Whole-Brain Cerebral Microvascular Pulsatility Imaging Using Dynamic pCASL MRI With Segmented Stack-of-Stars Golden-Angle Radial Acquisition.

PURPOSE: To develop an efficient dynamic pseudo-continuous arterial spin labeling (pCASL) MRI technique for whole-brain microvascular flow pulsatility imaging.

METHODS

The proposed technique combines pCASL with a segmented stack-of-stars golden-angle radial acquisition. All collected k-space segments were retrospectively binned into multiple cardiac phases according to the recorded pulse signal timings and reconstructed using compressed sensing. In vivo experiments were conducted in 25 participants on a 3 T MRI system. The segmented radial sampling scheme was evaluated and optimized, and results were cross-validated with the reference using a single-shot 3D GRASE acquisition. Test-retest experiments were conducted to evaluate the reproducibility of pulsatility index (PI) measurements. Further optimization, including reconstructions at different in-plane spatial resolutions, shortening the total acquisition time through retrospective down-sampling, and whole-brain imaging with increased through-plane resolution, was performed to investigate improvements in scan efficiency and whole-brain PI delineation.

RESULTS

Consistent PI measurements were obtained using the proposed acquisition scheme with different numbers (M) of radial spokes per segment, enabling improved scan efficiency with a higher M. Test-retest analysis demonstrated good reproducibility in PI measurements (ICC = 0.89, p < 0.001). Higher spatial resolution improved spatial PI delineation, resulting in a systematic increase in mean gray matter PI values, presumably due to reduced partial volume effects. Further optimization, including retrospective down-sampling and whole-brain coverage at 3.5 mm isotropic resolution, demonstrated robust whole-brain PI measurements within a shorter acquisition time (< 7 min).

CONCLUSION

The proposed technique enables reliable and efficient measurement of whole-brain microvascular PI using segmented stack-of-stars radial acquisition.