Carotid Artery Disease

In 2006, the UK Health Technology Assessment (HTA) Programme published a meta-analysis of studies evaluating the accuracy of noninvasive imaging (but not including multidetector computed tomographic angiography [MDCTA]) that concluded that although contrast-enhanced MRA (CEMRA) appeared the most accurate imaging modality (Table 98-2), its “clinical effectiveness” was limited by (in)accessibility, (un)availability, and delays.² As a consequence, the researchers, who “constituted a panel of experts in stroke, imaging, vascular surgery, statistics and health economic modelling,” concluded that DUS should remain the first-line imaging modality for identifying patients with 70% to 99% ICA stenoses because of (1) low cost, (2) the much higher number of strokes likely to be prevented in the long term (because patients are scheduled more rapidly for surgery), (3) robustness in sensitivity analyses (see Table 98-2), and (4) the ability of DUS to match the needs of current surgery (i.e., patients were selected for surgery through attendance at “single visit clinics” incorporating DUS with a low probability of the surgeon encountering unexpected findings at operation that might otherwise compromise patient safety).³ The panel did, however, highlight concerns about DUS accuracy for diagnosing 50% to 69% stenoses. Because the benefit conferred by CEA in patients with 50% to 69% stenoses falls significantly with any delay to surgery (see Table 98-1), the panel recommended that if 4 weeks or more have elapsed after the index event, it would be advisable to perform corroborative imaging (CEMRA or MDCTA) in order to confirm stenosis severity. If, however, DUS has been performed within 4 weeks of the index event, the panel judged it reasonable to proceed to CEA on the basis of DUS findings alone because the number of strokes prevented through rapidly performed surgery exceeds the potential risk to patients with less than 50% stenoses undergoing inappropriate surgery.

Vertebrobasilar Disease

Extracranial DUS offers limited information about flow dynamics in the vertebral arteries (VAs), but nothing about the basilar system (which requires color transcranial DUS). It is not always possible to image the vertebral artery origins, and only those sections running between the transverse processes can be imaged more distally. Accordingly, if a patient presents with suspected vertebrobasilar symptoms, DUS assessment can exclude coexistent carotid disease, but diagnostic imaging with MRA/MDCTA is necessary. DUS is, however, useful for demonstrating complete or partial vertebral artery flow reversal, which suggests the possibility of proximal subclavian or innominate artery disease.

Although the Society for Vascular Surgery advocates carotid screening in patients older than 55 years with cardiovascular risk factors, the 2007 U.S. Preventive Services Task Force recommends against it.⁴ A reluctance to screen also exists in Canada, the UK, and Scandinavia. For those committed to screening, DUS is the first-line modality. Portable DUS machines are available, but there is little data about their accuracy. Accordingly, users of portable machines are advised to undertake a corroborative DUS study in an accredited vascular laboratory (or use MDCTA/CEMRA) if a greater than 50% stenosis is suspected. DUS remains the preferred screening modality prior to coronary artery bypass grafting (CABG) surgery. Screening everyone is not cost-effective (5% yield for 80%-99% stenosis⁵) but the yield is increased in selected patient cohorts (left mainstem disease, carotid bruit, previous stroke/transient ischemic attack [TIA]). Any patient who is being considered for CABG using the internal mammary artery (IMA) and who has weakness or absence of the radial pulses should undergo assessment of the subclavian, innominate, and vertebral arteries. Where inflow disease or vertebral artery flow reversal is suspected, further imaging (MDCTA/CEMRA) is essential, and the cardiac surgeon should be warned that unless the flow reversal is corrected, subsequent use of the ipsilateral internal mammary artery (as a conduit) could cause postoperative coronary steal syndrome (Fig. 98-1).

Because of their vascularity, DUS is ideally suited for diagnosing carotid body tumor (CBT) and glomus vagale tumor (GVT). Given the rapid availability of DUS, it is unwise to perform biopsy of a “lymph node” in the jugulodigastric region without undertaking DUS. Once such a tumor is suspected, MDCTA or MRA should be performed to exclude bilateral lesions (5% incidence) and to image the upper and lateral extents of the lesion. Large, highly placed lesions require a totally different operative strategy, which must be anticipated preoperatively (temporomandibular subluxation cannot be performed once the operation is under way; see Chapter 104). DUS findings can also alert the surgeon to the possibility that a GVT might be present (Fig. 98-2), thus enabling the patient to be counseled about vagus nerve division (hoarseness, swallowing difficulties). Carotid body tumors cause splaying of the carotid bifurcation. GVTs do not splay the bifurcation, but they do cause displacement of the distal ICA (see Fig. 98-2). In addition, GVTs tend to have vascular serpiginous feeding vessels tracking down the proximal vagus nerve. These usually lie separately from the bifurcation, internal jugular vein, and common carotid artery (CCA).

