Doppler Velocity Measurement Techniques

High-frequency sound waves (2 to 10 MHz) penetrate soft tissues and are reflected by the different interfaces encountered. Reflection from a moving interface results in the reflected frequency being increased if the motion is toward the point of observation and decreased if the motion is away from it. The magnitude of the shift is determined by the following equation:

     [1]

where f_s is the frequency shift, V the velocity, f₀ is the transmitted frequency, φ is the angle between the ultrasound beam and the velocity vector, and C is the speed of sound in tissue (1540 m/s). For a given velocity, a greater frequency shift is obtained with a higher transmitting frequency. In contrast, tissue penetration varies inversely with probe frequency, so the selection of a frequency for a given application is a balance between depth and velocity requirements.

Continuous-wave detectors are the simplest systems. The probe has two separate crystals, one transmitting and one receiving continuously. This system detects all velocities within the intersecting paths of the sound beams. If this zone includes more than one vessel (e.g., an artery and a vein), the resulting signal represents a combination of both velocities. Pulsed Doppler systems use a single crystal that repeatedly transmits a short burst of sound followed by a waiting period, during which the crystal functions in a receiving mode. By selecting the time and duration of the listening phase, one can define a sample volume, or the portion of the vessel from which velocity is to be measured. Modern duplex scanners use complex scan probes made up of many elements in an array, but the principle of focal sampling is the same.

The shifted frequency obtained from a vessel is within the audible range, so the data can be presented to the examiner as an audio signal. Although qualitative interpretation is helpful in some patient examinations, quantitative measurements provide objective testing. Spectral analyzers are used to determine the main frequency components obtained from a given vessel. This information is usually displayed on a sonogram, which shows the frequency content in time (Figure 14-1).

FIGURE 14-1 Comparison of sonograms from continuous-wave (CW) and pulsed Doppler systems. Sonograms display the different frequency contents detected at each point in time. The CW system detects all velocities across the vessel, whereas the pulsed system detects only those velocity vectors within the sample volume, indicated by the marks on the ultrasound beam. A, Normal arterial signals. CW has more low-frequency content, because it detects flow near the walls as well as in the center stream. B, Within stenosis, there is increased peak frequency, and the frequency distributions of both types of Doppler systems are similar, because the sample volume encompasses the entire flow stream. C, Beyond stenosis, peak frequency is elevated, with increased frequency distribution resulting from turbulent flow. Spectral width is greater with CW systems.

In some applications, it is more useful to have a measure of velocity rather than the raw frequency data. If the probe angle relative to the direction of flow can be measured, the velocity is estimated using the Doppler equation. The accuracy of the estimate depends greatly on the accuracy of the angle measurement. Errors are greatest when the probe is at a right angle to flow and least when it is at a low angle. Whenever possible, velocities should be measured with an angle less than 60 degrees.

Duplex Scan

During the 1960s, B-mode ultrasound imaging was used for visualization of soft tissue structures. Although early devices had only crude resolution, equipment has improved to the point that clear, detailed images of vessels can be produced in real time (Figure 14-2). In general, experience shows that when high-quality imaging is obtained, the diagnostic accuracy is very high; however, in patients with advanced atherosclerosis, it is difficult to obtain optimal studies, and diagnostic accuracy is lower. A common problem is incomplete imaging of the vessel wall as a result of calcification, which is present in varying degrees in up to half of patients studied. The extent of interference may be limited, but in some vessels there is no visualization of substantial portions of the artery. Although calcified plaques stand out sharply in the ultrasound image, some atheromas are visualized poorly or not at all. A major source of error is that recent thrombus may have the same echo density as flowing blood, so that an occluded vessel may look normal on the ultrasound image.

FIGURE 14-2 B-mode images from carotid duplex scan. A, Normal bifurcation. B, Moderate heterogeneous plaque with varying echogenicity. C, Calcified lesion with high echogenicity and shadowing below the plaque. CCA, Common carotid artery; ECA, external carotid artery; ICA, internal carotid artery.

To overcome the limitations of ultrasound imaging, the research team at the University of Washington developed the duplex scanner, combining a real-time B-mode ultrasound image system with a pulsed Doppler detector.¹ The ultrasound image shows not only the vessel under study but also the location of the sample volume of the Doppler beam so that the examiner can position it to study velocity patterns at specific locations in the vessel. The device can study calcified vessels by analyzing the Doppler velocity signal distal to the areas of calcification. The evaluation of the Doppler signal from the common carotid artery and its branches is performed using spectral analysis (Figure 14-3). Based on the peak systolic velocity, end-diastolic velocity, velocity ratios, and degree of spectral broadening, a category of stenosis is assigned to the vessel segment.

