Accelerated aortic 4D flow MRI with wave‐CAIPI

Purpose The aim of this study was to investigate the acceleration potential of wave‐CAIPI (controlled aliasing in parallel imaging) for 4D flow MRI, provided that image quality and precision of flow parameters are maintained. Methods The 4D flow MRIs with acceleration factor R = 2 were performed on 10 healthy volunteers, using both wave‐CAIPI and standard Cartesian/2D‐CAIPI sampling for reference. In addition, 1 patient with known aortic valve stenosis was examined. The flow rate (Q), net flow (Qnet), peak velocity vmax, and net average through‐plane velocity (v¯⊥) were calculated in eight analysis planes in the ascending and descending aorta. The acquisitions were retrospectively undersampled (R = 6), and deviations of flow parameters and hemodynamic flow patterns were evaluated. Results Flow parameters measured with an undersampled wave‐CAIPI trajectory showed considerably smaller deviations to the references than the 2D‐CAIPI images. For vmax, the mean absolute differences were 6.02±2.08 cm/s versus 14.36±5.68 cm/s; for Qnet, the mean absolute differences were 3.67±1.40 ml versus 5.87±1.91 ml for wave‐CAIPI versus 2D‐CAIPI, respectively. Noise calculations indicate that the 2D‐CAIPI sampling exhibits a 43±38% higher average noise level than the wave‐CAIPI technique. Qualitative discrepancies in hemodynamic flow patterns, visualized through streamlines, particle traces and flow velocity vectors, could be reduced by using the undersampled wave‐CAIPI trajectory. Conclusion Use of wave‐CAIPI instead of 2D‐CAIPI sampling in retrospectively 6‐fold accelerated 4D flow MRI enhances the precision of flow parameters. The acquisition time of 4D flow measurements could be reduced by a factor of 3, with minimal differences in flow parameters.