White Matter Cerebrovascular Reactivity: Effects of Microangiopathy and Proximal Occlusions on the Dynamic BOLD Response

J Michael Gee,Xiuyuan Wang,Siddhant Dogra,Jelle Veraart,Koto Ishida,Seena Dehkharghani

DOI: https://doi.org/10.1101/2023.05.29.23290700

2023-06-03

MedRxiv

Abstract:Introduction: Cerebral microangiopathy often manifests as white matter hyperintensities (WMH) on T2-weighted MR images and is associated with elevated stroke risk. Large vessel steno-occlusive disease (SOD) is also independently associated with stroke risk, however, the interaction of microangiopathy and SOD is not well understood. Cerebrovascular reactivity (CVR) describes the capacity of cerebral circulation to adapt to changes in perfusion pressure and neurovascular demand, and its impairment portends future infarctions. CVR can be measured with blood oxygen level dependent (BOLD) imaging following acetazolamide stimulus (ACZ-BOLD). We studied CVR differences between WMH and normal-appearing white matter (NAWM) in patients with chronic SOD, hypothesizing additive influences upon CVR measured by novel, fully dynamic CVR maxima ( CVR max ). Methods: A cross sectional study was conducted to measure per-voxel, per-TR maximal CVR ( CVR max ) using a custom computational pipeline in 23 subjects with angiographically-proven unilateral SOD. WMH and NAWM masks were applied to CVR max maps. White matter was subclassified with respect to the SOD-affected hemisphere, including: i. contralateral NAWM; ii. contralateral WMH iii. ipsilateral NAWM; iv. ipsilateral WMH. CVR max was compared between these groups with a Kruskal-Wallis test followed by a Dunn-Sidak post-hoc test for multiple comparisons. Results: 19 subjects (age 50±12 years, 53% female) undergoing 25 examinations met criteria. WMH volume was asymmetric in 16/19 subjects with 13/16 exhibiting higher volumes ipsilateral to SOD. Pairwise comparisons of CVR max between groups was significant with ipsilateral WMH CVR max lower than contralateral NAWM (p=0.015) and contralateral WMH (p=0.003) when comparing in-subject medians and lower than all groups when comparing pooled voxelwise values across all subjects (p<0.0001). No significant relationship between WMH lesion size and CVR max was detected. Conclusion: Our results suggest additive effects of microvascular and macrovascular disease upon white matter CVR, but with greater overall effects relating to macrovascular SOD than to apparent microangiopathy. Dynamic ACZ-BOLD presents a promising path towards a quantitative stroke risk imaging biomarker. Background: Cerebral white matter (WM) microangiopathy manifests as sporadic or sometimes confluent high intensity lesions in MR imaging with T2-weighting, and bears known associations with stroke, cognitive disability, depression and other neurological disorders 1-5 . Deep white matter is particularly susceptible to ischemic injury owing to the deprivation of collateral flow between penetrating arterial territories, and hence deep white matter hyperintensities (WMH) may portend future infarctions 6-8 . The pathophysiology of WMH is variable but commonly includes a cascade of microvascular lipohyalinosis and atherosclerosis together with impaired vascular endothelial and neurogliovascular integrity, leading to blood brain barrier dysfunction, interstitial fluid accumulation, and eventually tissue damage 9-14 . Independent of the microcirculation, cervical and intracranial large vessel steno-occlusive disease (SOD) often results from atheromatous disease and is associated with increased risk of stroke owing to thromboembolic phenomena, hypoperfusion, or combinations thereof 15-17 . White matter disease is more common in the affected hemisphere of patients with asymmetric or unilateral SOD, producing both macroscopic WMH detectable by routine structural MRI, as well as microstructural changes and altered structural connectivity detected by advanced diffusion microstructural imaging 18, 19 . An improved understanding of the interaction of microvascular disease (i.e., WMH) and macrovascular steno-occlusion could better inform stroke risk stratification and guide treatment strategies when coexistent. Cerebrovascular reactivity (CVR) is an autoregulatory adaptation characterized by the capacity of the cerebral circulation to respond to physiological or pharmacological vasodilatory stimuli 20-22 . CVR may be heterogeneous and varies across tissue type and pathological states 1, 16 . Alterations in CVR are associated with elevated stroke risk in SOD patients, although white matter CVR, and in particular the CVR profiles of WMH, are only sparsely studied and not fully understood 1, 23-26 . We have previously employed blood oxygen level dependent (BOLD) imaging following a hemodynamic stimulus with acetazolamide (ACZ) in order to measure CVR (i.e. ACZ-BOLD) 21, 27, 28 . Despite the emergence of ACZ-BOLD as a technique for clinical and experimental use, poor signal-to-noise characteristics of the BOLD effect have generally limited its interpretation to coarse, time-averaged assessment of the terminal ACZ response at arbitrarily prescribed delays following ACZ administration (e.g. 10-20 minutes) 29 . More recently, we have introduced a dedicated computational pipeline to overcome historically intractable signal-to-noise ratio (SNR) limitations of BOLD, enabling fully dynamic characterization of the cerebrovascular response, including identification of previously unreported, unsustained or transient CVR maxima ( CVR max ) following hemodynamic provocation 27, 30 . In this study, we compared such dynamic interrogation of true CVR maxima between WMH and normal appearing white matter (NAWM) among patients with chronic, unilateral SOD in order to quantify their interaction and to assess the hypothesized additive effects of angiographically-evident macrovascular stenoses when intersecting microangiopathic WMH.