Traditionally, multiple sclerosis (MS) has been diagnosed on the basis of clinical evidence of dissemination in time and space. Previously, it could not be diagnosed in patients with single clinical episodes of demyelination known as clinically isolated syndromes. New diagnostic criteria from the International Panel of McDonald and colleagues incorporate MRI evidence of dissemination in time and space to allow a diagnosis of MS in patients with clinically isolated syndromes. From clinical and MRI examinations performed prospectively at baseline, 3 months, 1 year, and 3 years of follow-up, the frequency of developing MS was ascertained by the application of both the new McDonald criteria and the Poser criteria for clinically definite MS. The specificity, sensitivity, positive and negative predictive value, and accuracy of the new criteria for the development of clinically definite MS were assessed. At 3 months, 20 of 95 (21%) patients had MS with the McDonald criteria, whereas only 7 of 95 (7%) had developed clinically definite MS. After 1 year, the corresponding figures were 38 of 79 (48%) and 16 of 79 (20%), and after 3 years, they were 29 of 50 (58%) and 19 of 50 (38%). The development of MS with the new MRI criteria after 1 year had a high sensitivity (83%), specificity (83%), positive predicative value (75%), negative predictive value (89%), and accuracy (83%) for clinically definite MS at 3 years. Use of the new McDonald criteria more than doubled the rate of diagnosis of MS within a year of presentation with a clinically isolated syndrome. The high specificity, positive predictive value, and accuracy of the new criteria for clinically definite MS support their clinical relevance.