Magnetic resonance imaging in dementia: a study of brain white matter changes

Non-specific white matter changes (WMC) in the brain are common findings in the elderly population. Although they are frequently seen in non-demented persons, WMC seem to be more common in demented patients. The significance of these changes, as well as their pathophysiological background, is incompletely understood. The aim of this thesis was to study different aspects of WMC using MR imaging (MRI) and to investigate the clinical significance of such changes in subjects with mild cognitive impairment or dementia. In study I post-mortem MRI of the brain was compared to corresponding neuropathology slices. WMC were quantified and found to be more extensive on neuropathology. The areas that appeared normal on MRI but not on histopathology represented only minor changes with increased distance between the myelinated fibres but with preserved axonal network and glial cell density. Study II evaluated the blood-brain barrier (BBB) integrity to investigate if an increased permeability could be shown in WMC. A contrast-enhanced MRI technique was used to detect small degrees of enhancement. No general increase in BBB could be detected in the WMC areas. In study III the relation between WMC and apolipoprotein E (APOE) genotype was explored in patients with Alzheimer's disease (AD). Results showed that AD patients, who were homozygous for the APOE epsilon 4 allele had more WMC than patients with other genotypes. This was most significant for changes in the deep white matter. Results also indicated that in AD patients carrying the epsilon 4 allele, WMC are not age-related phenomena, but might be related to the aetiology of the disease. Study IV aimed to investigate if WMC in a specific brain region affect cognitive functions related to that area. Periventricular WMC in the left frontal lobe predicted a decrease in initial word fluency, a test though to reflect left frontal lobe functioning. This indicates that WMC might have specific effects in different brain regions. In study V we evaluated the prognostic significance of WMC in patients with memory impairment, regarding the rate of further global cognitive decline. There was no difference in outcome between patients having extensive WMC and a matched control group, during 2-4 years of follow up, and assessed by the "Mini-Mental State Examination". In conclusion, this work has shown and characterised pathological changes in the white matter not visible on conventional MRI. We have also shown that there is no major general increase in BBB permeability in areas of WMC. In addition, homozygosity with regard to the APOE epsilon 4 gene allele implies an increased extent of WMC in AD patients. In AD patients carrying this gene allele, WMC are not merely age-related phenomena, but might be related to the aetiology of the disease. We also claim that WMC in a specific location might impair cognitive functions that rely on those specific pathways. In contrast, WMC do not seem to have any prognostic value in predicting the rate of global cognitive decline in patients at a memory clinic.