Abstract
Purpose :
The measurement of various corneal parameters using anterior segment optical coherence tomography (AS-OCT) has been shown to reliably detect early keratoconus. The purpose of this study is to compare the reliability and reproducibility of the semi-automated EyeMark Python program to manual ImageJ software measurements of anterior segment structures in healthy compared to keratoconus eyes.
Methods :
AS-OCT imaging of 25 eyes from 14 healthy patients and 25 eyes from 15 patients with keratoconus between the ages of 20 and 80 were attained. Pregnant, nursing, and patients with any other anterior segment pathology were excluded from this study. Selection of the eye was randomized. A 15 line AS-OCT scan raster was performed and one image that included the fixation light beam encompassing visual axis was selected. Two trained observers used ImageJ software and EyeMark Python program to measure anterior corneal radius of curvature (ACRC), posterior corneal radius of curvature (PCRC), central corneal thickness (CCT), and anterior chamber depth (ACD). MedCalc Statistical Software was used to calculate the intraclass correlation coefficient (ICC) and generate Bland-Altman plots (BAPs).
Results :
The values of ICC were > 0.7 for ACRC, PCRC, CCT, and ACD for ImageJ versus Python measurements. The average ICC for comparing control group ImageJ and Python measurement of CCT was 0.952 (0.892-0.979, 95% CI), ACRC was 0.998 (0.996-0.999, 95% CI), PCRC was 0.997 (0.994-0.999, 95% CI), and ACD was 0.956 (0.900-0.981, 95% CI). The average ICC values for keratoconus group ImageJ versus Python measurement of CCT was 0.998 (0.996-0.999, 95% CI), ACRC was 0.999 (0.999-1.0, 95% CI), PCRC was 0.999 (0.997-0.999, 95% CI), and ACD was 0.999 (0.997-0.999, 95% CI). The BAPs comparing the ImageJ and Python measurements of anterior segment structures showed no systematic proportional bias and the average differences were near zero and within 95% of the limits of agreement.
Conclusions :
Overall, semi-automated Python measurements of the corneal parameters are in high agreement with manual ImageJ measurements and are reproducible. This study suggests that EyeMark Python may be a reliable tool for measuring anterior segment structures in keratoconus and optimizes AS-OCT corneal imaging protocol development, an area of active keratoconus research.
This is a 2021 Imaging in the Eye Conference abstract.