Personalizing treatments to deliver the right therapy at the right time to the right patient is vital to the future of heart care. Research is refining personalized care for arrhythmias, with artificial intelligence (AI) as a force multiplier.

Duke Health electrophysiologist Sana M. Al-Khatib, MD, MHS, Heart Rhythm Society (HRS) president, highlights AI advancements to personalize electrophysiology care. “In implantable devices, AI has enabled us to decrease false positives significantly, which has led to improved efficiencies in clinic when we interpret results,” she says. “AI in defibrillators enables them to terminate ventricular tachycardia, adjusting the antitachycardia pacing to increase the likelihood of termination. This lowers the need for shocks.”

As the science continues to advance, AI can be used for even more powerful prediction algorithms. “We want to predict who’ll have an arrhythmia,” says Al-Khatib. “For example, among patients with atrial fibrillation, which are at the greatest risk for stroke, heart failure, or cognitive decline? Once such algorithms and models are developed, they should be validated. Randomized clinical trials remain the best way to validate findings and determine which tools should be adopted in practice.”

Although the incidence of sudden cardiac death (SCD) is greatest among groups at high risk, including those with heart failure or sustained ventricular arrhythmia, research has found that approximately half of all fatal SCD cases occur in people not previously diagnosed with cardiovascular disease.

“We want to use AI to identify patients in those high-risk groups who are low risk for SCD, so we can focus on implanting ICDs [implantable cardioverter defibrillators] in those who will benefit,” says Al-Khatib. “At the same time, we want to identify and treat the patients who aren’t indicated for ICDs but are actually at risk of SCD among the low-risk general population.”

As part of this effort, Al-Khatib has formed a Sudden Cardiac Death Prediction and Prevention Task Force within HRS. The task force will work to promote research in screening and treatment protocols and using AI to advance the science. For example, the Multimodal AI for ventricular Arrhythmia Risk Stratification (MAARS) model, developed with neural networks, has already been shown to be effective in SCD risk stratification for patients with hypertrophic cardiomyopathy and cardiac sarcoidosis. HRS supports the development of AI-supported decision tools to reduce SCD risk.

Published in the New England Journal of Medicine, the CHAMPION AF trial compared left atrial appendage closure (LAAC) with the Watchman device (Boston Scientific, Marlborough, MA) to oral anticoagulation (OAC) among patients with atrial fibrillation (AFib) over three years. The trial found that LAAC was noninferior to OAC in cardiovascular-related death, stroke, or systemic embolism. In addition to the decreased risk of bleeding, however, there was a modest increased risk of ischemic stroke among the LAAC arm of the study.

“I’m convinced there’s an important and clinically relevant reduction in bleeding with the device,” says Duke cardiologist Christopher B. Granger, MD, who served on the study steering committee. “For patients not eligible for oral anticoagulation or who aren’t adherent or willing to take the drugs, this is an important procedure that reduces overall stroke risk.”

He notes that LAAC is most suitable for patients with lower stroke risk or who have had ablation. “This trial helps us to know more about in which patients this therapy is most beneficial,” Granger concludes.

Electrophysiologist Sean Pokorney, MD, MBA, director of the Duke Clinical Research Institute Arrhythmia Core Laboratory, copresented the preliminary findings of the Centers of Excellence Optimal Management Pathways for Atrial Fibrillation Specialty Services (COMPASS) quality improvement study in a late-breaking session at Heart Rhythm 2026.

The study tested different care pathways in three health systems to deliver early rhythm control for patients identified with atrial fibrillation (AFib). Duke’s care pathway included routing patients to a rapid-access AFib clinic within 48 hours of referral. Quick access to care reduced hospitalizations and emergency department visits.

As Duke advances further research, diagnosis and treatment in arrhythmia will continue to improve. “We hope to leverage AI to understand which patients will respond best to which treatments and which might face complications from interventions,” Al-Khatib says. “That will aid us in selecting the best approach for every patient, providing personalized care.”