Accuracy and Repeatability of an Anterior Segment Swept-Source Optical Coherence Tomographer.

Abstract

To assess accuracy and repeatability of the CASIA swept-source optical coherence tomographer (SS-OCT) in measuring contact lens (CL) radii of curvature and thicknesses compared with verified CL parameters and to investigate intrasession and intersession repeatability of the CASIA SS-OCT in measuring epithelial and total corneal thicknesses. Rigid CLs with varying radii of curvature (front, rf; back, rb) and thicknesses were imaged with the CASIA SS-OCT across two sessions. Contact lens parameters were measured from processed images using an automated MATLAB program and were compared with parameters verified using standard techniques. Central epithelial and total corneal thicknesses of 43 normal subjects were measured to assess intrasession and intersession repeatability of the CASIA SS-OCT. No significant differences (P>0.05) were found in rigid CL rf and central and peripheral thickness measurements compared with verified CL parameters. However, the rb values obtained from the CASIA SS-OCT were consistently 0.2 mm flatter than the verified rb values (P<0.001) along horizontal and vertical meridians. Bland-Altman analysis revealed excellent intrasession agreement with mean differences of 0.004 and 0.003 mm for rf and rb, 0.000 mm for CL thickness and 0.372 and 0.395 μm for epithelial and total corneal thicknesses, respectively. Similarly, mean intersession differences of 0.020 and 0.000 mm for rf and CL thickness and 0.100 and 0.984 μm for epithelial and total corneal thicknesses were found, respectively. Ninety-five percentage confidence intervals across one or two sessions indicated insignificant undermeasurement or overmeasurement for CL rf, and corneal thicknesses, but significant bias toward overmeasurement for CL rb was found across two sessions. The CASIA SS-OCT produces accurate measurements of CL parameters compared with verified values. Inbuilt distortion correction in this instrument necessitated no further correction of scanned images and provided high intrasession and intersession repeatability in measuring both CLs and corneal thicknesses. Further investigation of discrepancies in rb measurements is warranted.