Understanding Doppler - SD

Introduction

Hi, I am Dr. Christopher Merritt, professor of radiology at the Thomas Jefferson University in Philadelphia.

Topic of the presentations that we'll follow is Doppler.

As you know, doppler is now an integral part of almost all ultrasound examinations and in order to use doppler effectively, it's important to understand some of the basic principles of Doppler, how these relate to the information that Doppler provides for diagnosis, as well as some of the limitations and artifacts that arise in the Doppler examination.

The presentation is divided into eight parts, dealing with the basic principles of Doppler, ultrasound and later on with some of the clinical applications.

Doppler is an important part of diagnostic ultrasound and an understanding of basic doppler principles is important in utilizing Doppler effectively understanding and recognizing artifacts and optimizing the quality of information obtained from the Doppler examination.

The next eight segments will review some of the basic principles of Doppler, ultrasound, their applications and important artifacts and quality control issues, and obtaining accurate doppler information.

The Doppler Effect

We'll begin with a discussion of the Doppler effect.

Ultrasound, of course, is pressure waves that are transmitted through matter.

We can depict these as alternate areas of increased and decreased pressure that move through tissue over time.

The ultrasound image is based on differences in the amplitude of the back scattered echoes reflected from interfaces within the body, but the back scattered ultrasound signal also contains information related to changes in frequency of the back scattered ultrasound signal.

And this difference in frequency provides us information used to evaluate the presence of blood flow, its direction and velocity.

If we look at an ultrasound image with a doppler obtained from a segment within the image, the image information in this case the liver surrounded by CES is a display of the amplitude of the back scattered ultrasound energy, stronger reflectors, producing brighter echoes and weaker reflectors producing darker echoes.

This is all a display of the amplitude in the back scattered ultrasound signal.

The spectral display, however, is based on the detection and analysis and display of frequency differences in the back scattered signal.

These frequency differences arise as a result of the dopper effect.

Consider a transducer on the left insulating an immobile object on the right.

The frequency of the sound transmitted to the object and that reflected are the same.

We can depict the transmitted frequency as F sub T and the reflected or returning frequency as F sub R, and in this case, if the object is stationary, the reflected and transmitted frequencies are equal, or we can say that the difference between the reflected and transmitted frequencies is equal to zero.

Now if on the other hand the object is moving in this case toward the transducer, the reflected frequency is greater than the transmitted frequency or the difference between the reflected and transmitted frequency is greater than zero.

It's the difference between the reflected and transmitted frequencies that is referred to as the doppler frequency shift.

If we look at an object which is moving away from the transducer, in this case, the reflected frequency is less than the transmitted frequency or the Doppler frequency shift, which is the difference between the reflected and transmitted frequencies is left than zero.

So by looking at the frequency information in the back scattered ultrasound signal, we can immediately determine two things.

If the frequency is the same as that transmitted, we can infer that the target is not moving.

If the frequency is greater or smaller than the transmitted frequency, then we can infer that the target is moving and by looking at the sign of the difference that is whether it is greater than zero or less than zero, we can determine the relative direction of movement toward or away from the transducer.

Here's an image of a Doppler sample paste in soft tissue and we can see on the right the spectrum displays no motion, no frequency shift.

This is because the target is stationary, the transmitted and reflected frequencies are equal and there is no doppler frequency shift.

If however, we sampled the frequency shift from within a vessel where blood is flowing.

In this case a component of motion is towards the transducer, the scan line indicated by the dotted line, then the reflected frequency will be greater than the transmitted frequency.

That is to say there will be a positive doppler frequency shift or a Doppler frequency difference of greater than zero.

In this case, the frequency information tells us two things.

One, that there is motion of the target in this case the blood, and secondly that there is relative component of motion toward the transducer.

A particular interest to us in diagnostic ultrasound is the relationship between the doppler frequency shift, which we can measure with a ultrasound device and the velocity of the target that we are intonating usually blood.

The relationship of the Doppler frequency shift F sub D is simply that of twice the transmitted frequency times the velocity of the target divided by the propagation velocity of sound in the medium, which in human tissues is assumed to be 1,540 meters per second.

Let's look at a clinical example.

On the top we have a transducer, which is transmitting at a commonly used Doppler frequency of three megahertz and we're intonating a blood cell moving directly to toward the transducer.

Since the target is moving toward the transducer, we expect that the returning frequency will be slightly higher than the transmitted frequency, and in this case, we measure the reflected frequency to be 3.0039 megahertz.

