Pathology

On review of the arteries supplying the brain, different types of vascular and vessel wall alteration as well as different preferential sites may be observed in the large arteries. These changes can be physiologic in nature or may represent true pathologic findings.

In atherosclerosis, nonlaminar flow in the carotid bulb as well as in the area of bifurcations favors the formation of atheromas and stenoses, which explains their preferential sites. Other pathologies are strictly extracranially located, affecting the proximal segments as in Takayasu’s arteritis or the distal segments as in the dissections, the latter in part explained by the vessel segments prone to motion stress and in part by the vessel wall composition which may play a role in fibromuscular dysplasia. In contrast, moyamoya strictly involves intracranial vessels, again presenting a typical pattern at the carotid-T, but also in the proximal posterior cerebral artery (PCA) (Fig. A5.1).

Fig. A5.1 Distribution pattern of macroangiopathic lesions in the brain-supplying arteries. Pink = atherosclerosis (unilateral or bilateral), green = fibromuscular dysplasia (bilateral), dark blue = Takayasu’s arteritis (bilateral), light blue = dissection (mostly unilateral), gray = carotidynia (unilateral), yellow = moyamoya (mostly bilateral carotid-T, rare in the proximal PCA), Dotted lines indicate the junction of the extracranial and intracranial vessel segments.

Fig. A5.2 DSA, selective CCA filling, lateral view, of three patients with distinct coiling (arrows) of the extracranial ICA: (A) in the distal segment (not accessible to ultrasound); (B) in the middle (sometimes accessible to ultrasound); (C) proximal ICA (readily accessible to ultrasound).

Fig. A5.3 Extracranial duplex, longitudinal view. Left: B-mode insonation revealing an S-shaped ICA elongation. Right: Color-mode image of the same vessel segment.

Tortuosity

Vessel tortuosity is a frequent finding in extracranial brain-supplying arteries (Fig. A5.2) and can often be visualized by duplex ultrasound in proximal locations (Fig. A5.3). The general reported incidence varies between 10% and >40%, depending on the population studied (Ballotta et al 2005, Huber 1982, La Barbera et al 2006, Perdue et al 1975, Togay-Işikay et al 2005, Weibel and Fields 1965). Age and continuous strain due to long-term arterial hypertension as well as vessel wall alterations within the tunica media of the affected arteries have been suggested as potential underlying precipitants (Del Corso et al 1998, La Barbera et al 2006). Vessel tortuosity can be differentiated into three main groups (Fig. A5.4):

C– or S-shaped elongation with angles >90°.
Coiling of the artery up to 360°.
Kinking, considered as a variant of coiling with an angle <90°, likely to result in a lumen reduction and subsequent vessel narrowing.

Within the internal carotid artery (ICA), a straight vessel course is seen in 65–70% of individuals, a curved course in 23%, and coiling (including kinking) in 9–16%. Following Weibel and Fields (1965) an initial tortuous C– or S-shaped elongation over ~4 cm in length occurs twice as often bilaterally as unilaterally, usually affecting individuals older than 50 years. Coiling is usually found 4–8 cm distal of the bifurcation without side-to-side preferences and equal frequency bilaterally and unilaterally without any reported age dependency. Routine ICA ultrasound insonation should therefore be performed over the whole visible vessel length and as far distally as possible. Kinking predominantly involves the ICA, 2–4 cm distal of the bifurcation and is more common in elderly people. Unilateral kinking is twice as frequent as bilateral. There are no extensive analyses on the prevalence of kinking and the related grade of stenosis. The clinical importance of these vessel changes is under debate. Some authors regard elongations as an anatomic variant without any clinical implications (Togay-Işikay et al 2005), whereas others consider them to be responsible for symptomatic cerebrovascular disease (Ballotta et al 2005). Particularly problematic in this context is that elongations and atherosclerotic vessel changes frequently coexist, so a confident differentiation of causal proportion is difficult. Whenever elongations are found in symptomatic patients, careful consideration should be given to whether the symptoms match the affected vessel and if other potential causes are not being overlooked. Kinking and coiling seem to be related to carotid artery dissection (Saba et al 2014) whereas vessel elongations or tortuosity are not (Dittrich et al 2011).

Fig. A5.4 Left: Schematic of the three types of vessel tortuosity: (A) C– or S-shaped elongation with angles >90°, (B) coiling of the vessel up to 360°, (C) kinking with angles <90°. Right: Corresponding color-mode images.

Fig. A5.5 Extracranial duplex, B-mode insonation, longitudinal view. Top left and right: Single-point manual measurement of a normal (0.5 mm) and a pathologically raised IMT (1.4 mm) of the CCA. Bottom left and right: Multipoint automated IMT measurements of the CCA over a 2-cm distance given as mean ± standard deviation (normal 0.459 ± 0.013 mm, pathologic 1.326 ± 0.095 mm).

