Rhythm AI has announced results from the STAR Apollo study evaluating the company’s Stochastic Trajectory Analysis of Ranked signals (STAR) mapping approach in redo ablation procedures of patients with persistent atrial fibrillation (AF). The results were simultaneously published in the journal Heart Rhythm and presented by Dhanunjaya Lakkireddy (Kansas City Heart Rhythm Institute, Overland Park, USA) during the late-breaking clinical trials session at HRX Live 2026 (18–20 September, Atlanta, USA).
“After a first ablation fails, the hardest question is knowing where to go next, beyond the pulmonary veins,” Lakkireddy commented. “STAR Apollo gives us a way to answer that using a catheter and mapping system already in the lab. It’s not a black box but built on the same activation-timing principles electrophysiologists have relied on for decades, just applied automatically and at a scale no one could manage by hand. In this study, that translated into 90% of patients in sinus rhythm at the one-year follow-up visit, which is an encouraging signal.”
The prospective, single-arm STAR Apollo study was conducted across four US centres in 50 patients undergoing redo ablation for persistent AF that had recurred after a prior pulmonary vein isolation (PVI).
One in three patients will undergo a repeat ablation for AF and re-isolation of the PVI alone is often insufficient, as detailed in a Rhythm AI press release. The STAR Apollo mapping system provides a precise, patient-specific map of AF drivers by identifying repetitive patterns of activation (RPAs) and early sites of activation (ESAs), and is compatible with the two most widely used 3D mapping platforms in clinical practice.
The late-breaking one-year data for the 50-patient STAR Apollo study showed that the primary endpoint—procedural feasibility—was met in 100% of patients Additionally, 90% of patients were free of AF at the one-year follow-up. Very few patients had reinterventions, with analyses showing that 73% of patients had no recurrence of arrhythmia at all. Furthermore, a limited number of sites were targeted per patient, and the study showed a strong safety profile with no procedure-related deaths, stroke or atrio-oesophageal fistulas. Finally, STAR mapping added very little time to the cases, which had a median time of just over two hours.
“These are the kind of data that matter to us both clinically and commercially. STAR Apollo is already FDA [US Food and Drug Administration]-cleared and compatible with the two mapping platforms present in EP [electrophysiology] labs performing AF ablations,” said Mike Dineen, president and chief executive officer (CEO) of Rhythm AI. “This study shows clear evidence that our approach to locating AF drivers is effective and safe for this difficult-to-treat AF population, and the results exceeded our expectations. That’s the evidence base we need as we prepare to expand physician access and conduct further evaluations of the STAR Apollo technology.”
