Color Doppler Artifacts - SD

Introduction

Hello, my name is Bill Middleton.
I'm a professor of radiology at the Mallinckrodt Institute
of Radiology in Washington University School
of Medicine that's in St.
Louis, Missouri.
And today I wanted to discuss color Doppler artifacts.
There are four artifacts that are fairly basic
that I'll go over today.

Aliasing

And the first one is aliasing. Sampling theory says
that in order to sample a periodic phenomenon,
you have to sample it at twice its own frequency.
In Doppler imaging,
the sampling rate is the pulse repetition frequency,
and the thing
that you're sampling is the Doppler frequency shift.
So in order to sample correctly, you have
to have a sampling rate
or a pulse repetition frequency
that is twice the Doppler frequency shift.
If you sample at less than twice the Doppler frequency
shift, you're going to get components
that are artifact low.
Now, this diagram shows a sine wave
that's being sampled quite frequently,
and you can see each arrow indicates a sampling point in time.
If you reproduce that sine wave below,
based on this type of sampling, frequent sampling, you see
that it's a very good representation of
what the true sine wave was.
On the other hand, if you sample at less than twice the sine wave frequency, you can see how it would be easy
to generate artifactually low components
as shown in the waveform down below.
Now, with pulse Doppler, what happens with aliasing is that
as your pulse repetition frequency gets lower
and lower, the arterial signal or any signal that you have,
will become difficult to display.
This is an example of an arterial signal
that sampled at a pulse repetition frequency
that produced a Doppler scale of plus
or minus 80 centimeters per second.
So a fairly high sampling rate.
And even though the velocities are high,
they're displayed correctly
and there's no aliasing on this waveform.
Now notice as we go to a lower sampling rate,
lower pulse repetition frequency,
and a lower Doppler scale of plus
or minus 60,
the scale no longer can accommodate the high frequency
shifts and peak systolic flow gets alias into the negative part of the Doppler scale.
This is the pulse Doppler equivalent of aliasing.
Now, if we continue to go lower
and lower in the scale here, it's plus
or minus 30, then that aliasing goes from the negative part
of the Doppler scale back into the positive part
of the Doppler scale.
And if we continue to go even worse than that,
then it gets multiple wraparound so
that you no longer even recognize
that this is an arterial signal
and it just looks like Doppler noise.
So that's what aliasing does with pulse Doppler.
Now, what about with color Doppler?
We really wanna focus on the color Doppler appearance today with Doppler frequency shifts
as they get higher, originally the signal start
or the color assignment gets lighter and lighter,
but then you reach a point called the Nyquist limit
where you can no longer accommodate the
higher frequency shifts.
And at that point, the signal wraps around
to the negative part of the Doppler scale.
And the transition in areas of aliasing is from light shades
of red or yellow to light shades of blue
or green, as in this example.
Now that is different from true flow reversal.
With true flow reversal,
what happens is the transition occurs from the dark shades
of red to the dark shades of blue,
and it passes through an area of
filtering that is black.
So this is the type of transition that you get
with true flow reversal.
Let's look at some examples.
Here's a patient that's got a soft plaque at the origin
of the internal carotid producing a stenosis
and producing a flow jet at the site of stenosis.
Notice that in the jet, the transition is from yellow
to light green because of aliasing.
Now, on the other hand, there is an area
of flow reversal right here indicated by the curved arrow
where the transition is from dark red to dark blue
with a little black line in between the two.
This is another example of a plaque producing stenosis in the common carotid artery,
but you see the exact same phenomenon aliasing in the area
of the flow jet, and then true flow reversal just next to it.
But see the difference in the transition between flow reversal
and aliasing. This is a TIPS shunt very magnified view of proximal TIPS.
And in the origin
or the proximal portion of the TIPS you see aliasing.
In the mid portion of the TIPS you see flow reversal.
Now in this case, the aliasing is a very useful artifact
because it tells you where the highest frequency shifts are,
and many times that also corresponds to
where the highest velocities are.
That was the case in this patient
who had a stenosis producing high velocities in
that proximal stent.
And this is by seeing the aliasing that allows you
to concentrate your pulse Doppler analysis right in
that area where you'll have your highest yield.
Now this is another example of a TIPS shunt,
a magnified view showing what looks like very disordered flow within that shunt.
But notice the Doppler scale here.
It's plus or minus three centimeters per second. TIPS
shunts have maximum velocities in the 100
to 200 centimeter per second range,
and mean velocities probably in the 50
to 150 centimeter range.
So this Doppler scale is much too low
and it's resulted in extensive aliasing.
Remember on the pulse Doppler waveform
how we had multiple wraparounds made it look like noise?
Well, the same thing is happening here on the color Doppler
side with multiple wraparounds making it look like there's
extensive flow disturbance.
When we improve our Doppler scale to plus
or minus 72 centimeters per second,
the flow looks entirely normal
because we've eliminated the aliasing.
Here's a cine clip showing aliasing both with transitions from yellow to green during diastole,
but showing the mosaic pattern that you get
with extreme aliasing during systole.
These are transverse views of the splenic vein in a patient that had normal splenic venous flow
and with the Doppler scale set at plus or minus 20, the flow in
that splenic vein was displayed correctly.
When we drop the Doppler scale to half that value plus or minus 10, then the amount of aliasing makes it look like the flow in the splenic vein is reversed.

