Variability of different reference bodies in normal, keratoconus, and collagen crosslinked corneas

To evaluate the reproducibility of different Scheimpflug imaging-derived reference bodies in normal, keratoconus (KC), and crosslinked (CXL) eyes.

In this prospective, observational study, 40 participants populated the control group (CG), while 33 and 34 patients formed the KC group and the CXL group, respectively. One eye was randomly selected when both were eligible. Elevation measurements were obtained using Scheimpflug camera, by applying the following reference bodies: the best fit sphere (BFS), the best fit toric ellipsoid (BFTE), the best fit toric ellipsoid with fixed eccentricity of 0.4 (BFTEF), the best fit ellipsoid (BFE), and the best fit torus, by 2 operators in 2 sessions. The variability of different reference bodies' parameters was assessed using intraclass correlation coefficient (ICC).

Differences between operators were nonsignificant for all reference bodies' parameters in all study groups (p>0.1). Regarding CG, BFS for both operators and BFTE for operator 1 presented the best intrasession repeatability (majority of ICCs >0.90), while average interobserver reliability was recorded for the majority of reference bodies' parameters. Regarding KC and CXL groups, BFS, BFTEF, and BFE reference bodies demonstrated the best intrasession and interobserver reproducibility (majority of ICCs >0.90) for both groups. Steep and flat radius parameters presented the best overall reproducibility, with the majority of ICCs for all reference bodies ranging above 0.90, in all study groups.

High variability was encountered for most reference surfaces. Ellipsotoric surfaces presented acceptable repeatability in KC and post-CXL corneas.