A new digital optic disc stereo camera: intraobserver and interobserver repeatability of optic disc measurements

Abstract

Aims: To investigate the intraobserver and interobserver repeatability of optic disc measurement using a new digital optic disc stereo camera.

Methods: 112 consecutive new patients presenting to a glaucoma service had dilated optic disc photography performed using a new digital stereo camera (Discam, Marcher Enterprises Ltd, Hereford). The images were analysed by two masked observers using a stereo viewer and computer simulated stereopsis. Vertical and horizontal cup:disc ratios (CDR), cup area:disc area, and cup circumference:disc circumference were computed. Intraobserver and interobserver repeatability analyses were performed. Intraclass correlation coefficients (ICC) and 95% tolerance for change (TC) were computed.

Results: 220 optic discs were photographed, of which 196 were suitable for analysis (10 were of poor image quality and 14 had anomalous discs). Mean age of patients was 65 years, 60 were male and 48 female. For intraobserver measurements of: horizontal CDR, ICC = 0.94, TC = 0.11 (15% of range); vertical CDR, ICC = 0.92, TC = 0.14 (16% of range); cup area:disc area, ICC = 0.95, TC = 0.10 (13% of range), and cup circumference:disc circumference, ICC = 0.95, TC = 0.09 (14% of range). For interobserver measurements of: horizontal CDR, ICC = 0.89, TC = 0.14 (19% of range); vertical CDR, ICC = 0.90, TC = 0.14 (16% of range); cup area:disc area, ICC = 0.92, TC = 0.13 (16% of range), and cup circumference:disc circumference, ICC = 0.90, TC = 0.12 (17% of range). Systematic bias between observers was within acceptable limits.

Conclusions: Digital stereo disc photography and analysis provide repeatable measures of optic disc variables. The results compare favourably with ophthalmoscopic and stereophotographic methods of assessment of the optic disc.

Figure 1

Two examples of the stereo images obtained.

Figure 1

Two examples of the stereo images obtained.

Figure 2

Scatter plots of intraobserver and interobserver vertical CDR measurement difference plotted against measurement average demonstrating that the magnitude of any difference is inversely related to the size of the measurements. (A) Intraobserver scatter plot of difference in vertical CD ratios plotted against the average CD ratio. (B) Interobserver scatter plot of difference in vertical CD ratios plotted against the average CD ratio. (C) Intraobserver scatter plot of difference in horizontal CD ratios plotted against the average CD ratio. (D) Interobserver scatter plot of difference in horizontal CD ratios plotted against the average CD ratio.