MRI signatures associated with active ischemia and disease severity in cerebral small vessel disease

Abstract

Objective

Cerebral small vessel disease, a leading cause of stroke and cognitive impairment, manifests on neuroimaging with white matter hyperintensities (WMH) and disrupted microstructure in normal-appearing white matter. WMH, by definition have high T2 FLAIR signal; however, both T2 FLAIR and T1 signal in WMH are highly variable. We hypothesized that signal intensity parameters would differ in cerebral small vessel disease compared to healthy controls and that signal heterogeneity would be associated with more severe ischemia.

Methods

In this case-control cross-sectional study, participants with cerebral small vessel disease (n = 27) and controls (n = 35) underwent T1-weighted and T2 FLAIR MRI for signal intensity quantification as well as pseudocontinuous arterial spin labeling and asymmetric spin echo to measure cerebral blood flow, and oxygen extraction fraction, respectively, and diffusion tensor imaging to assess white matter microstructure. Following signal intensity normalization, we quantified white matter T1 and T2 FLAIR mean and heterogeneity and correlated them to biomarkers of disease severity and physiology (cerebral blood flow and oxygen extraction fraction) in order to understand how signal variability relates to tissue hypoxia-ischemia.

Results

The cerebral small vessel disease group had increased T2 FLAIR intensity (P = 0.006) and heterogeneity (P = 0.017) in normal-appearing white matter compared to controls. Within those with WMH, normal-appearing white matter T2 FLAIR intensity (P = 0.0016) and heterogeneity (P = 0.00036) showed significant relationships with lesion burden. Focal voxel-wise analyses within individual WMH demonstrated that T1 and T2 FLAIR signal intensities were highly variable within lesions, with greater variability in larger lesions. Moreover, the combination of regionally high T2 FLAIR and low T1 intensities was associated with elevated oxygen extraction, suggesting active underlying ischemia. Cluster analysis of lesion signal properties revealed a cluster of lesions that had low T1 intensity, high T2 FLAIR intensity, elevated oxygen extraction and mean diffusivity, representing a specific group of lesions characterized by ischemic physiology.

Interpretation

In conclusion, we found evidence that T1 and T2 FLAIR signal is heterogeneous in cerebral small vessel disease and is associated with disease severity. Within WMH, focal T1 hypointensity and variability, as well as T2 FLAIR hyperintensity and variability is associated with ischemia, but not infarction, particularly in larger lesions, suggesting that these patterns of MRI signal follow both disease severity and aberrant physiology. Finally, WMH clustered by structural properties align with specific physiologic patterns suggesting that imaging appearance may reveal underlying ischemic vulnerability. Although these data are exploratory, they suggest that certain categories of WMH with highly ischemic but non-infarcted physiology may represent high-risk tissue that is reversibly injured. Future work will investigate the role these imaging parameters have on clinically relevant outcomes such as incident stroke and cognitive impairment.