The doppler frequency shift therefore is 0 0 3 9 megahertz or 3.9 kilohertz.

Since we know the assumed propagation velocity of sound and tissue is 14, 1,540 meters per second, a simple arithmetic allows us to calculate the velocity of the target, which in this case turns out to be one meter per second.

This concludes a very brief introduction to the basic concept of the Doppler effect and how it is used to measure velocity.

Analysis of the Doppler Frequency Signal

The next segment we'll discuss the analysis of the Doppler frequency signal.

Having reviewed the basic concept of the Doppler effect will now turn to the method by which we can analyze the Doppler frequency shift, which arises when we intonate moving targets with ultrasound.

The doppler frequency shift that is detected when a target is moving can be analyzed in several ways.

In the early days of Doppler, it was analyzed by listening to the Doppler frequency shift on headphones or through a loud speaker.

This is because the frequency shifts produced by clinical ultrasound devices.

Intonating clinical targets are actually in the audible range currently, however, most analysis is performed by evaluation of the Doppler spectral waveform.

As you know, ultrasound is part of the continuum of sound, which includes audible sound as well as some industrial applications with medical ultrasound utilizing the highest frequencies.

It just so happens that the Doppler frequency shifts that are generated when clinical ultrasound devices intonate arterial or venous blood are in the audible range.

And therefore, simply by listening to the doppler sounds, we can get some idea of the amount of frequency shift that is present.

And in fact, when you hear the sounds being produced during a Doppler examination, you are actually hearing a representation of the actual Doppler frequency shifts that are being measured by the ultrasound device.

Now, although audible analysis was used in the very early days of ultrasound, modern ultrasound devices provide a display of the Doppler spectrum and I'd like to comment a moment about how we go from the measurement of a Doppler frequency into the typical Doppler frequency spectrum display shown at the bottom of this image.

Remember, ultrasound is simply a change in pressure amplitude over time.

The way that we convert, a frequency displayed in terms of amplitude over time into a spectral display in which we are looking at, frequency shifts displayed over very short intervals over time is by a process called the Fourier transformation.

If we consider a frequency of a given amplitude, we can transform the information contained in the amplitude time display into the display shown on the bottom in which we represent a given frequency at a specific point along a scale of frequencies with the height of the bar, related to the amplitude of the waveform.

If we add a higher frequency in this case with a somewhat lower amplitude, we can display this as a bar of somewhat lower height, reflecting the lower amplitude of the waveform as well as further to the right on the frequency scale reflecting its higher frequency.

If we sample all of the information in the at all of the frequencies that are returned to the ultrasound device and perform this transformation, we then can generate the information needed to provide the typical spectral waveform.

Now you notice on top, after the transformation we have a display of the relatively relative frequency components of the doppler signal and their amplitude, whereas on the bottom we show frequency displayed on the vertical axis and we're showing it as it changes over very short intervals of time.

But where's the amplitude information gone?

Well, the amplitude information is actually reflected in the gray scale of the spectral display.

The bright components of the spectral display represent the high amplitude frequency components, whereas the faint dark gray components represent the low amplitude frequency components and you can see in the spectral display that at any given instant in time a range of frequencies are present, which vary in amplitude.

At this point, we've talked a little bit about the source of the Doppler information that is processed and the way that we analyze it in form of the Doppler spectrum.

The Importance of the Doppler Angle

The next segment discusses the importance of the Doppler angle and our estimation of velocity using Doppler data.

When doppler is used to evaluate blood flow, it's usually not possible to direct the transducer down the axis of the blood vessel and therefore blood flow is usually detected by insulating the vessel at some angle.

This is very important in performing accurate estimates of velocity, which is one of the goals of the Doppler examination.

If we consider an object which is moving directly towards the transducer as in this case, then the doppler frequency shift, which is measured, is reflected in the simple doppler equation in which the frequency shift is proportional to twice the transmit frequency, the velocity of the target and the propagation velocity of sound.

In clinical situations, however, this relationship of the transducer to the moving object usually doesn't exist and the blood is usually moving in a vessel at some angle to the direction of intonation.

In this case, a component of motion is directed towards the transducer, the large area error and a smaller component is directed at 90 degrees to the direction of intonation.

As we change the angle, we can see that the relative component of motion towards the transducer and at 90 degrees to the transducer changes and in a clinical situation in which the transducer may be at a fairly large angle with respect to the blood vessel, a relatively small component of motion is directed toward the transducer and most of the motion is directed at 90 degrees.

In these situations, the basic doppler equation does not apply.

