Repeatability of measuring corneal nerve migration rate in individuals with and without diabetes

Purpose: To assess the repeatability of measuring the corneal nerve migration rate in individuals with and without neuropathy. Methods: Wide-field montages of the subbasal corneal nerve plexus were generated at baseline and after 3 weeks for 14 participants. Montages were manually examined side by side to identify a referent landmark in the inferior whorl region (origin) and throughout each montage. A software program was developed to measure nerve migration of all identified points relative to the origin. Repeatability was determined by measurement of nerve migration for within observer (one researcher on 2 occasions, 5 days apart) and between observers (2 observers) within 4 different zones based on the distance from the origin and in the vertical section of the widefield montage. The impact of images being montaged with fully automated software on repeatability was also investigated. Results: The mean difference between observations 1 and 2 for observer 1 was 0.02 ± 1.3 μm/wk (P = 0.94), with an intraclass correlation coefficient (ICC) of 0.99 [95% confidence interval (CI) = 0.99-1.00], and the mean difference between observer 1 and 2 was 0.3 ± 1.2 μm/wk (P = 0.41), with an ICC of 0.99 (95% CI = 0.99- 1.00). The mean difference between observations 1 (images montaged by semiautomated software) and 2 (images montaged by fully automated software) was 1.2 ± 4.9 μm/wk (P = 0.41), with an ICC of 0.96 (95% CI = 0.87-1.00). Conclusions: Measuring corneal nerve migration rate is highly repeatable for within and between observers and when using different methods of image montaging.