Objective: Spinal cord (SC) lesions are frequently found in multiple sclerosis (MS), but are rare in healthy aging and cerebrovascular patients.Our aimwas to analyze the contribution of SCinvolvement in clinically isolated syndrome (CIS) in diagnosing MS according the McDonald 2010 criteria and in predicting conversion to clinically definite MS (CDMS). Methods: We prospectively followed monofocal, relapsing onset CIS patients with either SC or brain symptom onset (including optic neuritis). MRI of the brain and SC were performed shortly after onset and patients were followed for 24 to 119 months (median 64months). SCMRI findings were assessed for their contribution to theMcDonald 2010 diagnostic criteria and their effect on conversion toCDMS. Results: One hundred twenty-one patients were included (63 spinal CIS). Based on the brain scan only, 36 patients fulfilled the McDonald criteria; by including SC findings, 6 additional patients fulfilled these criteria. To diagnose 1 additional nonspinal CIS patient, the number needed to scan is 7. In nonspinal CIS patients that did not fulfillMcDonald brainMRI criteria (n = 42), presence of an SClesionwas associated with a higher risk of conversion to CDMS (odds ratio: 14.4; 95% confidence interval: 2.6-80.0) and shorter time to conversion to CDMS (hazard ratio: 51.4; 95% confidence interval: 5.5-476.3). Conclusions: Presence of SC lesions facilitates diagnosing MS and is predictive for conversion to CDMS, especially in patients with nonspinal CIS who do not fulfill brainMRI criteria.We therefore recommend performing an SC scan in patients with nonspinal CIS who do not fulfill McDonald brain MRI criteria. © 2012 American Academy of Neurology.