Advanced Neuroimaging to Unravel Mechanisms of Cerebral Small Vessel Diseases

M. Edip Gurol, MD, MSc; Geert J. Biessels, MD; Jonathan R. Polimeni, PhD


Stroke. 2020;51(1):29-37. 

In This Article

Physiological Neuroimaging

As discussed extensively in the respective sections of this text, cSVDs cause not only hemorrhagic but also significant ischemic consequences. Even if occlusion of individual small vessels might explain the lacunar infarcts and microinfarcts, the more widespread pathological changes (leukoaraiosis, atrophy, and decreased connectivity) are more likely related to vascular dysfunction, that is, decreased vessel reactivity. Most of the existing data again come from advanced neuroimaging studies that focused on CAA. The first-in-human proof-of-concept study compared arterial flow responses in patients with CAA to healthy controls using functional transcranial doppler. CAA patients showed a blunted increase in mean PCA flow velocities in response to a visual stimulus (8.0±6.1% versus 17.4±5.7%, P=0.002).[55] Lower visual evoked mean flow velocity increase correlated with higher CMB counts and WMH volume. The PCA pulsatility index, a marker of distal vascular resistance, was also higher in CAA than healthy controls. The reported differences between patients with CAA and healthy controls, and the correlations of the flow velocity changes with markers of disease severity, suggested the presence of vascular dysfunction in CAA, which might mediate ischemic injury.

As functional transcranial doppler has significant limitations in terms of operator dependence and the ability to measure flow velocities only in the larger vessels in and around the circle of Willis, a functional MRI paradigm and analysis approach were developed to characterize temporal features of blood-oxygenation-level-dependent (BOLD) responses to a visual stimulus to obtain estimates of vascular reactivity from the occipital cortex. Here, techniques originally designed for measuring neuronal activity through tracking changes in blood flow and oxygenation were repurposed for the measurement of impaired vascular reactivity. BOLD-weighted time-series data of subjects' block responses were fit to a trapezoidal function with parameters to describe the time to reach peak response, the response amplitude, and the time to return to baseline. The 25 patients with CAA had significantly lower amplitude as well as prolonged time to reach peak and time to return to baseline when compared to 12 healthy controls.[56] Absolute resting blood flow in visual cortex obtained using arterial spin labeling MRI was identical between the groups, again suggesting that decreased vascular reactivity is the predominant physiopathological alteration rather than a static decrease in perfusion. Within the CAA group, longer time to reach peak values also correlated with higher WMH volume, again suggesting that vascular dysfunction can mediate CAA-related ischemia. The findings of decreased amplitude of BOLD response in CAA and its association with WMH volume were reproduced in a separate cohort that used a comparable fMRI approach.[57] This second study also included visual evoked potentials and the lack of difference in visual evoked potential P100 amplitudes between CAA and controls suggested that CAA-related structural changes may not influence the propagation of the neuronal signal resulting from a visual stimulus. The fMRI paradigm that includes a visual stimulus and modeling of the BOLD response from occipital cortex has been used in multiple CAA studies to date. As discussed under the structural imaging section above, this paradigm established vascular dysfunction as the mediator of ischemic change culminating into cortical atrophy in CAA.[35] Use of the same fMRI paradigm also helped demonstrate worse vascular reactivity in symptomatic compared to presymptomatic hereditary CAA patients.[58] The rising slope of the BOLD response, calculated from the time to reach peak and amplitude based on this fMRI paradigm, also served as the main outcome measure in a Phase 2, randomized, double-blind trial of a monoclonal antibody against Aβ1–40 (Ponezumab) to improve vascular dysfunction in CAA.[59] Although the study did not meet the prespecified efficacy criteria, the results obtained using this fMRI approach suggested that it is feasible to use the technique in a consistent manner across multiple study timepoints and sites.

A recent major advance in the field was the use of phase-contrast MRI at 7T to measure time-resolved blood flow velocity at high spatial resolution in the cerebral perforating arteries at the level of the centrum semiovale and the basal ganglia (Figure 5).[60,61] To our knowledge, these submillimeter perforators have not been targeted in any prior physiological MRI study. Using single-slice, 2-dimensional velocity-encoded phase-contrast MRI, the authors compared the number of detected arteries, the pulsatility index and the mean flow velocities between the patient groups (lacunar infarction and deep HTN-ICH) and healthy controls. At the level of basal ganglia and centrum semiovale, both patient groups with symptomatic cSVD had decreased counts of detected perforators and higher pulsatility indices when compared to controls. No velocity differences were observed. This novel physiological MRI approach holds the promise to record flow velocities from submillimeter deep perforating arteries of the brain, allowing assessment of vascular function from vessels that have never been studied in living humans previously (Figure 5). Extensions of this technique to measuring pulsatility indices from parenchymal microvasculature are currently under development.

Figure 5.

Blood flow velocity measurements in the centrum semiovale white matter perforators of a healthy human subject, obtained with high-resolution 2-dimensional phase-contrast magnetic resonance imaging (MRI). A, Mean magnitude image obtained over the cardiac cycle shows the perforators as hyperintense dots. Enlarged detail: magnitude and corresponding velocity map at single cardiac time point. B, Mean velocity-time curve over the cardiac cycle (average over all detected perforators [n=55]) and corresponding 95% CI. Maximum and minimal flow over the cardiac cycle can be used to calculate pulsatility index, a potential marker of small vessel stiffness. Reproduced from Zwanenburg and van Osch60 with permission. Copyright ©2016, Wolters Kluwer Health, Inc.

In addition to the methods briefly reviewed above, other studies that involve physiological stimuli, such as CO2 challenges, are commonly used to assess cerebrovascular reactivity. The incorporation of these techniques into cSVD research will increase the mechanistic yield of the studies.