A Multimodal Predictive Model for Chronic Kidney Disease and Its Association With Vascular Complications in Patients With Type 2 Diabetes: Model Development and Validation Study in South Korea and the U.K.
Authors' conclusion
Our multimodal AI model integrates fundus images and clinical data from binational cohorts to predict the risk of new-onset CKD within 5 years and associated vascular complications in patients with T2DM.
Study context
- Population
- Patients with T2D; Chronic kidney disease (CKD) defined by ICD-9/ICD-10 codes and/or estimated glomerular filtration rate (eGFR) ≤60 mL/min/1.73 m²
- Participants
- 7,028 vs 1,544
- Comparison
- Discovery vs Validation
- Discoveryn = 7,028
- Validationn = 1,544
- Outcome
- Prediction of CKD (AUC)
Methodological detailsmultimodal, deep learning ensemble, external validation, binational cohorts
- Multimodal
- Deep learning ensemble
- External validation
- Binational cohorts
Performance
As reported in the abstract
AUC for predicting CKD
| Test or model | Value | AUC |
|---|---|---|
| Discovery | 0.880 (95% CI 0.806–0.954) | 0.880 |
| Validation | 0.722 | 0.722 |
Value
Discovery0.880 (95% CI 0.806–0.954)
Validation0.722
AUC
Discovery0.880
Validation0.722
Further findings
Macrovascular complications by model probability tertile
- T1: aHR
- 1.42 (1.06-1.90) (95% CI)
- T2: aHR
- 1.59 (1.17-2.16) (95% CI)
- T3: aHR
- 1.64 (1.20-2.26) (95% CI)
Microvascular complications by model probability tertile
- T3: aHR
- 1.30 (1.02-1.67) (95% CI)
Original publication
A Multimodal Predictive Model for Chronic Kidney Disease and Its Association With Vascular Complications in Patients With Type 2 Diabetes: Model Development and Validation Study in South Korea and the U.K. Diabetes Care. Published 1 July 2025. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →