Artificial intelligence-guided screening for atrial fibrillation using electrocardiogram during sinus rhythm: a...

Area of Science:

• Cardiology
• Medical Informatics
• Artificial Intelligence in Healthcare

Background:

• Previous atrial fibrillation (AF) screening trials indicate a need for more targeted strategies.
• Undetected AF poses a significant stroke risk, necessitating improved diagnostic approaches.

Purpose of the Study:

• To evaluate the effectiveness of an artificial intelligence (AI) algorithm-guided targeted screening approach for identifying previously unrecognized atrial fibrillation.
• To assess the yield of AI-driven risk stratification in AF screening.

Main Methods:

• A pragmatic, non-randomized interventional trial involving patients with stroke risk factors and no known AF.
• Continuous ambulatory heart rhythm monitoring for up to 30 days, with AI algorithm applied to electrocardiograms (ECGs) for risk stratification (high-risk vs. low-risk).
• Propensity-score matching to real-world controls for secondary analysis.

Main Results:

• Atrial fibrillation was detected in 7.6% of high-risk patients and 1.6% of low-risk patients.
• AI-guided screening significantly increased AF detection compared to usual care (10.6% vs. 3.6% in high-risk group, p<0.0001).
• The AI algorithm demonstrated a significant association with higher AF detection rates (OR 4.98, p=0.0002).

Conclusions:

• An AI-guided targeted screening approach leveraging existing clinical data enhances the detection rate of atrial fibrillation.
• This AI-driven strategy holds potential for improving the overall effectiveness of atrial fibrillation screening programs.