Forecasting Progressive Trends in Keratoconus by Means of a Time Delay Neural Network.

Marta Jiménez-García,Carina Koppen,Jos Rozema,Ikram Issarti,Sorcha Ní Dhubhghaill,David Varssano,Elke Kreps

doi:10.3390/jcm10153238

Abstract

Early and accurate detection of keratoconus progression is particularly important for the prudent, cost-effective use of corneal cross-linking and judicious timing of clinical follow-up visits. The aim of this study was to verify whether a progression could be predicted based on two prior tomography measurements and to verify the accuracy of the system when labelling the eye as stable or suspect progressive. Data from 743 patients measured by Pentacam (Oculus, Wetzlar, Germany) were available, and they were filtered and preprocessed to data quality needs. The time delay neural network received six features as input, measured in two consecutive examinations, predicted the future values, and determined the classification (stable or suspect progressive) based on the significance of the change from the baseline. The system showed a sensitivity of 70.8% and a specificity of 80.6%. On average, the positive and negative predictive values were 71.4% and 80.2%. Including data of less quality (as defined by the software) did not significantly worsen the results. This predictive system constitutes another step towards a personalized management of keratoconus. While the results obtained were modest and perhaps insufficient to decide on a surgical procedure, such as cross-linking, they may be useful to customize the timing for the patient’s next follow-up.

Highlights

Keratoconus (KC) is a progressive, ectatic corneal disorder presenting with central or paracentral stromal thinning associated, among others, with corneal protrusion, structural changes, and transparency loss [1,2,3]
Stabilization via CXL may be performed based on other factors, many national CXL protocols require proof of KC progression to qualify for reimbursement [12]
After the filtering (Figure 2), 629 eyes from 421 patients measured with Pentacam remained with a mean age of 26.8 (6.4) years

Summary

Introduction

Keratoconus (KC) is a progressive, ectatic corneal disorder presenting with central or paracentral stromal thinning associated, among others, with corneal protrusion, structural changes, and transparency loss [1,2,3]. Definitions for progression have been established based on the repeatability of the clinical devices, which determines the minimum change measurable at a pre-established significance level, keeping in mind that repeatability is worse in keratoconic corneas [6]. This approach was recently introduced in the Pentacam. Progression in KC presents considerable interindividual variability: from cases with significant progression in 3 months [8] to cases that remain stable for more than a decade [9] These clinical variations, as well as the effectiveness of corneal cross-linking (CXL) [10], have led to a renewed interest in assessing the risk factors for a faster KC evolution [11]. Age and maximum keratometry are mainly used to schedule the follow-up [11], but additional features to stratify progression risk would be of added value

Objectives

Results

Discussion

Conclusion