Diagnostic Accuracy of the Optical Coherence Tomography in Assessing Glaucoma Among Filipinos. Part 2: Optic Nerve Head and Retinal Nerve Fiber Layer Parameters
Keywords:
Optical coherence tomography, Standard automatic perimetry, Glaucoma, Optic nerve head, Retinal nerve fiber layerAbstract
Objective: To determine the accuracy of the optic nerve head (ONH) and retinal nerve fiber layer (RNFL) parameters using the Stratus OCT in the diagnosis of glaucoma and to determine the validity of these measurements.
Methods: Glaucoma suspects undergoing glaucoma diagnostic tests were recruited consecutively. The numerical results of the Stratus OCT fast optic disc and fast RNFL protocols were analyzed against an independent assessment by glaucoma experts who were blinded as to the results of the OCT. An ROC curve analysis was applied to derive estimates of diagnostic accuracy and multi-level likelihood ratios.
Results: A total of 119 subjects assessed as glaucoma and 397 subjects assessed as normal were included. All ONH and RNFL OCT parameters showed statistically significant differences in mean values between the 2 groups. The ROC curve identified the vertical integrated rim area (AUC: 0.822), the cup-disc area ratio (AUC: 0.816), and the horizontal integrated rim width (AUC: 0.794) as the best optic disc parameters; and the RNFL average thickness (AUC: 0.827), the superior quadrant (AUC: 0.807), and the inferior quadrant (AUC: 0.804) as the best RNFL parameters. Multi-level likelihood ratios for ONH and RNFL parameters were calibrated using a projected posttest probability of 70% for a positive test result (therapeutic threshold) and a 10% posttest probability for a negative result (diagnostic threshold).
Conclusion: The results showed that statistically significant mean differences in ONH and RNFL parameters did not translate into a high predictive ability for each individual parameter. Single cut-off value for each OCT parameter based on the best sensitivity and specificity combination did not result in high predictive values for any single parameter. Multi-level likelihood ratios for the best ONH and RNFL parameters were derived to increase the diagnostic capability of the Stratus OCT.
Downloads
Published
Issue
Section
License
Copyright (c) 2012 Philippine Journal of Ophthalmology

This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.

