Imaging of Cerebrovascular Function in Chronic Traumatic Brain Injury

Abstract

Traumatic cerebrovascular injury (TCVI) is a common pathologic mechanism of traumatic brain injury (TBI) and presents an attractive target for intervention. The aims of this study were to assess cerebral blood flow (CBF) and cerebrovascular reactivity (CVR) using magnetic resonance imaging (MRI) to assess their value as biomarkers of TCVI in chronic TBI, characterize the spatial distribution of TCVI, and assess the relationships between each biomarker and neuropsychological and clinical assessments. Forty-two subjects (27 chronic TBI, 15 age- and gender-matched healthy volunteers) were studied cross-sectionally. CBF was measured by arterial spin labeling and CVR by assessing the MRI-blood oxygen level-dependent signal with hypercapnia challenge. A focused neuropsychological battery adapted from the TBI Common Data Elements and neurobehavioral symptom questionnaires were administered at the time of the imaging session. Chronic TBI subjects showed a significant reduction in mean global, gray matter (GM), and white matter (WM) CVR, compared with healthy volunteers (p < 0.001). Mean GM CVR had the greatest effect size (Cohen's d = 0.9). CVR maps in chronic TBI subjects showed patchy, multifocal CVR deficits. CBF discriminated poorly between TBI subjects and healthy volunteers and did not correlate with CVR. Mean global CVR correlated best with chronic neurobehavioral symptoms among TBI subjects. Global, GM, and WM CVR are reliable and potentially useful biomarkers of TCVI in the chronic stage after moderate-to-severe TBI. CBF is less useful as biomarker of TCVI. CVR correlates best with chronic TBI symptoms. CVR has potential as a predictive and pharmacodynamic biomarker for interventions targeting TCVI.

Keywords: cerebral blood flow; cerebrovascular reactivity; hypercapnia; magnetic resonance imaging-blood oxygen level dependent; neurovascular unit; traumatic brain injury; traumatic cerebrovascular injury.

Conflict of interest statement

No competing financial interests exist.

Figures

FIG. 1.

Time course of end-tidal CO2 (EtCO2) and blood oxygen level–dependent (BOLD) signal in the thalamus and average among the 15 healthy controls. The BOLD signal follows the EtCO2 data with a temporal delay of approximately 15 sec.

FIG. 2.

Examples of cerebral blood flow (CBF) and cerebrovascular reactivity (CVR) maps from one healthy volunteer (A) and two TBI patients (B and C) acquired with 3 magnetic resonance imaging (3 Tesla) sequences (fluid-attenuated inversion recovery [FLAIR], ASL and MRI-BOLD with hypercapnia challenge: FLAIR (A-1, A-2, B-1, B-2, C-1, C-2), CBF (A-3, A-4, B-3, B-4, C-3, C-4) and CVR (A-5, A-6, B-5, B-6, C-5, C-6). The white numbered circles indicate focal CBF and CVR deficits. The TBI patient in the middle three panels had no visible abnormalities on standard anatomical imaging, while the TBI patient in the last three panels had a lesion in the left frontal subcortical white matter (C-2). For example, the CVR map in B-5 displays three regions with focal CVR deficit, which are not evident in the CBF map (B-3). This mismatch can also be appreciated in B-6 and C-6. In C-6, we observed post-traumatic encephalomalacia in the corresponding FLAIR image (1C-2) and, in the vicinity of the lesion, we noted a decrease of CVR (C-6, region 2) but not of CBF (C-4).

FIG. 3.

Mean distribution of cerebrovascular reactivity (CVR) values (top) and cerebral blood flow (CBF; bottom) for the healthy control (HC) and traumatic brain injury (TBI) population in the whole brain (union between the gray matter and the white matter). The HC histogram (blue) is the average of 15 patients and the TBI histogram (red) is the average of 27 patients. The error bar is the standard deviation of CVR (or CBF) for one group within a given bin. HC have a median CVR value of 0.173 ± 0.014 and TBI have a median CVR value of 0.140 ± 0.230 (p value <0.001). HC have a median CBF value of 38.0 ± 12.8 and TBI a median CBF of 330 ± 8.1 (p value = 0.04).

FIG. 4.

Map of abnormal cerebrovascular reactivity (CVR) z-score values (color) overlay on the Magnetization Prepared Rapid Gradient Echo (MPRAGE; top) and map of abnormal cerebral blood flow (CBF) z-score values (color) overlay on the MPRAGE (bottom). The subject is a 51-year-old male with traumatic brain injury whose non normalized images are shown in Figure 2-B2 to 2-B4.

FIG. 5.

Individual measures of cerebrovascular reactivity (CVR; x-axis) versus cerebral blood flow (CBF; y-axis). The blue dots represent healthy controls (HC) and the red dots traumatic brain injury (TBI) subjects. The Spearman's correlation between CBF and CVR values is 0.04 (p = 0.7).