Multiple sclerosis (MS) is a chronic, demyelinating disease affecting the central nervous system. Diagnosis of MS is based on the proof of disease dissemination in time (DIT) and dissemination in space (DIS) and excluding other disorders that can mimic multiple sclerosis in laboratory tests and clinical manifestation. Over the years the diagnostic criteria have evolved; the introduction of magnetic resonance in the McDonald's 2001 criteria was revolutionary. Since then, the criteria have been modified up to the currently used McDonald 2017. The aim of this review is to analyse the 2017 McDonald criteria, assess what has changed from the 2010 criteria, and present the impact of revised criteria on rapid and accurate diagnosis of MS. The main differences are as follows: inclusion of oligoclonal bands in cerebrospinal fluid as a DIT criterion, and symptomatic and cortical lesions in magnetic resonance imaging are counted in the determination of DIS and DIT. We present also the newest recommendations of the Polish Medical Society of Radiology and the Polish Society of Neurology and international group of North American Imaging in Multiple Sclerosis and Consortium of Multiple Sclerosis Centers, as well as future directions for further investigations. A proper diagnosis is crucial for the patient's quality of life, to give the possibility of early treatment, and to help avoid misdiagnosis and unnecessary therapy.