Intima-media Thickness

Within the process of developing atherosclerosis the first observable sign of vascular alteration may be an increasing intima-media thickness (IMT). However, increased IMT may not only reflect atherosclerosis but may also occur in other conditions that lead to smooth muscle cell hyperplasia or fibrocellular hypertrophy (Touboul et al 2012). IMT can only be analyzed with duplex ultrasound. High insonation frequencies increase the spatial resolution and are therefore recommended. According to the Mannheim consensus criteria, IMT is defined as follows: IMT is a double-line pattern visualized by B-mode sonography on both walls of the carotid arteries in a longitudinal image. It is made up of two parallel lines, which consist of the leading edges of two anatomic boundaries: the interface between the lumen and the intima and the interface between the media and adventitia (Fig. A5.5). IMT measurements should be performed in regions without atherosclerotic plaque, preferably in the CCA, but may also be performed in the carotid bulb as well as the ICA. If possible, measurements should be made on the far wall as near-wall evaluations are less reliable. Either manual measurements or an automated system can be used for IMT determination (Fig. A5.5). The latter allows repeated measurements within a predefined vessel segment (preferably ≥10 mm) in a short time, the former requires rigorous quality control to keep the intra- and interobserver variability low (Touboul et al 2012). No consensus exists regarding the question of whether the maximum IMT or a mean IMT should be used and if the right and the left side should be averaged, as IMT values seem to be higher on the left side (Rodríguez Hernández et al 2003). Population-based values of IMT vary depending on age, gender, and ethnicity (Table A5.1) (Howard et al 1993). There is no clear accepted definition of a raised IMT but a size >1 mm is generally considered abnormal. Raised IMT values have been associated with several classic vascular risk factors such as hypertension, smoking, cholesterol, homocysteine levels, C-reactive protein, and the presence of a metabolic syndrome or coronary artery disease (Crouse 2006). Prospective analyses have demonstrated that raised IMT values are associated with an increased number of myocardial infarctions and stroke in the elderly population, indicating a predictive value for future vascular events (Lorenz et al 2007, O’Leary et al 1999, Polak et al 2011, Silvestrini et al 2013).

Atherosclerotic Plaques

Loss of laminar flow at a bifurcation or vessel widening like the carotid bulb leads to increased vessel wall shear stress and subsequently causes initiation and promotion of vessel wall arteriosclerosis and plaque formation in conjunction with vascular risk factors and age. It seems that the left carotid bulb is more vulnerable than the right for the development of atherosclerosis. In a large stroke-free population studied using MRI, bilateral plaques were seen in 85% but unilateral plaques were twice as prevalent on the left side as on the right (67% versus 33%). Also plaque thickness was greater on the left side (3.1 ± 1.2 versus 2.9 ± 1.3 mm), whereas grade of stenosis was similar (Selwaness et al 2014). According to the Mannheim consensus criteria, atherosclerotic plaques are defined as follows: Plaque is a focal structure encroaching into the arterial lumen of at least 0.5 mm or 50% of the surrounding IMT value, or demonstrates a thickness >1.5 mm as measured from the media–adventitia interface to the intima–lumen interface (Touboul et al 2012). Atherosclerotic plaques usually involve only one segment of the wall circumference; at the carotid bifurcation they are often found on the posterior wall. Atherosclerotic plaques can be further characterized by the following criteria (Figs. A5.6–A5.11):

Table A5.1 Common carotid artery intima-media thickness distribution in a population-based ultrasound analysis: Maximal wall thickness, combined data from the left and right side. (From Howard et al 1993). Values above the 95th percentile can be considered as pathologic.

Number: Singular, multiple.
Location: Affected vessel, anterior/posterior wall, lateral or medial wall.
Extension: Circumscribed, longish.
Form: Marginal, concentric/circular, eccentric/ semicircular.
Size: Length in longitudinal section and thickness in cross-section (in mm).
Texture: Hyperechoic, isoechoic, hypoechoic, anechoic (isoechogenicity = intima brightness), homogenous or heterogeneous pattern, calcifications with or without acoustic shadowing.
Surface: Regular smooth, irregular with recess/ulcerated.
Vessel lumen reduction: Grade of stenosis.
Response of the plaque to contrast agents: Enhancement versus no enhancement.

Because of accessibility to duplex ultrasound and clinical relevance, plaque description focuses mainly on the CCA and ICA. The vessel walls should be studied in both planes. Transverse images are mandatory to delineate the real extension of the plaque (Fig. A5.6 and Fig. A5.7). A vessel ulcer may be detected in larger lesions (Fig. A5.8). Plaques of low echogenicity may be difficult to detect in routine B-mode examination, but color-mode imaging uncovers its extent and intraluminal effect (Fig. A5.9). A homogenous hypoechoic plaque may appear similar to a fresh thrombus (Fig. A5.10). In steno-occlusive disorders of the CCA and ICA, description of plaque and stenosis in the external carotid artery (ECA) becomes more relevant. The vertebral arteries (VAs), especially their most frequently affected proximal part, are only rarely accessible for B-mode analysis of plaque composition and even plaque presence (see also Video A5.1).

Fig. A5.6 Extracranial duplex. (A–C) B-mode insonation. (A) Longitudinal view of the carotid bifurcation. Large near-wall hype rechoic plaque (arrowhead) with pronounced acoustic shadowing (arrows). (B) Cross-sectional view of the CCA: Semicircular eccentric plaque of largely homogenous isoechogenicity and a smooth surface (arrows). Note the circumscribed anechoic lesion within the plaque (middle arrow) which might be intraplaque hemorrhage or necrotic core. (C) Longitudinal view of the CCA: Multiple, predominantly hyperechoic plaques (arrows). (D) Color-mode insonation, longitudinal view of the carotid bifurcation, composite image: Large hypoechoic material (arrow).

