CHAPTER 90 Ultrasound Evaluation of the Carotid Arteries

Leslie M. Scoutt, Jonathan D. Kirsch, Ulrike Hamper

Stroke remains the third leading cause of death and is a major cause of morbidity in the United States.¹ Most strokes are due to thromboembolic events rather than to ischemia or reduced perfusion. Whereas the heart is the number one source, 20% to 30% of strokes are believed to be secondary to embolus from plaque or thrombus at the carotid bifurcation.² Carotid endarterectomy has been convincingly shown in several prospective multicentered, randomized, double-blind trials to significantly reduce the risk of stroke and death in patients with stenoses of the internal carotid artery (ICA) of more than 60% to 70% compared with optimized medical therapy.³^–⁶ However, at the time these trials were published in the 1990s, statin therapy was not a part of the standard medical regimen, and double-blind trials comparing carotid endarterectomy with medical management including statin therapy are currently ongoing.

Thus, the identification of patients with ICA stenoses of 60% to 70% is clearly important for patient management, allowing appropriate referral for carotid endarterectomy. Risk factors for disease at the carotid bifurcation include atherosclerosis, hypertension, diabetes mellitus, hyperlipidemia, hypercholesterolemia, obesity, and smoking. Patients with risk factors for carotid plaque, carotid bruits, and symptoms of stroke or transient ischemic attacks are typically referred for evaluation of the carotid arteries, which can be performed with ultrasonography, computed tomographic angiography, magnetic resonance angiography, or conventional angiography. Of these potential screening modalities, carotid ultrasound examination is the most readily available, least invasive, and least expensive. Numerous studies have shown that when it is performed appropriately, ultrasound examination of the carotid arteries is highly accurate for detection of surgical lesions (i.e., ICA stenoses ≥70%),⁷^–¹⁰ and additional confirmatory studies such as CT angiography, magnetic resonance angiography, and conventional angiography are usually unnecessary except in complex cases with discordant findings or poor visualization. The precise method of grading stenoses of the ICA changed with the publication of the North American Symptomatic Carotid Endarterectomy Trial (NASCET).³ Before the NASCET, the percentage stenosis of the ICA was typically calculated by comparing the width of the residual lumen with the estimated diameter (outer wall to outer wall) of the ICA at the site of the stenosis (Fig. 90-1A). However, because the outer wall of the ICA can be seen on an angiogram only if it is calcified, angiographic measurements of ICA diameter are only estimates in many cases. Hence, this method has significant interobserver and intraobserver variability as well as poor reproducibility. Thus, when the NASCET was designed, the measurement of percentage ICA stenosis was standardized angiographically by comparing the width of the residual lumen at its narrowest point with the diameter of the lumen of the distal normal ICA (beyond any post-stenotic dilation; Fig. 90-1B) because the vessel lumen can be accurately and reproducibly measured on angiographic images. For a given residual lumen, the percentage stenosis is generally higher with the pre-NASCET method of calculating an ICA stenosis.

FIGURE 90-1 Schematic diagrams demonstrating the methods of calculating percentage internal carotid artery stenosis. A, The pre-NASCET or traditional formula for calculation of an ICA stenosis compares the diameter of the residual lumen with the distance between the outer walls of the ICA at the site of stenosis. On angiograms, this method is less accurate than the NASCET method because of difficulty in identifying the outer wall of the ICA. B, Percentage ICA stenosis in the NASCET trial was calculated by comparing the diameter of the residual lumen with the luminal diameter of the distal normal ICA beyond any post-stenotic dilation. This is the currently accepted method of calculating percentage stenosis of the ICA. In general, percentage stenosis is greater when it is calculated by the pre-NASCET or traditional formula than by the NASCET method. CCA, common carotid artery; ECA, external carotid artery; ICA, internal carotid artery.

TECHNICAL REQUIREMENTS

Doppler ultrasound examination of the carotid arteries requires an ultrasound machine with high-resolution gray-scale imaging as well as color Doppler and spectral Doppler capability. A high-frequency 5- to 7.5-MHz linear array transducer should be used to optimize spatial resolution. However, if the carotid arteries are too deep to visualize with the linear array transducer in a patient with a short, thick neck, a lower frequency curved array transducer may be necessary for adequate penetration.