In many centers, up to 95% of CEAs are undertaken on the basis of DUS findings, and there is no evidence that reliance on DUS compromises safety or operability.³ More importantly, patients can be identified and scheduled for surgery faster than with any other imaging modality, thereby optimizing the long-term benefit conferred by surgery (see Table 98-1). There are, however, important caveats. First, the vascular laboratory should be accredited and should have clear and validated criteria for measuring carotid stenosis. Second, European vascular units must be absolutely clear about which measurement method is being used (NASCET/ECST; see Chapter 16); evidence suggests that there may still be confusion on this matter.⁶ Third, the physician must be aware of DUS findings that suggest tandem inflow/outflow problems or subocclusion (see later). For either situation, corroborative imaging is mandated (MDCTA, CEMRA). In practice, one of the most frequent pathologic lesions to be missed with DUS is the coiled distal ICA (see Chapter 104). In most situations the coil does not cause undue problems for the surgeon, but it can be a cause of shunt failure (and intraoperative hemodynamic stroke) if the distal shunt tip becomes occluded against the distal loop and the problem is not recognized. This can be avoided by using intraoperative transcranial Doppler ultrasound (TCD) to assess shunt flow. DUS can also be useful in identifying unusual causes of cerebral vascular symptoms, such as the ICA stump syndrome (Fig. 98-3). In this rare condition, an occluded ICA stump (identifiable on DUS) acts as a reservoir for fresh thrombus that can then embolize up the external carotid artery (ECA) and into the brain via retrograde flow through the supraorbital and infraorbital vessels. Finally, some surgeons utilize DUS in the operating room to localize the carotid bifurcation and thus vary the level and orientation of the incision.

By contrast, CAS cannot be planned on the basis of DUS findings alone. Table 98-3 details anatomic and morphologic features that either are associated with difficulties in performing CAS or might influence the choice of cerebral protection device (CPD) or stent. For example, choice of CPD depends on ECA patency, integrity of the circle of Willis, and whether the distal ICA is straight and nontortuous. Stent selection is determined by varying combinations of vessel tortuosity and plaque friability. DUS can demonstrate (1) the presence of an appropriate ipsilateral stenosis, (2) ECA patency, (3) extent of contralateral disease, (4) tortuosity around the bifurcation, (5) plaque morphology, and (6) the possibility of coexistent inflow and outflow problems. However, DUS cannot provide information regarding (1) type of aortic arch, (2) extent of calcification at the great vessel origins, (3) proximal common carotid artery tortuosity, (4) whether there is an adequate distal landing zone for a filter CPD, or (5) the status of the circle of Willis. If the circle of Willis is not intact (Fig. 98-4), CAS practitioners tend not to deploy the flow reversal type of CPD. This type of information can be provided by only MDCTA or CEMRA (see later). However, it is important to remind CAS practitioners that reliance on additional imaging will introduce delays in planning the management of symptomatic patients. If delays are likely to be excessive (i.e., >2 weeks), patients may be better treated by CEA.

DUS enables interpretation of carotid plaque morphology (Table 98-4)^7,8. Most practitioners would agree that better methods are needed to identify vulnerable or “at risk “ plaques in asymptomatic patients (as opposed to stenosis severity alone), and high-resolution DUS will likely prove useful to be one of them. In the Asymptomatic Carotid Surgery Trial (ACST), there was no evidence that the Gray-Weale classification (see Table 98-4) predicted outcome in medically treated asymptomatic patients,⁹ but the Asymptomatic Carotid Stenosis and Risk of Stroke (ACSRS) study showed that if computerized “image normalization” was under­taken, 94% of strokes destined to occur in a cohort of 1000 asymptomatic patients with 50% to 99% stenosis severity had Geroulakis type 1 to 3 plaques^8,10 (see Table 98-4).

