Measuring visual field progression in the central 10 degrees using additional information from central 24 degrees...

Area of Science:

• Ophthalmology
• Medical Imaging
• Data Science

Background:

• Glaucoma management relies on monitoring visual field (VF) progression.
• Accurate prediction of VF mean deviation (MD) is crucial for timely intervention.
• Current methods may have limitations in precisely tracking parafoveal VF changes.

Purpose of the Study:

• To enhance the accuracy of measuring longitudinal 10-2 VF mean deviation (MD) progression.
• To integrate information from 24-2 VFs into 10-2 VF MD progression analysis.
• To utilize Lasso regression for improved predictive modeling of VF changes.

Main Methods:

• A training dataset of 138 eyes and a testing dataset of 40 eyes with glaucoma or ocular hypertension were used.
• Lasso regression predicted 10-2 VF total deviation values using 24-2 VF data.
• The 10-2 VF 'Lasso MD' (LMD) was calculated and incorporated into linear regression models to predict future MDs.

Main Results:

• The novel method incorporating LMDs demonstrated significantly smaller average absolute prediction errors (1.6–1.8 dB) compared to the standard approach (1.7–3.4 dB).
• Statistical analysis (ANOVA test) confirmed the significance of the improved prediction accuracy (p<0.05).

Conclusions:

• Deriving 10-2 VF MD values from 24-2 VFs significantly improves the accuracy of predicting glaucomatous progression.
• This enhanced predictive approach aids clinicians in better forecasting visual function changes in the parafoveal region.
• The study highlights the potential of data-driven methods like Lasso regression in clinical ophthalmology.