Lesion detection characteristics of triple-head 360° versus adjacent dual-head 180° myocardial SPECT acquisitions

In this dissertation, 99mTc-MIBI myocardial SPECT images were evaluated for differences in lesion detection between two acquisition protocols: triple-head 360° and adjacent dual-head 180° SPECT scans using equal scan-time.; Two initial studies were performed. First, an experimental male phantom study was performed to evaluate cardiac uniformity and lesion contrast for these two protocols. One head of a triple-head SPECT camera was used to acquire projections to simulate 20-min scans for each protocol.; Second, observer studies were performed with the Channelized Hotelling Observer (CHO) and 5 human observers. Projections were simulated using voxelized male and female cardiac-torso digital phantoms.; After these initial studies, a comprehensive receiver operating characteristics (ROC) study was performed for a more complete and complex experimental paradigm. Physical male and female anthropomorphic cardiac-torso phantoms were used to acquire relatively noise-free projections on a triple-head SPECT camera. Five nuclear medicine physicians, four nuclear medicine physicists, and one certified nuclear medicine technologist participated as observers. The ROC results for physicians and for physicists (including technologist) were analyzed both separately and together.; The combined physicians, the combined physicists, and the combined group of physicians and physicists all showed a higher area under the ROC curve (A_z) for the 360° scan. For the male phantom case, the difference for the combined physicians is marginally significant (p = 0.0809), for the combined physicists is significant (p = 0.0118), for the combined physicists and physicians is significant (p = 0.0285). For the female phantom case, the difference for the combined physicians is marginally significant ( p = 0.0924), for the combined physicists is marginally significant (p = 0.0879), for the combined physicists and physicians is significant (p = 0.0023).; This study demonstrated that the comparison between adjacent dual-head 180° and triple-head 360° scan depends on the image reconstruction technique. In general, the results showed that for images reconstructed without ASC, the adjacent dual-head 180° scan has higher lesion detectability; for images reconstructed with ASC, the triple-head 360° scan has higher lesion detectability. (Abstract shortened by UMI.)...