Technique

The examination begins with evaluation of plaque burden. Plaque echotexture should be characterized as hypoechoic, heterogeneous, or echogenic (Fig. 90-2). The surface contour of the plaque should be described as smooth or irregular (Fig. 90-3), and the percentage reduction of the arterial diameter by the plaque should be estimated.

FIGURE 90-2 Carotid plaque: echotexture. A, Color Doppler image demonstrating a large amount of homogeneous hypoechoic plaque causing a “string sign” (arrow) in the left ICA. B, Note focal hypoechoic area (arrow) within a heterogeneous plaque in the left carotid bulb. C, Note homogeneous echogenic shadowing plaque (arrows) in the proximal right ICA. D, Color Doppler image demonstrating homogeneous echogenic shadowing plaque (arrows) in the left proximal ICA.

FIGURE 90-3 Carotid plaque: surface contour. A, Longitudinal gray-scale image of the distal left CCA demonstrating plaque with a smooth surface contour and heterogeneous echotexture. B, Longitudinal gray-scale image reveals a large amount of hypoechoic plaque with a smooth surface contour (arrows) along the posterior and anterior walls of the right ICA and bulb. C and D, Longitudinal gray-scale images of plaque with irregular surface contour from two different patients. Note hypoechoic area within the plaque (arrow in D). E, Longitudinal power Doppler image of the right ICA reveals an irregular residual lumen, indicative of irregular plaque surface.

Optimal evaluation of plaque requires both gray-scale and color or power Doppler imaging in the longitudinal and transverse planes of the entire CCA and ICA as well as the origin of the external carotid artery (ECA). The focal zone should be set at the level of the far wall of the vessel. The use of spatial compounding and harmonic imaging will also improve gray-scale resolution. The gray-scale gain should be adjusted such that plaque and the vessel wall are easily depicted but artifactual echoes are not present within the vessel lumen. If the gain is set too low, plaque will look artifactually hypoechoic. The color gain should be optimized by slowly increasing the gain until color speckles are noted in the surrounding soft tissues. The gain is then decreased until the color pixels are visible only within the vessel lumen. If the gain is set too high, the color pixels will overwrite or “bleed” over plaque, obscuring visualization of the true extent of plaque burden, particularly during systole, and stenoses may be overlooked (Fig. 90-4). If the gain is too low, sensitivity to blood flow will be decreased, and false-positive diagnoses of occlusion or stenosis will be made. The color velocity scale should be adjusted such that color fills the lumen reaching to the vessel wall. Therefore, the color velocity scale may have to be changed slightly as one interrogates the different vessels in the neck. If the scale is set too low, aliasing will occur, making it more difficult to detect flow disturbances at the site of a stenosis. In addition, color motion artifact, which may obscure visualization of the vessel, is more prominent when the velocity scale or pulse repetition frequency is set too low. The wall filter should be set as low as possible without degradation of the image by motion artifact. Angling of the color box will facilitate evaluation of the direction of blood flow, and a straight, small color box will increase sensitivity to flow. If the vessel lumen cannot be visualized because of shadowing from plaque, an approach to the vessel from different angles (including the transverse plane) or use of a curved array transducer may be helpful.

FIGURE 90-4 Color “blooming” artifact. A, Note extensive echogenic, irregularly surfaced plaque in the mid left CCA on a longitudinal gray-scale image. B, On the color Doppler image, however, the color overwrites and obscures the plaque.

Once the gray-scale and color or power Doppler images have been obtained, the sonographer should be able to describe the amount and type of plaque in the carotid arteries as well as the relationship of the ICA and ECA to the carotid bulb. Whereas the degree of stenosis can often be estimated on the gray-scale or color Doppler images, precise quantification of stenosis requires accurate measurement of peak systolic velocity (PSV) in the ICA and CCA on the spectral Doppler tracing.