For follow-up studies plaque size can be measured by recording the longest diameters in different planes. As well as the finding of plaque progression, plaque regression can also be observed under medical therapy. A large study in 4,378 ultrasound-assessed patients even reported a change in the number of patients with observed plaque regression from 25% before initiation of an intensified medical treatment, to 50% of patients after initiation of treatment (Spence and Hackam 2010). The authors concluded, “treating arteries without measuring plaque would be like treating hypertension without measuring blood pressure” and recommended validation of the findings in a randomized clinical trial.

Fig. A5.7 Extracranial duplex. (A) Longitudinal view of the proximal ICA revealing an almost homogeneous hypoechoic near-wall plaque (arrows). (B) Doppler spectrum analysis revealing a mildly turbulent flow within normal velocities ranges (82/36 cm/s). (C,D) Cross-sectional color-mode image of the proximal ICA: Circular, almost concentric mixed plaque with hyper- and hypoechoic areas. Diameter reduction, 55%. Area reduction of the same location, 80%.

Fig. A5.8 Extracranial duplex. (A) Color-mode image, longitudinal plane. (B) Simultaneous Doppler spectrum analysis. Proximal carotid artery stenosis with a flow velocity of 211/48 cm/s. (C) B-mode image demonstrating a mild hyperechoic plaque. Note the anechoic round structure (arrow) within the plaque, which might be overlooked if the gain is not carefully adjusted. (D) Same area in color-mode imaging with reduced PRF. Note that the anechoic structure is now delineated as a plaque ulcer comprising blood flow with low flow velocities (arrow).

Fig. A5.9 Extracranial duplex, longitudinal view of the carotid bifurcation. (A) B-mode image with optimal gain setting revealing a homogeneous mild hypoechoic near-wall plaque. (B) Color-mode image helps to delineate the real size of the plaque or stenosis. (C) Low gain leads to underestimation or overlooking of plaque. (D) High gain leads to a loss of B-mode details in the surrounding tissue, but delineation of the plaque remains good.

Fig. A5.10 Extracranial duplex, fresh thrombus in the proximal ICA caused by a distal ICA occlusion. (A) Color-mode, cross-sectional view showing an eccentric hypo- to isoechoic structure resembling thrombotic material (arrows) with incomplete filling of the vessel lumen. Note the color signal in the ECA (arrowhead). (B) B-mode, longitudinal view delineates a complete filling of the proximal ICA (arrows) by the fresh thrombus.

In our laboratories we measure the diameter of a plaque in relation to the vessel in the cross-sectional plane. Starting up with a local diameter reduction of 30% we grade a local stenosis in 10% steps up to a stenosis of 50% (Fig. A5.11). A diameter reduction below 30% is not reported as “stenosis” but described as severe atherosclerotic macroangiopathy.

Area measurements are an additional tool for assessing a lumen reduction. This approach may become of greater relevance as high-resolution CT angiography (CTA) also allows diameter and area measurements for comparison. Both methods, however, are hindered by severe calcifications. In stenoses, the sonographer should note if the distal end is detectable as it may have implications for the vascular surgeon. If shadows caused by calcified plaque impede the assessment, this should also be noted. Homogenous hypo echoic plaques are sometimes difficult to detect and require optimal gain adjustments and the use of color-mode imaging. In young patients without visible calcifications it may be impossible to distinguish a dissection with wall hematoma from a large plaque of low echogenicity. In these cases transverse planes should be studied. An enlargement of the total vessel circumference indicates a dissection.

Atherosclerotic vessel wall changes in the brain-supplying arteries show a specific distribution pattern. According to a four-vessel catheter angiographic analysis of 3,788 patients in the chronic phase after cerebral ischemia performed in the 1960s, stenoses are most frequently found at the extracranial ICA origin followed by the VA origin, the subclavian artery (SA), and the intracranial ICA whereas vessel occlusions may be found at slightly different locations (Fig. A5.12). Note that the number of intracranial stenoses may be an underestimate as no subtraction technique was applied.

Like IMT enlargement, carotid plaques are associated with several vascular risk factors and are strong predictors of stroke, death, or myocardial infarction even after adjustment for classic risk factors such as hypertension, smoking, and cholesterol levels (Silvestrini et al 2013, Spence 2006).

A large number of studies have tried to identify “high-risk” (vulnerable/complicated) plaques by using morphologic ultrasound criteria. Calcium as a sign of regressive plaque modulation seems to lower the risk. In contrast, heterogeneous echogenicity and ulcerated plaques, i.e., with an irregular surface and/or ulceration, are postulated to be less stable and more likely to cause embolic ischemic events. Also, hypoechoic plaques seem to be more likely to become symptomatic than hyperechoic plaques (Lal et al 2002, 2006, Sabetai et al 2000). More recently, plaque analy sis by means of contrast-enhanced (ce) ultrasound has been introduced. The observable ultrasound perfusion characteristics seem to correlate well with histological findings of plaque neovascularization (Li et al 2014). The latter is more frequently seen in symptomatic patients with carotid artery stenosis (Xiong et al 2009). Histopathology findings such as fibrous cap configuration, necrotic core, or intraplaque hemorrhage are beginning to be investigated by ultrasound but are currently not satisfactorily accessible (Funaki et al 2011) (for further details, see Case 1 and Video A5.2).

