Abstract:Purpose This study evaluated the velocity‐selective (VS) MRA with different VS labeling modules, including double refocused hyperbolic tangent, eight‐segment B1‐insensitive rotation, delay alternating with nutation for tailored excitation, Fourier transform–based VS saturation, and Fourier transform–based inversion. Methods These five VS labeling modules were evaluated first through Bloch simulations, and then using VSMRA directly on various cerebral arteries of healthy subjects. The relative signal ratios from arterial ROIs and surrounding tissues as well as relative arteria‐tissue contrast ratios of different methods were compared. Results Double refocused hyperbolic tangent and eight‐segment B1‐insensitive rotation showed very similar labeling effects. Delay alternating with nutation for tailored excitation yielded high arterial signal but with residual tissue signal due to the spatial banding effect. Fourier transform–based VS saturation with half the time of other techniques serves as an efficient nonsubtractive VSMRA method, but the remaining tissue signal still obscured some small distal arteries that were delineated by other subtraction‐based VSMRA, allowing more complete cancelation of static tissue. Fourier transform–based inversion produced the highest arterial signal in VSMRA with minimal tissue background. Conclusion This is the first study that angiographically compared five different VS labeling modules. Their labeling characteristics on arteries and tissue and implications for VSMRA and VS arterial spin labeling are discussed.

Direct angiographic comparison of different velocity‐selective saturation, inversion, and DANTE labeling modules on cerebral arteries