Background: Magnetic resonance imaging diagnostic criteria for paediatric multiple sclerosis have been established on the basis of brain imaging findings alone. The 2010 McDonald criteria for the diagnosis of multiple sclerosis, however, include spinal cord imaging for detection of lesion dissemination in space. The new criteria have been recommended in paediatric multiple sclerosis.
Objective: (1) To evaluate the 2010 McDonald multiple sclerosis criteria in children with a clinically isolated syndrome and to compare them with recently proposed magnetic resonance criteria for children; (2) to assess whether the inclusion of spinal cord imaging provided additional value to the 2010 McDonald criteria.
Methods: We performed a retrospective analysis of brain and spinal cord magnetic resonance imaging scans from 52 children with a clinically isolated syndrome. Sensitivity, specificity and accuracy of the magnetic resonance criteria were assessed.
Results and conclusion: The 2010 McDonald dissemination in space criteria were more sensitive (85% versus 74%) but less specific (80% versus 100%) compared to the 2005 McDonald criteria. The Callen criteria were more accurate (89%) compared to the 2010 McDonald (85%), the 2005 McDonald criteria for dissemination in space (81%), the KIDMUS criteria (46%) and the Canadian Pediatric Demyelinating Disease Network criteria (76%). The 2010 McDonald criteria for dissemination in time were more accurate (93%) than the dissemination in space criteria (85%). Inclusion of the spinal cord did not increase the accuracy of the McDonald criteria.