Spectral Doppler tracings are obtained from longitudinal images with the sample volume placed centrally within the vessel lumen or at the brightest spot in any area of focal color aliasing. To optimize the spectral Doppler tracing, the velocity scale and baseline should be adjusted such that the tracing fills the velocity spectrum or scale. If the scale is too high, the waveform will be too small to easily measure or analyze. However, if the scale is too low, the tracing will be too large, and wraparound or aliasing of the systolic velocity peak will occur, making it impossible to accurately measure PSV. The angle of spectral Doppler insonation should be kept between 45 and 60 degrees to minimize error in the calculation of velocity from the Doppler frequency shift. If at all possible, the angle should be kept constant on follow-up studies. The angle cursor should be placed parallel to the direction of blood flow in the color jet or vessel lumen rather than parallel to the vessel wall. The direction of blood flow will, in fact, parallel to the vessel wall in most cases, but the jet of blood may travel tangentially or obliquely in relationship to the vessel wall if plaque is irregular or asymmetric (Fig. 90-5). In such cases, the angle correction cursor should be placed parallel to the jet of blood as seen on color Doppler imaging. The sample gate should be kept small, between 1.5 and 2.5 mm in width, and placed in the center of the vessel because PSV will be lower near the vessel wall owing to the geometry of laminar flow and drag from the vessel wall. If the sample gate is too wide, a wider range of velocities will be depicted, and there is the potential risk of falsely creating the appearance of spectral broadening and turbulent flow. Finally, the spectral Doppler gain should be optimized; the gain should be increased until background speckles appear on the spectral tracing and then readjusted downward until the background is homogeneously black. Although incorrect spectral Doppler gain settings rarely cause shifts in PSV of more than 20 cm/sec, such variation in the distal CCA can be important when the PSV in the distal CCA is used as the denominator in calculation of the peak systolic velocity ratio (PSVR). Apparent spectral broadening with “fill in” of the spectral envelope can also be spuriously created if the spectral Doppler gain is set too high.

FIGURE 90-5 Angle of insonation. There is a large amount of both hypoechoic and echogenic plaque in the right ICA causing tortuosity and angulation of the vessel lumen in this patient who presented with a carotid bruit. The Doppler angle should be calculated in reference to a vector parallel to the direction of blood flow in the residual lumen (yellow line) rather than parallel to the vessel wall.

Protocol

• Longitudinal gray-scale and color or power Doppler images of the proximal and distal CCA, bifurcation, bulb extending into the ICA, and bulb extending into the ECA

• Transverse gray-scale and color or power Doppler images of the proximal and distal CCA, bulb, bifurcation, proximal ICA, and proximal ECA

• Spectral Doppler tracings from the proximal CCA, distal CCA, proximal ICA, mid ICA, distal ICA, proximal ECA, and mid vertebral artery. PSV is recorded for each segment, and end-diastolic velocity (EDV) is noted if it is more than 100 cm/sec. The PSVR is calculated by dividing the highest PSV in the proximal ICA or area of stenosis by the PSV in the distal CCA. PSV in the distal CCA should always be measured 2 to 3 cm proximal to the level of the carotid bulb. Several measurements, optimally three, of PSV should be obtained and the highest velocity recorded.

Additional gray-scale and color or power Doppler images as well as spectral Doppler tracings should be obtained as necessary wherever extensive plaque burden, vessel narrowing, or color aliasing is seen.

Normal Findings

The normal carotid arteries have a thin, regular echogenic wall without focal areas of calcification, intraluminal plaque, or thrombus. Color should fill the vessel lumen homogeneously with a slight central increase in color intensity consistent with normal parabolic or laminar flow. Where the carotid bulb widens, a helical blood flow pattern or reversal of peripheral flow is a normal finding, particularly in younger patients, and is believed to be due to boundary layer separation.

Normal PSV in the CCA is variable and depends on numerous factors, including cardiac output or stroke volume, heart rate, systolic blood pressure, and age. In general, however, PSV in the normal CCA ranges from 70 to 100 cm/sec and decreases gradually as one samples distally.

The CCA, ICA, and ECA demonstrate distinct characteristic waveform patterns. Whereas all segments of the extracranial carotid arteries normally demonstrate a sharp systolic upstroke and thin spectral envelope, the amount of diastolic flow varies in each vessel, reflecting the oxygen consumption and peripheral vascular resistance of the vascular bed supplied. The ECA, which supplies the muscular bed of the scalp, typically demonstrates completely absent or very low velocity end-diastolic flow. Although the amount of diastolic flow in the ECA may vary from patient to patient, it should be symmetric right to left and less than the diastolic flow in the ICA or CCA (Fig. 90-6). An early diastolic notch followed by a short reversal of flow in early diastole is often seen in the ECA. The ICA, which supplies blood to the brain (high oxygen consumption), has a low-resistance waveform pattern with continuous forward, relatively high velocity diastolic flow (Fig. 90-7). The waveform of the CCA demonstrates an intermediate amount of diastolic flow and often demonstrates a brief reversal of flow in early diastole (Fig. 90-8). The vertebral artery has a waveform pattern similar to that of the ICA, characterized by a sharp systolic upstroke and continuous forward diastolic flow (Fig. 90-9).