Eliminating Aliasing

Now, how do you eliminate aliasing?
Well, first of all, remember with color Doppler,
you may not want to eliminate aliasing
because it's a useful artifact to show you areas
of abnormal flow, but if you do need to diminish aliasing
or decrease it, go back to your Doppler equation.
Doppler equation says that the Doppler frequency shift
FD is equal
to the transmitted frequency f 10 times the velocity times
the cosine of the Doppler angle times constants,
which are the inverse of the speed of sound.
And two, so if you decrease the transmitted frequency,
the frequency shift will decrease and you'll decrease
or eliminate aliasing.
You can also decrease the cosine of the Doppler angle
by increasing the Doppler angle itself,
getting it closer to 90 degrees.
That will get your cosine function closer to zero,
and you can do that by either re-steering the Doppler beam
if you're using a linear probe
or by moving the probe to a different position.
If you are using a curvilinear
or a sector probe,
you can also increase the sampling rate
by adjusting your Doppler scale.
You will reach a certain limit
where you can't increase your sampling rate
or your Doppler scale any higher
unless you decrease the distance
between the probe and the vessel.
And the only way to do that is to move your probe
to a different location.
Now, there is something with pulse Doppler
that's called high pulse repetition frequency imaging,
which is probably worth being familiar with.
In the normal situation, when you sample a vessel,
you send a pulse down to that vessel
and then you wait for the pulse to return
before you send your next pulse.
By doing that, you're able to tell exactly where that Doppler signal comes from,
and there's a single sample volume.
With high pulse repetition frequency imaging.
The first pulse is sent down, but
before it returns to the transducer, a second pulse is transmitted by the transducer.
So at some point, those two pulses intersect
and the transducer sees two pulses
returning to the transducer.
This creates ambiguity in where the sample volume is so
that it may be where the vessel is originally
or there may be a second potential site of flow in the mid portion of the Doppler line of sight.
And here's an example of that.
This is a stenotic left renal artery.
You can see with a pulse repetition frequency of
about 6,250 pulses per second,
there's a Doppler scale up to 200 centimeters per second
and extensive aliasing of the peak systolic flow.
So we can't accurately measure
that flow velocity at this type of a scale
or PRF setting.
If we shift into high pulse repetition frequency imaging
indicated by the HPRF,
then the pulse repetition frequency is more than 14,000 pulses per second.
This has resulted in a Doppler scale
of 500 centimeters per second,
and now the aliasing has been eliminated
and we can accurately measure the velocity. Notice,
however, on the image
that an additional sample volume has been created
by going into this high pulse repetition frequency mode.