Plaques of intracranial vessels cannot be visualized by ultrasound. An unenhanced CCT scan can be suggestive of relevant atherosclerosis, as calcified plaques will delineate the affected vessel segments. Appropriate contrast window settings are essential when reading the CT scan, as vessel calcifications appear exaggerated in the typical window setting for brain parenchyma. Intracranial plaques are highly prevalent in the elderly general population. In a study population of 2,500 participants with a mean age of 70 years, they were found in over 80% of cases (Bos et al 2012). Usually they are present in the carotid siphon, followed by the VA and BA. In the MCA, ACA, and PCA calcifications are rarely seen (Fig. A5.13). Carotid siphon calcifications are correlated with advanced atherosclerosis of the carotid bulb and severe carotid siphon calcification is correlated with stroke (Bos et al 2014, Fisher et al 1965b). Calcification of the vertebrobasilar arteries is more often found in patients with stroke history (Pikija et al 2014). In acute vessel occlusion, a fresh embolus can also be depicted by noncontrast CT, e.g., in proximal M1-MCA occlusion in the form of a positive MCA sign, in M2-MCA occlusion as a so-called “dot sign” (Fig. A5.14). However, in the differential diagnosis between a thrombus/embolus or a calcification clear differentiation may be difficult.

Fig. A5.11 Extracranial duplex, B-mode, cross-sectional view of the ICA with different grades of local stenosis (lumen reduction) by eccentric plaques in the CCA. (A) Atherosclerotic plaque without considerable lumen reduction. (B) Atherosclerotic plaque with 30% vessel diameter reduction. (C) Atherosclerotic plaque with 40% vessel diameter reduction. (D) Atherosclerotic plaque with 50% vessel diameter reduction.

Fig. A5.12 Distribution pattern of stenosis >50% (A) and vessel occlusion (B), assessed by conventional angiography in patients in the chronic phase following cerebral ischemia. Preferential sites are indicated in yellow, followed by turquoise (adapted from Hass et al 1968).

Fig. A5.13 Unenhanced CCT scans, axial plane. Arrows indicate calcifications. (A,B) Distinct VA calcification. Note the better delineation of calcium formation in the conventional setting. However, if differentiation of calcification from bony structures is difficult, visualization using the bone window settings may be helpful. (C) Calcifications of all vertebrobasilar segments. (D) Distinct bilateral carotid siphon calcifications in the bone window setting. (E) Rare finding of MCA calcifications.

CT is particularly suitable for analyzing extracranial calcifications and other vessel wall pathologies of the brain-supplying arteries. Since the introduction of multislice CT in 1999 high-resolution plaque assessment has become available and even small ulcerations can be seen as pits filled by contrast medium. CTA-determined mean soft plaque thickness (assessed in the source data images) seems to be a useful marker for symptomatic ICA plaques and is also correlated with intraplaque hemorrhage (Gupta et al 2015a, 2015c). Conventional CTA, however, is less meaningful in the analysis of plaque composition or surface description in noncalcified plaques if compared with histopathologic findings (Denzel et al 2005, Oliver et al 1999, Saba et al 2007). Dual-source CT, allowing removal of hard plaque by using different tube voltages, seems more promising for plaque definition. Surface irregularity or ulceration was more frequently detected with dual-source CTA than with TOF-MRA and digital subtraction angiography (DSA) (Lv et al 2014). Cone beam CTA has recently been reported to provide high spatial resolution images of plaque morphology that might give additional information on intracranial atherosclerosis compared to DSA (Safain et al 2014).

Fig. A5.14 Unenhanced CCT, axial plane, parenchymal contrast settings. Left: Positive right-sided “media sign” resembling a fresh M1-MCA clot which extends into the M2 segments (arrow). Right: Right-sided positive “dot sign” (arrow) indicating cross-sectional imaging of embolic M2-MCA branch occlusion.

As with duplex sonography, the behavior after contrast agent administration may help to detect unstable plaques and stenoses. In one study analyzing symptomatic and asymptomatic carotid stenoses the presence of carotid wall enhancement was higher in symptomatic stenoses whereas the presence of either calcified plaque or no wall enhancement was more often observed in asymptomatic patients (Romero et al 2013).

Unlike CCT, MRI is unable to depict plaque calcification. However, high-resolution MR—using blood-suppressed T1-, T2-, and proton density-weighted fast spin echo, gradient echo, and time-of-flight sequences—is able to visualize carotid plaque components such as hemorrhages, the necrotizing core, and fibrous components in vitro and in vivo (Makris et al 2015, Millon et al 2013, Puppini et al 2006). Using 3D T1-weighted MRI of carotid plaques with histopathologic validation sensitivity and specificity of 100% was reported for discriminating vulnerable from stable plaques (Narumi et al 2015). Gadolinium enhancement was also detected in vulnerable plaques as has been shown after histologic analysis of surgical specimens after carotid endarterectomy (Millon et al 2012). A new approach for the detection of lipid-rich necrotic plaque cores is the diffusion-prepared turbo-spin-echo (DP-TSE) technique which allows plaque analysis with high spatial resolution (Xie et al 2014). The use of MRI techniques may be particularly valuable in the determination of high-risk plaques, for instance in patients with high-grade asymptomatic carotid stenosis (Crouse 2006, Gupta et al 2013, Nighoghossian et al 2005, Saam et al 2006) or in patients in whom CTA cannot be performed because of a renal insufficiency, often seen in vascular patients.

