Course of Glaucomatous Visual Field Loss Across the Entire Perimetric Range

Area of Science:

• Ophthalmology
• Computational Biology

Background:

• Glaucomatous visual field (VF) loss progression is crucial for patient management.
• Understanding the full spectrum of VF decay, from normal to blindness, is essential for accurate prognostication.

Purpose of the Study:

• To model the complete process of glaucomatous visual field decay.
• To assess VF loss trajectory across the entire perimetric range, from normal to perimetric blindness.

Main Methods:

• Retrospective analysis of Advanced Glaucoma Intervention Study and UCLA data.
• Application of linear, exponential, and sigmoid regression models to threshold sensitivity data.
• Evaluation of model fit using Root Mean Squared Error (RMSE) for VF locations with significant sensitivity change.

Main Results:

• Sigmoid regression demonstrated the best fit for 80.3% of analyzed VF locations.
• The sigmoid model yielded the lowest mean RMSE (4.1 ± 1.9), indicating superior accuracy.
• Linear and exponential models showed higher RMSE values, suggesting a poorer fit to the observed data.

Conclusions:

• Pointwise sigmoid regression accurately models glaucomatous VF decay across the full range of vision loss.
• Glaucomatous VF loss exhibits nonlinear progression, with deterioration rates potentially varying throughout the disease course.
• Findings support a nonlinear model for predicting glaucoma progression to perimetric blindness.