Mirror Artifacts

Another common artifact
or mirror artifacts, acoustic mirrors occur at sites
where there are highly reflective, smooth interfaces.
Now, gas is the strongest reflector that occurs in the body.
And when you have a gas interface that's very smooth,
it acts as a wonderful mirror
and that tends to occur most often around the lung.
Since a lung is filled with air,
This schematic shows how mirror images are created.
This is a schematic showing the liver with a focal lesion in the liver up against
the base of the lung.
Now the sound pulse is transmitted to the base of the lung.
It reflects back into the liver, interacts with that.
With that focal liver lesion then is transmitted back
to the base of the lung and back to the transducer.
Well, the transducer doesn't know that it's taken
that detour, it assumes that sound goes in a straight line,
so it duplicates that lesion up above the diaphragm.
We all have seen this with the gray scale imaging,
but the same thing can happen with Doppler imaging.
This is another schematic showing a vessel
with a blood flow indicated by the arrow.
The sound pulse that goes into the vessel lumen reflects
back to the transducer
and localizes that flow correctly,
but sound that goes to a mirror deep to that vessel
will reflect back into the vessel from below
and then back to the mirror and then back to the transducer.
Well, the transducer again, doesn't know
that it's taken that detour.
It assumes that the sound's gone in a straight line.
So it duplicates that velocity deep to the mirror
and recreates another vessel.
And this is an example of this in real life
where we're looking at the internal mammary artery,
which is located immediately adjacent to the pleural surface of the lung.
And this is the pleural interface right here.
Everything deep to this is gas
and that gas is acting as a mirror
to completely duplicate the true vessel
into a mirror image artifact deep to the gas interface.
Now, when you're dealing with Doppler mirror images,
the waveform that you get from the original vessel should be
identical to the waveform from the mirror.
The shape and the size should be identical as in this example, another area
where you've got large gas interface is in the air
column of the trachea.
And this is an example of a thyroid nodule
where we see the nodule on gray scale.
Here is the air interface.
This bright line, these little dark areas are the tracheal
cartilaginous rings,
and this is the thyroid nodule duplicated deep
to the trachea.
Well, the same thing occurs with color Doppler
and all this color Doppler signal gets duplicated as well.
This is a mirror image of a hepatic vein caused
by the base of the lung.
Gonna just mag up on that.
The sound pulse that originally interacts with the flow
produces a negative frequency shift
because the flow in that vein is going away from the
direction of the sound pulse.
However, the reflected pulse off
of the diaphragm is going in the opposite direction
of the original pulse, and
therefore the frequency shift is different.
It's positive. So the mirror image artifact is a positive
frequency shift and is displayed in red as opposed
to the original vein, which is a negative frequency shift
and it's shown in blue.
This is a TIPS shunt.
The back wall of the TIPS in this case
is acting as a mirror.
It's not a strong enough mirror
to see anything on gray scale,
but when you look on Doppler, you can see duplication
of the Doppler signal deep to the lumen
of that TIPS shunt.
The back wall of the carotid can act as a weak mirror
and reproduce flow in the carotid lumen actually deep
to the carotid wall.
That's called the carotid ghost.
And in this case, since it's such a weak mirror,
the mirror image signal is going
to be weaker than the original signal.
So that's the mirror. This is the
waveform from the true vessel.
And here's the waveform from the mirror.
Now notice that the mirror waveform is similar in size
and shape, but
because it's a weak mirror,
it's much weaker in its signal strength.

Tissue Vibration

Okay, the next artifact is tissue vibration.
Turbulent blood flow causes pressure fluctuations in the
lumen of the vessel, which result in vibration
of the vessel wall.
That vessel wall vibration, if it's significant enough,
will get transmitted into the perivascular tissues
and cause them to vibrate.
Here's a schematic showing a stenotic vessel
with turbulent flow distal to the stenosis vibration
of the wall of the vessel being transmitted into vibration
of the tissue interfaces next to the vessel wall.
Well, since those tissue interfaces are moving back
and forth, they generate a random Doppler signal that ranges
between red and blue or alternates between red and blue.
And this is the schematic representation of that.
This is what it looks like in real life.
This is a stenotic dialysis graft
and in the region of the stenosis, the velocities
and the turbulent flow within that lumen have produced a lot
of perivascular tissue vibration.
And you see that as random flashes of red
and blue in the tissues immediately surrounding the vessel.
This is a pseudoaneurysm and you see tissue vibration near the neck.
Here's the artery supplying the pseudoaneurysm.
This is the neck of the pseudoaneurysm
and here's the lumen of the pseudoaneurysm,
but near in the origin of the neck, the flow is very fast and turbulent
and is producing this tissue vibration.
This is another dialysis graft,
and you can see the vibration
of the soft tissues near the anastomosis of the graft.
Now, if you do a sample volume of the tissues themselves,
they're very strong reflectors,
so they give a strong Doppler signal,
but they're not moving very fast.
So it's not a very tall Doppler signal.
And because you have reflectors
that are moving both towards the transducer
and away from the transducer simultaneously,
you get symmetric flow above and below the baseline.
This is a renal transplant with stenosis near the hilum,
and you can see the tissue vibration in the hilum
of the transplant and the vibrational signal
on the pulse Doppler waveform.
Similar example in a stenotic carotid.
The upper left hand image shows the carotid
during in diastole when flow is slow
and there's really not much turbulence,
but in systole the flow accelerates.
You get turbulence and perivascular tissue vibration,
and then again, you see the vibrational signal in the soft
tissues on the Doppler waveform shown below.
Now, sometimes you'll see the vibrational signal
embedded within the arterial envelope.
This is an example where we've got a stenotic
left renal artery.
Notice that the arterial velocities are very high close
to 500 centimeters per second, and that's fine.
But look at this signal deep to the arterial signal.
This is the vibrational component coming from the vessel wall
and from the tissues around the vessel.
It's not nearly as high as the arterial signal itself,
but it's a stronger signal.
And as we've shown in the earlier examples, it's symmetric
above and below the baseline.
So when you see this type
of a embedded vibrational component in your arterial signal,
realize that you're dealing with turbulent flow
and some sort of a vascular lesion.
Now, every now and then, this vibrational flow will produce
a musical quality on the audio output,
and it will, it's called a musical bruit.
When that happens, you'll see little harmonic lines embedded
in the arterial signal,
and that's what these are shown here.
This is an arteriovenous fistula in
a renal transplant.

