RESUMO
OBJECTIVE: We aimed to evaluate magnetic resonance imaging (MRI) previously used criteria (Matthews's criteria, MC) for differentiating multiple sclerosis (MS) from neuromyelitis optica spectrum disorders (NMOSD) in Caucasian and non-Caucasian populations (Argentina, Brazil and Venezuela) with positive (P-NMOSD), negative (N-NMOSD), and unknown (U-NMOSD) aquaporin-4 antibody serostatus at disease onset and to assess the added diagnostic value of spinal cord MRI in these populations. METHODS: We reviewed medical records, and MRIs were assessed by two blinded evaluators and were scored using MC. Short-segment transverse myelitis (STM) was added as a new criterion. MC sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were determined. RESULTS: We included 282 patients (MS = 188 and NMOSD = 94). MC applied to the entire cohort showed 97.8% sensitivity, 82.9% specificity, 92.0% PPV, and 95.1% NPV for differentiating MS from NMOSD. A subanalysis applied only to non-Caucasian (MS = 89 and NMOSD = 47) showed 100% sensitivity, 80.8% specificity, 90.8% PPV, and 100% NPV. Similar sensitivity, specificity, PPV, and NPV of MC for MS versus P-NMOSD (n = 55), N-NMOSD (n = 28), and U-NMOSD (n = 21) were observed. CONCLUSION: MC distinguished MS from NMOSD of all serostatus in a Latin American cohort that included non-Caucasian populations. Addition of STM to MC did not raise the accuracy significantly.
