Analysis of MR phase-contrast measurements of pulsatile velocity waveforms

Errors in the measurement of the mean velocity of pulsatile velocity waveforms with ungated phase-contrast techniques were studied theoretically and experimentally. Waveforms consisting of a constant and two sinusoidal components were analyzed. Variations in magnitude and phase of the vascular magnetic resonance (MR) signal resulted in errors, the severity of which increased when either factor increased. Magnitude variations always resulted in overestimation. The general shape of the waveform greatly influenced the error, with certain waveforms producing greater inherent error than others. Experimental measurements were performed, validating the predicted sensitivity of these errors to changes in imaging parameters, including TR and flow-encoding sensitivity. Errors generally became more severe with increased flow-encoding sensitivity. The theoretical and experimental results suggest that accurate mean velocity measurements in many vessels of the body--with acquisition times of less than 15 seconds--should be attainable with ungated imaging techniques and with careful selection of relevant imaging parameters.