Twinkle Artifact

Okay, the last example of artifact
and color Doppler is the twinkle artifact.
This occurs secondary to ring down.
When a sound pulse enters the body, it can interact
with certain structures and cause them to resonate
or ring the ringing structure then transmits a continuous
stream of sound at a different frequency back
to the transducer.
Now that returning sound is interpreted by the transducer
as a Doppler shift since it's a different frequency from the
original sound pulse, and it also interprets it
as arising from interfaces deep
to the structure that's resonating.
Now, this phenomenon tends to occur
with crystalline material, calcification, gas,
and metallic structures.
This is a schematic trying to illustrate why this happens.
So the white object in the middle
of the screen is a tuning fork, so
that represents the resonating structure in the body.
When sound hits that, the sound reflects off of it,
but it also causes it to ring.
So as that original sound pulse returns to the transducer,
that resonating structure sends a continuous stream
of sound back to the transducer as well.
And that continuous stream just follows the original pulse.
Well, when the original pulse gets back, it is timed
and the transducer recognizes
that this resonating structure is a certain depth
and it writes an echo there.
But as the other pulses from the resonating structure
themselves return, the machine recognizes a sound delay
and it assumes that the signal is coming from a structure deep
to the actual resonating structure.
So as those pulses go back,
the machine writes additional echoes deep
to the resonating structure.
Here's a classic example in the right upper quadrant showing
the ring down artifact from gas with a loop of bowel.
Well, you can get the same sort of thing
with a color Doppler.
This is a patient that has some crystalline material in a
tiny cyst in a kidney.
And you can see as we magnify up on that, that this
material produces a short ring down or comet tail artifact.
Now when you turn the Doppler on,
you see the twinkle artifact.
Now notice that the color shades are very bright shades
of red and very bright shades of blue shifting towards yellow and towards green.
That's because the magnitude
of the frequency shift is very high
because the frequency, the resonant frequency
of the crystals is very different from the transmitted
frequency of the transducer.
Here is a polycystic kidney,
yet crystals frequently in the cysts
of these polycystic kidneys don't see much here on gray
scale, but when you look on color Doppler,
you can see the twinkle artifact associated with some of
that crystalline material within these cysts.
This is a pancreatic stone within a dilated pancreatic duct.
It looks pretty benign and easy to understand on this gray scale image.
But notice when we put on color Doppler,
it produces a very strong twinkle artifact
and that just reflects
that the crystalline matrix within this stone is resonating
at a different frequency than the transmitted sound
and being interpreted as a frequency shift.
Now if you do pulse Doppler from these twinkle artifacts, you get a waveform
that is completely unphysiologic.
Sometimes it will look like this a very strong signal
that just looks like Doppler noise.
These are examples of ring down from cholesterol crystals in the wall
of the gallbladder in a patient with adenomyosis.
Notice that you see the twinkle artifact
as well on both the color Doppler and on power Doppler.
Now, most of the time the twinkle artifact is not a problem, and in fact, in identifying kidney stones it can be
a useful artifact, but if you don't interpret it correctly,
you can mistake twinkle artifacts for true flow.
This is a patient with an endoluminal stent graft
and a large aortic aneurysm,
and you can see the flow within the two limbs
of the stent here,
but you also see signal within the aneurysm
that could potentially be mistaken as an endoleak.
However, this isn't real flow.
This is a twinkle artifact likely arising from a little bit
of calcification in that anterior aneurysm wall.
Now, as before, when you do pulse Doppler on something like
this, you get a signal that is non physiologic.
This is just kind of a real scattered high amplitude signal, but it doesn't look like any true flow
that you would see in an endoleak.
This is an example of a patient that had a calcified complex renal cyst.
You can see it on gray scale here.
Now, anytime you have a renal mass, you want
to know if there's internal vascularity.
If there is, then you can be fairly certain
that you're dealing with a neoplasm.
So when we did Doppler analysis on this lesion, you can see
that there is a little bit of flow
or apparent flow within the anterior aspect of that mass.
Well, actually that isn't flow.
That's a twinkle artifact coming from some of the calcium
within the wall of this cyst.
When we did a CT scan on this patient, you see
that the non-contrast scan
and the contrast enhanced scan are identical
and there was no contrast enhancement within this lesion.
So we feel comfortable that this is a benign,
although complex, renal cyst
and that this is artifactual signal arising from twinkle
rather than true signal arising from internal tumor vascularity.

Conclusion

So in conclusion, it's important
to understand Doppler principles in order to recognize these different Doppler artifacts.
And by recognizing Doppler artifacts,
you will avoid making mistakes.
And for some Doppler artifacts, you'll be able
to improve your interpretation by properly interpreting the artifact itself.
Thank you.