FIGURE 14-3 Doppler sonograms from carotid duplex scans. A, Normal study with normal spectral width. B, Moderate stenosis with spectral broadening but no increase in peak frequency. (Note that frequency scales are different in the three records.) C, Severe stenosis with high peak velocity and extensive spectral broadening.

In the past 20 years, there has been extensive improvement of duplex scanners in terms of both image resolution and Doppler signal processing. The early devices were limited to the study of superficial vessels; however, the availability of low-frequency probes (2.0 to 3.5 MHz) permits the evaluation of abdominal vessels, including the aorta, vena cava, and main visceral branches. Study of intracranial artery branches is also possible.

An important subsequent development is the color-coded Doppler system. A linear array transducer composed of many separate elements is used to produce a grid of sample volumes encompassing the area covered by the B-mode image (Figure 14-4). A portion of the grid is selected for color coding of velocity information. Each of the sample volumes within the area is examined. If the returning ultrasound signal has no change in phase or frequency, the amplitude information is used to create the gray-scale image at that point in the matrix. If there is a change in phase or frequency, the information is analyzed in terms of velocity. A color is assigned to represent an approximate mean velocity occurring at that point in the field. Red and blue show flow toward and away from the transducer, respectively. The magnitude of the velocity is represented by the hue of the color: a dark shade indicates slow flow, and a lighter shade or white indicates high flow. The aggregate of the color representation from the sample volumes detecting motion produces a real-time representation of the flow patterns within the vessels superimposed on the gray-scale image of the stationary tissue. Figure 14-5 (see color plate) illustrates examples of the advantages of color duplex scans. A more recent development is color coding of the Doppler power (as opposed to velocity) detected. Power is proportional to the square of the velocity; therefore this measurement provides more sensitive detection of very slow flow or flow in small vessels. A good example of the benefit of power imaging is the detection of an internal carotid string sign. Squaring the velocity eliminates the positive or negative value, so that power values have no directional representation. Power is represented in a single color, usually orange.

FIGURE 14-4 Color-coded duplex system. A linear array transducer is used to create a matrix of sample volumes. A gray-scale image is created within area A. Most examinations are performed with color coding of velocities limited to a portion of the image (area B). Within this portion of the matrix, ultrasound signals from sample volumes with a change in phase or frequency are interpreted as velocity data. Otherwise, the data are coded as part of the gray-scale image.

FIGURE 14-5 A, Advantages of using color duplex scanning. Normal carotid bifurcation, illustrating the reverse flow occuring in the bulb during peak systole (arrow). B, Marked tortuosity of internal carotid artery (ICA) is easily demonstrated with color scan. C, Conventional gray-scale image does not provide clear identification of large ulcer. D, Blood flow within plaque confirms the ulcer. E, Pseudoaneurysm originating from the superficial femoral artery (SFA) resulting from catheterization. Note the bidirectional flow recorded in the neck. F, Type 2 endoleak with flow detected in the aneurysm sac outside of prosthesis. G, Longitudinal view of same aorta showing leak is posterior to graft. H, Blood flow around partial occluding thrombus. I, Thrombus extending into common femoral vein, seen 2 days after endovascular ablation of great saphenous vein.

Duplex Scan

The routine examination covers as much of the common carotid artery (CCA) and its branches as can be visualized with the configuration of the transducer used. In some patients, the origins of the CCAs can be visualized. Figures 14-2A and 14-5A (see color plate) show normal carotid bifurcations. The color image demonstrates the reverse velocity detected in the carotid bulb as a result of the complex flow pattern at the bifurcation. Many older patients have tortuosity that precludes the CCA, the bulb, and the branches from being visualized in a single plane; in such cases, careful scanning is required to obtain satisfactory imaging. Figure 14-5B (see color plate) shows an example of tortuosity in an elderly patient. Although such arteries can be studied with a conventional scanner, the color-coded unit simplifies the examination. The scan usually identifies the pathologic regions, but with advanced atherosclerosis, it is often difficult to get an adequate image to accurately estimate the degree of stenosis. Much of the classification of stenosis is based on interpretation of the Doppler signal. The two branches are distinguished by the image and the velocity signals. The ICA has a low peripheral resistance at all times, resulting in forward flow throughout diastole, whereas the high resistance in the external carotid artery results in a diastolic flow of zero. Stenoses produce an increased velocity at the site of the lesion and turbulence beyond it (see Figure 14-1). The turbulence is identified as spectral broadening on the sonogram (see Figure 14-3). Mild stenoses may not produce a significant increase in peak systolic velocity, but are identified by a moderate degree of spectral broadening.

