Visual field-based machine learning model for predicting disease stage in glaucoma

Glaucoma is a leading cause of irreversible blindness, requiring early detection to prevent vision loss. This study evaluates non-linear machine learning (ML) models, such as AdaBoost, for forecasting glaucoma stage, comparing it to Logistic Regression (LR) using visual field data from three independent cohorts (Bern, Rotterdam, Washington). AdaBoost outperformed LR, achieving a kappa value of 0.85 (Bern) compared to 0.73 for LR. Notably, AdaBoost reached an AUC ROC of 0.93 (early glaucoma, Bern), surpassing LR’s 0.47. Similar trends were observed across other datasets. These results demonstrate the ability of non-linear ML models to detect glaucoma progression, highlighting their potential to improve early intervention and treatment planning.