Automated MRI Segmentation of Patients with Multiple Sclerosis: Prospects for Technology Development

Multiple sclerosis is an important medical and social problem of the modern health care system. The use of instrumental diagnostic methods, such as magnetic resonance imaging, can significantly improve the quality of diagnosis and management of patients with multiple sclerosis. At the moment, machine learning methods for analyzing magnetic resonance imaging data are not actively used. At the same time, many authors point to the prospects of using machine learning in the management of patients with multiple sclerosis and improving the outcomes of the disease in the form of reduced disability due to timely correction of its therapy with drugs that change the course of multiple sclerosis. We have proposed a method of segmentation of separate classes of demyelination foci detected by magnetic resonance imaging. The obtained parameters of the machine learning algorithm reveal the prospects of its clinical application with certain refinement, both in the diagnosis of the disease and in the assessment of the quality of its therapy.