Purpose: To test the predictive accuracy of the Belin-Ambrósio deviation index (BAD-D), the stiffness parameter A1 (SPA1), the Corvis biomechanical index (CBI), and the tomographic and biomechanical index (TBI) assessments for distinguishing subclinical and mild keratoconic eyes from normal eyes.
Setting: Medical Research Foundation, Sankara Nethralaya, Chennai, India.
Design: Retrospective case series.
Method: In this cross-sectional clinical study, the following 3 groups of eyes were analyzed: very asymmetric ectasia with normal topography (very asymmetric-normal topography), mild keratoconus, and normal control. All eyes had comprehensive assessment with corneal topography (TMS-IV), Scheimpflug tomography (Pentacam HR), and dynamic Scheimpflug biomechanical analysis (Corvis ST). The outcome measures were the BAD-D, SPA1, CBI, and TBI. Receiver operating characteristic (ROC) curve analysis was performed to determine each parameter's predictive accuracy in distinguishing between eyes with subclinical or mild keratoconus and control eyes.
Results: The area under the curve (AUC) ROC values for the very asymmetric-normal topography and normal control comparison were 0.81 (BAD-D), 0.76 (SPA1), 0.78 (CBI), and 0.90 (TBI). The TBI (using cutoff value 0.16) showed the highest diagnostic accuracy (85%), with 84% sensitivity and 86% specificity. The AUC ROC values for the mild keratoconus and normal control comparison were 0.998 (BAD-D), 0.91 (SPA1), 0.97 (CBI), and 0.999 (TBI). The TBI (with a 0.63 cutoff) showed the highest accuracy (99.5%), with 99% sensitivity and 100% specificity. The TBI also showed the weakest correlation with mean keratometry, biomechanically corrected intraocular pressure, and central corneal thickness in normal eyes and keratoconic eyes.
Conclusions: The TBI best distinguished eyes with mild ectasia from normal eyes and had the weakest correlation with biomechanical confounding factors, reinforcing the hypothesis that the TBI represents corneal biomechanical susceptibility.
Copyright © 2018 ASCRS and ESCRS. Published by Elsevier Inc. All rights reserved.