What question did this study set out to answer?

To evaluate the performance of an AI algorithm in detecting paroxysmal atrial fibrillation from sinus rhythm ECG data.

January 10, 2026Advances in Therapy0 citations

Artificial Intelligence Software for Detecting Paroxysmal Atrial Fibrillation from Sinus Rhythm Monitor ECG: Development and Clinical Trial

YTYuichi TamuraMita Hospital TTTomohiro TakataEngineering Advancement Association of Japan HTHirohisa TaniguchiEngineering Advancement Association of Japan

Key Result

An AI algorithm achieved 91.7% sensitivity and 65.0% specificity for detecting paroxysmal atrial fibrillation from sinus rhythm Holter ECGs within 7 days.

Key Points

To evaluate the performance of an AI algorithm in detecting paroxysmal atrial fibrillation from sinus rhythm ECG data.
Developed a deep-learning algorithm using 30-second Holter ECG data.
Conducted tenfold cross-validation for model training.
Validated with a separate set of patients, using a 4/10 positive rule for monitoring.
Included concurrent 7-day patch monitoring for pAF outcomes assessment.
Achieved mean sensitivity of 84.2% and specificity of 66.2% during cross-validation.
Best tuned model showed 84.9% sensitivity and 69.9% specificity on the validation set.
In the clinical trial, sensitivity was 91.7% and specificity was 65.0% in detecting pAF.
No adverse events related to the device were reported.

Structured PICO

Does an artificial intelligence algorithm analyzing short sinus rhythm Holter segments accurately detect paroxysmal atrial fibrillation within 7 days in adult patients?

Population

44 adults (age 20-85 years) in the prospective clinical trial, comprising 24 patients with paroxysmal atrial fibrillation documented within 7 days and 20 controls without AF history.

Intervention

Deep-learning artificial intelligence (AI) software analyzing ten consecutive 30-second sinus rhythm blocks from Holter ECG using a 4/10 positive rule.

Comparator

Concurrent 7-day patch monitoring (reference standard).

Outcome

Patient-level sensitivity and specificity for detecting paroxysmal atrial fibrillation within 7 days.surrogate

A deep-learning algorithm analyzing short sinus rhythm Holter segments can identify patients who will develop paroxysmal atrial fibrillation within 7 days, potentially serving as a triage tool for intensified monitoring.

Main Result

Absolute Event Rate: 0% vs 0%

Abstract

INTRODUCTION: Detecting paroxysmal atrial fibrillation (pAF) from sinus rhythm could enable earlier intervention and stroke prevention. We developed a deep-learning Holter electrocardiograph (ECG) algorithm and prospectively evaluated its patient-level performance against 7-day AF outcomes. METHODS: We curated 20,000 30-s sinus rhythm blocks (125 Hz) from Holter ECG data of patients with and without pAF, trained convolutional models with tenfold cross-validation, and assessed a separate validation set (n = 54; 27 pAF, 27 controls) to select the operating threshold. A multicenter prospective study then evaluated the algorithm using ten consecutive 30-s sinus rhythm blocks per patient with a 4/10 positive rule; patients with pAF underwent concurrent 7-day patch monitoring to anchor outcomes. RESULTS: Cross-validation during development yielded mean sensitivity 84.2% and specificity 66.2%; the best tuned model achieved 84.9% sensitivity and 69.9% specificity on the separate set. In the clinical trial, among 24 patients with AF documented within 7 days and 20 controls, the device showed sensitivity 91.7% (95% confidence interval (CI) 73.0-99.0) and specificity 65.0% (40.8-84.6). No device-related adverse events occurred. CONCLUSION: An artificial intelligence (AI) analyzing short sinus rhythm Holter segments can identify patients who develop pAF within 7 days, supporting use as a triage tool for intensified rhythm monitoring. TRIAL REGISTRATION: UMIN-CTR UMIN000047182.

Ask AI

Helpful

Bookmark

View Full Paper