The weakness of DUS plaque morphology studies has been a lack of objectivity. This has been partially addressed by gray-scale median (GSM) measurement,¹¹ in which high-resolution B-mode images are combined with a quantita­tive computer-assisted measurement of plaque echogenicity (Fig. 98-5). The GSM is a numerical measure representing the average echogenicity of the plaque (low values are observed in predominantly echolucent plaques, whereas higher values are observed in more fibrous lesions). Evidence from the Imaging in Carotid Angioplasty and Risk of Stroke (ICAROS) study suggested that a GSM value of 25 or less was associated with a 7.1% stroke rate after CAS, in comparison with 1.5% for a GSM value higher than 25.¹² This study was one of the first to suggest that an objective assessment of plaque morphology using DUS could be used to plan management (i.e., when to avoid CAS).

Ultrasonography shows considerable promise in helping identify a cohort of asymptomatic patients with a higher risk for stroke in whom to prioritize CEA or CAS.¹³ In one study, the presence of three or more plaque ulcers on high-resolution three-dimensional (3D) ultrasound was associated with a significantly higher risk of late ipsilateral stroke (3-year risk 20%) than absence or the presence of up to two micro-ulcers (3-year risk 2%).¹⁴ This study also corroborated data from the Asymptomatic Carotid Emboli Study, which showed that plaque morphology predicted late, ipsilateral stroke in asymptomatic patients and that the combination of spontaneous emboli and plaque morphology conferred a greater predictive value than either imaging modality alone.¹⁵

The ACSRS group, undertaking a number of outcome analyses in a cohort of 1000 asymptomatic patients with 50% to 99% stenoses who were treated medically, has proposed a relatively simple clinically and DUS-derived scoring system for predicting late stroke on the basis of (1) GSM, (2) stenosis severity, (3) a history of contralateral TIA/stroke, and 4) plaque area.¹⁶ This scoring system, shown in Figure 98-6, illustrates how predicted annual rates of ipsilateral stroke might vary from less than 1% to more than 10%.¹³ This scoring system still requires independent validation but could emerge as a very useful method for predicting individual patient risk in the outpatient clinic.

DUS should be repeated immediately prior to CEA if more than 4 weeks has elapsed since the initial DUS. This policy ensures that the ICA has not become occluded (i.e., CEA is unnecessary). It is also the final opportunity to detect distal disease extension (last chance for consideration of temporomandibular subluxation or alternative exposure techniques) and fulfills an important quality control/internal validation role. Intraoperative DUS can also offer completion assessment following flow restoration, by showing (1) residual filling defects (luminal thrombus), (2) undue turbulence, and (3) evidence of intimal flaps/residual stenoses (see Chapter 100). No randomized trial has determined whether a completion assessment by DUS reduces procedural risk. One overview suggested that it made no difference,¹⁷ but this assertion should be interpreted with caution because all events (fatal cardiac, intraoperative/postoperative strokes, all stroke etiologies) were combined in the analyses.

Postoperative Roles

Because of its versatility, DUS is invaluable in guiding the management of postoperative neurologic deficits. These often occur when other imaging modalities are unavailable and DUS can be brought to the patient’s bedside. Air in the deep tissues can interfere with insonation early after surgery, but it is usually possible to exclude thrombosis. Diagnosing the exact cause of the deficit can be difficult, and the role of DUS is enhanced if combined with TCD (see later).

The role of serial DUS surveillance after CEA and CAS remains enduringly controversial. As with screening, such surveillance is rarely undertaken in the UK and Scandinavia but remains common practice in North America, Australasia, and mainland Europe. It might be argued that an early postoperative DUS to assess the technical result of the operation and a follow-up study at 1 year would identify the majority of patients who will experience significant intimal hyperplastic recurrent lesions. However, the contrary opinion is that the identification of these lesions is of little clinical utility because this pathology is usually a benign phenomenon (see Chapters 100 and 101). On a practical note, those who offer DUS surveillance after CAS must be aware that stents change the physical properties of the carotid artery, this requiring different diagnostic criteria.¹⁸

DUS is, however, useful in warning about the possibility of late prosthetic patch infection (Fig. 98-7). DUS is the first-line investigation in patients returning with pulsatile neck swellings, discharging sinus tracts, or neck pain, and the recognition of patch “corrugation” may precede onset of clinical symptoms by up to a year¹⁹ (i.e., recognition of this phenomenon can alert the surgeon to the possibility of patch infection and antibiotics can be started early). DUS can also identify “leaks” into early false aneurysms (see Fig. 98-7).

