Noninvasive panoramic mapping of human atrial fibrillation mechanisms: a feasibility report.

INTRODUCTION

Recent developments in body surface mapping and computer processing have allowed noninvasive mapping of atrial activation responsible for various cardiac arrhythmias with increasingly greater resolution. We developed specific algorithms to identify localized sources and atrial propagation occurring simultaneously during ongoing atrial fibrillation (AF).

METHODS AND RESULTS

We report the feasibility of noninvasive panoramic mapping of human AF mechanisms and its validation by successful ablation. We used a commercially available mapping system using an array of 252 body surface electrodes and noncontrast thoracic CT scan to obtain high-resolution images of the biatrial geometry and the relative electrode positions. On the surface unipolar electrograms acquired during AF we developed specific signal-analysis process combining filtering, wavelet transform, and phase mapping. At least 5 windows with spontaneous, long ventricular pauses were selected for mapping. The incidence, location and characteristics of localized sources (foci and rotors) were assessed on the cumulative duration of all recorded windows. In a patient with paroxysmal AF, noninvasive maps showed multiple single or repetitive discharges from 3 pulmonary veins (PVs), a rotor meandering along the right venous ostia, and their mutual interplay. All areas outside the left posterior wall were passively activated. AF terminated during isolation of right PV. In a patient with persistent AF for 7 months, a rotor was identified recurrently, drifting in the left atrial inferior and posterior wall and in the roof. It was not stationary for more than 2 rotations. The right atrial free wall was activated over the Bachman's bundle by a passive wavefront propagating in a counterclockwise pattern. Ablation at the rotor locations abruptly converted AF into atrial tachycardia after 10 minutes of radiofrequency application. Further mapping and ablation confirmed a counterclockwise cavotricuspid isthmus-dependent flutter.

CONCLUSIONS

This report demonstrates the feasibility of noninvasive panoramic mapping of AF in identifying active sources, which include unstable rotors and PV foci, and its validation by ablation results.