SMOTE-Enhanced Explainable Artificial Intelligence Model for Predicting Visual Field Progression in Myopic Normal...

Area of Science:

• Ophthalmology
• Artificial Intelligence
• Medical Diagnostics

Background:

• Myopic normal-tension glaucoma (NTG) presents challenges in predicting visual field progression.
• Accurate prediction is crucial for timely intervention and management.

Purpose of the Study:

• To develop and validate a Synthetic Minority Over-sampling Technique (SMOTE)-enhanced artificial intelligence (AI) model for predicting visual field progression in myopic NTG patients.
• To identify key predictors of disease progression using SHapley Additive exPlanations (SHAP) analysis.

Main Methods:

• A retrospective cohort study of 100 eyes from myopic NTG patients with a mean follow-up of 10.3 years.
• A SMOTE-enhanced AI model was developed to address class imbalance.
• Model performance was assessed using ROC analysis, cross-validation, and calibration plots.
• SHAP analysis was employed to determine the importance of predictive factors.

Main Results:

• Visual field progression was observed in 28% of patients.
• The AI model achieved an AUC of 0.83, with sensitivity of 0.81 and specificity of 0.77.
• Key predictors identified included baseline mean deviation (MD), age, axial length, baseline intraocular pressure (IOP), and visual field index (VFI).
• Higher predicted risk scores correlated significantly with increased observed progression rates.

Conclusions:

• The SMOTE-enhanced AI model demonstrates promising predictive performance for visual field progression in myopic NTG.
• The model has potential clinical utility for personalized risk stratification and early intervention.
• Further validation in larger cohorts is recommended.