Motion‐Compensated Diffusion Imaging With Phase‐Contrast for Robust Quantification of Regional Cerebral Blood Flow

PURPOSE: To develop and evaluate a motion-compensated diffusion imaging with phase-contrast (MC-DIP) technique for mitigating errors in regional cerebral blood flow (rCBF) quantification caused by physiological brain motion. METHODS: Diffusion-weighted images were acquired in 11 healthy volunteers on a 3.0 T MRI system using three gradient schemes: second-order motion-compensated (2nd-MC), first-order motion-compensated (1st-MC), and non-compensated (non-MC) diffusion gradients. Absolute rCBF maps were generated for each scheme by calibrating intravoxel incoherent motion-derived relative perfusion maps with total cerebral blood flow measured by phase-contrast MRI. The rCBF values from the DIP methods were compared in gray and white matter with a reference arterial spin labeling (ASL) measurement. RESULTS: Both motion-compensated schemes (2nd-MC and 1st-MC) provided significantly better biexponential fitting accuracy in gray matter compared with the non-MC scheme (p < 0.05). In white matter, however, only the 2nd-MC scheme resulted in a significant improvement over the other methods (p < 0.05). While rCBF values from all three DIP methods showed a strong positive correlation with ASL in gray matter (ρ ≥ 0.82, p < 0.05), only the 2nd-MC-DIP method demonstrated a significant positive correlation in white matter (ρ = 0.69, p < 0.05). CONCLUSION: The implementation of second-order motion compensation within the DIP framework improves fitting accuracy, enabling robust rCBF quantification.