Ai-Assisted Early Detection Of Keratoconus Using The Ocusmart Platform: Pilot Clinical Experience
Published 2026 - 44th Congress of the ESCRS
Reference: POCOR326 | Type: Poster | DOI: 10.82333/wyep-mn76
Authors: Dr Ghada Atta* 1 , A/Prof. Alfredo Di Giovanni 2
1ophthalmology,Clinic,Cologne ,Germany, 2ophthalmology,universty,Naples,Italy
Purpose
To evaluate the feasibility of using an AI-assisted digital platform (OcuSmart) for early detection of keratoconus and risk stratification prior to refractive surgery.
Setting
Single tertiary eye care center / refractive surgery clinic
Pediatric and adult patients presenting for routine refractive assessment or pre-surgery evaluation
Study period: January 2025 to December 2025
Ethics approval obtained; informed consent from all participants
Methods
Prospective pilot study including 50 patients (100 eyes) evaluated for refractive surgery or routine screening
Collected parameters: corneal topography indices (e.g., Kmax, corneal asymmetry), corneal thickness (pachymetry), age, AI-generated alerts for suspect keratoconus patterns
Primary outcomes: detection rate of early or forme fruste keratoconus, agreement with clinician assessment, intervention planning, or further imaging triggered by AI alerts
Results
50 patients (100 eyes), mean age 24 plus/minus 6 years
AI detected early keratoconus in 12 eyes (12 percent)
Agreement with clinician assessment: 10 out of 12 eyes (83 percent)
Intervention planning or further imaging was initiated in 10 eyes (83 percent) flagged by AI
Mean corneal thickness of flagged eyes: 485 um plus/minus 12 um
Mean Kmax in flagged eyes: 47.8 plus/minus 1.2 D
Conclusions
AI-assisted digital screening using the OcuSmart platform is feasible and effective for early keratoconus detection, showing high agreement with clinician assessment. Implementation of AI alerts may enhance patient safety in refractive surgery and enable proactive management of at-risk corneas.