Prediction of perimetric progression in ocular hypertension and open angle glaucoma based on corneal biomechanics

Abstract

Purpose

To identify parameters that are significant risk predictors of visual field (VF) progression in patients with ocular hypertension (OHT) or early primary open-angle glaucoma (POAG), using Goldmann applanation tonometry intraocular pressure (IOP-GAT), ultrasound pachymetry and biomechanical indices measured with the Corvis ST and the Ocular Response Analyser (ORA) and create a prediction model.

Methods

A dataset of 65 eyes was analyzed. VF progression was determined using event analysis, the AGIS progression criteria, and experts’ opinion. Progression was defined when 2 of the 3 criteria agreed on VF progression. The artificial intelligence training pipeline to predict the consolidated target consisted of an initial outlier removal process, feature selection, and training using the leave-one-out (LOO) cross-validation. Techniques based on decision trees with the XGBoost (Xtreme Gradient Boosting) algorithm were employed.

Results

The results reveal that the most accurate predictors were those registered at 1-month follow-up, predicting glaucomatous VF progression with an accuracy of 86.2%. The main variables involved in the prediction were HC dArc length, HC Deflection Time, and the HC Deflection Length 1 month after PG therapy.

Conclusion

Using AI models, glaucomatous VF progression can be predicted with relatively high accuracy using Corvis ST parameters registered one month after initiating PG treatment. Highest concavity parameters after 1 month of treatment are associated with an increased risk of perimetric progression.