JoVE - Journal of Visualized Experiments
JoVE Visualize
Contact Us

Comparison of Machine Learning Algorithms and Nomogram Construction for Diabetic Retinopathy Prediction in Type 2 Diabetes Mellitus Patients

  • 0South Campus Outpatient Clinic, Sun Yat-sen Memorial Hospital, Sun Yat-sen University, Guangzhou, China.
Ophthalmic Research +

|

September 4, 2024

|

September 4, 2024

View abstract on PubMed

Summary

This summary is machine-generated.

Logistic regression best predicts diabetic retinopathy (DR) in type 2 diabetes mellitus (DM) patients. A developed nomogram aids in early DR detection and timely treatment for better patient outcomes.

Area Of Science

  • Medical Informatics
  • Ophthalmology
  • Machine Learning

Background

  • Diabetic retinopathy (DR) is a significant complication of type 2 diabetes mellitus (DM).
  • Accurate prediction models are crucial for early detection and intervention in DM patients undergoing DR screening.

Purpose Of The Study

  • To compare machine learning algorithms for DR prediction in type 2 DM patients.
  • To develop a predictive nomogram based on the optimal algorithm.

Main Methods

  • A cross-sectional study involving type 2 DM patients undergoing DR screening.
  • Comparison of six machine learning algorithms: Logistic Regression (LM), Decision Tree (DT), K-Nearest Neighbors (KNN), Random Forest (RF), Support Vector Machine (SVM), and XGBoost (XGB).
  • Model performance evaluated using ROC, calibration, and Decision Curve Analyses (DCAs).

Main Results

  • Logistic Regression (LM) showed the highest predictive performance (AUC 0.894, recall 0.92) in the validation set.
  • Key predictors identified: disease duration, diabetic polyneuropathy (DPN), insulin dosage, urinary protein, and albumin (ALB).
  • The developed nomogram demonstrated robust discrimination (AUC 0.856-0.868) and clinical applicability.

Conclusions

  • Logistic Regression (LM) outperformed other machine learning algorithms for DR prediction in this cohort.
  • A validated logistic regression-based nomogram is established for predicting DR in type 2 DM patients.
  • The nomogram can serve as a valuable tool for early DR detection and timely management.

Keywords: