Optimization of 4D combined angiography and perfusion using radial imaging and arterial spin labeling

link to paper

Optimization of 4D combined angiography and perfusion using radial imaging and arterial spin labeling

Thomas W. Okell, Mark Chiew

Purpose

To extend and optimize a non-contrast MRI technique to obtain whole head 4D (time-resolved 3D) qualitative angiographic and perfusion images from a single scan.

Methods

4D combined angiography and perfusion using radial imaging and arterial spin labeling (CAPRIA) uses pseudocontinuous labeling with a 3D golden ratio (“koosh ball”) readout to continuously image the blood water as it travels through the arterial system and exchanges into the tissue. High spatial/temporal resolution angiograms and low spatial/temporal resolution perfusion images can be flexibly reconstructed from the same raw k-space data. Constant and variable flip angle (CFA and VFA, respectively) excitation schedules were optimized through simulations and tested in healthy volunteers. A conventional sensitivity encoding (SENSE) reconstruction was compared against a locally low rank (LLR) reconstruction, which leverages spatiotemporal correlations. Comparison was also made with time-matched time-of-flight angiography and multi-delay EPI perfusion images. Differences in image quality were assessed through split-scan repeatability.

Results

The optimized VFA schedule (2–9°) resulted in a significant (p < 0.001) improvement in image quality (up to 84% vs. CFA), particularly for the lower SNR perfusion images. The LLR reconstruction provided effective denoising without biasing the signal timecourses, significantly improving angiographic and perfusion image quality and repeatability (up to 143%, p < 0.001). 4D CAPRIA performed well compared with time-of-flight angiography and had better perfusion signal repeatability than the EPI-based approach (p < 0.001).

Conclusion

4D CAPRIA optimized using a VFA schedule and LLR reconstruction can yield high quality whole head 4D angiograms and perfusion images from a single scan.