Predicting the Extent of Damage in the Humphrey Field Analyzer 24-2 Visual Fields Using 10-2 Test Results in Patients...

Area of Science:

• Ophthalmology
• Medical Technology
• Data Science

Background:

• Glaucoma management relies on visual field testing to monitor disease progression.
• Humphrey Field Analyzer (HFA) 10-2 and 24-2 tests provide crucial visual field data.
• Predictive models can aid in managing cases where standard testing is limited.

Purpose of the Study:

• To develop and validate a method for predicting Humphrey Field Analyzer 24-2 test results using Humphrey Field Analyzer 10-2 test data.
• To assess the accuracy of these predictions in patients with advanced glaucoma.

Main Methods:

• Prospective follow-up of 175 advanced glaucoma eyes over five years using HFA 10-2 and 24-2 tests.
• Least Absolute Shrinkage and Selection Operator (LASSO) regression was employed to construct a predictive formula using paired HFA 10-2 and 24-2 results from a training dataset.
• Prediction error was evaluated by comparing actual and predicted values in a separate testing dataset.

Main Results:

• The prediction error for mean deviation (MD24-2) was 2.98 dB (SD = 1.90 dB), and for its slope was 0.32 dB/yr (SD = 0.33 dB/yr).
• Prediction errors for hemifield mean total deviation (hemi-mTD) were also within acceptable ranges (e.g., 3.02 dB for superior hemi-mTD).
• These errors were comparable to the repeatability of stable HFA 24-2 visual field parameters.

Conclusions:

• Humphrey Field Analyzer 24-2 test results can be predicted with reasonable accuracy from Humphrey Field Analyzer 10-2 test results using LASSO regression in advanced glaucoma.
• This predictive approach is valuable when the HFA 24-2 test is not readily available for patients with advanced glaucoma.