- Vienna '18
- Athens 2019
- ESCRS Player
- On Demand
- ESCRS iLearn
- ESCRS YO's
First Author: I.Ramos BRAZIL
Co Author(s): B. Lopes R. Correa F. Faria-Correia B. Valbon M. Salomão R. Ambrósio Jr.
Back to previous
To develop objective methods to detect preoperative ectasia risk (susceptibility) among LASIK candidates considering clinical data, front surface curvature (topometric) data, and 3-D pachymetric and elevation (tomographic) data.
Instituto de Olhos Renato Ambrósio, Rio de Janeiro-RJ, Brazil; Hospital de Olhos Santa Luzia, Maceió-AL, Brazil; Rio de Janeiro Corneal Tomography and Biomechanics Study Group, Rio de Janeiro-RJ, Brazil.
A retrospective nonrandomized study involved 23 eyes that developed ectasia after LASIK and 266 eyes with stable LASIK outcomes (minimal follow up of 12 months). Preoperative clinical data and Oculus Pentacam data were available for all cases. Classic ERSS (Ectasia Risk Score System) was calculated based on age, spherical equivalent, residual stromal bed, central thickness, and subjective classification of corneal topography (front surface axial curvature map). Front surface curvature (topometric) and tomographic (thickness profile and front/back elevation) indices were assessed. Non parametric Mann-Whitneys test was performed to assess differences between the groups. Different combinations that best distinguished ectasia and stable LASIK groups were created using Fisher"s linear discriminant analysis (LDA) based on clinical parameters plus topometric data, and on clinical parameters plus tomographic data. The area under the ROC curve (AUC) were calculated for each LDA functions with pairwise comparisons.
Statistically significant differences were found among the groups for all tested parameters (p<0.001), but preoperative spherical equivalent (SE). ERSS was equal or higher than 3 on 12 eyes from the ectasia group (sensitivity=52.17%) and on 48 eyes from the stable group (specificity=81.95%). The best LDA function including clinical parameters and topometric indices used IHD (index of height decentration), with AUC of 0.980 (sensitivity=100%; specificity=93.23%). The best LDA function combining clinical and tomographic parameters, included the Belin-Ambrósio Deviation (BAD-D), achieving 100% sensitivity and 97.4% specificity, with a statistically better AUC (0.994) than all individual parameters (p<0.001), and LDA functions.
Integrated analysis of clinical data and objective topometric parameters were superior than the score that consider clinical data and the classic subjective classification of topography. Tomographic data significantly enhanced the ability for screening ectasia risk among LASIK candidates. Similarly, clinical parameters significantly improve the ability to detect ectasia susceptibility based on tomographic data. Further validation of the LDA functions are necessary.