Fig. A5.15 MRI, axial plane. (A,B) Enlarged T2-weighted image. (A) Normal flow void in both A1-ACA and M1-MCA segments. (B) MCA stenosis causing a reduced flow void within the right M1-MCA segment (arrow). (C) MRI, axial plane. Enlarged T2-weighted image: Absent flow void within the left M1-MCA segment in M1 occlusion (arrow). (D) Corresponding 3D TOF-MRA with absent M1-MCA signal (arrow).

Fig. A5.16 Extracranial duplex ultrasound, B-mode imaging of the common carotid artery, longitudinal plane. (A,B) Automated measurements of local arterial stiffness in form of mean arterial distensibility in two different patients. Note the excellent pulse curve reproduction (blue), derived from the moving vessel walls in the ultrasound images. (A) High mean CCA distensibility of 531 ± 37 μm compared with (B) showing a reduced distensibility value of 367 ± 30 μm. Note the correspondingly different amplitudes between both pulse curves.

Despite advances in the description of extracranial carotid plaques, the current spatial resolution of MRI prevents a detailed plaque evaluation of intracranial stenoses. Here, high-resolution MRI with and without gadolinium administration using at least 3–7 T potentially allows the graduation of stenosis, the assessment of plaque presence, and composition of intracranial wall pathologies similar to the extracranial brain-supplying arteries (Bodle et al 2013, Degnan et al 2012, Majidi et al 2013, Swartz et al 2009). Noncontrast sequences, in particular T2-weighted images, however, permit the assessment of vessel patency by analysis of the intravascular flow void (Fig. A5.15). We strongly recommend that if a CT scan or MRI has been performed before the ultrasound examination, the available information about vessel wall pathology should be taken into consideration.

Arterial Stiffness

Arterial stiffness is a relatively new biomarker which has recently gained attention, in particular within an epidemiological context of early vascular risk assessment (Bruno et al 2014, Laurent et al 2012, Tomiyama and Yamashina 2010). Besides the measures of IMT, arterial stiffness seems to be an additional and even earlier observable parameter, related not only to general cardiovascular morbidity but also to cerebral microangiopathy and impaired cognitive brain function (Singer et al 2014). The analysis principle is based on the assessment of arterial distensibility, i.e., the Windkessel function (for further reading also see also Chapter 3, “Analysis of Cerebral Blood Flow”) which is defined by the elastic properties of the arterial vessel walls. These elastic properties decrease with age and are related to an individual’s systemic blood pressure. B-mode ultrasound makes it possible to measure vessel wall movements and to analyze their extent and temporal pattern, resulting in typical pulse wave curves (Fig. A5.16). Comparison of time delay between pulse waves assessed in the carotid artery and in the femoral artery is then used to calculate the pulse wave velocity (PWV = distance/transit time) which is currently considered to be the “gold standard” for arterial stiffness assessments. Reference values derived from more than 16,000 subjects have been published (Mattace-Raso et al 2010) and can be used to identify individuals with an increased cardiovascular risk. A variety of systems for routine assessment of arterial stiffness are available. However, newly emerging techniques such as ultrafast ultrasound imaging, which even makes it possible to assess a local pulse wave velocity, may further simplify this diagnostic approach (Messas et al 2013; see also Chapter 1, “Ultrafast Imaging” under “Imaging Modalities, Parameters, and Settings”). Attempts to evaluate arterial stiffness in relation to medical therapy have been made by analyzing the effect of different antihypertensive medications in coronary artery disease. However, high-class evidence for treatment guidance is still lacking (Liao and Farmer 2014).

Dissection

Ultrasound as well as DSA, CTA, and MRI can be used in diagnosis and follow-up of patients with isolated dissections of the extracranial brain-supplying arteries in the ICA (Fig. A5.17, Fig. A5.18, Fig. A5.19, Fig. A5.20, Fig. A5.21) and vertebral arteries (Fig. A5.22 and Fig. A5.23; see also Videos A5.3 and A5.4), as well as in patients with dissections of the CCA continuing from lesions of the aortic arch (Fig. A5.24; see also Videos A5.5 and A5.6). Clear ultrasound signs of dissection are a large-vessel occlusion without atherothrombotic wall pathology, a cone-shaped occlusion, and a vessel wall hematoma with an anechoic or hypoechoic appearance. The latter leads to an obvious widening of the vessel which can be best seen if it occurs in the V3-VA segment at the atlas arch. If the dissection is in a more distal location the widening of the ICA may be missed. Other ultrasound findings are irregular vessel membrane with double lumen (Alecu et al 2007, Bartels and Flügel 1996, Lu et al 2000, Touboul et al 1988). As well as occlusion, a dissection may also cause stenosis. Dissections usually affect longer vessel segments and are mostly found in extracranial locations. In contrast to atherosclerotic vessel wall alterations, extra cranial ICA dissections are located distally near the base of the skull, which often limits their visualization by duplex sonography.

