Fully 4D reconstruction of gated cardiac images

In this paper we investigate the benefits of a spatiotemporal approach for the reconstruction of cardiac image sequences. To account for the temporal correlations among the gated frames, we introduce a temporal prior based on motion compensation in an image sequence. The image sequences are reconstructed simultaneously through a maximum a posteriori (MAP) framework. To evaluate the performance of our algorithm we used signal to noise ratio, time activity curves, and a channelized Hotelling observer (CHO). We simulated gated cardiac perfusion imaging with Tc-99m-sestamibi using the 4D gated mathematical cardiactorso (gMCAT) phantom. Our experiments show that the proposed algorithm can improve both the reconstruction accuracy of the myocardium and the detectability of perfusion defects without causing significant motion-blurring.