The Combined Utilization of Epithelial Thickness and Tomographic Parameters in Keratoconus Detection.

Abstract

Purpose: To evaluate the performance of the anterior segment optical coherence tomography (AS-OCT) and Scheimpflug-Placido tomographic (SPT) indices in differentiating suspect keratoconus (SKC) from normal (N). Methods: In this cross-sectional study, one eye of each patient was scanned on AS-OCT and SPT. Following, 5 regression analysis models were developed: Model 1: combined AS-OCT corneal thickness (CT); Model 2: combined AS-OCT epithelial thickness (ET); Model 3: Models 1 + 2; Model 4: SPT indices (symmetry index front [SIf], symmetry index back [SIb], keratoconus vertex back [KVb], and minimum CT [ThkMin]); Model 5: Models 3 + 4. The areas under the curve (AUC) of receiver operator characteristic (ROC) curves were compared across groups to estimate their performance accuracy. Results: Two hundred and five N eyes, 56 SKC eyes, and 89 keratoconus (KC) eyes were scanned with AS-OCT and SPT. For Models 1 and 2, no individual metric yielded an AUC > 0.84. For Model 3, the AUC was 0.96 (R2 Nagelkerke: 0.69) with 90.28% cases correctly identified as SKC. For Model 4, the AUC was 0.99 (R2 Nagelkerke: 0.87), with 95.58% of cases correctly classified as SKC. For Model 5, the diagnostic accuracy did not improve compared to Model 4. Conclusion: Both combined AS-OCT indices and combined SPT indices provided excellent diagnostic power to differentiate SKC from N eyes. The combined SPT model showed a superior value compared to the AS-OCT model. Furthermore, combined AS-OCT and SPT data did not perform better than the combined SPT data alone.