Introduction

Quantitative BOLD MRI^1,2 permits noninvasive evaluation of hemodynamic and metabolic states of the brain by quantifying parametric maps of deoxygenated blood volume (DBV) and hemoglobin oxygen saturation level of venous blood (Y_v), and along with a measurement of cerebral blood flow (CBF), the cerebral metabolic rate of oxygen (CMRO₂). One major challenge in qBOLD is to separate deoxyhemoglobin’s contribution to R₂′ from other sources modulating the voxel signal, for instance, R₂, R₂′ from non-heme iron (R₂′_,nh), and macroscopic magnetic field variations. Further, even with successful separation of the several confounders, it is still challenging to extract DBV and Y_v from the heme-originated R₂′ because of limited sensitivity of the qBOLD model³. To address these issues, we had previously proposed a prior-guided 3D qBOLD method⁴ in which a R₂′-sensitive, steady-state 3D pulse sequence (termed ‘AUSFIDE’)⁵ is employed for data acquisition, while a constrained qBOLD problem is solved using a plurality of preliminary parameters (R₂, R₂′, ΔB₀, and voxel susceptibility (Δχ)) obtained via AUSFIDE, along with additionally measured cerebral venous blood volume (CBV_v). It was shown that the method properly disentangled the above-mentioned confounding factors, and yielded the expected contrast for both Y_v and DBV maps across the entire brain⁵. The purpose of this work was to evaluate feasibility of the new qBOLD method in 3D whole-brain CMRO₂ mapping.

Methods

Constrained qBOLD mapping: The 3D qBOLD protocol in this study consists of two pulse sequences: AUSFIDE⁵ (Fig. 1a) yielding a set of volumetric maps (R₂, R₂′, ΔB₀, and Δχ; Figs. 1c, 1d), and additionally velocity-selective venous-spin-labeling (VS-VSL)⁶ (Fig. 1b) leading to voxel-wise CBV_v estimates (Fig. 1e) across the entire brain. The AUSFIDE-derived parameters employed to serve as preliminary estimates in solving the following problem:
$$min_\mathbf{\theta}\sum_{TE}|\textbf{y}(TE)-\mathbf{\Xi}(\mathbf{\theta},TE)|^2 + w|\Delta\chi-\Psi(\mathbf{\theta})|^2+p|{R_2}'-\mathbf{\Gamma}(\mathbf{\theta})|^2$$
where $$$\mathbf{\theta} = \{Y_v, DBV, R'_{2,nh}, \chi_{nb}\}$$$ is the set of unknown parameters. Definitions for the models, Ξ, Ψ, and Γ, along with sub-parameters therein are provided in Reference 4. In solving the above equation, VS-VSL-derived CBV_v maps were employed to initialize and constrain the solution of DBV. The regularization parameters w and p were empirically determined to 10 and 0.1, respectively. Resultant parametric maps are shown in Figs. 1g-1j.

Whole-brain 3D CMRO₂ mapping: 3D pseudo-continuous arterial spin labeling (pCASL) with stack-of-spirals readout⁷ was employed for CBF imaging. Control and tag pCASL images were realigned to the proton-density image acquired prior to the actual pCASL data collection using SPM12, and their pair-wise difference was averaged over multiple measurements, yielding ∆SI and subsequently CBF maps⁸. Given the CBF measurements along with qBOLD-derived Y_v across the entire brain, 3D CMRO₂ maps were computed using the following equation:
$$CMRO_2 = C_a \cdot CBF \cdot (Y_a - Y_v)$$
Here, C_a is the oxygen carrying capacity of arterial blood and Y_a is the hemoglobin oxygen saturation of arterial blood.

Experiments & data processing: Experiments were performed at 3T (Siemens Prisma) in seven healthy subjects using AUSFIDE, VS-VSL, and 3D pCASL. See References 5 and 6 for imaging parameters and data processing procedures for AUSFIDE and VS-VSL, respectively. Imaging parameters in 3D pCASL were: FOV=240x240x120mm³, matrix size=64x64x32, slice partial Fourier factor=6/8, labeling duration=1.8seconds, post labeling delay=1.5seconds, TR=4seconds, measurements=15 control/tag pairs, and scan time=4.5minutes. Additional data were collected using high-resolution MP-RAGE and T₂-relaxation under spin tagging (TRUST)⁹ for brain segmentation and global Y_v measurements. Whole-brain 3D maps of Y_v, CBF, and CMRO₂ in two representative subjects were presented in sagittal, coronal, and axial orientations. Regional averages of CBF and CMRO₂ were calculated in gray and white matter, along with whole-brain averages of Y_v obtained using the present qBOLD and TRUST.

Results

Figure 2 displays whole-brain 3D images in the three orthogonal planes in two study subjects: T₁-weighted magnitude (Fig. 2a), and CBF (Fig. 2b), Y_v (Fig. 2c), and CMRO₂ (Fig. 2d) maps. Consistent with the results in our previous study₄, Y_v obtained using the proposed qBOLD method is largely invariable across the entire brain. In contrast, CBF maps highlight GM/WM differences, thereby contributing predominantly to the contrast of the derived CMRO₂ maps. Table 1 compares TRUST’s global Y_v against the proposed method’s Y_v averaged across the entire brain voxels in the seven study volunteers, and summarizes individual averages of both CBF and CMRO₂ estimates in cortical GM and WM regions, respectively. Two-tailed, paired t-test suggests that the difference of whole-brain average Y_v values in TRUST and the proposed method is not statistically significant (p = 0.98).

Discussion and Conclusions

We introduce a new, MRI-based, regional oximetry technique by means of 3D constrained qBOLD parameter mapping. At the core of the present qBOLD method is the AUSFIDE pulse sequence that enables rapid, high-resolution 3D scanning for the full brain while providing prior information of voxel-averaged magnetic susceptibility (Δχ) and associated parameters at different scales (R₂, R₂′, ΔB₀), which is difficult to achieve with the currently practiced qBOLD techniques. In combination with CBF imaging, the new qBOLD method achieves 3D full-brain CMRO₂ mapping with the expected contrast as well as measurement values in good agreement with those reported in literature. The present 3D MRI oximetry protocol may find a range of clinical applications where tissue oxygen metabolism is regionally altered, for example, in diseases resulting from arterial stenosis/occlusion.

Acknowledgements

NIH grant P41-EB029460 and NRF 2021R1F1A1045621