Microstructural Brain Damage in Patients with SLE: An Analysis using Automated Segmentation of Nerve Tracts from Diffusion MRI

Sammanfattning: Systemic Lupus Erythematosus, SLE, is a disease with a large variation of symptoms. Some patients have neuropsychiatric symptoms, such as cognitive dysfunction and epilepsy. The classification of patients into NPSLE, with neuropsychiatric symptoms, and non-NPSLE, without these symptoms, is uncertain. Studying the white matter microstructure of SLE patients and comparing this with healthy controls can lead to better understanding of the disease and the neuropsychiatric symptoms. White matter microstructure can be characterized by the use of diffusion MRI. This imaging technique also enables segmentation of the brain into different nerve tracts. The data set used in this thesis consisted of 63 SLE patients and 20 healthy controls. Previous analysis on this data set was limited to three nerve tracts because of time-consuming manual work of the segmentation. The aim of this thesis was to extend the analysis of the white matter microstructure to include all the major nerve tracts of the brain and by that, enable group-comparisons of the whole brain. This was done by implementing the methods into a pipeline where no manual work was required. Group comparison based on the extracted values was done between healthy controls and SLE patients and between the two classifications of the disease. Significant differences of the microstructure were found when comparing SLE patients with healthy controls. In the comparison between the subgroups of SLE, with and without neuropsychiatric symptoms, the difference was not significant. This emphasizes the theory of the brain being affected in all SLE patients, and not only the ones having neuropsychiatric symptoms. The conclusion of this thesis is that the use of automated methods for extracting values of the white matter microstructure is a time-effective approach on analysis of large data sets, but it requires extensive quality control of each step.