Based on the peak systolic velocity and the degree of spectral broadening, the ICA is placed into one of six diagnostic categories. There are two sets of criteria that have been used for many years, and although some laboratories have made modifications or adjustments, the basic principles continue to be applied. The criteria developed at the University of Washington use primarily ICA velocity parameters (Table 14-1).² Further improvements in accuracy can be obtained using ratios of ICA velocities to CCA velocities in normal portions of the artery (Table 14-2).³ The diagnosis of ICA occlusion must be based on image and Doppler information, because the low flow found with some “string signs” is below the velocity detection threshold of many scanners. Newer color duplex devices have improved our ability to find small residual flow channels, especially using power flow mapping. Both the stippled appearance of chronic thrombus and a small diameter of the ICA indicate occlusion. Overall, low-grade plaques are best assessed with the image, whereas advanced lesions are best evaluated with the Doppler information.

TABLE 14-1 Categories of Internal Carotid Artery Stenosis:

University of Washington Criteria

TABLE 14-2 Categories of Internal Carotid Artery Stenosis: Bluth Criteria

There has been a rapid growth in the use of duplex scanning for carotid diagnosis. Different investigators have demonstrated that the technique can be highly accurate. Studies have shown rates of 92% to 96% accuracy in the identification of severe stenosis.^4–6 When these studies are analyzed in terms of correct category of stenosis, exact agreement is found in 77% to 87%, with poor agreement in only 1% to 2%. Of particular importance is the fact that experienced laboratories make few errors in separating severe stenosis from occlusion. Mansour and coauthors⁷ reported a 98% positive predictive value and 99% negative predictive value in the correct determination of ICA occlusion.

In addition to estimating the severity of a stenosis, scanners are now being used to study the plaque itself. Most investigators merely distinguish between homogeneous- and heterogeneous-appearing plaques and describe the surface as either smooth or irregular. More elaborate approaches to the description of morphology are being evaluated, but no single approach has been widely adopted.

Although the majority of attention has been focused on the carotid circulation, laboratories routinely investigate the status of the vertebral arteries as well. The examination seeks two types of problems: stenosis in the vertebral artery itself and the abnormal flow produced by subclavian steal. In the majority of cases of significant vertebral stenosis, the lesion is located at the origin of the vessel. In some cases of severe occlusive disease, there is sufficient asymmetry in the waveforms of the two vertebral arteries to point to the problem side. However, a more complete assessment is obtained by examining the origins. Because of its deeper location, the left vertebral artery is more difficult to study than the right. Ackerstaff and associates⁸ found that the status of the ostium could be studied satisfactorily in about 80% of patients. When adequate evaluation of the prevertebral portion was possible, a sensitivity of 80% and a specificity of 97% were achieved in the detection of reductions greater than 50% in diameter. Most clinical cases of subclavian steal are demonstrated by a reverse flow in the vertebral artery on the affected side. Von Reutern and Pourcelot⁹ demonstrated that in some cases of subclavian stenosis there is distortion of the waveform rather than complete reversal of flow. The abnormal waveforms may have attenuation of the systolic component or an alternating pattern with reverse flow in systole and forward flow in diastole (Figure 14-6). Such cases can be assessed more fully by recording the Doppler signal after arm exercise or the induction of reactive hyperemia. In the presence of advanced subclavian stenosis, this stress test produces full reversal of flow.

FIGURE 14-6 Vertebral artery Doppler waveforms. A is normal. B to D are signals recorded on the side of severe subclavian stenosis. B, Attenuated systolic flow. C, Reversed flow in systole, with forward flow in diastole. D, Complete flow reversal.

