Defining parameters for evaluation of keratoconus progression

Session Details

Session Title: Corneal Biomechanics

Session Date/Time: Monday 09/10/2017 | 08:00-10:30

Paper Time: 09:33

Venue: Room 2.1

First Author: : M.Shajari GERMANY

Co Author(s): :    G. Steinwender   K. Herrmann   K. Kubiak   I. Pavlovic   I. Schmack   T. Kohnen     

Abstract Details

Purpose:

Defining variables for evaluation of keratoconus progression and determining their cut-off values

Setting:

Department of Ophthalmology, Goethe-University, Frankfurt, Germany

Methods:

265 eyes of 165 patients who have been diagnosed with keratoconus underwent two Scheimpflug measurements (Pentacam) that took place one-year apart ± three months. Variables which are used for keratoconus detection were evaluated for their suitability to reflect keratoconus progression. Correlation analysis was performed between variables which are used for detection of keratoconus. By logistic regression analysis a keratoconus progression index (KPI) was defined. ROC analysis was performed and Youden-Index calculated to determine the cut-off value.

Results:

Variables used for keratoconus detection showed a weak correlation with each other (e.g.r=0.245 between RPIMin and KMax, P<0.001). For further analysis therefore parameters were used which take into consideration several variables (e.g. D-index, ISV, IHA, KPI). KPI was defined by logistic regression and consists of minimum pachymetry (P=0.001), maximum elevation of back surface and posterior curvature (both P <0.001). In the ROC analysis the two variables with the highest Youden-index were D-index and KPI : D-index had at a cut-off of 0.42(70.6%sensitivity) a Youden-index of 0.606. Cut-off for KPI was -0.78(84.7%sensitivity) and a Youden-index of 0.74, both 90% specificity.

Conclusions:

Keratoconus progression should be defined by evaluating parameters which consider several corneal variables (i.e. not simply Kmax but rather D-index, Keratoconus index etc.). We suggest D-index and KPI as variables to detect progression since they showed the best Youden-Index of all tested variables.

Financial Disclosure:

NONE