3 Basic Concepts of Doppler Frequency Spectrum Analysis and Ultrasound Blood Flow Imaging
If blood flow were continuous rather than pulsatile, if blood vessels followed straight lines and were uniform in caliber, if blood flowed at the same velocity at the periphery and in the center of the lumen, and if vessels were disease-free, then each blood vessel would produce a single Doppler ultrasound frequency. However, blood flow is pulsatile, vessels are not always straight or uniform in size, flow is slower at the periphery than in the center of the vessel, and the vessel lumen may be distorted by atherosclerosis and other pathology. For these reasons, blood flow produces a mixture of Doppler frequency shifts that changes from moment to moment and from place to place within the vessel lumen. Spectrum analysis is needed to sort out the jumble of Doppler frequencies generated by blood flow and to provide quantitative information that is critical for diagnosis of vascular pathology.
The word spectrum, as derived from Latin, means image. You may think of the Doppler spectrum as an image of the Doppler frequencies generated by moving blood.1–8 In fact, this image is a graph showing the mixture of Doppler frequencies present in a specified sample of a vessel over a short period of time.1–3 The key elements of the Doppler spectrum are time, frequency, velocity, and Doppler signal power. These elements are best described in pictorial form; therefore, this information is provided in Figure 3-1, rather than in the text. Please review this figure now, directing particular attention to the four key elements cited previously.
FIGURE 3-1 The Doppler spectrum display. The following information is presented on the display screen (A, Entire display; B, Magnified Doppler spectrum).Color flow image: The vessel, the sample volume, and the Doppler line of sight are shown in the color flow image at the top of the display screen. Color flow information: The “color bar” to the right of the image shows the relationship between the direction of blood flow and the colors in the flow image. By convention, the upper half of the bar shows flow toward the transducer. This is logical, as this part of the bar is nearest to the transducer in the image. The lower half represents flow away from the transducer. In this case, red/orange colors correspond to flow toward the transducers, and blue/green colors indicate flow in the opposite direction. A shift in color from red to orange or from blue to green represents increasing flow velocity. Doppler angle: The Doppler angle for the spectral Doppler appears at the upper right of the display screen, in this case 60°. Time: The time is represented on the horizontal (x) axis of the Doppler spectrum, at the base of the display. The lines represent divisions of a second, but typically a scale is not provided. Velocity: Blood flow velocity (cm/sec) is shown on the vertical (y) axis of the spectrum. In this case, velocity is shown on both vertical axes. On some instruments, the velocity is shown on one vertical axis and the Doppler-shifted frequency (KHz) on the other. The distribution of velocities within the sample volume is illustrated by the brightness of the spectral display (the z-axis). To better understand the z-axis concept, examine the magnified spectrum shown in B and imagine that the spectral display is made of tiny squares called pixels (for picture elements). You cannot see the pixels in this image because they are purposely blurred to smooth the picture. The pixels are there, however, and each corresponds to a specific moment in time and a specific frequency shift or velocity. The brightness of a pixel (its z-axis) is proportionate to the number of blood cells causing that frequency shift at that specific point in time. In this example, the pixels at asterisk 1 are bright white, meaning that a large number of blood cells have the corresponding velocity (about 41 cm/sec) at that moment in time. The pixels at asterisk 2 are black, meaning that no (or very few) blood cells have the corresponding velocity (about 12 cm/sec) at that moment. The pixels at asterisk 3 are gray, meaning that a moderate number of blood cells have the corresponding velocity (about 35 cm/sec) at that moment. Got it? If not, read this again and remember that the brightness of each pixel is proportionate to the relative number of blood cells with a specific velocity at a specific moment in time. Since the brightness of the pixels also shows the distribution of flow energy, or power, at each moment in time, the spectrum display is also called a power spectrum. Flow direction: The direction of flow is shown in relation to the spectrum baseline. In this case, flow toward the transducer is shown above the baseline, and flow away from the transducer is shown below the baseline. Note that the number 40 in the lower right corner is preceded by a minus sign. This is because the area below the baseline corresponds to flow away from the transducer, which would generate a negative Doppler shift. The relationship between the flow direction and the Doppler baseline may be reversed by the operator, but flow toward the transducer will always be represented by positive velocity or frequency values. Peak velocity envelope: The peak velocity throughout the cardiac cycle is shown by the blue line outlining the Doppler spectrum. Based on this envelope, a numeric output is provided at the bottom left, showing the peak systolic velocity (PSV) and the minimum diastolic velocity (MDV). In this case, the MDV also corresponds to the end-diastolic velocity, but this is not necessarily the case. The instrument also automatically calculates the resistivity index (RI) and the pulsatility index (PI), as shown below the velocity values. Pulse repetition frequency: A noteworthy number shown on the display is the pulse repetition frequency (PRF). The PRF for the color flow image is shown at the left of the image (1000 Hz, or cycles, or pulses per second). The PRF for the spectral Doppler is much higher (6250 Hz), as shown to the right of the color flow image. This difference illustrates the fact that the color flow image is based on the average Doppler frequency shift or velocity, while the spectral Doppler values are shown as absolutes, without averaging. A higher PRF is needed for the spectral Doppler to ensure that systolic velocities are shown accurately, without aliasing.
The Doppler frequency spectrum that you have just reviewed in Figure 3-1 is sometimes called a power spectrum,1–3 because the power, or strength, of each frequency is shown by the brightness of the pixels. The power of a given frequency shift, in turn, is proportionate to the number of red blood cells producing that frequency shift. If a large number of blood cells are moving at a certain velocity, the corresponding Doppler frequency shift is powerful, and the pixels assigned to that frequency are bright. Conversely, if only a small number of cells are causing a certain frequency shift, the pixels assigned to that frequency are dim. The power spectrum concept is important for understanding power Doppler flow imaging, which is discussed later in this chapter. The concept of the power Doppler spectrum is nicely illustrated in Figure 2-29.
The echoes that are reflected back to the transducer from moving cells in a sampled blood vessel contain only Doppler frequency shift information; yet the Doppler spectrum often displays both velocity (cm/sec or m/sec) and frequency (kHz) information. How does the instrument convert the Doppler frequency shift to velocity? This conversion occurs when the sonographer “informs” the duplex instrument of the Doppler angle,1,2,9 which is shown in Figure 3-2. If the instrument “knows” the Doppler angle, it can then compute the blood flow velocity via the Doppler formula (see Chapter 2). You may note in this formula that the frequency shift is proportional to the cosine of the Doppler angle, theta. When the operator informs the ultrasound machine of this angle, the frequency shift is proportional to blood flow velocity. Voila! The frequency spectrum becomes a velocity spectrum. A Doppler angle of 60 degrees or less is required to derive accurate frequency and velocity measurements. If the angle is greater than 60°, velocity measurements are unreliable. Although there is greater error in measurements obtained at higher angles, some applications (e.g., carotid examinations) are more easily performed at angles closer to 60°. It is generally recommended that the Doppler angle should be less than or equal to, but not greater than 60° for greatest accuracy.
FIGURE 3-2 The Doppler angle and sample volume. The nearly vertical line is the Doppler line of sight. The line in the center of the blood vessel indicates the axis of blood flow. The angle formed by these two lines is the Doppler angle (θ). The parallel lines (arrows) indicate the length of the Doppler sample volume.
In spite of potential measurement inaccuracy described in the previous paragraphs, it is desirable to operate the duplex instrument in the velocity mode rather than the frequency mode for two reasons.1,2,9 First, velocity measurements compensate for variations in vessel alignment relative to the skin surface. For instance, the Doppler frequency shift observed in a tortuous internal carotid artery might be radically different from one point to another, but angle-corrected velocity measurements will be similar throughout the vessel, in spite of dramatic changes in vessel orientation relative to the skin. Second, the Doppler frequency shift is inherently linked to the output frequency of the transducer, but velocity measurement is independent of the transducer frequency. For instance, if the output frequency goes from 5 to 10 MHz, the frequency shift is doubled. Imagine the clinical consequences of such frequency changes. If transducers with different frequencies were used to determine stenosis severity, different diagnostic parameters would be needed for each ultrasound transducer (e.g., 3.5, 5, or 7.5 MHz). This problem is eliminated when the instrument converts the “raw” frequency information to velocity data.
