AI-guided cardiac ablation improves 1-year outcomes in persistent atrial fibrillation
Click Here to Manage Email Alerts
Key takeaways:
- Tailored ablation guided by AI better prevented atrial fibrillation recurrence vs. standard pulmonary vein isolation.
- The patient population was those with persistent AF not responding to drug therapy.
Tailored ablation guided by AI was linked to better 1-year outcomes than pulmonary vein isolation alone for patients with persistent atrial fibrillation, according to the results of the TAILORED-AF trial presented at Heart Rhythm 2024.
“TAILORED-AF was a large-scale multinational randomized controlled trial designed to evaluate whether targeting areas of spatiotemporal dispersion detected by AI plus pulmonary vein isolation is superior to pulmonary vein isolation alone for the treatment of persistent and longstanding AF patients,” Isabel Deisenhofer, MD, professor of electrophysiology at the German Heart Centre Munich, said during a presentation. “Spatiotemporal dispersion is defined as a collection of electrograms that display atypical conduction characteristics across multiple electrodes of a mapping catheter, supposedly indicating a role in initiating or sustaining atrial fibrillation.”
The researchers randomly assigned 374 patients with symptomatic persistent or longstanding persistent AF refractory or intolerant to at least one antiarrhythmic drug (mean age, 66 years; 79% men) with a duration of more than 3 months but less than 5 years (< 1 year in U.S. patients) to receive pulmonary vein isolation (PVI) alone (anatomical approach) or PVI plus AI guidance (Volta AF-Xplorer, Volta Medical; tailored approach).
The AI guidance system
“The workflow for the tailored arm required biatrial mapping, and by starting with mapping the left atrium, physicians could employ a ‘smart PVI’ that informed their PVI such that they could incorporate dispersion zones to their [wide-area circumferential ablation] approach,” Deisenhofer said during the presentation. “Otherwise, operators ablated all left and right atrial dispersion zones. If termination occurs and all dispersion zones are ablated, the procedure would end. If no termination was achieved, the operators remapped and ablated the remaining identified dispersions. The AI algorithm of this device is based on data that have been collected over numerous cases where the physicians used dispersion assessment and ablation targeting approach. The resulting hundreds of thousands of electrograms are curated by Volta data scientists, who identify multiple features that take into consideration cycle length, activation, fractionation, voltage, spatial components and others. These are then meticulously annotated by expert electrophysiologists to identify patterns of dispersion that exhibited acute effects to the arrythmia. Finally, long-term outcome results are also taken into consideration based of the selectivity of the patient dataset used in the algorithm.”
Less recurrence in tailored group
The primary endpoint of freedom from AF after one procedure, on or off antiarrhythmic drugs, at 12 months occurred more often in the tailored group vs. the anatomical group in both the modified intention-to-treat and per-protocol populations (88% vs. 70%; P < .0001 for both), according to the researchers.
The secondary 1-year endpoints of freedom from AF/atrial tachycardia after one procedure (P = .09) and after one or two procedures (P < .01) both favored the tailored group in the per-protocol population, Deisenhofer said.
All three endpoints also favored the tailored group in the per-protocol population at 1 year when restricted to those who had AF for 6 months or more, she said.
In addition, she said, the tailored group was more likely to achieve acute AF termination by ablation (66% vs. 15%; P < .001) and acute sinus rhythm conversion by ablation (53% vs. 13%; P < .001) at 1 year vs. the anatomical group.
“A tailored cardiac ablation guided by AI significantly improves the long-term outcomes of persistent AF ablation,” Deisenhofer said during the presentation. “The use of artificial intelligence for the reproducible and reliable identification of ablation target areas was pivotal in achieving this difference and provides hope for further applications in interventional medicine.”
Perspective
Back to Top
The researchers observed a significant improvement in long-term freedom from AF with pulmonary vein isolation plus additional ablation guided by an AI/machine learning-assisted strategy that evaluates areas with electrical signals consistent with an observation/concept called spatio-temporal dispersion compared with the standard strategy.
The findings were very much remarkable in that this is one of the only studies that has ever demonstrated that a particular ablation strategy beyond pulmonary vein isolation shows benefit, just pulmonary vein isolation. For persistent AF, our field’s outcomes have been significantly inferior to those for paroxysmal AF, likely because persistent AF usually signifies more advanced/progressed disease, and hence abnormalities in myocardial tissues beyond just the pulmonary veins. However, finding the right location to ablate beyond the pulmonary veins has been challenging for our field. This is really the first time a strategy to tailor the ablation to areas that may be best candidates for ablation may be beneficial.
The treatment arm had notably much longer procedure times, although there was no increase in complications.
These procedures were all performed with radiofrequency ablation. Although this mapping strategy (Volta) is independent of the ablation energy source used, the use of pulsed field ablation to target these sites identified was not evaluated.
Finally, this initial late-breaking clinical trial presentation did not specify which regions of the left atrium were targeted and how often. That’s to be presented after additional analysis.
This has the potential to be incorporated into practice as a strategy to treat our most challenging AF patients, those who otherwise would not have benefited from just pulmonary vein isolation alone. The impact of the longer procedure times on a practical electrophysiology lab implementation perspective remains to be seen. And there are other companies/strategies that also incorporate some sort of AI/machine learning-based electrogram signal or other clinical/procedural characteristics analysis to guide ablation. It remains to be seen if additional inputs into the ablation decision will yield other benefits.
Collapse
Deisenhofer I, et al. Late-breaking clinical trials and science – innovation. Presented at: Heart Rhythm 2024; May 16-19, 2024; Boston.
Read more about
Click Here to Manage Email Alerts