Optic Nerve Cup-To-Disc Ratio Measurement Variability With Semi-Automated Digital Analysis.

Purpose: To evaluate the reliability of semi-automated digital analysis in
assessing optic nerve cup-to-disc ratio. Subjects: Four subjects (four eyes) were evaluated. All subjects had healthy, non-glaucomatous optic nerve heads.
Methods: The subjects' optic discs were captured, stored, and analyzed by two
examiners using the Helioasis Oigivid 2000 system. Phase 1- Intra-observer intra-image variability: One image from each disc was analyzed three times consecutively by one observer in order to determine the variability of multiple measures of the same image.
Phase 2- Intra-observer inter-image variability: Three images of each disc were analyzed once by one observer in order to assess the repeatability of measurements from different images of the same disc.
Phase 3- Inter-observer intra-image variability: Three images were taken of each disc and each image was analyzed one time by each examiner in order to assess the resultant variability of two different examiners evaluating each disc.
Results: Phase 1- Intra-observer intra-image variability: For vertical cup-to-disc (C/O)
measurement the standard deviation (SO) average was 0.0177. The coefficient of variation (COV) average was 4.07%. For the horizontal C/0 measurement, the SO average was 0.0263 and the average COV was 6.28%.
Phase 2:- Intra-observer inter-image variability: For vertical cup-to-disc measurement the average SO was 0.0222 and the average COV was 6.68%. For the horizontal measurement, the average SO was 0.0383 and the average COV was 8.65%.
Phase 3:- Inter-observer intra-image variability: For the vertical C/0 measurement the· average SO was 0.0376. The average COV was 14.68%. For the horizontal measurement, the average SO was 0.0817 and the average COV was 26.08%. After rounding the data to 0.05, the Paired Student's t-Test at 95% confidence showed that the vertical
measurements were repeatable between examiners 25.22% of the time and 29.52% of the time for the horizontal measurements. However, 63% of the measurements were within 0.05 of each other and 88% were within 0.15 of each other, which showed much better
repeatability.
Conclusion: Cup-to-disc ratio measurement with the Helioasis Oigivid 2000 system was found to be highly reproducible and would serve as a valid and reliable way of assessing current status and progression of optic nerve cup-todisc ratio over time.