Kim Annink

64 Chapter 3 edema was not associated with PC injury, so even without cytotoxic edema there still might be cerebellar injury Our study suggests that ADC values of the vermis are most reliable, it does show differences between the HIE and control group and it is associated with the normal PC count and percentage. Additionally, other MR sequences might be more useful to diagnose cerebellar injury in neonates with HIE. For example, Lemmon et al. showed that fractional anisotropy based on diffusion tensor imaging (DTI) was associated with cerebellar injury and outcome (24). There are some limitations to this study. First, this was a retrospective study, meaning that not all microscopic slices were cut at the same position in the cerebellar hemispheres or vermis and only one slice per structure was available. Also, different protocols for DWI were used. Both might have caused some variability in the results. In addition, the control group of infants with a surgical intervention is not a healthy control group. However, by verifying that their MRI showed no supratentorial or infratentorial abnormalities and their neurocognitive follow-up was normal at 24 months of age, it is justifiable to use this group as a control group. Furthermore, for the histopathological group, it would be useful to have a control group; however, there is no autopsy tissue of healthy term born neonates available. Finally, our study population was rather small, especially the number of slices of the vermis was small. A larger study population would have provided more statistical power and would have enabled us to investigate the effect of the postnatal age of death on the results. Nevertheless, to the best of our knowledge, this is one of the largest studies to date comparing MRI to histopathology in neonates with HIE. It is a unique dataset with neonatal human postmortem tissue and we were able to use quantitative MRI measurements, which was proven to be reproducible by different raters. Future research should focus on elaborate staining of the cerebellum to study the injury in infants with HIE in more depth. Additionally, it would be interesting to explore the pattern of PC injury in the cerebellum in more detail. Finally, more research is needed to study the association between ADC values of the cerebellum and neurocognitive outcome. Prospective research is necessary to answer this question.