Repeatability and clinical utility of a novel method to measure in-vivo corneal nerve migration in diabetic neuropathy

Purpose We developed a novel technique to measure in-vivo corneal nerve migration, and determined its repeatability and clinical utility in diabetic neuropathy. Methods Wide-field montages of the sub-basal corneal nerve plexus were generated at baseline and after three weeks for 14 participants (control participants (n=4); diabetic participants without (n=5) and with (n=5) neuropathy). Montages were manually examined side by side to identify a referent landmark in the inferior whorl region and 20 additional nerve landmarks throughout each montage. An algorithm was developed to measure nerve migration by quantifying the movement of the nerve landmarks relative to the inferior whorl landmark over the three-week period, with migration reported in microns per week. Intra- and inter-observer repeatabilities were determined by measurement of nerve migration on the pair of montaged images by one researcher on two occasions, five days apart, and by two observers respectively. A paired t-test was used to compare the difference between the two measurements. Bland–Altman plots were generated to assess agreement and the intra-class correlation coefficient (ICC) was calculated to estimate repeatability. To illustrate it’s clinical utility, we compared the average nerve migration between the three groups. Results The mean difference between observations 1 and 2 was 0.92 µm/week (p=0.10), with an ICC of 0.99 (95% confidence interval (CI)=0.99-1.00). The mean difference between observer 1 and 2 was 3.05 µm/week (p=0.04), with an ICC of 0.97 (95% CI=0.91-0.99). Movement of individual points was dependent on their distance from the inferior whorl, so the method was altered to keep the distance of the measured points at an average of 1,200±50µm from the inferior whorl. Using this method, the mean difference between observers was 0.38 µm/week (p=0.38), with an ICC of 0.97 (95% CI=0.99-1.00). The average nerve migration rate was 32.7, 53.4 and 45.3 µm/week for the diabetic individuals with and without neuropathy and the control participants, respectively (p=0.18). Conclusions This novel imaging technique allows repeatable and non-invasive measurement of in-vivo corneal nerve migration. The results indicate that patients with diabetic neuropathy may have a reduced rate of nerve migration, which should be explored in future studies.