Prognosis prediction model for multiple sclerosis patients by DEEP learning aproach

Multiple sclerosis (MS) is a chronic, autoimmune, inflammatory disease which characterized by the damage of the myelin sheaths surrounding the cells of the central nervous system, can progress from time to time with attacks and remissions. When considered global scale, the number of people who diagnosed with MS is increasing. Although the exact cause of MS is unknown, it is accepted that both genetic and environmental factors determine an individual's risk of the disease in a complex interaction that is not fully understood. It has been the main focus of the disease the fight against disability due to MS. The life expectancy of MS patients is increasing thanks to symptomatic treatments and modern rehabilitation practices. One of the main goals in the treatment processes of people diagnosed with MS is to prevent or minimize irreversible neurological damage during the treatment process. In this meaning, it is important to determine the prognosis of MS patients in the early. In the thesis study conducted for this purpose, using the data which demographic, clinical, MRGI and treatment information of the MS patients who 2 years of follow-up after the first diagnosis, a prognosis prediction model was developed for the 5th year EDSS score. Deep learning method was used in the development of the prediction model. Result of the study, calculated root mean square error between the results of the developed EDSS prognosis prediction model and the actual clinical results was obtained as 1.4.