Identification of Disease-Specific Neuroimaging Phenotypes in Childhood Inflammatory Brain Diseases

Tania Cellucci¹, Pascal N. Tyrrell², Shehla Sheikh³, Suzanne Laughlin⁴ and Susanne M. Benseler³, ¹Pediatrics/Rheumatology, The Hospital for Sick Children, Toronto, ON, Canada, ²Cardiology/Rheumatology, The Hospital for Sick Children, Toronto, ON, Canada, ³Rheumatology, The Hospital for Sick Children, Toronto, ON, Canada, ⁴Diagnostic Imaging, The Hospital for Sick Children, Toronto, ON, Canada

Meeting: ACR/ARHP Annual Meeting 2012

Keywords: central nervous system involvement, Inflammation, neuroimaging and pediatric rheumatology

Session Information

Title: Pediatric Rheumatology: Clinical and Therapeutic Disease III: Childhood Systemic Lupus Erythematosus and Other Vasculidities

Session Type: Abstract Submissions (ACR)

Background/Purpose:

Magnetic resonance imaging (MRI) is a key diagnostic modality of childhood inflammatory brain diseases (IBrainD). While overlapping clinical features contribute to significant diagnostic uncertainty, neuroimaging characteristics may provide guidance for a targeted diagnostic approach. The aim of this study was to identify specific MRI patterns at diagnosis of distinct childhood IBrainD.

Methods:

This single centre cohort study included children less than 18 years old who were diagnosed with an IBrainD between June 1989 and December 2010 and were enrolled in the BrainWorks cohort at SickKids Hospital. All patients also had a high-quality brain MRI based on institutional protocol at diagnosis. Demographic, clinical, laboratory, neuroimaging and histologic data at diagnosis were collected. Correspondence analysis was performed to obtain a multidimensional representation of the relationship between MRI findings and IBrainD diagnoses. Pearson residuals (PR) were calculated to estimate the strength of associations.

Results:

A total of 142 children (51% females, median age 8.8 years) with IBrainD were identified: 104 primary angiitis of the CNS (cPACNS), 11 secondary CNS vasculitis, 8 neuronal antibody syndromes, 6 post-infectious IBrainD and 13 unclassified IBrainD. Children with angiography-positive non-progressive cPACNS were likely to have unilateral, unifocal and ischemic lesions (PR 4.2, 3.4, and 3.0, respectively). Common sites for lesions were the basal ganglia (PR 3.5), middle cerebral artery territory (PR 1.8), and internal capsule (PR 1.3). Angiography-negative cPACNS was grouped together with multifocal lesions (PR 1.64) and involvement of the leptomeninges (PR 2.3), parietal or temporal lobes (PR 1.5, 2.8), and subcortical or deep white matter (PR 2.8, 1.5). Optic neuritis (PR 2.9), volume loss (PR 1.8) and frontal lobe lesions (PR 1.5) were more likely in neuronal antibody syndromes. Frontal lobe involvement was also seen in post-infectious IBrainD (PR 1.6). Symmetric lesions were grouped with secondary CNS vasculitis (PR 2.0) and unclassified IBrainD (PR 2.7).

Conclusion:

Childhood inflammatory brain diseases are characterized by distinct patterns of neuroimaging findings with respect to location and type of lesion. Identifying these patterns requires a standardized approach involving high-quality MRI for any child with suspected IBrainD. The neuroimaging phenotype may guide further diagnostic work-up and should be incorporated into diagnostic algorithms for childhood IBrainD in the future.

Disclosure:

T. Cellucci,
None;

P. N. Tyrrell,
None;

S. Sheikh,
None;

S. Laughlin,
None;

S. M. Benseler,
None.

ACR Meeting Abstracts - https://acrabstracts.org/abstract/identification-of-disease-specific-neuroimaging-phenotypes-in-childhood-inflammatory-brain-diseases/