A roadmap to success in AF ablation

Michel Haissaguerre, MD

Michel Haissaguerre, MD


Noninvasive mapping of atrial fibrillation has blossomed in recent years, and its use in cases of AF can provide insight into the driver mechanisms that propagate distinct subtypes of the condition, as well as delineating driver regions to guide ablation procedures.

Results of the mapping and ablation of 190 AF cases, and a demonstration of the technique, will be presented this morning by Michel Haïssaguerre (Hôpital Cardiologique du Haut-Lévêque and the Université Victor Segalen Bordeaux II, Bordeaux, France), who spoke to AF Symposium News to offer a glimpse as to his main messages.

Your group demonstrated the feasibility of noninvasive mapping in atrial fibrillation. How did you go about this?

Initially the noninvasive mapping technique, with an array of body surface electrodes and CT-based cardiac geometry, was used to map the origin of organized cardiac arrhythmias with a localization accuracy of 6 to 10 mm by the group of Yoram Rudy (Cardiac Bioelectricity & Arrhythmia Center (CBAC), Washington University in St. Louis, St. Louis, MO, USA). Our experience with now over 600 patients has confirmed this accuracy.

We use this system to locate difficult cases of atrial or ventricular tachycardias or accessory pathways, as well as to optimize cardiac resynchronization therapy. The developments in signal processing and computer power have allowed AF mapping using activation mapping,1 while our group has investigated mathematical computation of phase-based analysis of body surface potentials.2 We believe this is a more proper tool to identify re-entrant drivers (rotors). In this meeting we will present our results of 190 AF cases, and a live procedure to illustrate the technique.

Could you describe this mapping technique you have applied to patient cases?

The non-contrast thoracic CT scan (1.2 to 2 mSv irradiation) gives us high-resolution 3D images of the individual biatrial-pulmonary vein geometry, the location of esophagus during ablation and the relative position of the electrodes.

A technician performs the maps one to two days before ablation in persistent AF, thus saving procedural time. Atrial fibrillatory electrograms outside QRST are acquired during the longest (>1s) AF ventricular pauses, until a cumulative AF time of 10 to 30 seconds is obtained, which appears sufficient to capture most mapping scenarios happening over time.

Typically, the movies show multiple simultaneous wave propagation patterns and their beat-to-beat changes during AF. They require statistical analysis of active driver regions (versus passive wave regions) to create a spatiotemporal probabilistic map.

The AF drivers are either focal, when centrifugal activation originates from a point or an area (e.g., a PV), or re-entrant (‘rotor’), when at least one wave fully rotates around a center point. Importantly, we do not only rely on images, but confirm a rotor by looking for sequential activation of body surface potentials. And because rotors meander significantly, the term arrhythmogenic regions is preferable to sites to delineate targets of ablation.

What have your investigations revealed about the distinctions between paroxysmal and persistent cases of AF?

In paroxysmal AF (25 patients), the main drivers are by far the pulmonary vein regions, which display venous breakthroughs and ostial rotors. This is not a surprise for clinicians, but it replicates mechanistic results shown in animals using optical mapping.

In persistent AF, we found typically three to five driver regions with a widespread anatomical distribution in both atria. We found a confirmation of prior studies showing the importance of the regions of pulmonary vein antra, adjacent septum and left appendage. But there are wide inter-individual variations at other locations – the inferior left atrium, the coronary sinus, the right atrium – which is essential mapping information for ablation strategy.

How can we refine our knowledge of AF and ablation intervention using this method?

First of all, we have compared different subsets of persistent AF based on continuous AF duration. The number of targeted regions increases with the duration of continuous AF: from three to four in the first six months of AF, four to five in months six to 12 of AF, and six to seven in long-lasting AF.

The results of driver ablation alone (without lines) progressively deteriorate, with AF termination achieved in 75% of cases with less than six months of AF; to 47% in 7-12 months of AF; and 13% in long-lasting AF.

The patients presenting with sinus rhythm had particularly favorable results; after AF induction we could terminate 90% of AF by targeting only two or three driver regions. Clinical outcome at six months parallels these acute results, and this will be presented at the meeting.

This result provides conceptually and practically important information. While persistent AF is driven from a few regions in the initial months, it is tempting to hypothesize that this time period is followed by remodeling leading to progressive dissemination of AF drivers and structural substrate mechanisms (notably wavelets), which will require wide substrate or linear ablation. These results strongly argue for early intervention.

Secondly, the delineation of driver regions provides an ablation roadmap, which covers a smaller area than the conventional ablation, targeting pulmonary veins and complex atrial electrograms. This results in a significant reduction in RF delivery targeting driver domains than widespread stepwise ablation, as the conventional ablation targets pulmonary veins and complex atrial electrograms.

Professor Haïssaguerre will deliver his presentation ‘Mechanistic Insights from Noninvasive Mapping of AF - Implications for Catheter Ablation’ during this morning’s session: ‘Mechanisms of Atrial Fibrillation - Basic and Clinical Insights’; 7:30 - 9:20 AM



  1. Cuculich PS et al. Noninvasive characterization of epicardial activation in humans with diverse atrial fibrillation patterns. Circulation. 2010 Oct 5;122(14):1364-72.

  2. Haissaguerre M et al. Noninvasive Panoramic Mapping of Human Atrial Fibrillation Mechanisms: A Feasibility Report. J Cardiovasc Electrophysiol. 2013 February 1;24(6):711-7.

Tammy Griffin-Kumpey