FIGURE 90-6 Normal spectral Doppler tracing of the external carotid artery. The ECA has a higher resistance waveform with less diastolic flow than the ICA. Whereas the amount of diastolic flow may vary from patient to patient, it should be symmetric right to left in a given individual. A, This patient has no diastolic flow in the ECA. B, In another patient, a small amount of diastolic flow is present. Reversal of flow in early diastole or an early diastolic notch (arrows) is a normal finding in the ECA. Note sharp systolic upstroke and thin spectral envelope in the spectral tracings of both ECAs.

FIGURE 90-7 Normal spectral Doppler tracing of the internal carotid artery. Note that in comparison to the ECA, there is increased diastolic flow, the systolic peak is slightly blunted, and the spectral envelope is slightly widened in the ICA. The systolic upstroke is sharp, and velocity gradually tapers during diastole.

FIGURE 90-8 Normal spectral tracing of the common carotid artery. The systolic upstroke is sharp in the CCA, and there is an intermediate amount of diastolic flow in comparison to the ICA and ECA. An early diastolic notch may be present, but it is less pronounced than in the ECA.

FIGURE 90-9 Normal spectral tracing of the vertebral artery. The waveform of the normal vertebral artery is similar to the waveform of the ICA.

Differentiation of the ICA from the ECA is critically important to avoid misinterpretation of a stenosis in the ECA as a more clinically significant ICA stenosis. The best method of identifying the ECA is by visualization of branches arising from the vessel (Fig. 90-10A). The ICA virtually never gives rise to branch vessels in the neck. Temporal tapping over the ophthalmic artery will generate sharp, spike-like deflections in the waveform of the ECA during diastole (Fig. 90-10B). However, on occasion, transmitted pulsations to the ICA will be observed after temporal tapping, although the deflections are typically blunted in comparison to the deflections observed in the ECA.⁷^,⁸ In general, the ECA is smaller, is more medial, and has a higher resistance waveform pattern than the larger posterolateral ICA. However, location and vessel size are not reliable criteria for differentiation of the ICA from the ECA in all patients.

FIGURE 90-10 Identification of the external carotid artery. A, Note branches arising from the right ECA on this color Doppler image. Branches almost never arise from the ICA below the skull base. B, Note deflections in the spectral tracing from the right ECA due to temporal tapping. These deflections are most easily seen during diastole. Whereas temporal tapping may occasionally cause similar deflections in the ICA or CCA, they will be smaller and less sharply defined than in the ECA.

Grading Stenoses in the Internal Carotid Artery

Grading of stenosis in the ICA by Doppler velocity criteria is based on the simple precept that flow volume is equal to vessel area multiplied by PSV. Because flow volume is a relative constant, PSV is inversely related to vessel lumen diameter. Therefore, if the vessel diameter decreases, PSV must increase to maintain flow volume. As demonstrated by the Spencer diagram, once the vessel diameter decreases by more than 50%, PSV at the site of the stenosis rises exponentially, and this compensatory increase in PSV maintains flow volume until the stenosis reaches approximately 70%.⁹ When a stenosis exceeds 70%, flow volume will begin to drop despite continued exponential increase in PSV. Once a stenosis in the ICA becomes greater than approximately 96%, the gradient across the stenosis becomes too high for the pressure generated by the combination of myocardial contractility and vessel elasticity to force blood through the residual lumen, and PSV begins to drop until the vessel occludes and velocity is zero.⁹

