A new method to quantify epicardial fat in the human atria with potential relation to fibrillation dynamics

Abstract Background The contribution of epicardial fat to mechanisms maintaining persistent atrial fibrillation (PAF) is questioned. Purpose We aim to develop a method to quantify local epicardial fat and explore its significance in predicting PAF dynamics. Methods 20 patients with PAF were admitted for right and left atrial (RA & LA) mapping with the CARTO system. The biatrial geometry, bipolar Octa-Ray`s signals (30 sec long) and activation patterns were exported for off-line analysis (MatLab environment). A fast-Fourier transform was computed on 30 sec bipolar signal length segments to compute dominant frequency (DF) and represented on atrial shells. Segmentation of atrial anatomy and epicardial fat (-190,-30HU) was performed from CT scans. Exported geometries from CT scan (atrium and epicardial fat) were merged with CARTO geometries on MatLab environment with dedicated software. The epicardial fat width (in mm) was computed for every node of the atrial shell, as the module of the perpendicular vector projected outwardly from the epicardium up to the external surface of the epicardial fat geometry. To assign corresponding values, the epicardial fat width and the dominant frequency were interpolated to adjust dimensions of both matrixes. Results The panel A of the figure represents an example of epicaldial fat segmentation from the CT scan. The panel B of the figure represents an example from the same case of epicardial fat quantification on the CT shell. The color-coded bar represent the width of the epicardial fat at every node of the atrial shell. As can be observed by visual inspection, the case example displays an appropriate matching between panel A and panel B. The technique was reproducible, allowing the representation of epicardial fat width in all the analyzed cases. To analyze correlation of the reconstructed epicardial fat maps with PAF dynamics, the DF of nodes corresponding to the higher 10% percentile were compared with the DF of the rest of nodes. Nodes belonging to the higher 10% percentile of epicardial fat width displayed higher DF (5.8±0.4 Hz) than the DF of the rest of the nodes (5.7±0.5 Hz; P=0.014). Conclusions we describe a method to quantify epicardial fat in the atrium which is feasible and with potential link to fibrillatory dynamics in PAF in patients. The later pave the way to explore in dipper details the role of epicardial fat on PAF mechanisms.