Comparison of MRI Criteria for the Diagnosis of Multiple Sclerosis: Role of Cortical Lesions (S45.003)

Objective: To test the performance of different sets of imaging criteria, including the assessment of cortical lesions, for the development of multiple sclerosis (MS) in a multicentric cohort of clinically isolated syndrome (CIS) patients. Background: Since the publication of the revised McDonald 2010 criteria, new data regarding the application of MRI for the diagnosis of MS have become available. In a single-center study, adding the assessment of cortical lesions has been shown to modify the diagnostic algorithm resulting in higher specificity (Filippi 2010). Methods: Brain double inversion recovery and brain and cord T2-weighted and post-contrast T1-weighted sequences were acquired from 72 CIS patients from five European Centers within 3 months and after 12 months from disease onset. Patients were clinically followed for ≥24 months or until the development of clinically defined MS (median follow-up 24.2 months). Sensitivity, specificity and accuracy of the different DIS MRI criteria for the development of MS were tested. Results: At follow-up, 65 CIS patients (90[percnt]) had MS (clinically and /or radiologically). The sensitivity of all criteria was high (McDonald 2005 83[percnt], McDonald 2010 92[percnt], Filippi 2010 80[percnt]). Specificity of Filippi 2010 was higher (67[percnt]) compared to the others (50[percnt] for McDonald 2005 and 2010). The accuracy of all criteria was high (McDonald 2005 81[percnt], McDonald 2010 89[percnt], Filippi 2010 79[percnt]). Conclusions: Adding the evaluation of cortical lesions improves specificity of the diagnostic criteria preserving sensitivity and accuracy in a multicentric cohort of CIS patients.