Efficacy of ultrasound vector flow imaging in tracking omnidirectional pulsatile flow

Abstract Background Ultrasound vector flow imaging (VFI) shows potential as an emerging non‐invasive modality for time‐resolved flow mapping. However, its efficacy in tracking multidirectional pulsatile flow with temporal resolvability has not yet been systematically evaluated because of the lack of an appropriate test protocol. Purpose We present the first systematic performance investigation of VFI in tracking pulsatile flow in a meticulously designed scenario with time‐varying, omnidirectional flow fields (with flow angles from 0° to 360°). Methods Ultrasound VFI was performed on a three‐loop spiral flow phantom (4 mm diameter; 5 mm pitch) that was configured to operate under pulsatile flow conditions (10 ml/s peak flow rate; 1 Hz pulse rate; carotid pulse shape). The spiral lumen geometry was designed to simulate recirculatory flow dynamics observed in the heart and in curvy blood vessel segments such as the carotid bulb. The imaging sequence was based on steered plane wave pulsing (−10°, 0°, +10° steering angles; 5 MHz imaging frequency; 3.3 kHz interleaved pulse repetition frequency). VFI's pulsatile flow estimation performance and its ability to detect secondary flow were comparatively assessed against flow fields derived from computational fluid dynamics (CFD) simulations that included consideration of fluid‐structure interactions (FSI). The mean percentage error (MPE) and the coefficient of determination ( R 2 ) were computed to assess the correspondence of the velocity estimates derived from VFI and CFD‐FSI simulations. In addition, VFI's efficacy in tracking pulse waves was analyzed with respect to pressure transducer measurements made at the phantom's inlet and outlet. Results Pulsatile flow patterns rendered by VFI agreed with the flow profiles computed from CFD‐FSI simulations (average MPE: –5.3%). The shape of the VFI‐measured velocity magnitude profile generally matched the inlet flow profile. High correlation exists between VFI measurements and simulated flow vectors (lateral velocity: R 2 = 0.8; axial velocity R 2 = 0.89; beam‐flow angle: R 2 = 0.98; p < 0.0001 for all three quantities). VFI was found to be capable of consistently tracking secondary flow. It also yielded pulse wave velocity (PWV) estimates (5.72 ± 1.02 m/s) that, on average, are within 6.4% of those obtained from pressure transducer measurements (6.11 ± 1.15 m/s). Conclusion VFI can consistently track omnidirectional pulsatile flow on a time‐resolved basis. This systematic investigation serves well as a quality assurance test of VFI.