Registration and fusion of segmented left atrium CT images with CARTO electrical maps for the ablative treatment of atrial fibrillation

This study aims to extract the interior surface of the left atrium (LA) and pulmonary veins (PVs) from three dimensional tomographic data and to integrate it with LA CARTO electrical maps. The separation of LA and PVs from other overlapping structures of the heart was performed processing 3D CT data by marker-controlled watershed segmentation and surface extraction. CARTO maps were then registered on the LA internal surface by a stochastic optimization algorithm based on simulated annealing. The residual registration error resulted inferior to 3 mm. The integration between electrophysiological and high resolved anatomic information of LA results feasible and may constitute a significant support for mechanism investigation and treatment of atrial fibrillation.