April 2014
Volume 55, Issue 13
Free
ARVO Annual Meeting Abstract  |   April 2014
Identification of Keratoconus using Combined Layered Pachymetry and Tomography
Author Affiliations & Notes
  • Ronald H Silverman
    Columbia University Medical Center, New York, NY
    London Vision Clinic, London, United Kingdom
  • Raksha Urs
    Columbia University Medical Center, New York, NY
  • Timothy J Archer
    Riverside Research, New York, NY
  • Marine Gobbe
    Riverside Research, New York, NY
  • Michael Ha
    Columbia University Medical Center, New York, NY
  • Michele D Lee
    Columbia University Medical Center, New York, NY
  • Arindam RoyChoudhury
    Columbia University Medical Center, New York, NY
  • Dan Z Reinstein
    Riverside Research, New York, NY
  • Footnotes
    Commercial Relationships Ronald Silverman, Arcscan, Inc. (I), Cornell Research Foundation (P); Raksha Urs, None; Timothy Archer, None; Marine Gobbe, None; Michael Ha, None; Michele Lee, None; Arindam RoyChoudhury, None; Dan Reinstein, Arcscan, Inc. (I), Cornell Research Foundation (P)
  • Footnotes
    Support None
Investigative Ophthalmology & Visual Science April 2014, Vol.55, 2451. doi:
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Ronald H Silverman, Raksha Urs, Timothy J Archer, Marine Gobbe, Michael Ha, Michele D Lee, Arindam RoyChoudhury, Dan Z Reinstein; Identification of Keratoconus using Combined Layered Pachymetry and Tomography. Invest. Ophthalmol. Vis. Sci. 2014;55(13):2451.

      Download citation file:


      © ARVO (1962-2015); The Authors (2016-present)

      ×
  • Supplements
Abstract
 
Purpose
 

Keratoconus (KC) screening is presently accomplished by clinical signs and history, surface topography and pachymetry. Identification of subclinical KC, however, may be challenging. Remodeling of the corneal epithelium in response to stromal deformation produces a distinct pattern that is independent of conventional pachymetry and surface topography. In this study, we combine epithelial and stromal pachymetric maps with tomographic and anterior and posterior surface topographic data to develop a joint multivariate model for differentiation of normal and KC eyes.

 
Methods
 

Artemis very-high-frequency ultrasound and Pentacam Scheimpflug data were obtained on 126 normal and 28 independently identified KC eyes (one eye per subject). Maps of epithelial and stromal thickness were generated from the Artemis data and custom software used to derive variables characterizing features related to the thickness distribution of each layer. Pentacam analysis included characterization of pachymetry, curvature and difference-from-sphere of each surface as well as Belin/Ambrosio indices. Stepwise linear discriminant analysis (LDA), including leave-one-out cross-validation and ROC analysis, was performed separately on Artemis and Pentacam data and on the merged data set.

 
Results
 

LDA of Pentacam data resulted in a 9-variable model (Φ2=232.0). The area under the ROC curve (AUC) was 100%, indicative of complete separation of normal and KC eyes. Analysis of Artemis data yielded a 7-variable model (Φ2=238.0) with an AUC of 99.3%. An 11-variable model (Φ2=289.1) was produced from the merged database. The model contained 7 variables from Pentacam plus 4 from Artemis and produced an AUC of 100%.

 
Conclusions
 

Our findings demonstrate that both layered pachymetry and two-surface topography and tomography each are effective in distinguishing advanced KC from normal. In a multivariate analysis combining the data, the optimized classification function included parameters from both Artemis and Pentacam, indicating that both techniques provide significant and independent information. The application of such a combined model might facilitate early diagnosis by inclusion of parameters sensitive to epithelial remodeling in response to early stromal deformation, and this will be a topic for ongoing investigations.

 
Keywords: 574 keratoconus • 552 imaging methods (CT, FA, ICG, MRI, OCT, RTA, SLO, ultrasound) • 550 imaging/image analysis: clinical  
×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×