We have to take into account the doppler angle to adjust for the component of motion toward the transducer, and for that which is at 90 degrees, That means that we need to measure the angle between the presumed direction of flow and the direction of the ultrasound beam.

And we call this angle the doppler angle.

This adds another factor to the doppler equation so that the frequency shift measured in this situation is again proportional to twice the transmit frequency and the velocity of the target reduced in proportion to the cosine of this angle and divided by the propagation velocity of sound and tissue.

Here we can see an example in which we are intonating blood within a vessel.

In this case, the angle is 60 degrees, the velocity of the blood is a hundred centimeters per second, and we're intonating with a transducer at a frequency of three megahertz.

These are the same velocities and transmit frequencies that were shown.

In an earlier example in this series, when the doppler angle was zero, the resulting doppler frequency shift, which is detected by the transducer, was almost four kilohertz.

Whereas at 60 degrees the doppler frequency shift is approximately half that that's because the cosine of 60 degrees is 0.5 and therefore the Doppler frequency shift represented by a velocity of a hundred centimeters per second is reduced by half from what it would be if we were looking at an angle of zero degrees.

Some ultrasound devices actually display the spectral waveform and doppler frequency shift.

This is an image from an older machine in which a sample is obtained from the internal carotid artery.

The scale on the right is in kilohertz, and we can see that the peak systolic frequency in this case is about 3.25 kilohertz.

Well, this doesn't help us very much because what we would really like to measure is the velocity of the blood flowing in the tissue or in the blood vessel.

And in order to do this, we need to incorporate a measurement of the doppler angle.

We do this by placing a cursor, within the vessel in which we are sampling blood flow and align the cursor so that it is parallel to the walls of the vessel.

This then allows us to calculate the angle between the presumed direction of flow and the direction of the ultrasound pulse.

And in this example, as illustrated in the annotation provided on the right of the image, the angle is 60 degrees.

This then allows us to convert the frequency shift which is detected by the machine into an estimate of velocity.

And in this case, as you can see, the estimated peak systolic velocity in this vessel is 60 centimeters per second.

We can only obtain this estimate of velocity by indicating to the machine what the presumed angle of flow direction is, allowing the calculation and incorporation of the Doppler angle into the Doppler equation.

Continuous Wave and Pulsed Doppler Ultrasound

In this section, I'll speak briefly about the concept of continuous wave and pulse doppler ultrasound and how these are utilized.

It's possible to produce ultrasound from a transducer in two ways.

One is to produce a continuous stream of ultrasound emanating from the transducer, and the other is to produce the ultrasound in brief pulses of ultrasound delivered at very high rates.

Continuous wave doppler utilized for some simple blood pressure measurement devices and pulse detection devices and typically employs two transducers, one transmitting the ultrasound energy, the other acting as a receiver.

This is a sensitive way of detecting flow, but does not allow differentiation of the site of origin from which the flow is arising.

In this example, two vessels lying near one another both produce components and contributions to the Doppler frequency shift, but with a continuous wave device, it is not possible to determine from which vessel or vessels the Doppler signal arises.

Pulse ultrasound however, uses the range gating principle to allow us to selectively sample the site from which the Doppler information is collected and analyzed.

In fact, this is the basic principle of all of ultrasound, the principle of echo ranging the concept is a very simple one.

If we consider an object at unknown depth, we initiate an ultrasound pulse times zero.

It takes a finite time for the pulse to travel to the object and a similar amount of time for the echo to return to the transducer.

By measuring this short interval of time and doing simple arithmetic based on our knowledge of the propagation velocity of sound in tissue, we can then very easily calculate the distance from the transducer to the object in question.

Doppler works very much in a similar fashion.

If we have two vessels and are only interested in capturing the Doppler information from the deeper of these two vessels, we can send out an ultrasound pulse.

We can tell the machine not to process the echo from the first vessel, but only to process the later echo coming from the second vessel.

Simply by using controls to indicate the time interval that we wish to wait before we process the signal, we can selectively position a Doppler sample volume at any site, within a structure that we desire.

Evaluation of Vascular Stenosis

One of the most important applications of Doppler ultrasound is in the evaluation of vascular stenosis.

When we look at Doppler applications, we can consider the assessment of large vessel changes such as occlusion and stenosis as well as changes in small vessels.

We will first talk about large vessel stenosis and in a later segment we'll discuss waveform analysis and inferences related to small vessel disease and organ perfusion.

We can think about stenosis of a vessel in several ways.

We can measure the reduction in the diameter of the vessel.

