Comparison of Automated and Manual Measurements of Corneal Epithelial Thickness in Eyes with Limbal Stem Cell Deficiency

PURPOSE:The purpose of this study was to compare the outcomes of manual and automated corneal epithelial thickness (CET) assessments using anterior segment optical coherence tomography in eyes with limbal stem cell deficiency (LSCD). METHODS:Eighty-seven eyes from 64 patients with LSCD and 65 eyes from 39 healthy subjects were included in this cross-sectional study. OCT images of corneas were acquired using a Fourier-domain anterior segment optical coherence tomography. The automated measurement of CET was obtained using the manufacturer's program and the manual measurements were obtained by 2 masked observers. The average CET was obtained from 3 consecutive epithelial profile maps. The average CET of manual measurement was obtained from 3 consecutive cross-line scans using 3-point measurement as previously described. RESULTS:The central CET was thinner in patients with LSCD compared with healthy individuals by both manual and automated measurements. Automated CET measurements were similar to the manually obtained CET in the healthy controls (r = 0.911). However, there is a lower correlation between the automated and manually obtained CET in the LSCD group (r = 0.497). Manual measurements of CET (39.8 ± 11 μm) in eyes with LSCD were significantly lower than the automated measurements (48.8 ± 7.2 μm, P = 0.001, 95% CI -11.0 to -6.9). The correlations between the automated and manually obtained measurements in the mild, moderate, and severe stages of LSCD were 0.623, 0.632, and 0.378, respectively. CONCLUSIONS:There is a large discrepancy between the manual and automated measurements in eyes with LSCD but not in normal eyes. Automated measurements in LSCD exhibited a higher rate of errors and nonuniform measurements due to the presence of subepithelial scars. Automated CET measurements appear to be unreliable in eyes with corneal scars in LSCD.