Key result
A deep neural network analyzing resting 12-lead ECGs predicted new onset atrial fibrillation within 1 year with an AUROC of 0.85, identifying a high-risk group with a hazard ratio of 7.2 for long-term AF.
Why the study?
AF is associated with substantial morbidity, especially when undetected, and predicting new-onset AF from resting 12-lead ECGs could enable targeted population screening to find it early and help prevent AF-related stroke.
Does a deep neural network analyzing resting 12-lead ECGs predict incident atrial fibrillation and help identify patients at risk for AF-related stroke in patients without a history of AF?
Cohort (n=430,000)
Yes
Does a deep neural network analyzing resting 12-lead ECGs predict incident atrial fibrillation and help identify patients at risk for AF-related stroke in patients without a history of AF?
Hazard Ratio: 7.2 (95% CI 6.9–7.6)
A deep neural network can accurately predict incident atrial fibrillation within 1 year from a standard resting 12-lead ECG in patients without a history of AF, potentially enabling targeted screening to prevent AF-related strokes.
May support ECG-based AF risk stratification; leaves open prospective validation before clinical use.
Background Atrial fibrillation (AF) is associated with substantial morbidity, especially when it goes undetected. If new onset AF could be predicted, targeted population screening could be used to find it early. We hypothesized that a deep neural network could predict new onset AF from the resting 12-lead electrocardiogram (ECG) and that this prediction may help prevent AF-related stroke. Methods We used 1.6M resting 12-lead ECG voltage-time traces from 430k patients collected from 1984-2019 in this study. Deep neural networks were trained to predict new onset AF (within 1 year) in patients without a history of AF. Performance was evaluated using areas under the receiver operating characteristic curve (AUROC) and precision-recall curve (AUPRC). We performed an incidence-free survival analysis for a period of 30 years following the ECG stratified by model predictions. To simulate real-world deployment, we trained a separate model using all ECGs prior to 2010 and evaluated model performance on a test set of ECGs from 2010 through 2014 that were linked to our stroke registry. We used standard metrics to explore different prediction thresholds for the model and also calculated how many AF-related strokes might be potentially prevented. Results The AUROC and AUPRC were 0.83 and 0.21, respectively, for predicting new onset AF within 1 year of an ECG. Adding age and sex improved the AUROC to 0.85 and the AUPRC to 0.23. The hazard ratio for the predicted high- vs. low-risk groups over a 30-year span was 7.2 [95% confidence interval: 6.9 – 7.6]. In a simulated deployment scenario, using the F 2 score to select the risk prediction threshold, the model predicted new onset AF at 1 year with a sensitivity of 69%, specificity of 81%, and positive predictive value (PPV) of 12%. This model correctly predicted new onset AF in 62% of all patients that experienced an AF-related stroke within 3 years of the ECG. Conclusions Deep learning can predict new onset AF from the 12-lead ECG in patients with no prior history of AF. This prediction may prove useful in preventing AF-related strokes.
No takes yet. Share an insight, caveat, or question.
Raghunath et al. (2020) conducted a cohort in Atrial Fibrillation (n=430,000). Deep neural network prediction of high risk from 12-lead ECG vs. Predicted low risk group was evaluated on Incident atrial fibrillation over a 30-year span (HR 7.2, 95% CI 6.9-7.6). A deep neural network analyzing resting 12-lead ECGs predicted new onset atrial fibrillation within 1 year with an AUROC of 0.85, identifying a high-risk group with a hazard ratio of 7.2 for long-term AF.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: