MRI in the diagnosis of pediatric multiple sclerosis

• Published in: Neurology Vol. 72; no. 11; p. 961
• Main Authors: Callen, D J A, Shroff, M M, Branson, H M, Lotze, T, Li, D K, Stephens, D, Banwell, B L
• Format: Journal Article
• Language: English
• Published: United States 17.03.2009
• Subjects: Adolescent; Adult; Age Factors; Aged; Brain - pathology; Child; Diagnosis, Differential; Encephalomyelitis, Acute Disseminated - pathology; Female; Humans; Image Interpretation, Computer-Assisted; Lupus Erythematosus, Systemic - pathology; Magnetic Resonance Imaging - methods; Male; Middle Aged; Multiple Sclerosis - diagnosis; Multiple Sclerosis - epidemiology; Multiple Sclerosis - pathology; Myelitis, Transverse - pathology; Nervous System Diseases - diagnosis; Nervous System Diseases - pathology; Observer Variation; Optic Neuritis - pathology; Regression Analysis; Retrospective Studies; Sample Size; Young Adult
• ISSN: 1526-632X
• DOI: 10.1212/01.wnl.0000338629.01627.54

MRI diagnostic criteria have not yet been adopted for pediatric multiple sclerosis (MS). MRI plays a pivotal role in supporting the diagnosis of MS in adults. We sought to quantitatively define the MRI features of pediatric MS, to determine features that distinguish MS from nondemyelinating relapsing childhood neurologic disorders, and to propose MRI criteria for lesion dissemination in space in children with MS. A retrospective analysis of MRI scans from 38 children with clinically definite MS and 45 children with nondemyelinating diseases with relapsing neurologic deficits (migraine, systemic lupus erythematosus) was performed. For each scan, T2/FLAIR hyperintense lesions were quantified and categorized according to location and size. Mean lesion counts in specific locations were compared between groups to derive diagnostic criteria. Validation of the proposed criteria was performed using MRI scans from a second independent MS cohort (n = 21). MRI lesion location and size categories differed between children with MS and nondemyelinating controls with a medium to large effect size for most variables. The presence of at least two of the following-five or more lesions, two or more periventricular lesions, or one brainstem lesion-distinguished MS from other nondemyelinating disease controls with 85% sensitivity and 98% specificity. We propose modifications to the currently established McDonald MRI criteria for lesion dissemination in space that will enhance the diagnostic accuracy of these criteria for multiple sclerosis in children.