Comparison of the Variational Bayes Linear Regression Visual Field Test Algorithms and the Swedish Interactive...

Area of Science:

• Ophthalmology
• Medical Technology

Background:

• Glaucoma diagnosis relies on accurate visual field (VF) testing.
• Swedish Interactive Threshold Algorithm (SITA) Standard is a common VF testing method.
• Variational Bayes linear regression (VBLR) algorithms offer potential improvements in VF testing.

Purpose of the Study:

• To compare 10-2 visual field (VF) test results between SITA Standard and VBLR-VF algorithms in glaucoma patients.
• To evaluate the performance of VBLR-VF, VBLR-VF Fast, and VBLR-VF Fast+ against SITA Standard.

Main Methods:

• A multicenter prospective observational cohort study involving 133 glaucoma patients.
• Comparison of SITA Standard with three VBLR-VF algorithms (VBLR-VF, VBLR-VF Fast, VBLR-VF Fast+) using the 10-2 test program.
• Evaluation of mean deviation (MD), pattern standard deviation (PSD), pointwise VF sensitivity, repeatability (RMSE), and test duration.

Main Results:

• VBLR-VF algorithms significantly reduced test duration (12.2%–46.3%) compared to SITA Standard.
• No significant difference in test-retest repeatability (RMSE) was observed.
• Minor, sometimes significant, differences in VF sensitivity were noted between VBLR-VF algorithms and SITA Standard.

Conclusions:

• VBLR-VF algorithms substantially decrease testing time while maintaining VF test repeatability.
• These algorithms present a viable alternative for glaucoma visual field testing, with minor variations in sensitivity.