The human ear was the spectrum analysis instrument used initially for Doppler blood flow studies. The ear is a highly capable spectrum analysis instrument, which is evident in its ability to distinguish one person’s voice from another. Even though duplex ultrasound instruments are equipped with electronic spectrum analysis devices, an audible Doppler output is provided as well, to take advantage of the human ear’s capabilities. Certain features of the Doppler flow signal can be appreciated aurally that are difficult or impossible to display electronically, and as a result, the audible Doppler signal remains important in ultrasound vascular diagnosis. For instance, in very high grade carotid stenoses, a distinctive whining or whistling sound is heard. In spite of its abilities, however, the human ear has three major drawbacks. First, the ear is a purely qualitative device; second, it is not equipped with a hard copy output for permanent storage; and third, some ears work better than others—some cannot hear very high frequencies. Electronic spectrum analysis overcomes these obstacles.
The frequency spectrum shows blood flow information from a specific location called the Doppler sample volume, which is illustrated in Figure 3-2. You should be familiar with the following three characteristics of the Doppler sample volume: First, it is, in fact, a volume (three dimensions), even though only two of its dimensions are shown on the duplex image. The “thickness” of the sample volume cannot be shown on the two-dimensional spectrum display, and this can sometimes lead to errors of localization. Doppler signals may be obtained from vessels that are marginally within the sample volume but are not shown on the two-dimensional display. For instance, the ultrasound image may show the internal carotid artery, but you may actually be receiving flow signals from an adjacent external carotid branch. Second, the actual shape and size of the sample volume may be somewhat different from the linear representation shown on the duplex image. Third, and most important, the Doppler spectrum displays flow information only within the sample volume and does not provide information about flow in other portions of the blood vessel that are visible on the ultrasound image. Therefore, if the sample volume is positioned incorrectly, key diagnostic information may be overlooked.
The frequency spectrum shows blood flow relative to the transducer. Flow in one direction, toward the transducer, is displayed above the spectrum baseline, and flow in the opposite direction is shown below the baseline. One must always remember that the flow direction is relative to the transducer and is not absolute. The apparent direction of flow can be reversed by turning the transducer around or by pressing a button on the instrument that inverts the spectrum! The arbitrary nature of this arrangement can lead to significant diagnostic error. Clues to the correct direction of flow can be found by comparing the color (e.g., red or blue) in the vessel to the color bar or color velocity scale and by checking whether the velocity information on the spectrum is positive (toward the transducer) or negative (away from the transducer). Another method to check direction of flow is comparison with a reference vessel in which the flow direction is known (e.g., when working in the abdomen, the aorta is a handy reference vessel).
In arteries, each cycle of cardiac activity produces a distinct “wave” on the Doppler frequency spectrum that begins with systole and terminates at the end of diastole. The term waveform refers to the shape of each of these waves, and this shape, in turn, defines a very important flow property called pulsatility.1,2,10–28 In general terms, Doppler waveforms have low, moderate, or high pulsatility features, as illustrated in Figure 3-3. Please review this figure before proceeding to the material that follows.
FIGURE 3-3 Pulsatility. A, Low pulsatility is indicated by broad systolic peaks and persistent forward flow throughout diastole (e.g., the internal carotid artery). B, Moderate pulsatility is indicated by tall, sharp, and narrow systolic peaks and relatively little diastolic flow (e.g., the external carotid artery). C, High pulsatility is characterized by narrow systolic peaks, flow reversal early in diastole, and absence of flow late in diastole. In this classic triphasic example, the first phase (1) is systole, the second phase (2) is brief diastolic flow reversal, and the third phase (3) is diastolic forward flow. Triphasic flow is seen in normal extremity arteries at rest.
Low-pulsatility Doppler waveforms have broad systolic peaks and forward flow throughout diastole (see Figure 3-3, A). The carotid, vertebral, renal, and celiac arteries all have low-pulsatility waveforms in normal individuals because these vessels feed circulatory systems with low resistance to flow (low peripheral resistance). Low-pulsatility waveforms are also monophasic, meaning that flow is always forward, and the entire waveform is either above or below the Doppler spectrum baseline (depending on the orientation of the ultrasound transducer).
