AI-ECG Risk Stratification for Atrial Fibrillation: Real-World Performance and Explainability.

AI / informaticsJACC. Advances· Jul 2026

AI-ECG Risk Stratification for Atrial Fibrillation: Real-World Performance and Explainability.

Kouki Matsuo, Yoshihiro Sobue, Taiji Miyake … Hideo Izawa

Read paper

Prediction Models & Risk ScoresAI / informaticsDiagnosisQ1IF 2.9

AI-ECG alone shows modest ability to identify atrial fibrillation risk.

AI-ECG Risk Stratification for Atrial Fibrillation: Real-World Performance and Explainability.

Kouki Matsuo … Hideo Izawa

JACC. Advances · Jul 2026

Background

Artificial intelligence-enabled electrocardiography (AI-ECG) has emerged as a potential method for identifying atrial fibrillation (atrial fibrillation) from sinus rhythm.

Purpose

The objective of the study was to assess the performance and explainability of an AI-integrated ECG system for atrial fibrillation risk stratification in a multicenter real-world cohort.

Methods

We enrolled 665 patients aged ≥40 years who underwent 12-lead ECGs using an AI-enabled electrocardiograph (FCP-9900).

n = 665 patients

0.64-0.69

Results

0.64-0.69

The modest ability of AI-ECG alone to predict irregular heartbeat risk

n = 665 patients

More results

Atrial fibrillation prevalence increased across AI-ECG risk categories, with significantly higher odds in the mid-high and high groups vs low.

SHapley Additive exPlanations analysis showed CHA 2 DS 2 -VASc as the most influential predictor, whereas AI-ECG provided modest incremental value.

“

Conclusion · 1 of 2

AI-ECG provides rapid, low-cost atrial fibrillation risk estimation from a single sinus rhythm ECG; however, its predictive performance is modest compared with clinical scores.

Conclusion · 2 of 2

At present, AI-ECG may complement, but not replace, traditional risk stratification.

AI-ECG Risk Stratification for Atrial Fibrillation: Real-World Performance and Explainability.

AI-ECG Risk Stratification for Atrial Fibrillation: Real-World Performance and Explainability.

Related papers