Detection of subclinical keratoconus using a patient-specific geometric 3D modeling
Session Details
Session Title: Corneal Biomechanics
Session Date/Time: Monday 09/10/2017 | 08:00-10:30
Paper Time: 08:12
Venue: Room 2.1
First Author: : M.Tiveron Jr SPAIN
Co Author(s): : F. Cavas-Martinez L. Bataille D. Fernandez-Pacheco J. Alio
Abstract Details
Purpose:
To develop and investigate a new custom method for 3D model characterization capable to detect differences between normal corneas and corneas with abnormal topography associated to normal visual acuity.
Setting:
Vissum Corporation, Alicante, Spain. Technical University of Cartagena, Spain.
Methods:
This retrospective case control series study evaluated 120 eyes divided into 2 groups: normal eyes and subclinical keratoconic corneas. The first group consisted of 89 normal eyes of 89 patients (37.49 ± 15.11 years). The second group comprised of 31 eyes of 31 patients diagnosed as subclinical keratoconus (34.72 ± 11.12 years). For each cornea, reconstruction from raw data of the Sirius tomographer (CSO, Florence, Italy) and a posterior analysis through patient-specific 3D model were performed. ROC curves were established. A comparison with the Sirius' own neural network based algorithm for keratoconus screening by the posterior apex deviation was performed.
Results:
The early keratoconus group showed a statistically significant decrease in total corneal volume (p<0.05). Both anterior and posterior surfaces were statistically lower in subjects with normal corneas (p<0.05). The posterior apex deviation reached the best performance in discrimination between both groups (area: 0.883 ± 0.041, p<0.000, cutoff: 0.0655mm, sensitivity: 99.3%, specificity: 98.5%). When applying the same analysis for both studied eyes, the following results were found: the sensitivity of 85.9% and specificity of 97.8% were yielded from the standardized KC screening of Sirius.
Conclusions:
The overall studied metrics showed statistically significant differences between groups. Corneal volumetric analysis based on geometrical modeling can discriminate between subclinical keratoconic and normal corneas.
The analysis of corneal geometric custom modeling is a new useful tool for the practice of refractive surgery providing excellent accuracy to detect early corneal deformation in keratoconic patients with normal visual acuity.
Financial Disclosure:
NONE
