A clinically interpretable machine learning model for early detection of diabetic retinopathy in multiple community health centers - Takeaways - MDSpire
Coming Soon: Introducing MDSpire News. Learn more
Conexiant’s news site is now MDSpire News. Learn more

A machine learning approach for the early identification of diabetic retinopathy in various community health settings

  • By

  • Juncheng Tong

  • Aifa Tang

  • Lifang Liu

  • Luyuan Zhang

  • Hainan Wang

  • Mengyuan Qu

  • Bing Liu

  • May 4, 2026

Share

  • 1

    A predictive model was developed to identify diabetic individuals at risk of diabetic retinopathy using routine health data from community health facilities.

  • 2

    The study analyzed 1,475 diabetic individuals, dividing them into a development group of 1,177 and a validation group of 298.

  • 3

    GLMNET achieved an AUROC of 0.770 and an AUPRC of 0.452 in the validation cohort, indicating moderate discrimination capabilities.

  • 4

    Urine glucose was identified as the most significant predictor of diabetic retinopathy through SHAP analysis.

  • 5

    The model shows potential for risk stratification in primary care but requires external validation before clinical implementation.

Original Source(s)

A clinically interpretable machine learning model for early detection of diabetic retinopathy in multiple community health centers

  • by Juncheng Tong, Aifa Tang, Lifang Liu, Luyuan Zhang, Hainan Wang, Mengyuan Qu, Bing Liu

  • May 4, 2026

Related Content

Marked and reversible circulating insulin-like growth factor-1 elevation during teprotumumab N01 treatment for thyroid eye disease with limited correspondence to glycemic changes

Early Proteomic and Metabolomic Signatures in Diabetes Associated with Progression to Diabetic Retinopathy Over 1–2 Years

Phenotype-driven management of diabetic macular edema: multimodal imaging biomarkers and individualized therapy