Fig. A5.17 ICA dissection. (A) DSA, left CCA injection, lateral view: Proximal occlusion of the ICA caused by a dissection. Note the typical cone-shaped or “flame-like” occlusion (arrow). (B,C) Duplex ultrasound, longitudinal plane. (B) B-mode image demonstrating the cone-shaped vessel narrowing (arrows). (C) Color-mode image demonstrates absent flow in the distal ICA (arrows). Note that no clear wall hematoma is visible in this case.

Fig. A5.18 ICA dissection. (A) Contrast-enhanced 3D MRA, coronal MIP in acute dissecting ICA occlusion 3 days after symptoms onset (arrow). (B) MR T1-weighted fat-suppressed image, axial plane corresponding to the white dotted line in A showing a moderate hyperintense vessel wall hematoma signal because of the recent bleeding. The typical bright signal may appear days later. A fresh intraluminal thrombus may mimic a wall hematoma. The marked widening of the external vessel lumen (maximal 7.9 mm; arrowheads) is, however, confirmatory of dissection in case of doubt. Note the normal diameter of 4.5 mm in the unaffected ICA (arrows) with normal flow void indicating undisturbed flow.

Fig. A5.19 ICA dissection. (A) MR T1-weighted fat-suppressed image, axial plane showing subacute ICA dissection. The vessel wall hematoma appears brighter surrounding a patent lumen but the vessel diameter is clearly widened (arrowheads) in comparison to the normal contralateral diameter and flow void (arrows). (B) Color-mode image of the mid ICA, axial plane, corresponding to the MR image. A first look may suggest normal vessel appearance, but on a second look the eccentric homogenous hypoechoic wall hematoma becomes obvious (C). (C) Inner circle: residual lumen, hardly affected; outer circle: delineation of widened outer vessel circumference.

However, on rare occasions, dissections may also occur or extend intracranially. The site of dissection should be determined according to its beginning. Intrastenotic flow velocities are often lower than expected in vessel dissections if compared with short segment stenoses of atherosclerotic origin. A dissecting aneurysm may also be detected if accessible by ultrasound (Fig. A5.25). In the chronic phase after vessel recanalization a collapse-like appearance of the ICA may be observed. Flow analysis then shows that the main flow is into the ophthalmic artery (OA). In such a case the ipsilateral anterior circulation is perfused via one or both communicating arteries (Fig. A5.26). For a more detailed discussion, see Case 11, Case 18, and Case 19.

Fibromuscular Dysplasia (FMD)

The brain-supplying arterial segments which are most frequently affected by FMD, i.e., the distal segments of the extracranial ICA and VA, are not readily accessible by duplex ultrasound (see Fig. A5.1). FMD is therefore rather diagnosed by DSA, CTA, or ce-MRA. However, if proximal vessel segments are affected, the irregular arterial vessel walls may also be visualized by duplex ultrasound (Fig. A5.27). There is an overlap between FMD and vessel dissection, as patients with FMD are prone to dissecting lesions. For a more detailed discussion of FMD, see Case 13.

Fig. A5.20 ICA dissection. (A) Contrast-enhanced 3D MRA, coronal MIP in dissecting ICA stenosis after partial recanalization 6 weeks after symptoms onset (arrows). (B) MR T1-weighted fat-suppressed image, axial plane corresponding to the white dotted line in A showing vessel wall hematoma now revealing the typical crescent sign with an eccentric hyperintense rim indicating hematoma transformation. (C) Color-mode image of the midpart ICA, axial plane, corresponding to the MR image showing the almost hypoechoic and homogenous eccentric wall hematoma (arrows).

Fig. A5.21 ICA dissection. (A) Contrast-enhanced 3D MRA, coronal MIP in dissecting ICA stenosis after partial recanalization 6 weeks after onset of symptoms. The white dotted line indicates the assumed widening of the outer wall due to intramural hematoma. (B) Color-mode image of the midpart ICA, longitudinal insonation plane. Corresponding to the MR image a large hypoechoic hematoma is detected. The fine hyperechoic line, indicated by the arrows corresponds to the outer vessel wall (arrows). 1 indicates the normal vessel diameter of 4.5 mm, 2 indicates that the outer diameter has almost doubled.

Fig. A5.22 VA dissection at the atlas arch (V3). (A) Color-mode image, longitudinal view demonstrating a normal diameter (4.0 mm), vessel wall, and color signal of the proximal V2-VA. (B) Longitudinal color-mode image at the atlas arch (V3) showing an enlarged vessel diameter (8.2 mm) and a remaining perfused lumen of 2.1 mm. (C) T1-weighted MRI, fat-suppressed image, axial plane showing the wall hematoma (arrow). (D) Doppler spectrum analysis with turbulent flow and raised intrastenotic flow velocity of 199/91 cm/s.

