Rhombencephalitis in Children: Diffusion Magnetic Resonance Imaging (MRI) Correlation With Clinical Outcomes

Abstract

In children with rhombencephalitis, neuroimaging abnormalities have been infrequently correlated with clinical outcome. We investigated whether magnetic resonance (MR) neuroimaging studies could predict clinical outcomes and disturbance of consciousness in patients with rhombencephalitis. We retrospectively analyzed the MR studies of 19 pediatric patients with rhombencephalitis (median age: 4.2 years, range 0.5-17; sex: 32% male). Fluid-attenuated inversion recovery imaging and diffusion-weighted imaging findings were graded to create imaging scores according to the extent of imaging abnormality. Clinical outcomes in the first week and 12th month were graded by using Glasgow Outcome Scale scores (1-5) and dichotomized to unfavorable or favorable outcome. Correlations of the imaging scores with the clinical outcomes and with disturbance of consciousness were assessed by using multivariate logistic regression analysis. No significant correlation was found between fluid-attenuated inversion recovery score or diffusion-weighted imaging score ( P = .608, P = .132, respectively) and disturbance of consciousness. In the first week, the unfavorable outcome group (n = 11) had significantly higher diffusion-weighted imaging score than did the favorable outcome group (n = 8) (Mann-Whitney U test, P = .005). Multivariate logistic regression analysis showed that the diffusion-weighted imaging score (odds ratio, 18.182; 95% confidence interval: 1.36, 243.01; P = .028) was significantly associated with unfavorable outcome. In the 12th month, the fluid-attenuated inversion recovery score or diffusion-weighted imaging score ( P = .994, P = .997, respectively) were not significantly associated with unfavorable outcome. Patients with rhombencephalitis who have a higher diffusion-weighted imaging score are more likely to have an unfavorable 1-week clinical outcome.