A Novel Modeling System For Keratoconus Prediction Based On Artificial Intelligence

Keratoconus diagnosis with the help of a modeling system would be considered a promising approach for eye care professionals. We aimed to develop a novel model based on a chaotic system to extract the non-constant-coefficient in every individual eye. So based on a review of literature it is the first time that a chaotic model is being designed and applied for ophthalmic disease.

Bina Eye Hospital, Tehran, Iran.

For corneal properties modeling, Pentacam images with the model of chaotic were used. The designed model was implemented on two subgroups with 65 healthy cases and 48 keratoconus patients who participate in our study are between 23 to 74 years old (62 men and 51 women). To examine the accuracy of results, all cases are set into two subgroups. In subgroup one, 34 men and 27 women were included and in subgroup two, 28 men and 24 women were analyzed. The designed model was applied to each subgroup and gathered data was analyzed.

The results of model analysis based on the chaotic system for Pentacam images indicate significant sensitivity and specificity. The sensitivity of 89% and 91% and also the specificity of 91% and 94% for subcategories of 1 and 2 respectively were reported (subgroup 1 p-value: 0.028 and subgroup 2 p-value:0.025). 

In this study, diagnosis of keratoconus with the help of a novel chaotic system with the characteristic of being exclusive for each individual eye (Iran model). We have developed a novel chaotic artificial intelligence-based system that can be utilized successfully to diagnose the keratoconus in a very individual way. Despite the results being very promising in terms of accuracy and repeatability but it seems that a larger study might give even greater credibility to this novel model