A machine learning model for diabetic retinopathy risk stratification using routine blood and urine parameters: insights into kidney-eye crosstalk - Takeaways - MDSpire
Coming Soon: Introducing MDSpire News. Learn more
Conexiant’s news site is now MDSpire News. Learn more

Development and External Validation of a Machine Learning Model for Assessing Diabetic Retinopathy Risk Utilizing Standard Blood and Urine Biomarkers: Exploring Kidney-Eye Interactions

  • By

  • Yong Wang

  • Ling Yao

  • Tianpeng Chen

  • Qianyu Zhang

  • Xin Cao

  • Jiayi Gu

  • Sijie Bao

  • Xiaojuan Chen

  • Cheng Cao

  • June 17, 2026

Share

  • 1

    A machine learning model was developed for diabetic retinopathy risk assessment using standard blood and urine biomarkers.

  • 2

    The LightGBM algorithm achieved an external validation AUC of 0.841, outperforming other classifiers in the study.

  • 3

    Fourteen key predictors related to glycemic control, renal function, and lipid metabolism were identified as significant for diabetic retinopathy.

  • 4

    A Bayesian Network revealed that renal impairment markers and chronic glycemic toxicity are direct drivers of diabetic retinopathy.

  • 5

    The model provides a web-based clinical decision support system for early diabetic retinopathy screening in resource-constrained settings.

Original Source(s)

A machine learning model for diabetic retinopathy risk stratification using routine blood and urine parameters: insights into kidney-eye crosstalk

  • by Yong Wang, Ling Yao, Tianpeng Chen, Qianyu Zhang, Xin Cao, Jiayi Gu, Sijie Bao, Xiaojuan Chen, Cheng Cao

  • June 17, 2026

Related Content

Diagnosis Gaps Mark Retinopathy Care

A US systematic review found substantial losses across diabetic retinopathy screening, diagnosis awareness, referral, and follow-up.

Diabetes Mellitus Coding in Retina

Age-related macular degeneration (AMD) is among the most common retinal diagnoses. Although the clinical features of the disease are well known, ICD-10 coding can be confusing if details such as stage, activity, and laterality are not clearly documented. Incomplete documentation may prompt questions about the medical necessity of diagnostic tests or intravitreal injections. This Q&A addresses common sources of confusion and offers practical guidance to support clear, accurate documentation.

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