Demonstration of Brain Tumor-Induced Neurovascular Uncoupling in Resting-State fMRI at Ultrahigh Field

Abstract

To demonstrate in a small case series for the first time the phenomenon of brain tumor-related neurovascular uncoupling (NVU) in resting-state blood oxygen level-dependent (BOLD) functional magnetic resonance imaging (fMRI) at ultrahigh field (7T). Two de novo (i.e., untreated) brain tumor patients underwent both BOLD resting-state fMRI (rsfMRI) on a 7T MRI system and motor task-based BOLD fMRI at 3T. Ipsilesional (i.e., ipsilateral to tumor or IL) and contralesional (i.e., contralateral to tumor or CL) region of interest (ROI) analysis was performed on both 3T motor task-related general linear model-derived activation maps and on 7T rsfMRI independent component analysis (ICA)-derived sensorimotor network maps for each case. Asymmetry scores (ASs) were computed based on numbers of suprathreshold voxels in the IL and CL ROIs. In each patient, ASs derived from ROI analysis of suprathreshold voxels in IL and CL ROIs in task-related activation maps and rsfMRI ICA-derived sensorimotor component maps indicate greater number of suprathreshold voxels in contralesional than ipsilesional sensorimotor cortex in both maps. In patient 1, an AS of 0.2 was obtained from the suprathreshold Z-score spectrum (voxels with Z-scores >5.0) of the task-based activation map and AS of 1.0 was obtained from the suprathreshold Z-score spectrum (Z-scores >5.0) of the ICA-derived sensorimotor component map. Similarly, in patient 2, an AS of 1.0 was obtained from the suprathreshold Z-score spectrum (Z-scores >5.0) of the task-based activation map and an AS of 1.0 was obtained from the suprathreshold Z-score spectrum (Z-scores >5.0) of the ICA-derived sensorimotor component map. Overall, decreased BOLD signal was noted in IL compared with CL ROIs on both task-based activation maps and ultrahigh field resting-state maps, indicating the presence of NVU. We have demonstrated evidence of NVU on ultrahigh field 7T rsfMRI comparable with the findings on standard 3T motor task-based fMRI in both cases.

Keywords: motor activation; neurovascular uncoupling; presurgical mapping; resting state fMRI; ultra-high field.

FIG. 1.

Demonstration of NVU on both 3T task-based sensorimotor activation map and 7T BOLD resting-state ICA-derived sensorimotor component map in Patient No. 1 with a right perirolandic low-grade oligoastrocytoma (grade II). The 3T tongue motor task fMRI (first row) and 7T rsfMRI (third row) maps were overlaid on T2 FLAIR images with IL ROI and CL ROI shown in white contours with suprathreshold voxels displayed in red. The second and fourth rows demonstrate the Z-score spectra of voxels within IL and CL ROIs in the 3T motor fMRI task and 7T rsfMRI maps, respectively. BOLD, blood oxygen level-dependent; CL, contralesional; FLAIR, fluid-attenuated inversion recovery; fMRI, functional magnetic resonance imaging; ICA, independent component analysis; IL, ipsilesional; NVU, neurovascular uncoupling; ROI, region of interest; rsfMRI, resting-state fMRI.

FIG. 2.

Demonstration of NVU on 3T task-based sensorimotor activation map and 7T BOLD resting-state ICA-derived sensorimotor component map in Patient No. 2 with a left frontoparietal opercular low-grade oligodendroglioma (grade II). The 3T tongue motor task fMRI (first row) and 7T rsfMRI (third row) maps were overlaid on T2 FLAIR images with IL ROI and CL ROI shown in white contours with suprathreshold voxels displayed in red. The second and fourth rows demonstrate the Z-score spectra of voxels within IL and CL ROIs in the 3T motor fMRI task and 7T rsfMRI maps, respectively.