Rigid-body transformation of list-mode projection data for respiratory motion correction in cardiac PET

Respiratory motion is a source of artefacts and quantification errors in cardiac imaging. Preliminary studies with retrospective respiratory gating in PET support the observation of other imaging modalities of a rigid-body motion of the heart during respiration. However, the use of gating techniques to eliminate motion may result in poor count statistics per reconstructed image. We have implemented a motion correction technique which applies rigid-body transformations on list-mode data event-by-event on the basis of a geometric model of intersection of the lines-of-response with the scanner. Pre-correction for detector efficiencies and photon attenuation before transformation are included in the process. Projection data are acquired together with physiological signal (every ms) from an inductive respiration monitor with an elasticised belt at chest level. Data are retrospectively sorted into separate respiratory gates on an off-line workstation. Transformation parameters relating the gated images, estimated by means of image registration, can be applied on the original list-mode data to obtain a single motion-corrected dataset. The accuracy of the technique was assessed with point source data and a good correlation between applied and measured transformations, estimated from the centroid of the source, was observed. The technique was applied on phantom data with simulated respiratory motion and on patient data with C/sup 15/O and /sup 18/FDG. Quantitative assessment of preliminary C/sup 15/O patient datasets showed at least 4.5% improvement in the recovery coefficient at the centre of the left ventricle.