Interexpert agreement on epithelial thickness mapping for refractive surgery screening: variability highlights need for standardization.

Abstract

Purpose: To evaluate interexpert agreement on epithelial thickness map interpretation and its influence on eligibility decisions for keratorefractive surgery candidates.

Setting: Multicenter international collaboration involving 4 experienced refractive surgeons.

Design: Retrospective, masked case review study.

Methods: A total of 103 cases were retrospectively evaluated, including patients with normal (n = 44), borderline (n = 44), and keratoconic (n = 15) corneal tomography, as defined by the Belin-Ambrosio display (BAD). 4 independent experts (M.D., J.L.A.d.B., A.S., and N.L.) first assessed surgical eligibility based on Scheimpflug tomography and subsequently re-evaluated the same cases with the addition of anterior segment optical coherence tomography-based epithelial thickness maps. To ensure comparability, a standardized questionnaire with 9 predefined epithelial thickness characteristics was used. Interrater agreement was quantified with Fleiss κ, and the relative impact of each imaging modality on decision-making was ranked.

Results: Incorporating epithelial mapping changed the eligibility decision in 19% (77/412) of cases. Specifically, 9% (16/176) of normal, 20% (35/176) of borderline, and 3% (2/60) of keratoconic cases were reclassified as eligible, while 2% (3/176), 6% (10/176), and 3% (2/60), respectively, were deemed ineligible. Interrater agreement on eligibility was moderate before (κ = 0.46) and after (κ = 0.43) adding epithelial maps but remained poor in borderline cases (κ = 0.14 and 0.12, respectively). Agreement on epithelial profile features ranged from poor to fair (κ = -0.78 to 0.35). Corneal topography and BAD were ranked as most influential, followed by epithelial mapping.

Conclusions: Epithelial mapping frequently shifted decisions toward eligibility. However, poor interexpert agreement highlights the need for standardized, safety-oriented interpretation to ensure consistent ectasia risk assessment before refractive surgery.