Assist Diagnostic System of Multiple Sclerosis Through Analysis of Optical Coherence Tomography Records

There is no biomarker for the diagnosis of multiple sclerosis. The McDonald criteria are used to make the definitive diagnosis of the disease, which are updated according to the evolution of knowledge. The next revision of these criteria will propose the use of retinal structural information obtained with optical coherence tomography. In this work, two methods of explainable artificial intelligence are implemented over a support vector machine classifier for the diagnosis of a group of patients (n=79) and a control group (n=69). The results obtained by both methods are analyzed through a comparative study. In the best case, a diagnostic accuracy of 92.57% and the features characteristics with greater diagnostic capability are identified. The results suggest the usefulness of artificial intelligence to contribute to the diagnosis of multiple sclerosis in its initial stages.