Likelihood ratios for glaucoma diagnosis using spectral-domain optical coherence tomography

Area of Science:

• Ophthalmology
• Medical Diagnostics
• Biomedical Engineering

Background:

• Glaucoma diagnosis relies on various tests, but integrating continuous measurements like retinal nerve fiber layer (RNFL) thickness can be challenging.
• Spectral-domain optical coherence tomography (spectral-domain OCT) provides precise RNFL measurements, offering potential for improved diagnostic accuracy.

Purpose of the Study:

• To develop and present a methodology for calculating likelihood ratios for glaucoma diagnosis using continuous RNFL thickness data.
• To establish a framework for interpreting spectral-domain OCT RNFL measurements in glaucoma assessment.

Main Methods:

• An observational cohort study included 262 eyes with glaucoma and 190 healthy control eyes.
• Retinal nerve fiber layer (RNFL) thickness was measured using Spectralis spectral-domain OCT.
• Likelihood ratios were estimated using a method based on receiver operating characteristic (ROC) curve tangents for continuous RNFL values.

Main Results:

• A methodology was established to calculate likelihood ratios for continuous average RNFL thickness measurements.
• Average RNFL thickness values below 86 μm were associated with positive likelihood ratios (indicating increased glaucoma probability).
• Values above 86 μm yielded negative likelihood ratios (indicating decreased glaucoma probability), with a modified Fagan nomogram provided for clinical use.

Conclusions:

• The developed methodology enables the calculation of likelihood ratios for specific RNFL thickness values.
• This approach avoids arbitrary test result categorization, potentially enhancing the integration of diagnostic data into clinical decision-making for glaucoma.