The enhanced ectasia susceptibility score (EESS) for predicting ectasia risk prior to LASIK

Session Details

Session Title: Primary and Secondary Ectasia

Session Date/Time: Sunday 23/09/2018 | 08:00-10:00

Paper Time: 08:12

Venue: Room A4

First Author: : R.Ambrosio BRAZIL

Co Author(s): :    B. Lopes   S. Schallhorn   J. Schallhorn   D. Reinstein   L. Jose Marcelo   A. Machado     

Abstract Details

Purpose:

To compare the surgical LASIK parameters, along with the preoperative clinical, topometric and tomographic features of the stable LASIK cases and cases that developed progressive post-LASIK ectasia; and to develop an enhanced tool to quantify ectasia risk using artificial intelligence.

Setting:

The Rio de Janeiro Corneal Tomography & Biomechanics Study Group

Methods:

In a multicentered retrospective study, one eye randomly selected from 3,278 patients that had LASIK with no ectasia after a minimal follow up of two years (Stable-LASIK [SL] group) were compared to 105 eyes that developed post-LASIK ectasia (PLE). All cases had preoperative Pentacam (Oculus, Wetzlar, Germany) raw data available, for generating topometric and tomographic indices. Clinical data, including age, refraction, flap thickness and maximal ablation depth were computed to calculate the residual stromal bed (RSB) and percentage of tissue altered (PTA). Different artificial intelligence approaches were used to combine the data to best separate the SL and PLE groups.

Results:

All metrics analyzed had significant differences among the groups (Mann-Whitney; p<0.05), but having area under the ROC (Receiver operating characteristic) curve (AUC) lower than 0.8, with exceptions of the Pentacam Random Forest Index (PRFI; AUC 0.921; 75.24% sensitivity and 92.98% specificity; cut off 0.068), and the Belin/Ambrósio Deviation (BAD-D; AUC 0.807; 59.05% sensitivity and 90.15% specificity; cut off 1.5). The cut offs of 300µm for RSB and 0.4 for PTA had respectively 24.8% and 25.7% sensitivity, and 94.1% and 88.2% specificity. The EESS, calculated by 10% internal validation/random forest, had AUC 0.97, 93.33% sensitivity and 90.91% specificity.

Conclusions:

The significant overlap limits the ability of most topometric and tomographic metrics as a screening tool. A combined approach significantly enhances the ability to quantify ectasia risk prior to LASIK. However, some mysterious cases still are present. While these cases may occur due to excessive eye rubbing, this may also suggest the need for further characterization of the cornea beyond 3D Scheimpflug tomography.

Financial Disclosure:

... gains financially from product or procedure presented, ... is employed by a for-profit company with an interest in the subject of the presentation