Accuracy and Limitations in Clinical Practice

The most important determinants for the diagnostic accuracy of DUS are the experience and skill of the operator. Second is awareness of the wide array of stenosis thresholds being used in individual centers (>50%, >60%, >70%) and the spectrum of peak systolic velocity (PSV) and velocity ratios being used to diagnoses stenosis subgroups. This uncertainty has been lessened by publication of North American DUS consensus criteria.²⁰

Third is establishing the ECST or NASCET measurement criteria used. UK experience suggests that many vascular laboratories remain unsure about this matter.⁶ Interestingly, the North American consensus document provides no advice for grading the large carotid bulb that contains large amounts of atherothrombotic material. This could score 0% with the NASCET measurement method (i.e., no need for CEA) but 70% stenosis with the ECST method (i.e., CEA justified). Future guideline makers must address this anomaly.

Fourth, in some situations, caution should be exercised in interpretation of DUS findings, and corroborative imaging should be considered. These include (1) inability to image high bifurcations, (2) inability to image above the plaque, (3) damped waveforms in the proximal common carotid artery suggesting inflow stenosis, (4) high-resistance flow in the distal ICA suggesting a distal tandem lesion, (5) the presence of a contralateral occlusion or very severe stenosis, because peak systolic velocity values may be increased in the contralateral ICA (due to hyperemic collateralization), leading to the potential for overestimation of degree of ipsilateral stenosis on the basis peak systolic velocity mea­surement, (6) excessive calcification (which prevents accurate velocity measurement because of acoustic shadowing), and (7) suspicion of subocclusion. The last condition has aroused considerable controversy. Although subocclusion (i.e., a severe ICA stenosis that does not open out into a normal-caliber lumen) was previously considered to be a marker of increased stroke risk, the Carotid Endarterectomy Trialists’ Collaboration observed that it is not a high-risk predictor for stroke and that CEA does not confer long-term benefit for patients with the condition.¹

MRA has assumed an increasingly important role in the evaluation of patients with carotid artery disease. Table 98-2 summarizes the sensitivity analyses from the HTA systematic review for time-of-flight (TOF) MRA and CEMRA. The sensitivity and specificity for diagnosing a 70% to 99% stenosis with TOF MRA is identical to that for DUS and similarly poor for diagnosing 50% to 69% stenoses. Two-dimensional (2D) TOF MRA provides strong vascular signals, even when flow is low (useful for differentiating occlusion from subocclusion), whereas 3D TOF MRA offers better spatial resolution for measuring severity of stenosis and also enables assessment of flow directionality (useful in evaluating steal phenomena). However, in order to minimize the long acquisition times and artifacts (flow-related, motion and boundary), the field of view has to be restricted. Accordingly, although it is possible to image both carotid bifurcations simultaneously, two further data acquisitions are required to image the aortic arch/great vessels and the circle of Willis with TOF MRA. This limits its overall versatility and accessibility and means that, for the most part, TOF MRA offers little in addition to DUS.

CEMRA uses the paramagnetic agent gadolinium and obtains images more rapidly than TOF MRA. CEMRA incurs fewer flow-related artifacts and provides a much greater field of view that enables high-resolution imaging from the aortic arch (Fig. 98-8) up to the circle of Willis while retaining the ability to also evaluate flow directionality. The HTA systematic review and meta-analysis (see Table 98-2) concluded that CEMRA was the best (noninvasive) imaging modality (although MDCTA was not evaluated) and it has completely superseded TOF MRA. However, some centers still do not have rapid access to CEMRA imaging. Accordingly, if it is going to take several weeks for CEMRA to be performed, the symptomatic patient will gain greater clinical benefit by undergoing expedited CEA on the basis of DUS findings alone (see Table 98-1). Another advantage of CEMRA is that it can be combined with MR functional brain imaging, although (unlike MDCTA) doing so requires an additional period of data acquisition. The use of blood-pool contrast agents now facilitates high-resolution image acquisition by allowing equilibrium-phase imaging, thereby enlarging the temporal window.

There is emerging evidence that the presence of infarction on CT/MRI can identify patients at increased risk of suffering an early recurrent stroke after presenting with a TIA or minor stroke (Table 98-5). The ABCD2 score is a clinically based scoring system that predicts the 48-hour and 7-day risk of stroke on the basis of age, blood pressure, clinical presentation, duration of symptoms, and diabetes.²¹ The maximum score is 7, and the risk of recurrent stroke increases as the score increases (see Table 98-5). Accordingly, the presence of infarction on CT/MRI in a patient with a recent TIA or minor stroke should alert the surgeon about an increased risk of early recurrent stroke, prompting expedited CEA. Interestingly, one study found that where no infarction was present, a high ABCD2 score no longer predicted an increased risk of recurrent stroke.²²