Fig. A5.23 VA dissection at the atlas arch (V3). Left: T1 weighted MRI, fat-suppressed image, coronal plane showing the wall hematoma (arrow). Right: Color-mode image, longitudinal view with initial hypoechoic enlarged vessel wall (top and middle). Complete normalization after 2 months (bottom).

Vasculitis/Arteritis

Ultrasound, CTA, and MRI may be used to diagnose vasculitis of large- and medium-sized arteries and DSA of all types including small-vessel arteritis. Conditions accessible to duplex ultrasound are intracranially the primary intracranial arteritis and extracranially Takayasu’s arteritis and the giant cell arteritis (Fig. A5.1, Fig. A5.28, Fig. A5.29). For a more detailed discussion of giant cell arteritis, see Case 16; for Takayasu’s arteritis, see Case 23.

Carotidynia and Carotid Artery Vasospasm

Carotidynia is a rare clinical diagnosis. Patients report mild to moderate neck pain with point-tenderness over the anterolateral aspects of the neck, without history of trauma. The pain responds quickly to steroids or nonsteroidal anti-inflammatory drugs (NSAIDs). Duplex sonography reveals hypoechoic wall thickening of the carotid bulb in the region of tenderness leading to a mild lumen narrowing and an outward extension of the vessel wall. Regression or normalization is the usual finding in follow-up examination after several weeks (Arning 2005, Schaumberg et al 2011). MRI corresponds with duplex sonography describing a hyperintense lesion in the distal CCA and carotid bifurcation. Gadolinium administration usually leads to an enhancement surrounding the carotid bifurcation without presence of stenosis (Burton et al 2000). Also in MRI the findings normalize over time (Fig. A5.30 and Fig. A5.31). The underlying pathology remains unclear (Taniguchi et al 2008): A variant of vasculitis has been suggested. Dissection has to be excluded and therefore MRI is mandatory. In all image modalities the alterations may resemble a plaque (Woo et al 2008). Knowledge of this rare and distinct self-limiting syndrome is important and can avoid unnecessary and invasive procedures.

Fig. A5.24 Aortic arch dissection DeBakey type I, extending into the CCA. Duplex ultrasound. (A) Cross-sectional image of the CCA. Top: B-mode image revealing a hyperechoic membrane within the vessel lumen. Bottom: Color-mode image demonstrates flow in both lumina. (B,C) Longitudinal color-mode image and simultaneous Doppler spectrum analysis of the CCA. (B) Sample volume within the true lumen with antegrade unidirectional flow. (C) Sample volume within the false lumen with bidirectional flow.

Fig. A5.25 V1-VA segment dissection before entering the first transverse process at C6. Left: Ce-MRA, coronal MIP showing a dissecting aneurysm (white circle). Top: Ce-MRA, coronal MIP, 90° counterclockwise rotated to correspond with the ultrasound image. Bottom: Duplex sonography, B-mode compound image: Normal V1-VA diameter at the midpart segment (d1 = 3.9 mm), diameter widening of the distal V1-VA before entering the C6 transverse process (d2 = 6.5 mm), corresponding color-mode image above demonstrating extended filling of a dissecting aneurysm, normal distal V2-VA diameter between transverse processes C4 and C5 (d3 = 4.1 mm).

Fig. A5.26 ICA dissection. (A) Contrast-enhanced 3D MRA, coronal MIP in dissecting ICA stenosis after partial recanalization one year after symptoms onset. The ICA reveals no wall irregularities but the diameter of 2.8 mm is still lower than normal (arrowheads). (B) Extracranial duplex, color-mode image, and Doppler spectrum of the ICA origin, longitudinal insonation plane. Note the low velocity and mildly increased pulsatility (flow velocity 52/14 cm/s, PI = 1.4). (C) TCCS, transorbital approach, color-mode image, and Doppler spectrum of the ipsilateral ophthalmic artery (OA) revealing comparable flow parameters (flow velocity 33/10 cm/s, PI = 1.3). In the presence of a patent intracranial collateral flow through one or both communicating arteries the ICA may be responsible for the OA flow only, i.e., become an “extended OA,” also termed “OA-like ICA.”

Fig. A5.27 Fibromuscular dysplasia: (A) Contrast-enhanced 3D MRA, coronal MIP revealing marked elongations and caliber irregularities in the ICA (arrows) and even more in the VA (arrowheads). (B,C) Extracranial duplex, longitudinal color-mode image of the ICA showing proximal (B) and midpart (C) segmental narrowing and dilation without raised flow velocity (spectrum not shown).

Fig. A5.28 Takayasu’s arteritis. Duplex ultrasound of the CCA. 11-MHz linear probe. (A,B) Cross-sectional plane. (A) B-mode image: Homogenous, isoechoic to mildly hyperechoic circular vessel wall thickening. (B) Color-mode image: Color filling of the remaining arterial vessel lumen. (C) Doppler spectrum analysis of the affected CCA. (D) Color-mode, longitudinal plane of the same vessel segment.

