Before the 2001 McDonald criteria were adopted, the diagnosis of multiple sclerosis (MS) was based on clinical findings; as a result, the times to diagnosis and to starting disease-modifying therapy (DMT) were substantial. The time spent waiting at the clinically isolated syndrome (CIS) stage until the diagnostic criteria for relapsing-remitting MS (RRMS) are fulfilled exposes patients to the risk of more disease activity, tissue damage, and disability. The 2001 McDonald criteria were the first to include imaging requirements1; their implementation (and that of the revisions made in 2005, 2010, and 2017) led to a reduction in time to diagnosis and an earlier initiation of DMT. But these criteria are not perfect. Whereas the majority of patients with CIS develop clinically definite MS (CDMS), some never experience another clinical demyelinating event and thus would not be considered to have CDMS. In the London CIS cohort, for example, a clinical cohort of people presenting with typical CIS, a third did not develop CDMS after 30 years of follow-up.2 Ideal diagnostic criteria therefore need high sensitivity—the ability to identify the disease when it is present—but also high specificity—the ability to rule out MS when it is not present. Specificity is important to avoid creating false positives and mislabeling stable CIS as RRMS.