Estimation and Correction of Rigid and Non-Rigid Respiratory Motion of the Heart for SPECT

One of the factors limiting the diagnostic accuracy of cardiac SPECT perfusion imaging is the respiratory motion of the heart. We present two algorithms for estimating and correcting heart motion in SPECT slices. One is a rigid-body motion estimator which involves prior extraction of a region-of-interest (ROI) around the heart for patient studies. The second is a non-rigid, constrained free-form deformation (FFD) motion estimation and correction method. The rigid registration was applied to two MCAT phantom datasets with realistic noise and rigid motion (SI translation of 1.5 cm). The maximum translational motion was estimated to within 0.7 mm and 0.3 mm for the two datasets respectively. Both the rigid and FFD algorithms were applied to a respiration gated patient study with phased binned datasets and compared.