Moderate-pulsatility Doppler waveforms have an appearance somewhere between the low- and high-resistance patterns (see Figure 3-3, B). With moderate flow resistance, the systolic peak is tall and sharp, but forward flow is present throughout diastole (perhaps interrupted by early-diastolic flow reversal). Examples of moderate pulsatility are found in the external carotid artery and the superior mesenteric artery (during fasting).
High-pulsatility Doppler waveforms have tall, narrow, sharp systolic peaks and reversed or absent diastolic flow. The classic example of high pulsatility is the triphasic flow pattern seen in an extremity artery of a resting individual (see Figure 3-3, C). A sharp systolic peak (first phase) is followed by brief flow reversal (second phase) and then by brief forward flow (third phase). High-pulsatility waveforms are a feature of circulatory systems with high resistance to blood flow (high peripheral resistance).
Pulsatility and flow resistance may be gauged qualitatively, either by visual inspection of the Doppler spectrum waveforms or by listening to the auditory output of a Doppler instrument. Qualitative assessment of pulsatility is often sufficient for clinical vascular diagnosis, but in some situations (e.g., assessment of renal transplant rejection), quantitative assessment is desirable. A variety of mathematical formulae can be used for this purpose, but the most popular measurements are the pulsatility index (of Gosling), the resistivity index (of Pourcelot), and the systolic/diastolic ratio,24,26,28,29 all of which are illustrated in Figure 3-4.
Normal values for pulsatility measurements vary from one location in the body to another. Furthermore, both physiology and pathology may alter arterial pulsatility. For example, the normal high-pulsatility pattern seen in extremity arteries during rest converts to a low-resistance, monophasic pattern after vigorous exercise (because the capillary beds open and flow resistance decreases). Although this monophasic pattern is normal after exercise, it is distinctly abnormal in a resting patient and, in that circumstance, indicates arterial insufficiency resulting from obstruction of more proximal arteries. The point to be made here is that proper interpretation of pulsatility requires knowledge of the normal waveform characteristics of a given vessel and the physiologic status of the circulation at the time of examination. The status of cardiac function is also important; slowed ventricular emptying, valvular reflux, valvular stenosis, and other factors may significantly affect arterial pulsatility.
Acceleration is another important flow feature evident in Doppler spectral waveforms.24,25 In most normal situations, flow velocity in an artery accelerates very rapidly in systole, and the peak velocity is reached within a few hundredths of a second after ventricular contraction begins. Rapid flow acceleration produces an almost vertical deflection of the Doppler waveform at the start of systole (Figure 3-5, A). If, however, severe arterial obstruction is present proximal (upstream) to the point of Doppler examination, systolic flow acceleration may be slowed substantially, as shown in Figure 3-5, B and C. Quantitative measurement of acceleration is achieved by measuring the acceleration time and the acceleration rate (index), as illustrated in Figure 3-6. These measurements are used, for example, in evaluating renal artery stenosis.
FIGURE 3-5 Acceleration and damping. A, The acceleration time (0.03 sec) is normal in the right kidney. B, The acceleration time is prolonged (0.15 sec) in the left kidney due to severe proximal renal artery stenosis. (A and B are from the same patient.) C, Severely damped dorsalis pedis artery waveform distal to common iliac and superficial femoral artery occlusion. Normally, this waveform should look like Figure 3-3, C. Acceleration is severely delayed, and a large amount of flow is present throughout diastole, consistent with severe ischemia.
As you may have already surmised, vessels can be identified by their waveform pulsatility features.1,2,14,21–23,26 For example, Doppler waveforms readily differentiate between lower extremity arteries, which are distinctly pulsatile, and veins, which have gently undulant flow features. Doppler waveforms are particularly helpful in distinguishing the internal and external carotid arteries, which have low and moderate pulsatility, respectively. Pulsatility is also of value within the liver for differentiating among the portal veins, hepatic veins, and hepatic arteries, as discussed in Chapter 30.
