Rapid whole-brain venous cerebral blood volume mapping using velocity-selective venous-spin-labeling with 3D GRASE

Venous cerebral blood volume (CBVv) is an important neurophysiological parameter and contributor to the BOLD signal mechanism. However, MRI-based, noninvasive methods for CBVv mapping remain scarce, and existing approaches are limited by estimation errors resulting from a large-scale induced magnetic field or long scan times. To address these challenges, we have developed a rapid whole-brain 3D CBVv mapping technique that combines velocity-selective venous-spin-labeling (VS-VSL) and a 3D gradient-and-spin-echo (GRASE) readout. In the GRASE configuration, variable refocusing flip angles were deployed for a long echo train and a short scan, with a segmented-linear center-out k-space trajectory to minimize signal modulations from alternation of gradient- and spin-echo sampling. Simulations were performed to optimize GRASE parameters for optimal scan efficiency with minimal loss of estimation accuracy, followed by experimental validations on the optimized protocol. The performance of the proposed method was evaluated in healthy subjects in terms of repeat reproducibility, and sensitivity to breath-hold and caffeine-intake challenges, stimuli known to alter blood volume in opposite directions. The optimized VS-VSL 3D GRASE sequence allowed whole-brain 3D CBVv mapping in under one minute scan time with high test-retest reproducibility (R2/ICC = 0.88/0.93 in gray matter (GM) and 0.76/0.87 in white matter (WM)). Furthermore, the method captured the expected CBVv changes in response to both breath-hold (+26.8% (GM) and +31.2% (WM)) and caffeine-intake (-16.3% (GM) and -14.1% (WM)), all presenting statistical significance (pv across the whole brain.