We can attempt to measure the cross-sectional area of the vessel, or we can attempt to measure the flow passing through the vessel.

If we consider a vessel, with a native diameter of eight millimeters, we can see that if the diameter is symmetrically reduced to four millimeters, we achieve a 50% reduction in diameter.

But more important is the fact that we achieve a 75% reduction in cross-sectional area.

That's the cross-sectional area.

Of course, that limits the amount of blood that can flow through the vessel with angiography and other methods.

It's common to attempt to measure the diameter from longitudinal views of the vessels, and in this case, we may be able to achieve a fairly accurate estimate of diameter reduction.

However, a pitfall of this direct measurement approach is that we can achieve a same measurement of diameter reduction with a lesion which is, very severe as indicated on the left, or one which is really not, quite as significant depending on the configuration of the plaque or the narrowing lesion.

With ultrasound, we depend on neither of these methods to evaluate stenosis.

We depend rather on an assessment of the change in velocity of the blood as it passes through the stenotic segment.

This looks like a rather complicated diagram, but is one that is very important in understanding how doppler ultrasound contributes to the assessment of stenosis.

If you look along the bottom of the graft, you'll see a series of images which represent a vessel which goes from a diameter of five millimeters on the right to completely occluded on the left.

This same series of images is shown at the top of the image, but here we're depicting changes in cross-sectional area rather than diameter.

And if you look carefully, you'll see that a diameter of two millimeters, which represents a 60% decrease in diameter, the cross-sectional area has been decreased 84% and a diameter of one millimeter corresponding to an 80% diameter reduction corresponds to a 96% reduction in cross-sectional area.

Now the vertical scale on the right shows the peak systolic velocity measured within the vessel at the point of stenosis, and you'll notice that this value, which is depicted by the yellow curve as well, less than a hundred centimeters per second and gradually increases as the vessel decreases in diameter from 20 to 30 to 40% at around 60% reduction in diameter.

There's a very rapid increase in velocity up to greater than 400 centimeters per second.

And then as the vessel approaches complete occlusion, the velocity declines again, reaching zero at the time of complete occlusion.

Now superimposed upon this graph is a blue curve, which is indicative of blood flow in milliliters per minute.

This scale is shown on the left.

You can see that for mild degrees of narrowing 20, 30, 40% diameter reduction, the blood flow is unchanged, remained steady at about 300 milliliters per minute, but beyond a diameter reduction of 60%, we begin to see a decline in the blood flow increasing rapidly as we go from 60 to 80 to greater than 90% diameter reduction.

These relationships are displayed in this fashion, which is called a Spencer diagram.

The sampling of velocity for this assessment is generally made at the point of greatest narrowing within the vessel.

Sometimes when calcified plaque obscures the lumen of the vessel at the point of suspected greatest narrowing, the measurement is taken immediately before or immediately distal to the shadowing produced by the plaque.

Let's look at an example here.

We start with a normal vessel with a diameter of five millimeters.

The diameter is a hundred percent of normal.

The cross-sectional area is a hundred percent of normal.

We measure the peak systolic velocity and find it to be 40 centimeters per second.

The blood flow to the end organ is normal at 300 millimeters per minute.

At a 30% reduction in diameter, We only have 70% of our residual diameter remaining.

Our cross-sectional area, however, has been reduced by 50%.

We see a little increase in velocity, but we have no change in the blood flow to the end organ at around 60% diameter reduction.

The cross-sectional area that remains is only 16% of what we began with.

And here now we begin to see a increase in velocity that is quite noticeable.

And at this point, when the velocity begins to increase, we also begin to see an effect on the end organ and that the volume flow is beginning to decrease.

The important observation here is that for mild degrees of stenosis 2030 or 40% diameter reduction, there's really very little change of velocity and there's no change on in organ blood flow.

And in essence, staler is, virtually useless in evaluating these mild degrees of stenosis.

But at around 60% diameter reduction, things begin to get interesting and Doppler really begins to play a role in evaluating these types of stenosis.

And this of course is important because these are the levels of stenosis that become hemodynamically significant and begin to deprive the end organ of its blood supply.

As we go from 60 to 80% diameter reduction, we can see that there's a huge increase in velocity as the diameter decreases to only 20% of its original value, and the cross-sectional area is reduced to only 4% of its original value.

The end organ at this point is receiving only two thirds of its expected blood flow.

Finally, at complete occlusion, we have no velocity.

We have no flow.

Now this, theoretical model displayed in the Spencer diagram, which we've just reviewed, actually works very well when, clinical data are applied to it.

