4-D motion field estimation by Combined Multiple Heart Phase Registration (CMHPR) for cardiac C-arm data

In interventional cardiology three-dimensional anatomical and functional information of the cardiac chambers would have an important impact on diagnosis and therapy. With the technology of C-arm CT it is possible to reconstruct intraprocedural 3-D images from angiographic projection data. In order to generate accurate and artifact-free reconstructions from dynamic cardiac projections, the motion needs to be taken into account. We present the novel Combined Multiple Heart Phase Registration (CMHPR) method. CMHPR is an iterative motion estimation and compensation algorithm that uses projection data acquired during a single C-arm sweep. Filtered-backprojection (FBP) volumes from electrocardiogram (ECG)-gated data are reconstructed for different motion states of the heart. According to an unknown 4-D motion vector field the ECG-gated FBP images are deformed and accumulated to a sum volume for representing the status of a particular heart phase. In an iterative optimization procedure the 4-D motion vector field is computed by registering the sum volume to a reference volume of the same heart phase. The negative normalized cross correlation (NCC) of both volumes is used as a cost function. In this paper, the reference image is generated using the prior image constrained compressed sensing (PICCS) algorithm combined with the improved total variation (iTV). First preliminary experiments on clinical porcine data sets show promising results. CMHPR reduces streak artifacts and simultaneously preserves sharp edges without producing the artificial comic-like appearance of the PICCS + iTV reference volume.