Pranav Bhagirath

81 Implementation of a standardized cardiac magnetic resonance based workflow for atrial fibrillation catheter ablation Ablation guidance For navigation purposes, the anatomical shell and LA scar-map were converted into the proprietary format for the EAM (EnSite Verismo™, St. Jude Medical) and fused with the invasively acquired anatomy using a landmark based registration process (movie 1). The fusion was performed by selecting landmarks (fiducials) on the invasively obtained anatomical shell and the LA shell. These fiducials usually consisted of the pulmonary veins and the mitral valve annulus. Comparison of scar-map and EAM For patients undergoing a redo ablation, the location and distribution of scar on the generated scar-maps was compared with the ablation sites from the index procedure. The gaps identified on the scar-maps were compared with the ablation sites of the redo procedure. RESULTS Five patients suffered from paroxysmal AF and three had long-standing persistent AF. Therewas ahistory of hypertension in5patients (63%). Onepatient suffered fromsystolic heart failure (LVEF 26%). Based on conventional imaging, all patients had LA dilatation and 5 patients (63%) suffered from mitral regurgitation (moderate). LAA thrombus was excluded in all patients using TEE and MRA. Additional patient characteristics are provided in table 2 . Image acquisition and analysis There were no patients with contra-indications for undergoing CMR. All scans were performed in sinus rhythm. Typical scan time for the CMR protocol was 36 ± 3 minutes, depending on subject respiratory- and heart rate. LA segmentations were constructed in 2.8 ± 0.8 minutes. The biplane area-length method took 52 ± 7 seconds to perform and underestimated the average LA volume (67 ± 16 ml) compared to ITK-Snap by 15 ml.