Objective:

To develop and externally validate an interpretable machine learning model for diabetic retinopathy risk stratification using routine clinical biomarkers and to explore interactions between these biomarkers and diabetic retinopathy pathogenesis.

Approach:

Key Findings:

• The LightGBM algorithm achieved an external validation AUC of 0.841 (95% CI: 0.809-0.862).
• Fourteen key predictors were identified, including markers of glycemic control, renal function, and lipid metabolism.
• The Bayesian Network revealed a hierarchical dependency structure.

Interpretation:

Limitations:

• The study may be limited by the generalizability of the findings to broader populations.
• External validation was conducted on a specific cohort.

Conclusion:

A high-performing, non-invasive LightGBM model for early diabetic retinopathy screening was developed.

Attribution Notice

This content is an AI-generated, fully rewritten summary based on a published scholarly article. It does not reproduce the original text and is not a substitute for the original publication. Readers are encouraged to consult the source for full context, data, and methodology.