Assessment of optic disc cupping with digital fundus photographs

Abstract

PURPOSE: To determine the agreement between the assessment of retinal digital images by using an overlay transparency sheet and the Heidelberg retinal tomograph (HRT) in determining cup-disk ratios greater than 0.6. DESIGN: Diagnostic test comparison. METHODS: Computerized topographic and monoscopic digital images of the optic disk of 628 people aged 70 to 79 years were assessed. A grader (M.C.) defined the disk margin on HRT images, and the operation software computed the area cup-disk ratio. The same grader also determined whether the vertical cup-disk ratio on retinal images was greater than 0.6 by superimposing a transparency overlay sheet over the images. Findings of a second grader (J.G.F.) were used to establish reliability measures. RESULTS:The intragrader reliability for the overlay method and HRT was almost perfect (intraclass correlation coefficient [ICC] = 0.96 and 0.99, respectively), whereas the intergrader reliability was good (ICC = 0.77 and 0.92, respectively). A perfect agreement was found on 28 (85%) of 33 eyes between the overlay and HRT methods in determining cup-disk ratios greater than 0.6. CONCLUSIONS: The overlay transparency method appears to be a reliable and promising alternative in determining cup-disk ratios greater than 0.6 in a community screening setting.