Lesion reversibility in small vessel disease: understanding changes contributing to vascular dementia

Abstract

Cerebral small vessel disease (SVD) affects the small perforating blood vessels in the brain and can be detected on neuroimaging as recent small subcortical infarcts, lacunes, perivascular spaces, brain atrophy, microbleeds, cortical superficial siderosis, cortical cerebral microinfarcts, incidental diffusion weighted imaging positive (DWI+) lesions and white matter hyperintensities (WMH). SVD and particularly WMH are associated with poorer clinical outcomes, cognitive decline and increased risks of stroke and dementia. In contrast, WMH regression might lead to better clinical outcomes, but, unlike WMH progression, WMH regression is understudied and underlying mechanisms to WMH changes are unknown. This thesis aims to examine longitudinal WMH changes, their severity, underlying tissue structures and the consequences of WMH progression and regression on long-term cognition and physical outcomes.

This thesis starts with an introduction to SVD, SVD features, cognitive domains, how WMH might affect cognition and methods to assess WMH change. This is followed by an overview of current literature on longitudinal WMH changes. I present results from a systematic review and meta-analysis of 41 papers including 12,284 participants, assessing longitudinal WMH volume changes within individuals with sporadic SVD. We showed that net WMH volume progression is common but around 30% of participants have net WMH volume regression. I examined longitudinal WMH in two populations: the Lothian Birth Cohort 1936 (LBC1936) (N = 675, mean age = 72.7 years [SD=0.72]) and the Mild Stroke Study 3 (MSS-3) (N=229, mean age = 65.9 years [SD=11.13]). These studies included community-dwelling older people born in 1936 and patients with minor ischaemic stroke, respectively. To understand the contribution of WMH changes to enlargement of the lateral ventricles I assessed WMH and ventricular volume changes over a six-year period in LBC1936. We found that lateral ventricle and WMH volume changes were moderately associated, independently of general brain atrophy.

To get more understanding of the role of the global white matter microstructure, I examined cross-sectional associations between Peak-width of Skeletonized Mean Diffusivity (PSMD) and global cognition at different time points in MSS-3. We found no association between global cognition and PSMD three months after stroke (N=229). At one year after stroke (N=173) we found cross-sectional associations between poorer global cognition and higher PSMD (worse global microstructural integrity). Higher baseline PSMD was also associated with poorer global cognition at one year post-stroke, suggesting that PSMD might predict long-term cognitive decline. I then assess longitudinal associations between WMH volume change and incidental DWI+ lesion with mobility (N=206) and dexterity (N=207) in MSS-3 to discover how WMH can affect physical functioning. We found associations between worsening WMH volumes and worsening mobility and dexterity, but no associations between incidental DWI+ lesions, mobility and dexterity. The in-person measures of dexterity found that the non-dominant hand was more affected by minor stroke than the dominant hand, but patients reported more impact of decline of the dominant hand on daily life.

To further understand the underlying processes of WMH change, I then examine the co-occurring WMH progression and regression within individuals. We created white matter change masks, i.e. stable normal appearing white matter (NAWM), stable WMH, regressing WMH and progressing WMH and assessed signatures of underlying tissue with diffusion tensor imaging, quantitative T1 and neurite orientation density and dispersion imaging. We found that regressing WMH have better underlying microstructure than stable WMH before any changes are visible on conventional MRI. While the underlying structure of stable NAWM, stable and progressing WMH worsen over time, the microstructure of regressing WMH might improve.

Finally, I explore different definitions of WMH regression, progression and stable WMH in the literature and apply them to MSS-3 data. I assessed whether people with WMH regression and WMH progression have different cognitive trajectories than people with stable WMH. WMH progression might be related to worse cognition and most WMH regression is associated with better executive functioning. Outcomes of cognitive trajectories related to WMH change seem to depend on what definitions of progression and regression are applied indicating a need to use consistent definitions in future studies.

To conclude, this work confirms that WMH can regress and areas of both WMH regression and progression co-occur in individuals. Regressing WMH tissue is less damaged than that of stable WMH and might improve over time. Even when acute stroke effects have settled down, WMH changes and WMH progression in particular, contribute to changes in cognition, mobility and dexterity in the long-term. WMH regression might benefit executive functioning but more research is needed and it is also crucial to reach a consensus on definitions of WMH progression and regression.