UCLA

Cardiac electroanatomical mapping (EAM) to locate myocardial scar substrate during ventricular tachycardia (VT) ablation therapy is currently faced with a major limitation. The presence of epicardial fat leads to false-positive low-voltage maps that may interpreted as myocardial scar. Pre-procedural cardiac magnetic resonance (CMR) imaging can produce images of the heart in which both cardiac geometry and epicardial fat are particularly conspicuous. The goal of the present work is to build CMR based geometric models composed of a shell of the cardiac anatomy and epicardial fat. Herein, we first defined a black-blood imaging protocol that is feasible in patients with implanted devices. Then a semi-automated image segmentation method was developed and applied to extract ventricular surface anatomy and epicardial fat maps from which three-dimensional surfaces were built. The results suggest that deriving maps of epicardial fat from MRI data is feasible, accurate when compared to expert observers, and suitable for integration in a clinical catheter-based ablation procedure.