Detecting Progression of Retinitis Pigmentosa Using the Binomial Pointwise Linear Regression Method.

Area of Science:

• Ophthalmology
• Medical Imaging
• Statistical Analysis

Background:

• Establishing a reliable method for evaluating central visual field (VF) progression in retinitis pigmentosa (RP) remains a challenge.
• Previous research suggested the utility of binomial pointwise linear regression (PLR) for glaucoma progression.
• This study aimed to assess the effectiveness of binomial PLR in eyes with RP.

Purpose of the Study:

• To investigate the usefulness of binomial PLR for evaluating central visual field progression in retinitis pigmentosa (RP).
• To compare the performance of binomial PLR with other methods in terms of accuracy and time to detection.

Main Methods:

• Retrospective analysis of 10 visual fields (VFs) from 196 eyes of 103 RP patients.
• Application of binomial PLR by regressing total deviation across all test points over 10 VFs.
• Comparison of binomial PLR with permutation analysis of PLR (PoPLR) and mean deviation (MD) trend analysis using accuracy metrics and time to progression detection.

Main Results:

• Binomial PLR demonstrated comparable positive predictive values (0.55-0.95), negative predictive values (0.67-0.92), and false-positive rates (0.01-0.05) to PoPLR and MD trend analyses.
• Binomial PLR achieved significantly earlier detection of VF progression (5.0 ± 2.0 years) compared to PoPLR and MD trend analyses (P < 0.01, P < 0.001).

Conclusions:

• Binomial PLR provides reliable and earlier detection of central VF progression in eyes with RP.
• The method is useful for assessing VF progression in RP, offering a potential advancement in patient management.