Spatial-Aware Transformer-GRU Framework for Enhanced Glaucoma Diagnosis From 3D OCT Imaging

Area of Science:

• Ophthalmology
• Medical Imaging
• Artificial Intelligence

Background:

• Glaucoma is a primary cause of irreversible blindness worldwide.
• Early detection and intervention are critical to prevent vision loss.
• Accurate diagnostic tools are essential for effective glaucoma management.

Purpose of the Study:

• To develop and evaluate a novel deep learning framework for automated glaucoma detection.
• To leverage 3D Optical Coherence Tomography (OCT) imaging for enhanced diagnostic accuracy.
• To improve clinical decision support systems for glaucoma management.

Main Methods:

• Integration of a pre-trained Vision Transformer for slice-wise feature extraction from retinal data.
• Utilized a bidirectional Gated Recurrent Unit to capture inter-slice spatial dependencies in 3D OCT scans.
• Developed a dual-component deep learning approach for comprehensive structural analysis.

Main Results:

• The proposed framework achieved high performance on a large dataset.
• Achieved an F1-score of 93.01%, Matthews Correlation Coefficient (MCC) of 69.33%, and AUC of 94.20%.
• Demonstrated superior performance compared to existing state-of-the-art methods.

Conclusions:

• The novel deep learning framework effectively utilizes 3D OCT data for automated glaucoma detection.
• The approach offers significant potential for improving patient outcomes in glaucoma care.
• The framework can enhance clinical decision support systems for ophthalmologists.