A computational model using artificial neural networking for predicting astigmatism following corneal surgery

Abstract

Corneal ecstatic disorder is a corneal thinning condition and Keratoconus is a specific form of ectasia when the cornea thins and starts to bulge. Corneal cross linking is the only available accepted treatment for prevent or decrease the progressive Keratoconus condition. Although this treatment named as a successful surgery, numerically it has a 7.6% failure rate and 2.9% complication rate according to the post-operative statistics. However, there is no prediction mode available to measure or capture the post-operative results before the surgery occurs. Accordingly, this research is framed by the pre-operative and post-operative variables in order to develop a computation model for future predictions before the surgery occurs. The cornea reassembles a watch glass, which is transparent, tissue. The cornea covers a one fifth of the eye globe, which is the main refractive element and contribute to two third of the refractive power for the eye There are several refractive errors related to the eye, such as Myopia, Hyperopia, Presbyopia and astigmatism. The research is focus about the astigmatism as the refractive error. Astigmatism is a type of a refractive error caused by the irregularity in the shape of a person’s cornea. In this condition, the eye fails to focus the light equally on the retina leading to blurred or distorted vision causing either myopia or hyperopia. The astigmatism can be categorized into two main sections. Regular astigmatism and the Irregular astigmatism. Most of the astigmatism is regular and can be corrected with the spectacles. Irregular astigmatism is the worst scenario resulted from scarring of the cornea following injury, due to eye surgery or corneal ectasia like keratoconus. The keratoconus is a progressive, noninflammatory, bilateral, asymmetric disease, characterized by paraxial stromal thinning irregular protrusion. One of the acceptable ways to prevent progression of the keratoconus is Corneal collagen crosslinking, which strengthens corneal tissue to halt progression of the eye's surface in keratoconus by removing the outer layer of the cornea (called the epithelium) is removed to allow entry of riboflavin, a type of B vitamin, into the cornea, which then is activated with UV light. According to the past literature, outcome of the corneal collagen crosslinking displaying a nonlinear behavior. The researcher used a novel idea of predicting the factors associated with collagen crosslinking for the outcome of pre-operative and post-operative astigmatism with the aid of an artificial neural network model. The researchers were able to discover the age and the gender as a related factor to the astigmatism changes of the post-operative outcome of the surgery. The researchers were also able to apply radial basis functions to suit neural network model. The dataset used to train the neural network is needed to be increased and improved to suit a better fitting of the model.the covid-19 outbreak has restricted the researchers from collecting and finetuning the dataset from the hospitals. The researchers have to use alternative source from the public domain to improve the error rate. The premilitary data visualization techniques and statistic showed that Cornea Cross link might Fail after year of the surgery for a age group of 15 to 19 years and after 30 years of age. More parameter testing is recommended for the future researchers to improve the outcome of the selected domain.