Prediction of diabetic retinopathy using machine learning and its association with dementia risk in older adults with type 2 diabetes mellitus
Authors' conclusion
ML-predicted DR was significantly associated with future dementia, highlighting its value in early risk stratification among patients with diabetes.
Study context
- Population
- Patients aged ≥50 years newly diagnosed with T2D without prior dementia or eye disease
- Participants
- 27,929 patients
- Comparison
- Machine Learning (ML)-predicted Diabetic Retinopathy (DR) vs non-DR
- Outcome
- DR occurrence (prediction) and dementia incidence
- Follow-up
- DR within one year; dementia over three years
Methodological detailsmachine learning, external validation across eight institutions, meta-analysis of pooled results
- Machine learning
- External validation across eight institutions
- Meta-analysis of pooled results
Performance
As reported in the abstract
DR prediction performance · within one year
XGBoost best-performing model, AUROC
0.746
External validation AUROCs
0.555 to 0.620
Further findings
| Outcome | HR (95% CI) | |
|---|---|---|
| All-cause dementia · three years | 1.32 (1.12–1.56) | |
| Alzheimer's disease · three years | 1.30 (1.07–1.58) | |
| Vascular dementia · three years | 1.38 (1.12–1.69) | |
| 0.512Hazard ratio (log scale) |
Original publication
Prediction of diabetic retinopathy using machine learning and its association with dementia risk in older adults with type 2 diabetes mellitus. Diabetes Res Clin Pract. Published 15 July 2025. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →