Posters

Fourier analysis for differentiating the ectasia phenotype

Poster Details

First Author: F.Faria Correia PORTUGAL

Co Author(s):    M. João   M. Simões   N. Franqueira   T. Monteiro           

Abstract Details

Purpose:

To determine the overall accuracy of Fourier analysis data derived from the Pentacam HR (Oculus, Wetzlar, Germany) in differentiating the different phenotypes of keratoconus (KC).

Setting:

Hospital de Braga, Braga, Portugal.

Methods:

Pentacam data from 63 normal (control group) and 97 keratoconus (KC group) was analyzed. The KC group was further subdivided into five phenotypes (20 Bow-Tie, 22 Croissant, 20 Duck, 21 Nipple and 14 Snowman). Receiver operating characteristic (ROC) curves were calculated, along with pairwise comparisons.

Results:

All variables expressed significant differences between the Duck and control groups (Kruskal-Wallis, p<0.05). All variables, except Peripheral Astigmatism, showed significant differences between the control group and Croissant, Nipple and Snowman phenotypes (Kruskal-Wallis, p<0.05). There were also significant differences in the Spherical Rmin for all comparisons involving the Nipple group (ANOVA, p<0.001). In addition, differences were also present in the Max Decentration data for all comparisons involving the Bow-Tie group (Kruskal-Wallis, p<0.05). The Max Decentration parameter was the most accurate in differentiating Croissant (AUC 0.99), Duck (AUC 0.99), Bow-Tie (AUC 0.971) and Snowman (AUC 0.996) KC eyes.

Conclusions:

Specific parameters derived from the Fourier analysis may help in characterizing the KC phenotype. When divided into phenotypes, Max Decentration remained the most accurate at differentiating Croissant, Duck, Bow-Tie and Snowman phenotypes.

Financial Disclosure:

None