Corneal Deformation Response with Dynamic Ultra-high-speed Scheimpflug Imaging for Detecting Ectatic Corneas

Abstract

ABSTRACT Purpose To test the ability of metrics derived from corneal response to noncontact tonometry (NCT) to distinguish between normal and ectatic cases. Materials and methods The prototype of CorVis ST (Oculus, Wetzlar, Germany) was used for assessing corneal biomechanical response using ultra-high-speed 8 mm horizontal Scheimpflug photography, taking 4,330 frames per second during NCT. Patients were stratified based on clinical data, including rotating Scheimpflug corneal tomography (Oculus Pentacam HR). Biomechanical data from one eye randomly selected of 177 patients with normal corneas (N) and from 79 patients with bilateral keratoconus (KC) were investigated. Group forme fruste keratoconus (FFKC) was composed of 20 eyes with normal topographic patterns from cases with ectasia detected in the fellow eye. Group keratoconus suspect (KCS) had 16 eyes from 16 patients with topographic patterns suspicious of KC but documented stability over 3 years and normal tomographic findings. A combination of deformation parameters using linear regression analysis (Prototype Factor 1, pF1) was created by the BrAIn (Brazilian Artificial Intelligence on Corneal Tomography and Biomechanics) study group in order to provide the best possible separation of KC and normals. Results Statistical significant differences were found for N × KC for several parameters, including first and second applanation times, deformation amplitude, and maximal concavity radius (Mann–Whitney, p < 0.001). However, the areas under the receiver operating characteristic curves (AUC) were lower than 0.90. The pF1 had AUC of 0.945 (IC 0.909–0.97; sensitivity = 87.3% and specificity = 89.3%). The pF1 had statistically significant differences between the ectatic (KC and FFKC) and nonectatic groups (N and KCS) (p < 0.05, Kruskall–Wallis Test with post hoc Dunn's test). Conclusion Corneal deformation response analysis by ultra- high-speed 8 mm horizontal Scheimpflug photography provides relevant data for distinguishing ectatic and nonectatic corneas but cannot be used independently to detect KC. This data may be integrated with corneal tomography data for enhancing sensitivity and specificity for screening ectasia. How to cite this article Salomão MQ, Correia FF, Ramos I, Luz A, Ambrósio R Jr. Corneal Deformation Response with Dynamic Ultra-high-speed Scheimpflug Imaging for Detecting Ectatic Corneas. Int J Kerat Ect Cor Dis 2016;5(1):1-5.