 kg/m]) were analyzed retrospectively. Images were reconstructed with and without TOF and with varying number of iterative updates (2-12 iterations). Convergence and clinical relevance of differences were assessed both visually and quantitatively (automated 17-segment scoring). An anthropomorphic torso phantom also was scanned in order to study how TOF affects reconstruction of a myocardial distribution. Time-of-flight imaging provided significant improvement in image quality and convergence rate compared with non-TOF imaging. Time-of-flight reconstruction typically required 2 to 4 iterations to converge versus 8 to 12 iterations for non-TOF reconstruction. Even at 12 iterations, non-TOF images occasionally had apparent differences in relative perfusion compared with TOF images that exceeded 10% and were considered likely to affect clinical interpretation. Time-of-flight reconstruction has a significant clinical impact on cardiac PET in obese patients. When interpreting cardiac studies from non-TOF-capable PET scanners, one must ensure to reconstruct images with sufficient iterations and to be aware of potential artifacts.