HomeFeatures › Keratoconus Screening
Early Detection

FIND THE CORNEA THAT SHOULD NOT BE LASERED.

Written by the RLES AI clinical rules team · Medically reviewed by Fırat Helvacıoğlu, MD · Berrin Toksü, MD · Ercüment Bozkurt, MD · Last updated: 18 August 2026

Post-refractive ectasia is the complication every surgeon plans against. RLES AI screens every cornea with a full battery of tomographic, topographic and biomechanical indices — interpreted together, never in isolation — and runs even when the device’s own software reports "Normal".

MANY INDICES, ONE DISCIPLINED READING

BAD-D, ART max, Kmax, ISK, IVA, posterior elevation, pachymetric progression[4], CBI/TBI, anterior I-S asymmetry[1] and the Sirius index family are evaluated against graded bands — normal, borderline, pathologic — and combined in a decision matrix that determines eligibility, follow-up interval and the place of cross-linking.

Crucially, the counting is correlation-aware: mathematically related indices (ART max is literally thinnest pachymetry divided by pachymetric progression)[2] are never counted as independent evidence. One geometric feature moving four derived indices is one finding — not four. This single discipline prevents a whole class of false-positive ectasia verdicts.

BORDERLINE EYES GET A CLINICAL PROTOCOL, NOT A GUESS

  • Retinoscopy — including the early, non-scissoring irregular reflex that naive yes/no questions miss
  • Slit-lamp verification set: Vogt striae, Fleischer ring, apical thinning
  • Epithelial thickness mapping, corneal biomechanics (Corvis ST / ORA), serial tomography at defined intervals
  • A positive clinical finding escalates the eye regardless of what tomography claimed — and a normal exam never downgrades borderline tomography

SCREENING THAT DOESN'T TRUST THE DEVICE'S VERDICT

In a documented RLES AI case, the topographer’s built-in keratoconus module reported "Normal" in both eyes of a patient whose corneas were deformed by contact-lens wear — the irregularity only appeared when indices and maps were read together. RLES AI performs its own analysis on every eye, every time. The system never rules that surgery "cannot be done"; it grades the risk,[3] explains the reasons, and leaves the decision with the surgeon.

FREQUENTLY ASKED QUESTIONS

Can RLES AI flag a cornea the device software called “Normal”?
Yes. It re-reads the indices together — BAD-D, ART max, posterior elevation, I-S asymmetry and more — with correlation-aware counting, and has flagged deformed corneas that built-in screening modules reported as Normal.
What happens with a borderline cornea?
It is routed to a clinical verification protocol — retinoscopy, slit-lamp signs, epithelial mapping, biomechanics and serial tomography — and is never silently approved.
SEE IT ON YOUR OWN CASES.
RLES AI runs as a mobile interface on your phone — nothing to install from a store, reports in minutes. The final clinical decision always rests with the surgeon.

Request Information
References — peer-reviewed sources
  1. Rabinowitz YS. Keratoconus. Surv Ophthalmol 1998;42:297-319. PMID: 9493273
  2. Ambrósio R Jr et al. Novel pachymetric parameters based on corneal tomography for diagnosing keratoconus. J Refract Surg. 2011;27(10):753-8. doi:10.3928/1081597X-20110721-01 PMID: 21800785
  3. Randleman JB et al. Risk assessment for ectasia after corneal refractive surgery. Ophthalmology 2008;115:37-50. PMID: 18083221
  4. Sezgin Akcay BI, Ozgurhan EB, Bozkurt E, Kurt T, Yildirim Y, Cosar MG, Yildirim A, Umurhan Akkan JC, Demirok A. Evaluation of pachymetric measurements with Scheimpflug photography-based system and optical coherence tomography pachymetry at different stages of keratoconus. J Ophthalmol. 2014;2014:719205. doi:10.1155/2014/719205 PMID: 25143828
Sources open on PubMed in a new tab.