- Vienna '18
- Athens 2019
- ESCRS Player
- On Demand
- ESCRS iLearn
- ESCRS YO's
First Author: S.Taneri GERMANY
Co Author(s): S. Oehler C. Schwethelm
Back to previous
Iatrogenic ectasia is one of the most severe complications of LASIK. Many parameters have been proposed to differentiate normal from ectasia susceptible eyes. However, the interpretation of diagnostic tests like the Orbscan (Bausch&Lomb Technolas, Munich, Germany) remains challenging for the clinician because of the multitude of parameters involved. Automated analysis of those maps is therefore desirable in order to help deciding whether LASIK is suitable for a given eye. We wanted to evaluate the accuracy of a new software for the objective evaluation of Orbscan examinations regarding the risk of corneal ectasia development.
Zentrum für Refraktive Chirurgie, Augenabteilung am St. Franziskus Hospital, Münster, Germany
Orbscan examinations (.OTE files) were retrospectively analyzed with a prototype of a new software (developed by Bausch & Lomb Technolas, Munich, Germany in collaboration with Gatinel and Saad, Rothschild Foundation, Paris, France). These historic exams were obtained using an acoustic factor of 0.96 for pachymetry instead of 0.92 as expected by the software. Only exams within the application range of the algorithm were included, i.e. exams of myopic or myopic astigmatic eyes of patients between the age of 18 and 50 years. Three groups of eyes were evaluated: 1. Normal eyes: only left eyes with unremarkable topography prior to a LASIK procedure at least 5 years ago with uneventful postoperative follow-up (n=46). 2. Early stage keratoconus: subjectively better asymptomatic eyes of keratoconus patients with corrected distance visual acuity of at least 1.0 (n=20). 3. Full keratoconic eyes: the more affected eyes of keratoconus patients (n=71).
Evaluation of Orbscan .OTE files was not possible with this software in 15.2% of normal eyes, 10.0% of early keratoconic eyes, and 49.3% of full keratoconic eyes, respectively. Of the remaining exams all normal eyes were classified as normal (i.e. specificity was 100%). Looking at true positive detection (sensitivity), 85.2% of early or full keratoconic eyes were correctly classified as pathologic by the software (LASIK is not recommended). Of the 8 pathologic eyes not correctly detected, 7 were clinically diagnosed as early stage keratoconic eyes (38.9%) and 1 as full keratoconic (2.8%).
This is a promising approach to make the detection of corneas at risk for ectasia after LASIK less examiner-dependent. However, sensitivity was not yet perfect in this version. Expanding the database of examinations and compensating for various acoustic factors may help to improve the performance of versions to come.