Comparison of a large keratoconic population with “normal” eyes with a new algorithm for detecting keratoconus using Scheimpflug imaging

Session Details

Session Title: Cornea Medical I

Session Date/Time: Monday 12/09/2016 | 17:00-18:30

Paper Time: 17:12

Venue: Hall C2

First Author: : A.Amit Gupta INDIA

Co Author(s): :                        

Abstract Details

Purpose:

To document demographics, morphometric, biomechanical parameters and quality of vision in large cohort of Indian keratoconic (250) eyes and compare them with ‘normative’ data (250 emmetropic eyes) in order to establish their significance. In addition, to generate a new algorithm based on the large amount of data available which may accurately diagnose both manifest as well as forme fruste keratoconus.

Setting:

Advanced Eye Centre, Postgraduate Institute of Medical Education and Research, Chandigarh, India

Methods:

Group I was 250 Keratoconus eyes (136 patients ) Group II: 250 'normal' eyes (125 patients). Demographic details and clinical findings were noted including visual, refractive status etc. Quantitative data generated by Pentacam analysis, Aberrometry and Ocular Response Analyser, specifically Keratometry, Keratoconic Indices, pachynetric values, Progression Index, Ambrosio relational thickness (ART max), Maxp5A and Maxp5P (Maximum anterior and posterior power at 5 mm radius), MaxAE6 and MaxAD6 (maximum anterior elevation and posterior depression at 6 mm radius), AEED and PEED ( anterior and posterior elevation depression difference) etc. . Data was subjected to Independent Student’s t Test and Mann-Whitney U test.

Results:

Logistic regression analysis was used to measure predictive capability of various algorithm models which were generated. In discriminating keratoconus from normal eyes MaxAD6, AEDD, AER, AEDD/BFS, PEDD/BFS had maximum AUC (≥ .99), followed by MaxAE6, ISV, IVA, PEDD which had AUC (0.98-.99). ORA- KMP had maximum AUC (Area under curve) and IOP cc had lowest AUC in discriminating keratoconus from normal eyes. Overall sensitivity and specificity of HOA was found to be less, of which horizontal coma [Z(3,1)] had maximum AUC with high specificity and sensitivity.

Conclusions:

We generated a new algorithm Combining Pentacam parameters with Corneal Biomechanics using multiple regression analysis. The new Algorithm is as follows- Eyes with Maxp5A > 45.65D , AEDD > 19.5 D, ARTmax > 351 and CH <9.315 have a 99.2% probability of being Keratoconic Eyes. In addition to the already established criteria for diagnosing keratoconus, this new algorithm will be extremely useful especially in eyes with forme fruste keratoconus or to study the progression of disease as well as interventions including collagen cross-linking in such eyes.

Financial Disclosure:

NONE