Peak velocity overestimation and linear-array spectral Doppler

Ultrasound instruments are used to evaluate blood flow velocities in the human body. Most clinical instruments perform velocity calculations based on the Doppler principle and measure the frequency shift of a reflected ultrasound beam. Doppler-only instruments use single-frequency, single-crystal transducers. Linear- and annular-array multiple-crystal transducers are used for duplex scanning (simultaneous B-mode image and Doppler). Clinical interpretation relies primarily on determination of peak velocities or frequency shifts as identified by the Doppler spectrum. Understanding of the validity of these measurements is important for instruments in clinical use. The present study examined the accuracy with which several ultrasound instruments could estimate velocities based on the identification of the peak of the Doppler spectrum, across a range of different angles of insonation, on a Doppler string phantom. The string was running in a water tank at constant speeds of 50, 100, and 150 cm/sec and also in a sine wave pattern at 100- or 150-cm/sec amplitude. Angles of insonation were 30, 45, 60, and 70 degrees. The single-frequency, single-crystal transducers (PC Dop 842, 2-MHz pulsed-wave, 4-MHz continuous-wave) provided acceptably accurate velocity estimates at all tested velocities independent of the angle of insonation. All duplex Doppler instruments with linear-array transducers (Philips P700, 5.0-MHz; Hewlett-Packard Sonos 1000, 7.5-MHz; ATL Ultramark 9 HDI, 7.5-MHz) exhibited a consistent overestimation of the true flow velocity due to increasing intrinsic spectral broadening with increasing angle of insonation.(ABSTRACT TRUNCATED AT 250 WORDS)