These are actual data points from a study, published by ED Grant in 2000 in which he compared Doppler, velocity assessments indicated in the vertical scale on the left with angiographic estimates of vessel diameter stenosis.

And if we take these data points and superimpose them over the Spencer diagram, you can see how well these actual measurements correlate, with the predictions in this model.

Pitfalls and Limitations of Velocity Measurement

The last segment emphasized the importance of velocity as a means of estimating the degree of vessel stenosis.

Therefore, it becomes extremely important to understand pitfalls and limitations of velocity measurement using ultrasound.

The major pitfalls in velocity measurement are the use of improper doppler gain and improper doppler angle correction.

In some cases, aliasing may complicate the assessment as well.

Here's an example of a doppler measurement obtained from the proximal internal carotid artery.

If you look carefully, you can see that the sample volume is placed near the origin of the internal carotid artery.

The angle cursor is aligned parallel to the vessel walls with an angle of 60 degrees, and the estimated peak systolic velocity from this measurement is 60 centimeters per second.

Now here's another slide taken just a moment later without any change in the position of the doppler sample or the angle correction and the velocity is estimated here to be only about 45 centimeters per second.

And here another image again, without any change in the position of the sample volume or doppler angle in which the velocity is estimated to be 70 centimeters per second.

The question here then is what is the correct velocity?

Is it 45, is it 70 or is it 60?

Well, the reason that we see these three different, estimates of velocity is because we've used different settings for the doppler gain.

And in fact, if you look at this slide, you can see that beginning on the left, as we gradually decrease the doppler gain, the peak systolic velocity decreases.

It becomes very important therefore to use a standardized method for setting the doppler gain when velocity measurements are being performed.

This is particularly important when serial measurements are being made.

For example, in the carotid artery to evaluate for disease progression, the proper way to establish the doppler gain is to first carefully position the Doppler sample volume in the desired location and to align the doppler angle correction parallel to the vessel walls and then to turn the doppler gain up to the point where noise is seen in the background.

This is illustrated in the left portion of this image.

Then the doppler gain is gradually turned down until the noise in the background completely disappears, but no further that results in the measurement obtained near the middle of this, tracing.

If the doppler gain is continued to be turned down even more, there will be an underestimate of the peak systolic velocity.

So going back to the prior slides, the correct doppler gain setting is shown here, the velocity of 60 centimeters per second.

Here the Doppler gain is too low resulting in an underestimation of the peak velocity, and here the Doppler gain is too high, clearly evident with all the noise in the background resulting in an overestimate of the velocity.

Now another very important, factor in estimating velocity, of course, is the doppler angle because the doppler frequency shift is reduced in proportion to the doppler angle and if the doppler angle is incorrect, the velocity calculated from the frequency shift will also be incorrect.

Here's an example in which we have placed a sample within the vessel, and the angle cursor is indicated by the dotted yellow line.

This angle is indicated in the graphics on the right of the screen as 70 degrees With this data, the peak velocity in this vessel is estimated to be 110 centimeters per second, but as you can see from the orange line, the actual direction of blood flow is really near an angle of 58 degrees to the direction of intonation rather than 70 degrees.

Clearly there's an error in the placement of the doppler angle here.

Does this really make much difference in the estimate of velocity?

Well, the answer is it makes a huge dis difference in the estimate of velocity.

'cause in this case, when the proper doppler angle is entered, that is 58 degrees, we see that the velocity is actually only 60 centimeters per second, not 110 centimeters per second.

An error of this type then can result in an incorrect estimation of the degree of stenosis within a vessel perhaps leading to improper management.

Now, one of the rules that, most individuals using doppler are quite familiar with is the rule that doppler angles greater than eight than 60 degrees should not be employed.

And this, graphic attempts to, indicate why that is such an important number.

What is displayed here in yellow is the measurement error in percent for a five degree error in the estimate of the Doppler angle, and the red line indicates the doppler angle of 60 degrees.

You can see to the left of the red line, the doppler angles of less than 60 degrees are associated with very small measurement errors in velocity, but at 60 degrees and beyond, even a small error in the assessment of the Doppler angle, particularly as the angle approaches, 70 or 80 degrees may result in a measurement error of 100, 200 or even 300%.

Therefore, it's important not only to measure the angle correctly, but to attempt to use doppler angles of 60 degrees or less so that a slight error in estimate of the angle will not result in a huge error in the estimate of velocity.

Another potential source of errand analysis of the Doppler signal is related to the use of the wall filter.

