Abstract
Purpose: :
Keratoconus (KCN) and form fruste keratoconus (FFKC) are contraindications to many refractive surgical procedures. Therefore, it is imperative to identify and exclude these conditions preoperatively. The Orbscan Scanning slit anterior segment analyzing system (Bausch and Lomb, Rochester, NY) recommends to clinicians several criteria to diagnose KCN based on a measurement. These criteria are a collection based on previous published reports and clinician experience, and have become commonly known as the "Orbscan KCN detection indices." The purpose of this study is to evaluate these detection indices.
Methods: :
Orbscan exams from 47 KCN and 41 non–diseased eyes were reviewed by a masked clinician, who classified the map as KCN or normal based on each of the screening criteria. This Orbscan–based classification was compared to the clinical diagnosis of whether KCN was present or not. The clinical diagnosis of KCN was based on the criteria established by the Collaborative Longitudinal Evaluation of Keratoconus (CLEK) study.
Results: :
The accuracies of the 13 individual tested indices ranged from 0.48 to 0.90. Sensitivities ranged from 0.02 to 1. Specificities ranged from 0 to 1. The indices with both sensitivities and specificities, as well as accuracies above 0.8 included: (1) mean power of >45.5 D, (2) an asymmetric bow–tie pattern observed on the axial map, (3) a highest point on the posterior elevation at least 50 microns above the best–fit sphere, (4) and inferior–temporal displacement of highest elevation on anterior and poster surfaces. Basing a diagnosis of KCN solely on the fact that the highest elevation on the posterior surface was greater than 50 microns above the best–fit sphere yielded a sensitivity of 0.85 and specificity of 0.98. Combining any 7 of the 13 tested criteria only slightly improved the diagnostic ability (sensitivity = 0.88, specificity = 0.98, accuracy = 0.92).
Conclusions: :
Although several of the corneal irregularity indices generated from the Orbscan performed quite poorly in detecting keratoconus, a few performed quite well. Performance of the best–performing single index could only be slightly improved if applied in combination with 6 other indices. As the single index with the best diagnostic ability was based on the posterior elevation data, having detailed knowledge of the posterior surface seems beneficial in detecting keratoconus.
Keywords: keratoconus • topography • cornea: clinical science