Purchase this article with an account.
Alain Saad, Alice Grise-Dulac, Damien Gatinel; The Use of Discriminant Function to Detect Corneas at Risk for Ectasia after Refractive Surgery. Invest. Ophthalmol. Vis. Sci. 2011;52(14):4196.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
To investigate if topography and tomography indices may help to detect early subclinical keratoconus non detectable with Placido based neural network program (Form Fruste Keratoconus-FFKC).
We retrospectively reviewed data of 173 eyes separated into 3 groups by the Nidek Corneal Navigator (NCN) System of the OPD scan: normal (N) (n=96), Forme Fruste Keratoconus (FFKC) (Classified as N by the NCN but with a KC in the fellow eye) (n=44) and Keratoconus (KC) (n=33).Topography and tomography indices were obtained with the Orbscan IIz. Discriminant analysis was used to determine the group of an observation based on a set of indices obtained from the anterior and posterior corneal surface and from the thickness spatial profile. The discriminant analysis constructs a set of linear functions of the indices. The output values of the discriminant function were compared by Kruskal Wallis test.
Posterior elevation of the thinnest point (TP), Percentage of Thickness Increase (PTI) for all distances from the TP, and Percentage of variation of Anterior Curvature (PVAK) from a 5- to 7-mm distance from the TP were significantly different in the FFKC compared with the N group. The mean output values of the discriminant function able to differentiate between the N group and the FFKC group with a sensitivity of 93% and a specificity of 92% were 0.55 +/- 0.78 for the N group, 3.55 +/- 1.32 for the FFKC group and 14.73 +/- 8.60 for the KC group. The discriminant functions between the FFKC and the N groups and between the KC and the N groups reached an area under the ROC curve of 0.98 and 0.99, respectively. PTI indices and maximum posterior central elevation were the most important contributors to the discriminant function.
Combining topography and tomography parameters may help in creating a sensitive and specific artificial intelligence for the detection of corneas at risk for refractive surgery. New charts and graphs exploring data derived from elevation and pachymetry maps should be generalized in future corneal topography software to help the clinician in screening patients for refractive surgery.
This PDF is available to Subscribers Only