Accuracy of Measurements of Flow Velocity by Pulsed and Continuous Wave Doppler Echocardiography: An In Vitro Study


Takahiro Shiota, Dag Teien, Robert Weintraub, Sage Ewing, You-Bin Deng, James C. Conti, Ph.D., Univ of Calif, San Diego, CA.

Presented by: Takahiro Shiota (David Sahn)

Flow velocity (VEL) estimation by Doppler echocardiography has been widely accepted clinically. To determine the accuracy of Doppler measurement of flow VEL compared to a precisely calibrated pulsatile Doppler phantom (EchoCal CD 10) as a "Gold Standard", we evaluated a total of 6 commercially available echo-Doppler systems with 9 pulsed (PW) and 4 continuous wave (CW) Doppler transducers. Flow VELs ranged from 30 cm/sec to 150 cm/sec for PW, 300 to 400 cm/sec for HPRF PW and 50 cm/sec to 525 cm/sec for CW evaluation. For the total of 113 measurements with PW measured from modal VEL and CW from peak VEL, excellent correlations existed between measured (y) and actual flow VEL (r=0.99, y = 0.98 x +4.1). The mean difference between measured and actual flow VEL was 1.4 cm/sec with a standard deviation of 5.5 cm/sec and 105 out of 113 measurements were within 2 SD of the mean. When measuring the lowest flow VELs using low frequency transducers, however, a tendency for overestimation was noted for all the systems (up to 14 cm/sec or 28% of VEL). On the other hand, when measuring high flow VELs using CW, a tendency for underestimation was observed for all systems (up to 19 cm/sec or 4% of VEL). High PRF Doppler showed substantial accuracy for mid-range flows. A generally satisfactory agreement between measured and actual flow VEL was found for all systems in our study. The minor discrepancies noted are probably of little clinical importance except when making absolute rather than ratio determinations defining low VEL (eg: pulmonary vein, mitral inflow) flows.