Diffusion-weighted imaging predicts cognition in pediatric brain injury

Abstract

Apparent diffusion coefficient maps from diffusion-weighted imaging predict gross neurologic outcome in adults with traumatic brain injury. Few studies in children have been reported, and none have used apparent diffusion coefficient maps to predict long-term (>1 year) neurocognitive outcomes. In this study, pooled regional and total brain diffusion coefficients were used to predict long-term outcomes in 17 pediatric brain injury patients. Apparent diffusion coefficient values were grouped into peripheral and deep gray and white matter, posterior fossa, and total brain. Regions of interest excluded areas that appeared abnormal on T(2)-weighted images. Apparent diffusion coefficient values from peripheral regions were inversely correlated with cognitive functioning. No significant correlations were apparent between the cognitive scores and apparent diffusion coefficient values for deep tissue or the posterior fossa. Regression analyses suggested that combined peripheral gray and white matter apparent diffusion coefficients explained 42% of the variance in the combined neurocognitive index. Peripheral gray diffusion coefficients alone explained an additional 20% of variance after accounting for clinical variables. These results suggest that obtaining apparent diffusion coefficient values, specifically from peripheral brain regions, may predict long-term outcome after pediatric brain injury. Discrepancies in the literature on this topic, as well as possible explanations, including sampling and clinical considerations, are discussed.