A wall filter is an electronic filter which removes low frequency shifts from the display.

This is generally intended to remove low frequency noise associated with vessel wall motion.

In this example, the wall filter is set at a hundred hertz and this basically suppresses the information in the spectral display below the level of the yellow dotted line in the waveform.

Here we can see the same waveform with the wall filter set at 400 hertz.

The dotted line again indicates the 400 hertz level, and you can see that the information, essentially all of the information in diastole is suppressed by this filter.

And those cases we were performing waveform analysis looking for changes in diastolic flow, perhaps as an indication of increased peripheral resistance.

The use of an excessively high wall filter may result in an inaccurate conclusion that, diastolic flow is reduced or absent when in fact it's perfectly normal and we simply electronically suppressed it in addition to the evaluation of peak systolic velocity as an indicator of stenosis.

Another component which is sometime utilized relates to the range of velocities present in the Doppler spectrum.

And this leads us to the discussion of spectral broadening.

As you can see in this illustration, at any given point in the cardiac cycle, blood cells are flowing at a range of velocities.

In this case, the highest velocity, cells and diastole are moving at a velocity of approximately 30 centimeters per second and the low velocity components in the vessels at around 10 centimeters per second.

The range between 30 and 10 centimeters per second is the spectral breadth.

This is generally not of great interest to us, except in cases of very severe stenosis where we may see extreme spectral broadening.

For example, in this vessel, in addition to very high peak systolic velocities, we can see that throughout the waveform there's a very large range between the highest moving and the lowest moving red cells as measured by doppler.

This is an indication of turbulence, and this feature of the waveform is important as we approach complete stenosis.

As we recall from this diagram, as the vessel narrows velocity increases, but as we approach complete occlusion, the velocity decreases again, finally coming to zero at the time of complete occlusion.

So as you can see, there's a point as the vessel approaches complete occlusion in which the peak systolic velocity may be very close to that which is seen in a normal vessel.

What is helpful in differentiating the normal peak systolic velocity from the waveform approaching complete stenosis is the presence of turbulence, which is reflected by extreme spectral broadening with very, very high velocities.

It's common to see a phenomenon which is recognized as aliasing.

Here we can see that the highest frequencies, the highest velocities are not displayed in the display but wraparound and appear below the baseline.

As you recall, ultrasound involves the use of brief pulses of sound which are delivered at high frequencies.

The frequency of these pulses is called a pulse repetition frequency and is often several thousand pulses per second.

The pulse repetition frequency is very important in the production of aliasing and even more important in color doppler in the way the information is displayed.

For example, here we have a sampled waveform with a frequency of two kilohertz, and we're sending ultrasound pulses as a frequency of five kilohertz.

In this situation, the high pulse repetition frequencies allows accurate sampling of the Doppler frequency, and we did not see sine if on the other hand, the pulse repetition frequency is small.

In comparison to the sample frequency, there's under sampling result in display of a lower frequency than the frequency being sampled, and this is responsible for the lower frequency components displayed in the, spectral display.

In general, in order to obtain accurate sampling without sine the pulse repetition frequency must be greater than twice the highest frequency being sampled and preceding segments.

Spectral Waveform Analysis

The evaluation of stenosis using velocity analysis was discussed.

Ultrasound is also capable of providing information about changes in sites proximal and distal to the Doppler sampling site.

This is by way of analysis of the spectral waveform.

This is an example of a dramatic change in the appearance of a spectral waveform over a single heartbeat.

On the left, we see a high resistance waveform with very little flow in diastole and in fact a little reversal of flow in early diastole.

On the right we see a typical low resistance arterial wave form in which flow continues at a fairly high rate throughout most of diastole.

This change between high resistance and low resistance can be produced very easily by simply in plating a blood pressure cuff over the forearm and sampling in the brachial artery central or proximal to the pressure cuff.

When the cuff is inflated above arterial pressure, the downstream resistance is of course very high, and the arterial waveform proximal to the side of pressure application will appear as in the left hand side of the image, a high resistance arterial waveform.

If after a few moments the blood pressure cuff is suddenly released, the peripheral arterial bed is vasodilated offering very little resistance, and the waveform changes to that shown on the right of the, spectral display.

There are several means that are used to attempt to, describe the differences in systolic and diastolic flow, which may arise from differences in downstream resistance.

The simplest of these is the systolic to diastolic or AB ratio, which is simply a ratio of the peak systolic amplitude, which can be either frequency or velocity and the end diastolic frequency or amplitude.

