Introduction
Cerebral small vessel disease (SVD) is a common disease process accounting for a quarter of all ischaemic strokes, around 80% of haemorrhagic strokes, and is the major contributor to vascular cognitive impairment and dementia. Despite it being a major public health burden, understanding of the natural history is incomplete and the specific pathophysiological processes involved have not been fully elucidated. Consequently, there are few effective disease modifying treatments. In part I of this thesis I aimed to clarify elements of the natural history of white matter hyperintensity lesions (WMHs) in SVD. These are a key radiological feature of SVD that are strongly correlated with clinical sequelae, and I further tested whether brain lesion volume can regress over time. In part II I investigated the role of two novel pathophysiological mechanisms (inflammation and the permeability of the blood-brain barrier), and their relationship with SVD severity and progression.

Methods
I performed a systematic review of WMH growth and used inverse variance-weighted meta-analysis to determine the expected WMH change over time in high-risk populations. I next used a novel timepoint-blind WMH marking technique to assess whether WMH volume regresses over time in three separate SVD cohorts. Finally I studied a cohort of patients with SVD undergoing PET-MRI imaging, phlebotomy and neuropsychometric testing.

Results
WMHs typically expand at 2.50 ± 3.02 cc/year in patients with SVD and this is significantly more likely in patients with hypertension and who currently smoke. I found only 12/417 participants (2.9%) who showed modest WMH regression on longitudinal imaging, and this was more likely in patients with less severe disease at baseline. I demonstrated significant differences between patient and control groups in both microglial signal and blood-brain barrier permeability, and associations between microglial signal and both clinical and radiological markers of SVD severity. These disease processes did not predict disease progression at one year.

Conclusions
I calculated the expected rate of WMH growth in relevant populations and how these data affect the sample sizes required to show a treatment effect, which should inform future trials. My results investigating WMH regression suggest that this is unlikely to be a significant factor in severe SVD.

I showed that both microglial signal and blood-brain barrier permeability are likely to be relevant in SVD, but whether they are disease causing remains unclear. I further discussed the ongoing interventional study in which the data collection for this thesis was nested (MINERVA). The MINERVA trial aims to answer this question by testing whether minocycline can inhibit activated microglia and stabilise the blood-brain barrier, and I presented baseline data from the trial.