Prediction and Detection of Glaucomatous Visual Field Progression Using Deep Learning on Macular Optical Coherence...

Area of Science:

• Ophthalmology
• Artificial Intelligence
• Medical Imaging

Background:

• Glaucoma is a leading cause of irreversible blindness.
• Early detection of glaucoma progression is crucial for timely intervention.
• Macular optical coherence tomography (OCT) provides detailed retinal imaging.

Purpose of the Study:

• To develop and validate a deep learning model for detecting and predicting glaucoma progression using macular OCT.
• To assess the model's diagnostic performance in identifying concurrent and future visual field progression.

Main Methods:

• A novel deep learning method, based on a vision transformer (ViT) model, was developed.
• The model was pretrained on a large dataset of unlabeled OCT images.
• Glaucoma progression was defined by specific rates of change in mean deviation (MD) from Humphrey visual field tests.

Main Results:

• The ViT model achieved high accuracy in detecting concurrent glaucoma progression (AUC 0.90) and rapid progression (AUC 0.92).
• The model demonstrated strong predictive ability for future glaucoma progression (AUC 0.85) and rapid progression (AUC 0.84).
• The model successfully distinguished between stable and progressing eyes.

Conclusions:

• Deep learning models trained on macular OCT imaging can effectively detect and predict glaucoma progression.
• This technology offers potential for earlier identification of patients at risk of progression.
• Early identification can significantly aid clinical decision-making and patient management.