Monday, December 10, 2012

Pulsed Wave Doppler and Aliasing

Pulsed Doppler is in distinction to continuous wave Doppler.

In continuous wave Doppler, reflectors anywhere in the ultrasound transducer beam contribute to the Doppler signal.

From Ref 2

In pulsed Doppler, a piece of the return signal can be selected, thereby only detecting the moving particles at a certain depth and in a particular volume (the "sample volume"). Even though the volume cannot be appreciated on 2D grayscale imaging, it can lead to localization artifact (Doppler signal can be acquired from nearby vessels, not necessarily in the field of view).

One of the advantages of pulsed wave Doppler is that other grayscale imaging can be carried on simultaneously (although there is a decrease in the frame rate as some of the information is being used for Doppler).

The farther the sample volume is from the transducer, the larger the volumes becomes because of divergence of the U/S beam.

So what is the best frequency to use for Doppler imaging?  As a general principle, the higher the frequency, the better since the intensity of U/S waves from small scatterers (such as RBCs) increased rapidly with higher frequencies (proportional to f^4 (!)). However, this advantage is compromised by the fact that higher frequency waves have less penetrating power (greater attenuation).  Frequencies as low as 2 MHz can be used for deeper vascular structures.

The main disadvantage of pulsed wave Doppler is aliasing. It has trouble detecting very high blood flow velocities... which, unfortunately, often occurs around areas of pathology.

Aliasing is intimately tied to the concept of PRF -- pulse repetition frequency.  Since Doppler signal is "sampled" out of a larger volume of ultrasound information, the amount of sampling (the PRF) has to be fast enough to accurately reproduce the movements in the target flow... or the sample waveform will be only bits and pieces of the real  waveform. If you sample multiple times during the real waveform's cycle, then no problem... but if the PRF is restricted (i.e. at a greater depth), the lower it must be.

At a minimum, the PRF must be at least twice the frequency of the Doppler signal to reconstruct it accurately.  When the PRF = 2 x Freq, this is called the Nyquist Sampling Rate.

When the PRF is too low, high velocities are converted into reversed flow at the point of aliasing ("wraparound"). This can be adjusted by changing the velocity scale and/or adjusting the baseline. (When the scale is decreased, the machine increases the PRF).  Aliasing can also be fixed by using a lower frequency transducer... a lower frequency transducer results in a lower frequency return Doppler signal, so the PRF can potentially fit this lower frequency return Doppler signal.

So isn't the highest possible PRF best for vascular imaging? Not necessarily.

If the PRF becomes too high "range ambiguities" occur because of overlap of returning frequencies.

At a certain distance from the transducer (d), the minimum time needed between pulses of the sample volume is twice the distance there and back (Time needed = 2 x d/c).  The maximum PRF before range ambiguity occurs is the inverse of this.

1. "Introduction to Vascular Ultrasound" Zweibel and Pellerito, ed. 5th ed. (2005)
2. "Echo in Context" Duke University SOM "Pulsed and Continuous Wave Doppler"
3. Boote EJ. "AAPM/RSNA Physics Tutorial for Residents: Topics in US Doppler US Techniques: Concepts of Blood Flow Detection and Flow Dynamics"