Real-Life Evaluation of the MOGAD Diagnostic Criteria: Application Challenges and Discrepancies

Myelin oligodendrocyte glycoprotein (MOG) antibody-associated disease (MOGAD) international panel criteria have been recently proposed to guide MOGAD diagnosis. The aim of this study was to evaluate the criteria performance and assess the discrepancies in their application in the clinical practice in an Italian multicenter cohort and to discuss some challenging aspects.

We applied the 2023 MOGAD criteria to patients who tested MOG-Abs positive on cell-based assays and were retrospectively recruited from 29 centers. Detailed clinical and paraclinical data were collected. Patients were classified as true positive/negative (TP/TN) in case of concordance between MOGAD criteria application and enrolling center final diagnosis, as false positive (FP) when MOGAD criteria were fulfilled but final diagnosis was different from MOGAD, and as false negative (FN) when MOGAD criteria were not fulfilled and final diagnosis of MOGAD was confirmed. Central revision of FN and FP cases was performed.

We included 214 patients (median age at onset 38.2 years [interquartile range 25.2-50.7], 60.3% female, 23 pediatric patients). Of these, 168 (78.5%) were classified as TP, 9 (4.2%) as FP, 23 (10.7%) as FN, and 14 (6.5%) as TN. The sensitivity of MOGAD criteria was 87.96% (CI 82.5%-92.2%), specificity 60.9% (CI 38.5%-80.3%), positive predictive value 94.9% (CI 91.8%-96.9%), negative predictive value 37.8% (CI 26.7%-50.2%), and accuracy 85.1% (CI 79.6%-89.5%). In 11 of 32 revised cases, available information did not allow a proper diagnosis. Independent revision changed the diagnosis in 17 of 21 remaining cases, increasing the performance of the MOGAD criteria. Of note, in 3 cases, diagnostic criteria were satisfied only at follow-up. The sensitivity and specificity after independent revision were 98.9% (CI 96%-99.9%) and 91.7% (CI 73%-98.9%), respectively. Moreover, 29 of 214 patients (13.6%) had 1 or more asymptomatic radiologic supportive features, and in 50% (3/6) of FP cases, independent revision did not confirm the presence of supportive features. Patients with clear positive serum titer or CSF-only MOG-Abs were those who received more commonly a MOGAD diagnosis.

MOGAD criteria demonstrate a good performance across different centers; however, controversial cases might benefit from collegial discussion and reassessment of MOGAD criteria during the follow-up. Main challenges include availability of proper radiologic data and interpretation of radiologic supportive features.