Blood generally flows through arteries in an orderly way, with blood in the center of the vessel moving faster than the blood at the periphery. This flow pattern is described as laminar, because the movement of blood is in parallel lines.1,2,4,14,15 When flow is laminar, the great majority of blood cells are moving at a uniform speed, and the Doppler spectrum shows a thin line that outlines a clear space called the spectral window (Figure 3-7).*
FIGURE 3-7 Laminar flow. A, Illustration of parallel lines of blood cell movement. B, Doppler spectrum during laminar flow. At all times, the blood cells are moving at similar velocities. As a result, the spectrum is a thin line that encloses a well-defined black “window” (W).
In disturbed flow, the movement of blood cells is less uniform and orderly than in laminar flow. Disturbed flow is manifested as spectral broadening or filling in of the spectral window.1,2,4,15–19 The degree of spectral broadening is proportionate to the severity of the flow disturbance, as illustrated in Figure 3-8. Although disturbed blood flow often indicates vascular disease, it must be recognized that flow disturbances also occur in normal vessels. Kinks, curves, and arterial branching may produce flow disturbances, as illustrated quite vividly in the carotid bulb, where a prominent area of reversed flow is a normal occurrence11,20,21 (Figure 3-9). In addition, a spurious disturbed flow appearance may be created in normal arteries through the use of a large sample volume that encompasses both the slow-flow area near the vessel wall and more rapid flow at the vessel center.16–19 The Doppler spectrum, in such cases, appears broadened because both the high-velocity flow at the vessel center and the slow flow at the periphery of the vessel are encompassed by the wide sample volume.
FIGURE 3-8 Disturbed flow. A, Disturbed flow illustration. B, Minor flow disturbance is indicated by spectral broadening at peak systole and through diastole. C, Moderate flow disturbance causes fill-in of the spectral window. D, Severe flow disturbance is characterized by spectral fill-in, poor definition of spectral borders, and simultaneous forward and reversed flow. The audible Doppler signal has a loud, gruff character when flow is severely disturbed.
Modern duplex instruments are capable of measuring the volume of blood flowing through a vessel (volume flow).1,2,30–32 This is done by measuring the average flow velocity across the entire lumen (slow peripheral flow and high central flow) through several cardiac cycles while simultaneously measuring the vessel diameter, which is converted mathematically into cross-sectional area. Knowing the average velocity and the vessel area, it is an easy matter for the Doppler instrument to calculate the blood flow (in mL/min), and this is done automatically by the ultrasound instrument. Although the ability to calculate volume flow has been available on duplex instruments for more than 20 years and measurement accuracy appears satisfactory, issues of reproducibility have kept volume flow measurements from routine use in a clinical setting.†
Now that we have covered the basic concepts of Doppler spectral analysis, we can turn to the “heart of the matter,” namely, how to use Doppler spectral analysis to diagnose arterial obstruction. Five main categories of information are used in this process: (1) increased stenotic zone velocity, (2) disturbed flow in the poststenotic zone, (3) proximal pulsatility changes, (4) distal pulsatility changes, and (5) indirect effects of obstruction, such as collateralization.‡ These categories are summarized in Table 3-1, and each is discussed in the following sections.
|Elevated flow velocity in the stenotic lumen|
|Poststenotic flow disturbance|
|Proximal (upstream) pulsatility changes|
|Decreased velocity overall, due to decreased flow|
|Distal (downstream) pulsatility changes|
|Slowed systolic acceleration|
|Broad systolic peak|
|Increased diastolic flow (reduced peripheral resistance)|
|Decreased velocity overall|
|Secondary (collateral) effects|
|Increased size, velocity, and volume flow in collateral vessels|
|Reversed flow in collateral vessels|
|Decreased pulsatility (flow resistance) in collateral vessels|
The term stenotic zone refers to the narrowed portion of the arterial lumen. For determining the severity of arterial stenosis, the single most valuable Doppler finding is increased velocity in the stenotic zone. Flow velocity is increased in the stenotic zone because blood must move more quickly if the same volume is to flow through the narrowed lumen as through the larger, normal lumen. The increase in stenotic zone velocity is directly proportional to the severity of luminal narrowing.