Log in to MyACC
Artificial Intelligence-Guided Screening for AF Using ECG
Quick Takes
- In this nonrandomized prospective study, 1,003 patients with a mean age of 74 years underwent artificial intelligence (AI)-enabled sinus rhythm ECG followed by 3-week ambulatory ECG.
- AF was detected almost 5 times more often with the AI algorithm than in controls, suggesting that an AI-guided targeted screening approach using existing clinical data may improve the effectiveness of AF screening.
Study Questions:
What is the effectiveness of an artificial intelligence (AI) algorithm-guided targeted screening for identifying previously unrecognized atrial fibrillation (AF)?
Methods:
The authors recruited patients with stroke risk factors but with no history of AF who had an electrocardiogram (ECG). Participants wore a continuous ambulatory heart rhythm monitor for up to 30 days. The AI algorithm was applied to the ECGs to divide patients into high-risk or low-risk groups. The primary outcome was newly diagnosed AF. In a secondary analysis, trial participants were propensity-score matched (1:1) to individuals from the eligible but unenrolled population who served as real-world controls.
Results:
There were 1,003 patients. Their mean age was 74 years. Over a mean 22 days of continuous ECG monitoring, AF was detected in six (1.6%) of 370 patients with low risk and 48 (7.6%) of 633 patients with high risk (odds ratio, 4.98; p = 0.0002). Compared with usual care, AI-guided screening was associated with increased detection of AF (high-risk group: 3.6% with usual care vs. 10.6% with AI-guided screening, p < 0.0001; low-risk group: 0.9% vs. 2.4%, p = 0.12) over a median follow-up of 9.9 months.
Conclusions:
The authors concluded that an AI-guided targeted screening approach based on clinical data increases the yield for AF detection.
Perspective:
The diagnostic yield of screening for AF is low in unselected populations. Such screening remains controversial resulting in differing guidelines for patients aged ≥65 years in European and US guidelines. Prior retrospective studies showed that AI algorithms could identify ECG signatures of AF risk present during normal sinus rhythm. The present study is the first prospective study showing that the AI algorithm adds to risk stratification beyond traditional clinical risk factors. The findings from this study suggest that an effective strategy for AF screening may be possible if confirmed in future randomized trials.
Clinical Topics: Arrhythmias and Clinical EP, Geriatric Cardiology, Prevention, Implantable Devices, SCD/Ventricular Arrhythmias, Atrial Fibrillation/Supraventricular Arrhythmias
Keywords: Algorithms, Arrhythmias, Cardiac, Arrhythmia, Sinus, Artificial Intelligence, Atrial Fibrillation, Diagnostic Screening Programs, Electrocardiography, Geriatrics, Risk Assessment, Risk Factors, Stroke, Secondary Prevention
< Back to Listings
