Qualitative comparison of the preprocessing pipelines of the inferior cerebellar peduncle in pre-operative paediatric oncology data

Abstract

Cerebellar mutism syndrome (CMS) arises in some patients after the surgical removal of posterior fossa tumours. The complications include speech retardation, emotional lability, various mobility impairments and developmental delays. Even though this form of cancer makes up a big part of the mortality rates in paediatric oncology, the evidence for what causes the CMS complication is still generally unknown.

The clinical interest in cerebellar investigation relies onto differentiating anatomical properties influencing CMS pathogenesis. Diffusion-weighted imaging (DWI) of a cerebellar region can help to identify the areas, which contribute to the development of CMS. Being based on the principle of diffusion anisotropy, this technique can provide an overview of the connectivity and volumes of white matter regions in the brain. After minimising artefacts and performing the registration, a cerebellar region can be investigated with fiber tractography. The tract of interest, which in the case of this project is the inferior cerebellar peduncle (ICP), is segmented. Then parameters like mean diffusivity (MD) and fractional anisotropy (FA) can be acquired as quantitative markers of white matter microstructural differences. Along with the segmented tract profiles, obtained FA and MD can be compared between the two groups: patients who developed CMS and who did not. Identifying the regions that contribute to the development of CMS can be used when considering the risks of CMS development in surgical planning.

This report compares two pre-processing pipelines, aiming to choose the more suitable one considering the presence of the tumour. It discusses the challenges associated with the visualisation of tracts near the cerebellum and sources of errors and proposes improvements that could be considered in terms of the future development of the project. It also describes the outcomes of a deterministic tractography method used for segmentation of the inferior cerebellar peduncle.

The study concluded that the FSL pre-processing pipeline resulted in more visual improvements in a particular cohort than the pre-processing performed in ExploreDTI. However, the reliability of tractography results was lowered due to the presence of remaining artefacts. Therefore, further pipeline development is required before the segmentation of ICP and assessment of tractography parameters can make it possible to discover reasons for CMS development.