ESCRS - FP06.12 - Epithelial Refractive Remodeling Score (Errs) Using Epithelial Refractive Power Mapping To Detect Early Keratoconus: Multicenter Validation

Epithelial Refractive Remodeling Score (Errs) Using Epithelial Refractive Power Mapping To Detect Early Keratoconus: Multicenter Validation

Published 2026 - 44th Congress of the ESCRS

Reference: FP06.12 | Type: Free paper | DOI: 10.82333/hymp-db66

Authors: Dr Neofytos Mavris* 1 , Dr Tomer Batash 1 , Dr Itay Lavy 1 , Prof. Dr Renato Ambrósio Jr. 2 , Dr Alexandre Neto 3 , Dr Aleksander Stojanovic 4 , Dr Yue Feng 5 , Prof. Dr David Smajia 1

1Ophthalmology,Hadassah Medical Center, Faculty of Medicine, The Hebrew University of Jerusalem,Jerusalem,Israel, 2Ophthalmology,Federal University of the State of Rio de Janeiro (UNIRIO), Department of Ophthalmology, Rio de Janeiro, Brazil,Rio de Janeiro,Brazil, 3Ophthalmology,Department of Ophthalmology, Federal University of the State of Rio de Janeiro (UNIRIO), Rio de Janeiro, Brazil,Rio de Janeiro,Brazil, 4Ophthalmology,Eye Department, University Hospital of North Norway, Tromsø, Norway,Tromsø,Norway, 5Community Medicine,Institute of Community Medicine, Faculty of Health Sciences, University in Tromsø (UiT), Tromsø, Norway,Tromsø,Norway

Purpose

To develop and validate a new screening score—the Epithelial Refractive Remodeling Score (ERRS), for the detection of early keratoconus (KC) using epithelial refractive remodeling metrics derived from epithelial thickness maps

Setting

Multicenter cornea/refractive clinics; retrospective–prospective dataset of normal, early–moderate KC, and advanced KC eyes.

Methods

High-resolution epithelial thickness maps were acquired with spectral-domain OCT (Heidelberg Engineering GmbH, Heidelberg, Germany) and converted to epithelial refractive power (ERP) maps using a proprietary VisionScan OCT analyzer. The features set included ERP, refractive pattern standard deviation, and spatial overlap of predefined areas of interest. The ERRS model (regularized logistic regression/gradient boosting) was trained with nested cross-validation and calibration. Performance was compared with standard tomographic indices using AUC, sensitivity/specificity, and decision-curve analysis.

Results

Across 1034 eyes (433 normal, 299 early–moderate KC, 302 advanced KC), ERRS showed a strong discriminative performance. All normal corneas (433/433) and all advanced KC (302/302) were correctly classified. Among early–moderate KC, 98.3% (294/299) were correctly identified, with 5 subclinical KC eyes (1.7%) misclassified as normal. Overall performance showed very high discriminative ability to identify normal and keratoconic eyes, correctly classifying 1029/1034 (99.5%) eyes.

Conclusions

The ERRS derived from epithelial refractive power mapping provides highly accurate discrimination of normal and keratoconic corneas. It represents a novel screening score enabling objective keratoconus assessment and standardized interpretation of epithelial thickness maps, supporting its role as a powerful adjunct to tomography for early KC detection.