Demonstration of lesion dissemination in space (DIS) and time (DIT) is necessary for the diagnosis of multiple sclerosis (MS) in clinically isolated syndromes (CIS). The McDonald criteria accepted two methods to demonstrate DIS. The fulfillment of at least three of four MRI Barkhof criteria (MRI-BC) or, alternatively, the finding of at least two MRI lesions on T2-weighted images (T2 lesions) plus the presence of oligoclonal IgG bands (OCGB) in cerebrospinal fluid (CSF). We aimed to evaluate the accuracy of both methods for DIS demonstration to predict conversion of CIS to MS using a new OCGB test. We studied fifty-eight CIS patients with OCGB detection and brain MRI, and followed them up during 6 years. Twenty-eight patients fulfilled MRI-BC. Twenty-five of them converted to MS during follow-up (sensitivity 73.53%, specificity 87.50%, accuracy 79.31%). Thirty-four patients had at least two T2 lesions plus oligoclonal bands. Thirty-three converted to MS during follow-up (sensitivity 94.29%, specificity 95.65%, accuracy 94.82%). The presence of oligoclonal IgG bands plus two T2 lesions accurately predicts CIS conversion to MS. MRI-BC criteria have a high specificity but less sensitivity and accuracy. These results reinforce the role of CSF study in MS diagnosis.