Flow MR fingerprinting

Purpose: To investigate the feasibility to quantify blood velocities within the magnetic resonance fingerprinting framework, while providing relaxometric maps of static tissue. Methods: Bipolar gradients are inserted into an SSFP-based MRF sequence to achieve velocity-dependent signal phases, allowing tri-directional time-resolved velocity component quantification. The accuracy of both relaxometric mapping and velocity quantification was validated in vivo and in phantom studies. Results: Simulations determined that even for strong cardiac cycle length variations (700–1400 ms) Flow-MRF determines accurate velocity maps deviating <0.1% from the ground truth on average. The cardiac cycle length variability only results in reduced velocity-to-noise ratios. Good agreement in the velocity quantification between a standard phase–contrast cine and the Flow-MRF sequence was reached in phantom experiments. Relaxometric phantom experiments determined mean deviations between Flow-MRF and spin-echo-based reference measurements of 89 ± 25 ms / 0.8 ± 2.5 ms over the range of 630–2630 ms / 49–145 ms for T₁ / T₂, respectively. The in vivo study of a human knee determined mean T₁ / T₂ values of 1383 ± 75 ms / 26 ± 4 ms for the gastrocnemius muscle that agree with literature values. Conclusion: Flow-MRF presents a novel way of quantifying velocities while simultaneously providing relaxometric maps of static tissue and it can potentially be a viable method to accelerate the inherently long acquisition times of time-resolved velocity quantification.