Use of standard gradients with compound oblique angulation for optimal quantitative MR flow imaging in oblique vessels

The earliest described phase-modulation techniques for flow quantification by MR imaging require a phase image obtained by modifying one of the imaging gradients and a reference phase image obtained without the modified gradient. However, by using the same gradients that are used for routine two-dimensional Fourier transform imaging, both anatomic and velocity-encoded images can be obtained in one scan. Although convenient, this technique is sensitive to flow both within and perpendicular to the imaging plane. Consequently, significant errors occur in the measurement of flow in vessels oblique to the image plane. To determine the relative accuracy and practicality of quantitatively measuring flow in oblique vessels, we used standard sequence gradients with (1) routine orthogonal plane imaging and (2) direct compound oblique plane imaging. Phantom studies of flow in a vessel aligned along the z axis showed a significant linear correlation (r = .999; p < .05) between the spin phase and spin velocity. However, studies of flow at relatively low physiologic rates (12-17 cm/sec) in vessels angled 0-30° off axis showed that obliquities of as little as 10° result in significant quantification errors. This is due to a larger phase shift per unit velocity along the frequency-encoding direction vs along the slice-select direction and to a mixture of velocities within a voxel that is oblique to the flow direction. In most instances, resolution of these errors can be achieved satisfactorily only by electronic plane rotation with compound oblique angulation so that the image plane and vessel are perpendicular. When so used, this technique potentially might provide important adjunctive quantitative flow data in oblique vessels during routine clinical imaging.