| Summary: | <h4>Background</h4> <p>MRI and laboratory features have been incorporated into international diagnostic criteria for multiple sclerosis (MS). We evaluated the pattern of MRI lesions and contributions of CSF and serum antibody findings that best identifies children with MS, and the applicability of the international diagnostic criteria in the pediatric context.</p> <h4>Methods</h4> <p>Detailed clinical evaluations, serum and CSF studies, and MRI scans were performed in youth with incident acquired demyelinating syndromes (ADS). Participants were examined prospectively to identify relapsing disease. All MRI scans were evaluated using a validated scoring tool. A random forest classifier identified imaging and laboratory features that best predicted an MS or monophasic outcome. Performance of the 2001, 2010, and 2017 International “McDonald” MS criteria, 2016 MAGNIMS criteria, and our prior proposed (Verhey) criteria were determined.</p> <h4>Findings</h4> <p>We included 324 participants with median follow-up of 73 months (range 6-150); 71 MS, 237 monophasic ADS, 14 relapsing non-MS, and 2 with alternative diagnoses. We scored 2391 brain, 444 spinal, and 67 dedicated orbital MRI scans. One or more T1 hypointense lesions plus one or more periventricular lesions (Verhey criteria) best predicted MS outcome. Performance of the 2017 McDonald criteria was comparable to the 2010 McDonald and was easier to adjudicate. The ability of CSF oligoclonal bands to substitute for the requirement for both enhancing and non-enhancing lesions improved 2017 criteria performance. MOG testing at baseline did not improve performance of the 2017 criteria.</p> <h4>Interpretation</h4> <p>The 2017 McDonald criteria for MS, as applied at the time of incident attack, perform well in identifying children and youth with MS- indicating that the same diagnostic criteria for MS apply across the age-span. The presence of “black holes” and periventricular lesions at baseline (Verhey criteria) also effectively distinguish children with MS from children with monophasic demyelination. The presence of CSF OCBs improve diagnostic accuracy. MOG antibodies identify children with ADEM, and those with relapsing non-MS, the vast majority of whom do not meet 2017 criteria at onset.</p>
|
|---|