Biomechanics-Function in Glaucoma: Improved Visual Field Predictions from IOP-Induced Neural Strains

Area of Science:

• Ophthalmology
• Biomedical Engineering
• Medical Imaging

Background:

• Glaucoma is a leading cause of irreversible blindness.
• Accurate prediction of functional loss is crucial for glaucoma management.
• Current prediction models often rely solely on structural data.

Purpose of the Study:

• To determine if neural tissue structure and biomechanics can predict functional loss in glaucoma.
• To evaluate the contribution of biomechanics to prediction accuracy.

Main Methods:

• A cross-sectional study involving 238 glaucoma patients (age >50).
• Optic nerve head (ONH) imaging using spectral-domain OCT under varying intraocular pressure (IOP).
• Deep learning (Point-Net) used to predict visual field defects from structural and biomechanical data (IOP-induced strains).

Main Results:

• The integrated model (structure + biomechanics) achieved a higher F1-score (0.76 ± 0.02) compared to the structure-only model (0.71 ± 0.02).
• The inclusion of IOP-induced neural tissue strains significantly improved predictive performance (p < 0.05).
• The study included a diverse range of glaucoma severity (Mean Deviation -1.8 to -25.2).

Conclusions:

• Biomechanical data, specifically IOP-induced neural tissue strains, significantly enhances the prediction of visual field loss in glaucoma.
• The biomechanics-function relationship is critical for accurate glaucoma progression modeling.
• This approach offers a more comprehensive understanding of glaucoma pathophysiology.