Wednesday afternoon grand ballroom AB papers 357–364. Flow imaging: Experimental techniques
✍ Scribed by G Laub; PR Moran
- Publisher
- John Wiley and Sons
- Year
- 1991
- Tongue
- English
- Weight
- 393 KB
- Volume
- 1
- Category
- Article
- ISSN
- 1053-1807
No coin nor oath required. For personal study only.
✦ Synopsis
To quantify accuracy of MR phase-velocity measurements without acceleration compensation, repeated studies were made of laminar and turbulent flows and compared with laser Doppler measurements. Measurements were made on a Philips 1.5-T imager with use of a head coil, a TE of 15 msec, FOV of 128 mm, and matrix size of 128 X 128. Velocity profiles were obtained in three phantom models: laminar straight-tube flow [A], turbulent straight-tube flow [B], and laminar stenotic flow [ C ] . Measurements over a Reynolds number range of 500 to 10,000 were compared with theoretical data for laminar flow and laser Doppler anemometer (LDA) data for turbulent flow on a point-by-point basis. Ten repeated measurements of velocity in model [A] were made at several flow rates on separate days. Linear regression analysis yielded a n r value of ,995, with a slope of 1 .OO and an intercept of -0.06 cm/sec, giving a n accuracy level of more than 99%. MR velocity measurements obtained for turbulent flow in model [B] yielded significant errors near the wall. The turbulent MR measurements versus LDA resulted in a much poorer correlation coefficient of .66. a slope of 1.06, and a n intercept of -5.0 cm/sec. In the upstream section of the stenosis model [C]. the measured values of velocity were greater by as much as 40% than the true velocity. The large velocity errors can be fully explained by an analysis of the convective acceleration. The MR velocity measurements demonstrate that phase-velocity mapping is precise and accurate for simple, laminar flow but that significant errors arise from turbulent and accelerating flows. In the presence of turbulent fluctuations, MR measurements are adequate for total flow or average velocity but not for velocity on a pixel-by-pixel basis. It may be possible to minimize acceleration-based phase errors by shape modifications in the basic velocity-encoding gradient waveform. Application of such sequences may improve velocity measurement for both turbulent and accelerating flows.