Applications

Segmental Extremity Pressure Measurement

Indirect measurement of extremity pressures has been performed since the beginning of the 20th century using a sphygmomanometer and auscultation of the Korotkoff sounds with a stethoscope. Although this technique is universally used to measure pressures in the brachial artery, its application in the lower extremity is less practical because of the difficulty of listening for Korotkoff sounds in the popliteal space. The technique is certainly not applicable in the distal portions of the extremity because of the small size of the vessels involved. Investigators overcame this limitation by using a variety of plethysmographic devices. In 1959, Winsor²⁰ first described the clinical measurement of arterial gradients using a plethysmograph. Systolic pressures in the lower extremity are normally higher than those in the upper extremity. He described the blood pressure index (blood pressure of arm/blood pressure of leg), which in normal persons is less than 1.0. A value greater than 1.0 indicates clinically significant occlusive disease proximal to the point of measurement. (Note that the ratio winsor described is the inverse of the currently used ankle-brachial index.) Likewise, a gradient between two sampling sites localizes the occlusive disease in the intervening segment. The main limitation of this method is that it detects only occlusive lesions that are sufficiently advanced to reduce the systolic pressure, so that it is not possible to detect early disease. Introduction of the Doppler velocity detector greatly simplified the indirect measurement of extremity pressures. For this application, the Doppler device is used merely to detect the presence or absence of the movement of blood in the artery. Measurements made by this method are reproducible, but do not provide diastolic pressure. Plethysmographic techniques are cumbersome and are not used routinely.

In clinical practice, simple screening can be performed by measuring the pressure at the brachial arteries and at the dorsal pedal and posterior tibial arteries on each side. The ankle-brachial index (ABI) is determined by dividing the ankle pressure on each side by the higher of the two brachial pressures. The resulting value reflects the severity of the occlusive disease for the entire extremity. Normally, the ABI should be greater than 1.0, and values less than 0.95 are abnormal. Figure 14-7 summarizes the general relation between ABI values and clinical status. It must be emphasized that this is only a rough correlation and that patients with similar values may have substantial differences in exercise tolerance. Likewise, the index at which rest pain appears varies considerably from patient to patient, ranging from 0.30 to 0.50.

FIGURE 14-7 Relation of ankle-brachial index to patient symptoms.

An important limitation of the indirect measurement of extremity pressure is seen in patients with abnormal stiffening of the vessel wall, most often due to heavy calcification. Such conditions occur with diabetes mellitus but can also be found with other disorders. In these cases, the systolic pressure measured reflects the cuff pressure required to collapse the vessel wall in addition to the pressure required to overcome the intraluminal pressure. In a few patients, it is not possible to stop the flow of blood at all. Error due to wall stiffness should be suspected whenever the ABI is greater than 1.3 or its value is out of proportion to the patient’s clinical status. In general, a leg with a normal ABI should have an easily palpable ankle pulse. In some patients with stiff arteries, it may be possible to obtain an accurate evaluation by measuring the toe pressure. In a healthy person, there is a gradient of 20 to 30 mm Hg between the ankle and the toe; therefore a correction must be made when toe pressures are being used.

Localization of occlusive disease can be obtained by measuring the pressures at different levels of the leg. Segmental pressure measurements are usually performed by applying cuffs at the thigh, the upper calf, and immediately above the ankle. A standard adult-sized cuff (12 cm wide) is satisfactory for calf and ankle determinations, but a thigh cuff (18 cm) should be used above the knee. (Using a narrower cuff above the knee results in artificially high pressure measurements as a result of the size discrepancy between the cuff and the diameter of the thigh. Thigh measurements with an arm cuff usually result in determinations that are 20 to 30 mm Hg higher than those obtained with the wider cuff.) A thigh pressure lower than the brachial pressure indicates obstruction proximal to the location of the thigh cuff. Gradients of more than 20 mm Hg between measuring sites are diagnostic of occlusive disease in the intervening segment, and higher gradients are usually associated with more severe lesions.

An important limitation of the use of the wide cuff for thigh measurement is that it is possible to make only a single thigh measurement. As a result, it is not possible to distinguish occlusive disease above the ligament from that in the proximal portion of the superficial femoral artery, because both conditions can result in the same thigh pressure measurement. To overcome this problem, some investigators have recommended using 12-cm–wide cuffs to obtain two separate thigh measurements. When this is done, it is necessary to take into account the 20- to 30-mm Hg artifact in the results. In a study comparing the wide cuff with the two narrow-cuff techniques in the same group of patients, Heintz and coworkers²¹ reported an increased accuracy in the localization of disease using the two-cuff technique. Both methods of thigh pressure measurement are still being used, so it is important to know which method is being reported when reviewing the results of patient studies. Although segmental pressures have been used extensively to detect proximal disease, diagnostic errors may occur in 25% of patients. Other techniques should be used when an accurate determination of the segmental localization is needed.