An evaluation of cardiac uniformity, contrast, and SNR with dual-head 180° and triple-head 360° SPECT scans

An experimental phantom study was performed to evaluate cardiac uniformity, contrast, and signal-to-noise ratio for two clinical cardiac SPECT imaging protocols: adjacent dual-head 180° and triple-head 360° scans. One head of a SPECT camera was used to acquire 180° and 360° projections with different times per step to simulate the clinical case where dual-head 180° and triple-head 360° each takes a total of 20 min. Scans were acquired with no lesion, anterior lesion, and posterior lesion in the myocardium. Maximum a posteriori reconstruction was done by an iterative coordinate descent algorithm using a quadratic convex prior. The L-curve method was used to obtain the prior strength. Some investigation was done on obtaining the L-curve and on the ways to fit the L-curve and to get the corner point. Images both with attenuation and scatter correction (ASC) and without ASC were compared. The 180° scan shows an intensity decrease in anterior apical and posterior basal regions. The 360° scan shows an intensity decrease in the posterior wall. For the anterior lesion, the 180° scan has slightly better contrast, while for the posterior lesion, the 360° scan has slightly better contrast. The difference between the 180° and 360° scans is subtle, and the comparison results depend on the lesion position and the view angle of the heart. A receiver operating characteristic study of 180° versus 360° acquisition designed based on these characterizations of contrast, uniformity, and noise will be necessary to evaluate overall performance