A more commonly used index is the resistive index, which is the difference between the peak systolic and end diastolic amplitudes divided by the peak systolic amplitude.

A third method, which may be used is simply the difference between the peak systolic and end diastolic divided by the mean amplitude.

In addition to the use of these indices to evaluate downstream resistance, we also need to look at the, waveform in terms of the acceleration of blood in systole because this may be a useful indicator of upstream or proximal stenosis.

As the vessel becomes severely narrowed, two changes occur in the ray form, which are illustrated here.

First, the acceleration in systole becomes less, and secondly the amplitude of the systolic P begins to diminish so that on the right we have a small hump shaped waveform, a classic tardis parvis.

Here's actually a clinical example, and a patient with a liver transplant shortly after transplant.

The baseline study shows a normal waveform with a high systolic amplitude and flow continuing throughout diastole eight weeks after transplantation, we see a somewhat different wave form and that the systolic amplitude is reduced.

There is spectral broadening indicating turbulence, and the ratio of systolic to diastolic amplitude indicates decreased to peripheral resistance.

At 16 weeks, we see delayed systolic acceleration in a very low amplitude waveform, a classic tardis parvis waveform.

These are all changes associated with significant and progressive hepatic artery stenosis.

Color Doppler

Of all the applications of doppler and ultrasound color, doppler is probably the most widely used and probably the least well understood in terms of its capabilities and limitations.

And in fact, color doppler is probably one of the most complicated imaging methods, available in radiology to interpret.

Going back to basics, ultrasound is, based on the detection of reflected, changes in pressure, mechanical pressure waves propagated through tissue differences in the back scattered echo amplitude that produce the information used to generate a gray scale image.

Whereas differences in frequency allow us to detect the presence of flow, determine its direction, and to calculate its velocity.

With color doppler, we're doing both of these things.

We are sending from our transducer short ultrasound pulses, which are processed for their amplitude content to produce the gray scale image component of the, doppler, color display, as well as sending longer pulses which collect doppler information and allow both these components of the amplitude image information and the Doppler frequency shift information to be combined and displayed in the final image.

The color doppler is, a very appealing image and it looks very simple to interpret, but in fact, the information contained in the image is simply the information of the Doppler frequency shift recorded in each pixel.

And as we'll review in a moment, the Doppler frequency shift at any given point is determined by number of important variables, and the color in which it is displayed in the image is displayed by equally complicated number of issues and variables.

So the important point to remember is that when you look at a colored Doppler image, you're simply looking at a map of the doppler frequency shift at each point in the image.

Now the convention is to use different colors to indicate the sign of the Doppler frequency shift.

As you'll recall, if the target is moving away from the transducer, we end up with a negative doppler frequency shift.

If the target is moving towards the transducer, we have a positive doppler frequency shift, and by convention we use colors to indicate whether the doppler shift is positive or negative.

Once we've determined the sign of the doppler frequency shift, we then look at the value itself and we can use a color map to show the relative doppler frequency shift.

Generally less saturated colors indicate higher doppler frequency shifts, darker colors indicate lower doppler frequency shifts, and as we'll see a little bit later, we can really assign any color to any doppler frequency shift.

And this in fact is a source of considerable potential confusion and error.

Now, the very early Doppler devices actually showed in their color map illustrated in the yellow circle, the doppler frequency shift, which was being reflected by the scale.

And in the illustration shown here, the very light blue color was meant to indicate a doppler frequency shift of about 1.25 kilohertz.

Unfortunately, modern machines have not, that tradition and we'll come back to that issue in a moment.

Now, we spoke briefly when we discussed spectral doppler about aliasing and color doppler.

If the doppler frequency shift is too great to be displayed in the color map that is selected, the color wrapper wraps around to the opposite end of the color spectrum as illustrated in this slide.

So that here in this view of the aorta near the left of the image, we see a small blue pixels, which are actually alias Containing higher doppler frequency shifts than the surrounding pixels.

Now, the frequency shift displayed in each pixel is of course affected by two things.

It's affected by the doppler angle, which determines the component of motion towards the transducer and by the velocity of the target of the red cell.

And this is where things begin to get complicated.

Let's look at this vessel.

This is a view of the common carotid artery, the bifurcation, the external and the internal carotid artery.

And here we can see a broad range of colors reflected at different sites within these vessels at the bifurcation and in the distal common carotid artery, we can see red, blue, red, blue.

And the problem here is that the doppler angle is almost 90 degrees.

In addition, because the doppler angle is very large, there's ambiguity in terms of the relative direction of these small doppler frequency shifts, and therefore the colors shift from red to blue.

