[Suitability of various topographic corneal parameters for diagnosis of early keratoconus].

BACKGROUND: For the detection of early stages of keratoconus (KC) several metrics have been developed in recent years. The aim of the study was to assess the suitability of keratometric indices to discriminate between eyes with early and subclinical KC from normal eyes.
METHODS: From 33 eyes with early KC (group 1), 16 eyes with subclinical KC (group 2) und 121 normal eyes (group 3) the following metrics were computed from axial keratometric data: central keratometry (cK), astigmatism (AST), paracentral inferior-superior keratometric difference (PISD), skew of the steepest axes index (SRAX), the KISA% index, a discriminant function from the KISA% parameters AST, cK, PISD and SRAX (DKISA), corneal C(3) (-1) and a discriminant function from corneal Zernike coefficients (1(st)-7(th) order, pupil diameter 6 mm). The discriminative ability of these metrics between KC and normal eyes was assessed using receiver operating characteristic (ROC) curves and measuring the area under the curve (A (z)ROC).
RESULTS: Applying the published criteria, the Rabinowitz-McDonnell test (cK and PISD) and KISA% lacked sensitivity. Adjustment of critical values using ROC curve analysis improved the discriminative ability. The PISD (A (z)ROC 1 versus 3: 1, 2 versus 3: 0.947) and C(3) (-1) (A (z)ROC 1 and 0.98, respectively) metrics were the two single parameters with the highest discriminative ability. By weighting KISA parameters and Zernike coefficients with discriminant analyses, 100% of group 1 eyes (DKISA) and 96.7% of group 2 eyes (DA) were correctly classified.
CONCLUSION: After lowering the critical values the keratometric indices yielded a high discriminative ability for the detection of early KC stages. However, the excellent classification rates for wavefront-based metrics were not achieved.