GlauCUTU-WebXR: Virtual Reality Visual Field Perimetry System Embedded with Enhanced Humphrey Field Analyzer...

This study presents GlauCUTU-WebXR, an advanced glaucoma screening system leveraging commercial virtual reality (VR) devices with WebXR for broader accessibility. Transitioning from custom 3D-printed hardware, GlauCUTU-WebXR integrates an improved prediction model that enhances sensitivity predictions using positional features (position index: PID) and severity classification metadata (SeverityScore). These additions help the model capture positional and severity-specific patterns, which, in turn, significantly improve prediction accuracy in converting GlauCUTU sensitivity to Humphrey Field Analyzer (HFA) sensitivity. The model trained and validated on 97 eyes (47 patients) and tested on 24 eyes (12 patients) with varying glaucoma severity, it achieved a mean absolute error (MAE) of 4.3398, root mean square error (RMSE) of 6.6614, and a Pearson correlation of 0.8033. The system significantly reduced testing time compared to HFA (p<0.001). GlauCUTU-WebXR shows no significant differences in results compared to HFA across severity levels: none (p=0.8750), mild (p=0.5888), moderate (p=0.9153), and severe (p=0.9958). By utilizing commercial VR hardware and enhanced prediction models, GlauCUTU-WebXR provides a cost-effective, efficient alternative for glaucoma screening, particularly in resource-limited settings, demonstrating its potential to improve accessibility and early detection efforts.