Multiparametric imaging of brain hemodynamics and function using gas-inhalation MRI
Peiying Liu,Babu G Welch,Yang Li,Hong Gu,Darlene King,Yihong Yang,Marco Pinho,Hanzhang Lu
DOI: https://doi.org/10.1016/j.neuroimage.2016.09.063
IF: 5.7
2017-02-01
NeuroImage
Abstract:Diagnosis and treatment monitoring of cerebrovascular diseases routinely require hemodynamic imaging of the brain. Current methods either only provide part of the desired information or require the injection of multiple exogenous agents. In this study, we developed a multiparametric imaging scheme for the imaging of brain hemodynamics and function using gas-inhalation MRI. The proposed technique uses a single MRI scan to provide simultaneous measurements of baseline venous cerebral blood volume (vCBV), cerebrovascular reactivity (CVR), bolus arrival time (BAT), and resting-state functional connectivity (fcMRI). This was achieved with a novel, concomitant O2 and CO2 gas inhalation paradigm, rapid MRI image acquisition with a 9.3min BOLD sequence, and an advanced algorithm to extract multiple hemodynamic information from the same dataset. In healthy subjects, CVR and vCBV values were 0.23±0.03%/mmHg and 0.0056±0.0006%/mmHg, respectively, with a strong correlation (r=0.96 for CVR and r=0.91 for vCBV) with more conventional, separate acquisitions that take twice the scan time. In patients with Moyamoya syndrome, CVR in the stenosis-affected flow territories (typically anterior-cerebral-artery, ACA, and middle-cerebral-artery, MCA, territories) was significantly lower than that in posterior-cerebral-artery (PCA), which typically has minimal stenosis, flow territories (0.12±0.06%/mmHg vs. 0.21±0.05%/mmHg, p<0.001). BAT of the gas bolus was significantly longer (p=0.008) in ACA/MCA territories, compared to PCA, and the maps were consistent with the conventional contrast-enhanced CT perfusion method. FcMRI networks were robustly identified from the gas-inhalation MRI data after factoring out the influence of CO2 and O2 on the signal time course. The spatial correspondence between the gas-data-derived fcMRI maps and those using a separate, conventional fcMRI scan was excellent, showing a spatial correlation of 0.58±0.17 and 0.64±0.20 for default mode network and primary visual network, respectively. These findings suggest that advanced gas-inhalation MRI provides reliable measurements of multiple hemodynamic parameters within a clinically acceptable imaging time and is suitable for patient examinations.