Detection and correction of patient motion in dynamic and static myocardial SPECT using a multi-detector camera.

We have developed a method for the detection and correction of translational patient motion in dynamic and static myocardial SPECT studies. The method uses a low activity 99mTc point source and is especially designed for multi-detector cameras. The source's centroid coordinates are measured or derived for all projection images in a temporal frame. The coordinate curves fitted to predicted distributions and the projection images shifted to realign measured to fitted values, with sub-pixel accuracy. In dynamic studies, the frame with the best fits serves as reference for all others. The accuracy of this method, measured with cardiac phantom experiments, was found to be +/- 0.37 mm and +/- 0.44 mm in the axial and transaxial dimension, respectively. By comparison, overall motion in 42 patients undergoing 99mTc-teboroxime dynamic cardiac SPECT studies was +/- 1.6 mm and +/- 1.2 mm, respectively (average on 39,272 projection views). Application of the method to phantom experiments, 99mTc-sestamibi and 99mTc-teboroxime human studies visually eliminated artifactual perfusion defects from simulated phantom motion and actual patient motion.