As we go a little bit further, we can see that the doppler angle becomes slightly smaller in the proximal internal carotid artery, we see somewhat less saturated colors which indicate higher doppler frequency shifts being detected at this point.

And then as we go a little bit further, we see that the doppler angle becomes even smaller and we begin to see aliasing within both the external and internal carotid arteries.

Now, the interesting thing about this image is that if we actually carefully measure with angle corrected spectral doppler, we find that the velocity at all of these sites is essentially identical.

The Doppler frequency shifts of course, are changing because the doppler angle is changing, but the velocity is constant throughout the entire image.

Now, just to, refresh your memory, if we have a target which is moving at some angle with respect to the direction of the ultrasound beam, only a small component of that motion is directed towards the transducer, and most of it is directed at 90 degrees, and it's this angle between the direction of flow and the direction of the ultrasound beam that determines the Doppler frequency shift, which is detected and measured by the ultrasound device.

Now, to make color doppler even more complicated is the fact that a single color can represent many doppler frequency shifts or for that matter any doppler frequency shift depending on the pulse repetition frequency.

But when we talk about spectral doppler pulse repetition frequency is pretty easy to understand how that produces wrap over of the spectral display and color.

However, the issue is quite a different matter.

To understand the impact of pulse repetition frequency on the color Doppler display, we need to look at this slide for a moment.

This is a coronal view of the aorta in which we see a range of colors, with some aliasing on the left, Medium colors of red in the middle of the aorta, and then red and blue as the vessel moves toward and away from the transducer as it bifurcates.

The color map on the left of the image indicates the values and frequency shift of the colors assigned to the image.

And in this particular case, the very lightest red colors are meant to represent a frequency of around 1.25 kilohertz and the very lightest blue colors a frequency shift of minus 1.25 kilohertz.

In order to change the assignment of these colors to higher velocities, we can simply change the pulse repetition frequency, for example, by increasing the pulse repetition frequency, we can extend the range of reds that apply to, frequency shifts, of greater than 1.25 kilohertz or to less than minus 1.25 kilohertz.

And here's an example. Here are two views of the common carotid artery.

The only thing that has changed in these two images is the pulse repetition frequency on the left, the pulse repetition frequency is very high, 5,000 hertz, and as a result, the highest frequency shifts displayed in the pixels that constitute this image are all mapped in the red end of the, color display.

Whereas on the right the pulse repetition frequency is 2,500 hertz, and in this case, the highest frequencies which are in the center of the vessel Are alias and wraparound to the other color display.

This can have considerable clinical importance.

For example, here are four images of the junction of the left subclavian vein and the innominate vein, And we can see that the color filling this vessel changes dramatically as we change the pulse repetition frequency from 2.5 kilohertz down to 0.9 kilohertz at the high pulse repetition frequency.

The low velocity components are all displayed in the black portion of the color map, giving the impression that the vessel does not completely fill with flow and perhaps even raising the question of, thrombosis.

As we decrease the pulse repetition frequency to 1800 hertz, more of the lower frequency pixels are included in the color map and begin to appear within the vessel.

At 1200 hertz, we see more fill in, but with the appearance of some aliasing and at 900 hertz, we can see that the entire, vein fills with flow.

However, with considerable alias, all of these are effects strictly of the pulse repetition frequency that we have selected in the color display.

Now, what about power doppler?

A couple of brief comments about power doppler with color doppler.

When we turn the gain up too high, we detect and display noise, and with a color doppler map, the noise appears at all the colors in the frequency map.

With power doppler, we look at the amplitude of the doppler information rather than the frequency.

Because noise tends to be very weak and low in amplitude, we can cause the noise to be displayed in the dark or invisible portion of the color map and only the true doppler signals to appear in the color portion of the map.

This effectively allows us then to use much higher gain to detect, Doppler information when we use a power doppler display than would be possible if we were using a frequency display.

This means that we can fill in vessels more completely.

We can reduce, noise in the image, but we pay a price in removing from the image, the frequency information, which provides us information about the direction of flow and the relative, velocity of flow, assuming that we correct for the Doppler angle.

In conclusion, then color doppler is a very important adjunct to spectral imaging for the evaluation of blood flow, but when critical assessments are made, it probably should not be used alone.

Color and spectral doppler are complimentary color localizes flow, and can help assess the presence and direction of flow.

But spectral doppler is essential to evaluate velocity and to exclude low velocity flow and vessels that are thought to be possibly occluded.