BACKGROUND: MRI allows demonstration of dissemination in space and time at the first demyelinating event. However, no pediatric MS study has investigated the validity of MS-specific outcomes described in MRI radiological reports that clinicians rely on to make the MS diagnosis and to assess the MS treatment effect. Our aim was to validate MS-specific outcomes in hospital MRI reports in pediatric MS by comparing MS-specific outcomes in MRI reports with secondary MRI review.
METHODS: A senior consultant and a resident neurologist extracted data on MS-specific outcomes from MRI reports at baseline and follow-up in children with MS onset during 2008-15 in Denmark. Gold standard was an expert neuroradiologist's secondary MRI review. We estimated percent agreement and Kappa values by comparing data extracted from hospital MRI reports (what we wanted to test) with results from the secondary MRI reviews (our gold standard).
RESULTS: Among 55 children with MS, we included 44 baseline and 48 follow-up MRIs. The median age at MS onset was 16.3 years (range 9.2‒17.9). Agreement between the MRI reports and the secondary MRI review ranged 68%-100% for MS-specific outcomes; agreement was higher when MRI outcomes were present. Kappa values ranged from 0.42 ("moderate") to 1.00 ("excellent"). Kappa for fulfillment of the McDonald 2017 criteria was 0.60 on baseline MRI, and 0.53 on follow-up MRI. Kappa for a new lesion on follow-up MRI was 0.41.
CONCLUSION: Agreement was moderate to good for most MS-specific outcomes between MS neurologists' data extraction from hospital MRI radiological reports compared with a neuroradiologist's secondary MRI review. The agreement was moderate for both fulfilling the McDonald 2017 criteria and acquiring a new lesion on follow-up MRI. We recommend structured MRI reporting in children suspected of acquired demyelinating syndromes to increase validity of hospital MRI reports and clinical use.