Low-dose 4D cardiac computed tomography with prior image constrained compressed sensing

Abstract

Cardiac computed tomography (cardiac CT) has received great attention since its advent due to the advantage of noninvasive method. However, in order to visualize the heart, higher temporal and spatial resolution are necessary, and other techniques to compensate the cardiac motions, such as respiration and heart beating, are required when reconstruction. In addition, large dose to patients is another limitation of cardiac CT. In this research, we proposed a new scanning method for more accurate reconstruction of the beating heart. We used orthogonally scanned two-cardiac-cycle data at the interested ranges of the reconstruction, whereas one-cardiac-cycle projections are used in the conventional CT systems. Prior image constrained compressed sensing (PICCS) algorithm was adapted for phase reconstruction, and we lower x-ray tube current in order to solve the dose problem. We evaluated the result images both qualitatively and quantitatively. The proposed scanning method with PICCS algorithm improved the accuracy of the reconstruction for each cardiac phase. Possibilities of 4D cardiac CT have been confirmed by the sequential image of the reconstructed images.