Numerous studies have attempted to correlate Doppler measurements of PSV, PSVR, and EDV criteria with angiographically calculated percentage stenosis. Widely variable results have been reported.¹⁰^–¹⁴ In general, velocity criteria have a higher accuracy and positive predictive value for high-grade stenoses (>70%) and are less reliable for accurate grading of more moderate stenoses (<50%).¹²^,¹⁴ Criteria are probably both machine (possibly even transducer) and laboratory specific and therefore should be validated for each laboratory and machine upgrade if possible. Studies have also demonstrated that vascular laboratories cannot reliably differentiate stenoses in the ICA by 10% increments.¹²^,¹⁴ Accuracy is best achieved by focusing on whether a stenosis is greater than or less than a specific percentage stenosis. In most institutions, carotid endarterectomy is recommended in patients with ICA stenosis of more than 70%. Hence, most vascular laboratories categorize ICA stenoses as less than 50%, 50% to 69%, 70% to 96%, and more than 96%.

A meta-analysis published by Grant and colleagues ¹² found that accuracy varied little over a wide range of threshold values for PSV and PSVR, although the sensitivity and specificity were inversely proportional. Hence, the authors recommend that if the Doppler examination is to be used as a screening test, lower thresholds with higher sensitivity should be used. However, if the Doppler examination is intended as a diagnostic test without anticipating confirmation by some angiographic imaging modality, then specificity should be emphasized and higher thresholds are recommended.¹²

In 2002, the Society of Radiologists in Ultrasound (SRU) convened a multidisciplinary panel of experts, including both radiologists and vascular surgeons, to develop a consensus for grading of ICA stenoses by Doppler ultrasonography. The published recommendations of this consensus conference are as follows ¹³:

PSV < 125 cm/sec and PSVR < 2.0 suggest a stenosis < 50%

PSV between 125 and 230 cm/sec and PSVR between 2.0 and 4.0 are most consistent with a stenosis of 50% to 69% (Fig. 90-11)

PSV > 230 cm/sec, PSVR > 4.0, and EDV > 100 cm/sec indicate a stenosis of ≥70% to 96% (Figs. 90-12 and 90-13)

FIGURE 90-11 ICA stenosis, 50% to 69%. A, Longitudinal color Doppler image of the right ICA in a patient with a right carotid bruit reveals echogenic plaque and a stenosis at the origin of the right ICA. Note color mosaic just distal to the narrowing of the vessel lumen indicative of increased velocity of flow in the post-stenotic jet. B, Spectral tracing reveals a PSV of 195 cm/sec just distal to the stenosis. C, PSV in the right CCA is 96 cm/sec. PSVR is approximately 2 : 1. By SRU criteria, this corresponds to a moderate stenosis of 50% to 69%.

FIGURE 90-12 ICA stenosis, 70% to 96%. A, Longitudinal duplex Doppler image of the right ICA in a 73-year-old man presenting with a transient ischemic attack demonstrates that the diameter of the proximal right ICA is narrowed approximately 80%. PSV is 305 cm/sec at the site of the stenosis. B, Spectral tracing of the distal right CCA demonstrates a PSV of 85 cm/sec, yielding a PSVR close to 4 : 1. By SRU criteria, these velocity measurements are consistent with a stenosis of the right ICA of more than 70%, which was confirmed by a CT angiogram.

FIGURE 90-13 ICA stenosis, 70% to 96%. Longitudinal color Doppler image of the proximal right ICA (A) demonstrates a marked degree of narrowing over a long segment from heterogeneous, largely hypoechoic plaque in a 68-year-old woman presenting with a stroke. By visual inspection, the percentage diameter reduction is estimated at nearly 90%. Spectral Doppler tracing reveals a PSV of 523 cm/sec in the stenosis (B), with a PSV of 58 cm/sec in the distal CCA (C). PSVR is 9 : 1. Although it is not possible to accurately classify stenoses at 10% increments by Doppler velocity criteria, these measurements are well above the SRU threshold for a stenosis of 70% or more, and therefore the degree of stenosis is likely in the higher range, approximately 90% diameter reduction, as suggested by the color Doppler image.

Once an ICA stenosis becomes greater than an approximately 96% diameter reduction, PSV velocity begins to drop (Fig. 90-14) until the vessel occludes and the velocity reaches zero.⁹ Hence, the panel also recommended that the estimation of diameter reduction by plaque on gray-scale and color Doppler imaging be correlated with the velocity-based estimate of percentage stenosis.¹³