Artificial intelligence identifies individuals with prediabetes using single-lead electrocardiograms
Authors' conclusion
This study establishes that an AI model, DiaCardia, can accurately identify individuals with prediabetes from an ECG alone, with performance that is robust across different patient cohorts and independent of major clinical confounders. Our highly generalizable, single-lead DiaCardia model offers a promising solution for scalable prediabetes screening via wearable devices, potentially enabling early, home-based detection and transforming diabetes prevention strategies.
Study context
- Population
- Health checkup records; prediabetes/diabetes defined as FPG ≥ 110 mg/dL, HbA1c ≥ 6.0%, or ongoing diabetes treatmentprimary cohort
- Participants
- 16,766 health checkup records
- Comparison
- DiaCardia vs Single-lead DiaCardia
- Outcome
- Identification of prediabetes from ECG (AUROC)
Methodological detailsArtificial intelligence (AI) model development and validation, internal held-out test dataset, independent external validation cohort
- AI model development and validation
- Internal held-out test dataset
- Independent external validation cohort
Performance
As reported in the abstract
Prediabetes identification, internal test dataset
| Test or model | AUROC | Sensitivity | Specificity |
|---|---|---|---|
| DiaCardia | 0.851 | 85.7% | 70.0% |
| Single-lead DiaCardia | 0.844 | 82.3% | 70.2% |
AUROC
DiaCardia0.851
Single-lead DiaCardia0.844
Sensitivity
DiaCardia85.7%
Single-lead DiaCardia82.3%
Specificity
DiaCardia70.0%
Single-lead DiaCardia70.2%
Further findings
| Outcome | DiaCardia |
|---|---|
| External validation cohort (n = 2,456) | |
| AUROC | 0.785 |
| After adjustment for six major confounders by propensity score matching | |
| AUROC | 0.789 |
Prominent predictors (SHAP)
- Higher R-wave amplitude in leads aVL and I, and smaller peak interval dispersion were prominent predictors
Original publication
Artificial intelligence identifies individuals with prediabetes using single-lead electrocardiograms. Cardiovasc Diabetol. Published 11 November 2025. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →