Arrhythmias & electrophysiology
Deep Neural Network Algorithm Using the Electrocardiogram for Detection of Obstructive Sleep Apnea
Authors' conclusion
Our AI-ECG model demonstrated good diagnostic performance as an ECG-based screening tool for OSA in a clinical population, particularly among females.
Study context
- Population
- Patients who completed sleep evaluation and underwent 12-lead electrocardiogram (ECG) at Mayo Clinic; Obstructive sleep apnea (OSA) defined as apnea-hypopnea index ≥57,170 OSA, 4,129 controls
- Participants
- 11,299 patients
- Outcome
- Discrimination of OSA (AUC)
Methodological detailsretrospective, single-centre
- Retrospective
- Single-centre
Performance
As reported in the abstract
Identification of OSA (test sample)
| Test or model | AUC | Accuracy | Sensitivity | Specificity |
|---|---|---|---|---|
| AI-ECG | 0.80 (95% CI: 0.77-0.83) | 73.7% | 77.0% | 68.6% |
AI-ECG
AUC0.80 (95% CI: 0.77-0.83)
Accuracy73.7%
Sensitivity77.0%
Specificity68.6%
Further findings
| Outcome | Females | Males | P value |
|---|---|---|---|
| AUC by sex (subgroup)Better discrimination in females | P < 0.001 | ||
| AUC | 0.82 (95% CI: 0.79-0.86) | 0.73 (95% CI: 0.68-0.78) |
Sensitivity analyses
- Predictive abilities robust across time intervals and with ECGs showing cardiac abnormalities
Original publication
Deep Neural Network Algorithm Using the Electrocardiogram for Detection of Obstructive Sleep Apnea. JACC Adv. Published 16 September 2025. doi:10.
Summarised from the published abstract; every number is checked against it