White Matter Integrity Differences in 2-year-old Children Treated with ECMO: A Diffusion-Weighted Imaging Study

Abstract

School-aged and adolescent survivors of neonatal extracorporeal membrane oxygenation (ECMO) treatment still suffer from neurodevelopmental delays such as verbal, visuo-spatial and working memory problems, motor dysfunction and sensorineural hearing loss, respectively, later in life. These neurodevelopmental delays are normally assessed by neuropsychological testing within follow-up programs. The purpose of this study is to demonstrate that diffusion-weighted imaging (DWI) in 2-year-old survivors of neonatal ECMO treatment might be a predictor of neurodevelopmental outcome. Therefore, 56 children underwent DWI at 3 T. Fractional anisotropy (FA), first fibre partial volume fraction estimate (F1) and radial diffusivity (RD) are compared using tract-based spatial statistics adapted to a paediatric brain atlas and whole-brain voxelwise statistics with age and gender as covariates of no interest. A significant difference in FA, F1 and RD between no-ECMO and ECMO group is seen in major white matter tracts and subcortical white matter in gyri leading to the conclusion that these differences are driven by alterations in axon coherence. Additionally, we examine individual diffusion measures by looking at masks from 50 brain regions taken from a paediatric brain atlas. We find left anterior corona radiata, left and right corpus callosum (genu, body and splenium), left and right crus of fornix, left anterior limb of internal capsule, left anterior commissure, left tapetum and right uncinate fasciculus to have significantly different means in no-ECMO compared to ECMO group which matches the reports of neuropsychological delays found in behavioural tests. To conclude, analysing diffusion measures at an early stage of life serves as a good tool to detect structural white matter changes in survivors of neonatal ECMO treatment like lacking axon coherence in fibre bundles which develop early in life. The advantage of DWI lies in looking only at the neurobiology, e.g. white matter integrity. Compared to neuropsychological testing, DWI in this age range is a very time-efficient method which does not depend on the child’s active participation. Additional targeted training could help to mitigate the neurodevelopmental deficits ECMO survivors face later in life.