MRI analysis and neurodevelopmental outcome in preterm infants

Abstract

With the increasing numbers of preterm infants who survive, it is of great benefit to be able to predict outcome to improve parental counselling and to identify neonates susceptible for neurological disabilities who could benefit from ‘developmental support’ programs. The aim of this thesis was to quantify magnetic resonance imaging (MRI) findings at term-equivalent age (TEA) in a cohort of preterm infants. Additionally MRI postprocessing techniques and diffusion tensor imaging (DTI) were implemented. It was assessed whether they could be used in the prediction of neurodevelopmental outcome at two years’ corrected age. We segmented the neonatal brain at TEA using an in-home developed, user-independent, automatic probabilistic segmentation method which is based on k-Nearest Neighbour. This is the first method that can segment cortical and central gray matter, myelinated and unmyelinated white matter (WM), cerebrospinal fluid in the ventricles and cerebrospinal fluid in the extra-cerebral space, brainstem and cerebellum separately. We found that relative volume changes of cerebral structures in reference to the intracranial volume (ICV) occurred in preterm infants around TEA. After correction for both ICV and postmenstrual age (PMA), ventricular volumes were negatively associated with cognition, fine motor and gross motor skills at two years’ corrected age in addition to neonatal parameters (gestational age, birth weight z-score, intraventricular haemorrhage and WM injury on conventional MRI) and maternal education. Additionally, the cerebellar volume corrected for PMA was related to cognitive skills after correction for the neonatal and maternal confounders. These findings revealed the possibility to use brain volumes at TEA as biomarkers for neurodevelopmental outcome in preterm infants. Besides the volume, the cerebellar N-acetylaspartate/Choline ratio obtained with proton MR spectroscopy was associated with cognition. Using DTI, we showed that fiber tracking of the corpus callosum (CC) and bundles through the posterior limb of the internal capsule (PLIC) can be performed in preterm infants at TEA. Bundle volumes and lengths and diffusion parameters in these bundles were associated with WM injury. Therefore quantitative fiber tracking can provide additional insight into the underlying changes in WM microstructure compared to conventional T1- and T2-weighted images. Fiber tracking parameters in the CC in girls and in the left PLIC for boys were associated with cognitive scores, fine motor and gross motor performance. Tract-based spatial statistics, optimized for the neonatal brain, can be used to assess the relationship between diffusion parameters at TEA and neurodevelopmental outcome at two years’ corrected age. An increase in fractional anisotropy en a decrease in radial diffusivity in specific WM regions are related to cognition, fine motor and gross motor outcome. The results of our study exhibited the potential use of diffusion parameters at TEA as biomarkers for neurodevelopment in preterm infants. We assessed the anatomy of the circle of Willis in our preterm cohort. The infants demonstrate a high prevalence of variant types of the circle of Willis at TEA. Variations in the anatomy of the circle of Willis are related to alterations in the flow through the internal carotid arteries and the basilar artery.