Arrhythmias & electrophysiology
Detecting Paroxysmal Atrial Fibrillation From an Electrocardiogram in Sinus Rhythm: External Validation of the Artificial Intelligence (AI) Approach
Authors' conclusion
The approach of using an AI-enabled electrocardiographic algorithm for the identification of patients with underlying paroxysmal AF from ECGs in SR was independently validated.
Study context
- Population
- In- and outpatients with at least 1 electrocardiography (ECG) in sinus rhythm (SR)494,042 ECGs in SR
- Participants
- 142,310 participants
- Outcome
- Identification of underlying paroxysmal AF from an ECG in SR (first ECG of each patient)
Methodological detailssingle-centre, external validation, Patients randomly allocated to training, internal validation and testing datasets in a 7:1:2 ratio
- Single-centre
- External validation
- Patients randomly allocated to training, internal validation and testing datasets in a 7:1:2 ratio
Performance
As reported in the abstract
Identification of underlying paroxysmal AF from first ECG in SR (testing dataset, AF prevalence 9.0%)
Accuracy
78.1%
95% CI 77.6%–78.5%
AUC ROC
0.87
95% CI 0.86–0.87
Area under the precision recall curve (AUPRC)
0.48
95% CI 0.46–0.50
Further findings
AUPRC in low-risk group (atrial fibrillation (AF) prevalence 3%)
- AUPRC decreased
- 0.21 (95% CI: 0.18-0.24)
- Secondary analysis with modified AF prevalence
AUPRC in high-risk group (AF prevalence 30%)
- AUPRC increased
- 0.76 (95% CI: 0.75-0.78)
- Secondary analysis with modified AF prevalence
External hospital validation
- Performance robust when validated in an external hospital
Original publication
Detecting Paroxysmal Atrial Fibrillation From an Electrocardiogram in Sinus Rhythm: External Validation of the AI Approach. JACC Clin Electrophysiol. Published 21 June 2023. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →