Statistics and behavior of clinically significant extrapulmonary vein atrial fibrillation sources - machine-learning-enhanced electrographic flow mapping in persistent atrial fibrillation

Aug 5, 2026

Machine-learning-enhanced electrographic flow (EGF) mapping is presented as a method to identify clinically significant extrapulmonary vein sources sustaining persistent atrial fibrillation (AF). The EGF model is trained on procedural outcomes from 199 fully anonymized retrospective patient datasets to learn an activity threshold and optimize model hyperparameters. Divergent wavefront patterns reconstructed from 64-electrode basket recordings are quantified for temporal prevalence. The threshold is validated in 85 prospective FLOW-AF trial patients, where persisting post-procedure sources predict higher recurrence and ON-state switching reduces AF cycle-length spatial variability by over 50%, suggesting an entraining effect.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages11

Reading Time28 min

AccessFree to download

Introduction

Electrographic flow (EGF) mapping and machine learning gap

ExtraPV mechanisms in persistent atrial fibrillation

Methods and Results

Patient datasets and model training

Activity threshold and validation in FLOW-AF

Clinical recurrence and cycle-length variability effects

Conclusions

Outcome-based optimization for actionable AF source detection

TYPE Original Research PUBLISHED 26 August 2025 DOI 10.3389/fcvm.2025.1517484 EDITED BY Rui Providencia, University College London, United Kingdom REVIEWED BY James P. Hummel, Yale University, United States Cristiano F. Pisani, University of São Paulo, Brazil *CORRESPONDENCE Peter Ruppersberg [pruppersberg@cortexep.com](pruppersberg@cortexep.com) RECEIVED 26 October 2024 ACCEPTED 03 July 2025 PUBLISHED 26 August 2025 CITATION Ruppersberg P, Castellano S, Haeusser P, Ahapov K, Kong MH, Spitzer SG, Nölker G, Rillig A and Szili-Torok T (2025) Statistics and behavior of clinically signiﬁcant extrapulmonary vein atrial ﬁbrillation sources: machine-learning-enhanced electrographic ﬂow mapping in persistent atrial ﬁbrillation. Front. Cardiovasc. Med. 12:1517484 . doi: 10.3389/fcvm.2025.1517484 COPYRIGHT © 2025 Ruppersberg, Castellano, Haeusser, Ahapov, Kong, Spitzer, Nölker, Rillig and Szili-Torok. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY) . The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Statistics and behavior of clinically signiﬁcant extrapulmonary vein atrial ﬁbrillation sources: machine-learningenhanced electrographic ﬂow mapping in persistent atrial ﬁbrillation Peter Ruppersberg1*, Steven Castellano1, Philip Haeusser1, Kostiantyn Ahapov1, Melissa H. Kong1, Stefan G. Spitzer2,3, Georg Nölker4, Andreas Rillig5 and Tamas Szili-Torok6 1Cortex, Inc. Menlo Park, CA, United States, 2Praxisklinik Herz und Gefäße Dresden, Akademische Lehrpraxisklinik der TU Dresden, Dresden, Germany, 3Brandenburg University of Technology Cottbus-Senftenberg, Institute of Medical Technology, Cottbus, Germany, 4Department of Cardiology, Heart and Diabetes Center North Rhine-Westphalia, Ruhr University Bochum, Bad Oyenhausen, Germany, 5Interventional Electrophysiology, University Heart Center, Hamburg, Germany, 6Department of Cardiology, University of Szeged Albert Szent-Györgyi Medical School, Szeged, Hungary Introduction: Electrographic ﬂow (EGF) mapping is an FDA 510(k)-cleared method for visualizing atrial activation wavefronts in atrial ﬁbrillation (AF) . Its clinical efﬁcacy was demonstrated in the FLOW-AF randomized controlled trial, and its fundamental principles have been previously described. However, the underlying machine learning strategy used to develop and reﬁne the EGF algorithm has not yet been detailed. Here, we present how our EGF Model—trained on procedural outcomes from 199 fully anonymized retrospective patient datasets—identiﬁes clinically signiﬁcant sources of AF and how this machine learning–driven hyperparameter optimization underlies its clinical effectiveness. We also examine the statistical characteristics of the identiﬁed sources and their impact on cycle length variability, offering insights into potential pathophysiological mechanisms. Methods and results: Unipolar electrograms were recorded from patients with persistent or long-standing persistent AF using 64-electrode basket catheters. The EGF Model processes these recordings to reconstruct divergent wavefront propagation patterns and quantify their temporal prevalence. We included 399 retrospective patients in total: 199 for training and optimizing 24 model hyperparameters, and 200 for subsequent analyses of source prevalence and characteristics. Our machine learning approach established an activity threshold, above which divergent wavefront patterns—termed “signiﬁcant sources” —predicted AF recurrence. This threshold was validated in 85 prospective patients from the published FLOW-AF trial. Signiﬁcant sources persisting post-procedure were associated with signiﬁcantly hig

FAQs

What does the EGF model learn to identify in persistent atrial fibrillation?

The EGF model learns how to identify clinically significant AF sources by using a machine-learning-driven activity threshold and hyperparameter optimization based on procedural outcomes.

How was the model validated in prospective patients?

The activity threshold was validated in 85 prospective patients from the published FLOW-AF randomized controlled trial.

What relationship did significant sources show with AF recurrence and cycle-length variability?

Significant sources persisting after the procedure were associated with higher recurrence rates, and when sources switched ON, spatial variability of AF cycle lengths in the atrium decreased by more than 50%.

Aug 5, 2026

Upload by Seraphina