Does white matter matter? Neuroimaging in dementia with Lewy bodies and related conditions

In this thesis, the role of white matter and white matter changes in dementia with Lewy bodies (DLB) is explored through neuroimaging. The overarching aim of this thesis project was to further increase the understanding of the role of cerebrovascular disease (CVD) in DLB. We wanted to elucidate the relationship between WMSA and markers of neurodegeneration as well as with clinical markers, such as cognition. Through the four projects included in the thesis, the role of white matter changes is evaluated using a range of research cohorts, imaging modalities, and statistical methods.

As we age, there is an increasing risk of developing a cognitive or dementia disorder. In clinical practice, cognitive disorders are mainly diagnosed through the clinical presentation whilst in research biomarkers of pathology are more often used. There is a challenge to accurately diagnose these conditions, particularly as the symptoms can overlap and it is common with multiple pathologies in patients. In DLB, which is a relatively common dementia, patients often present with concomitant Alzheimer's Disease (AD) and with signs of white matter changes on neuroimaging. These concomitant pathologies can also alter the presentation of DLB. These findings among patients with DLB may lead to a delayed diagnosis or delay in treatment.

Changes in the white matter can be evaluated through magnetic resonance imaging (MRI) and computer tomography (CT) through white matter signal abnormalities (WMSA). Furthermore, the white matter can be evaluated with a specific MRI sequence, diffusion-weighted imaging (DWI), which allows for the exploration of the microstructure of tissues. The images can then be evaluated through either clinically oriented visual rating scales or through more research oriented automatic methods. The WMSA are usually considered to be of a vascular origin.

In the first project, the aim was to evaluate the frequency of high WMSA in DLB compared to other diagnoses. To evaluate this, neuroimaging of over 4,000 participants across seven different diagnostic groups and four cohorts was used. Using a clinically available rating scale for WMSA, the Fazekas scale, which was binarised into a low (0-1) or high (2-3) rating, it was observed that patients with DLB had a higher likelihood of a high rating compared to patients with AD and with controls, but a lower likelihood of a high rating compared to patients with vascular dementia (VaD). Furthermore, in patients with DLB, a high rating of Fazekas was associated with a high rating of medial temporal lobe atrophy. However, among patients with DLB, a high WMSA rating was not associated with lower performance on the cognitive test the Mini-Mental State Examination (MMSE), however; in patients with AD, a high WMSA was associated with lower MMSE performance.

In the second project, the aim was to evaluate the regional placement of WMSA. In DLB, AD and other dementias, the cholinergic system is affected. The cholinergic system is important for attention and memory. Patients with DLB may also have more severe impairments of the cholinergic system compared to patients with AD. Therefore, the cholinergic system is an important target for investigation. Leveraging the finding that patients with DLB have more WMSA compared to controls, the aim was to evaluate if these WMSA were more likely to be placed in the cholinergic system in patients than in controls. Using an advanced imaging technique based on DWI, masks of the cingulate and external capsule pathways of the cholinergic system were created. The WMSA in the whole brain was assessed with an automated method and WMSA overlapping the mask of the pathways of the cholinergic system was considered cholinergic WMSA. We found that patients with DLB or prodromal-DLB had more WMSA in their cholinergic system pathways, both as a raw volume and as a proportion of all WMSA.

In the third project, the aim was to evaluate the microstructure of both WMSA and normal-appearing white matter (NAWM). The aetiology of WMSA is multifactorial and its correlates can vary in postmortem analysis. Using DWI, the differences in microstructural tissue composition of patients with DLB or prodromal DLB compared to controls from two cohorts were evaluated. The method used for evaluating the microstructure is called Single-Shell 3-Tissue Constrained Spherical Deconvolution (SS3T-CSD) and allows for a probabilistic assumption of tissue fractions in each voxel of WMSA or NAWM. The patients with DLB or prodromal DLB in both cohorts differed in terms of microstructural tissue composition in the NAWM, and the results were suggestive of a neurodegenerative process. The results in WMSA suggested group differences only in the WMSA closest to the ventricles, which could be suggestive of less an inflammatory process among patients.

In the fourth project, the aim was focused on clinically recruited patients with either mild cognitive impairment (MCI) or subjective cognitive impairment (SCI) from memory clinics in the Stockholm area. Using their results on neuropsychological test results, patients were subgrouped based on three different strategies. Two of the strategies were based on an a-priori-determined division based on the patient's performance on the tests whilst the third strategy was based on a data-driven clustering method. These subgroups were evaluated in terms of associations with abnormal ratings of WMSA. The only subgroup with a higher likelihood of WMSA was the data-driven subgroup with patients who had the most pronounced neuropsychological difficulties.

To summarise, in this thesis, WMSA in DLB has been evaluated. The conclusions are that a high degree of WMSA is common in DLB, the WMSA is associated with more atrophy, and is likely to be placed in the cholinergic system of patients with DLB or prodromal DLB. Further, we found that the microstructural tissue composition of WMSA and NAWM is different in patients with DLB or prodromal DLB than in controls.

List of scientific papers

I. Rennie, A., Ekman, U., Shams, S., Rydén, L., Samuelsson, J., Zettergren, A., Kern, S., Oppedal, K., Blanc, F., Hort, J., Garcia-Ptacek, S., Antonini, A., Lemstra, A. W., Padovani, A., Kramberger, M. G., Rektorová, I., Walker, Z., Snædal, J. G., Pardini, M., ... , & Ferreira, D. (2024). Cerebrovascular and Alzheimer's disease biomarkers in dementia with Lewy bodies and other dementias. Brain communications. https://doi.org/10.1093/braincomms/fcae290

II. Rennie, A., Nemy, M., Jerele, C., Rodríguez-Baz, Í., Montal V4., Bejanin, A., Kramberger, M., Aarsland, D., Fortea, J., Lleó, A., Westman E., Alcolea, D., Ferreira, D. Regional associations between cerebrovascular disease and cholinergic white matter pathways in the Lewy body continuum. [Manuscript]

III. Rennie, A., Nemy, M., Badji, A., Habich, A., Castellanos-Perilla, N., Firbank, M., Donaghy, P., Rodriguez-Baz, I., Bejanin, A., Lleó, A., Fortea, J., O'Brien, J., Thomas, A., Aarsland, D., Alcolea, D., Taylor, J-P., Westman, E., Ferreira, D. Tissue compositional analysis reveals microstructural alterations in the white matter of patients with prodromal and manifest dementia with Lewy bodies. [Manuscript]

IV. Rennie, A., Ekman, U., Wallert, J., Muehlboeck, J. S., Eriksdotter, M., Wahlund, L .- O., Ferreira, D., & Westman, E. (2023). Comparing three neuropsychological subgrouping approaches in subjective and mild cognitive impairment from a naturalistic multicenter study. Neurobiology of aging, 129, 41-49. https://doi.org/10.1016/j.neurobiolaging.2023.04.008