Fig. A5.29 Duplex ultrasound of the superficial temporal artery. 22-MHz linear probe. Left (A–C): Normal individual. Right (D–F): Giant cell arteritis. (A) Cross-sectional B-mode image: Isoechoic unremarkable vessel wall. (D) Cross-sectional B-mode image: Hypoechoic and enlarged vessel appearance. (B) Cross-sectional color-mode image. Normal vessel wall appearance. (E) Cross-sectional color-mode image. Large circular hypoechoic wall thickening in form of a typical halo sign. (C) Longitudinal color-mode image. Isoechoic B-mode aspect of the vessel bordering tissue. (F) Longitudinal color-mode image. Hypoechoic wall thickening over the total visible vessel length.

Fig. A5.30 Carotidynia. Images of a patient with typical subacute neck pains. Top: B-mode image longitudinal (left) and transverse plane (right). Note the mildly hypo-/isoechoic structure in the carotid bifurcation starting in the distal CCA and visible in both planes which might be mistaken for a homogenous atheroma (arrows). Right: Same patient with remitted symptoms after 2 weeks of treatment with nonsteroid analgesics. Note a mild residual structure on the transverse plane only (arrow).

Fig. A5.31 Carotidynia. MRI and ce-MRA of the same patient on the day of the first ultrasound examination. Top left: Circular but eccentric hyperintense wall thickening (yellow circle). Bottom left: Homogenous contrast enhancement after gadolinium administration (yellow circle). Middle: Ce-MRA, rotated coronal MIP. Mild proximal ICA irregularities (arrows) indicate extraluminal location the pathology. Right: Ce-MRA, rotated coronal MIP. MRA revealing normalization after 3 months (arrowhead) with only a mild residual vessel wall thickening and without contrast enhancement (not shown).

Another rare but distinct entity is recurrent extracranial carotid artery vasospasm. Since the first description by Lieberman et al (1984) only a few patients, some of them migraine sufferers, have been reported. The first patient diagnosed by duplex ultrasound presented in 1998. She was a 32-year-old woman suffering from cerebral ischemia due to recurrent stenoses of the ICA ~4 cm distal to its origin (Arning et al 1998). In 2006 two further patients were reported, clinically also suffering from manifest stroke (Janzarik et al 2006). Extracranial vasospasms as a cause of stroke might be considered in patients with recurring ischemic events and distal ICA stenosis without signs of dissection. Treatment with NSAIDs and/or calcium channel blockers seemed the best option to reduce the frequency of vasospasms.

Fig. A5.32 Extracranial duplex, cross-sectional view. (A) B-mode image: ICA filled with mildly hyperechoic material, leading to an eccentric lumen narrowing. (B) Color-mode image: Confirmation of the suspected stenosis. (C) Diameter measurements: D_total = 9 mm (green bar), D_stenosis = 3.4 (red bar), resulting in a 62% stenosis. (D) Area measurements: A_total = 52.4 mm² (green circle), D_stenosis = 10.1 mm² (red circle), resulting in an 81% stenosis.

Fig. A5.33 Calculated relationship of stenosis grade using the diameter and area method. Blue line: Relation in axisymmetric vessel narrowing. A 30% diameter stenosis is equivalent to a ~50% area stenosis, and a 70% diameter stenosis is equivalent to a ~90% area stenosis. Red line: More linear relation between diameter and area stenosis in severely asymmetric plaque formation and vessel narrowing.

Stenoses and Occlusions

In general all segments of the extracranial and the relevant parts of the intracranial brain-supplying arteries can be assessed by duplex ultrasound, provided the insonation conditions are good. The extent and order of ultrasound investigation should always be oriented according to the clinical picture and other relevant clinical data such as age, vascular risk factors, concomitant circumstances of the cerebral ischemia, and suspected etiology, based on a radiologically documented stroke pattern if this is available before ultrasound examination. For instance, in a Caucasian who has suffered an embolic ischemia in the MCA territory, a proximal ICA stenosis should be considered first, which can be assessed adequately by duplex ultrasound. If no relevant pathology is found, the distal extracranial ICA, the intracranial ICA in all its accessible segments, as well as the MCA in its M1 and M2 segments must be studied. In an Asian patient, a lesion in the MCA has to be considered first. In case of a cerebellar ischemia, the question of a VA stenosis at its origin should be the primary focus. If not found, a more distal VA stenoocclusive process, typically in the intracranial VA or BA, has to be looked for. In the case of PCA territory ischemia, the total visible length of the PCA has to be examined. A single-vessel pathology might be the cause of the recent ischemia but ipsilateral (tandem stenosis) or contralateral steno-occlusive lesions may also present a challenge to the interpretation of otherwise simple findings.

Ultrasound Criteria of Stenoses

Within the extracranial brain-supplying arteries, duplex ultrasound permits the morphologic analysis of the affected vessel segment. Direct (velocity within the stenosis) and indirect hemodynamic effects of stenoses (pre- and poststenotic velocities and waveform abnormalities) may be assessed in all extra- and intracranial brain-supplying arteries. For exact assessment of a stenosis all available criteria should be considered. The proximal extracranial ICA is not only the most commonly affected site (at least in Caucasians), but can also be used to explain the main principles of ultrasonography in vessel disorders. The following remarks are mainly based on studies of the extracranial ICA.

Direct Morphologic Assessment

Extracranial duplex ultrasound is able to visualize the arterial vessel lumen near and at the carotid bifurcation, the formation of intraluminal and vessel-narrowing plaques, or even complete vessel filling, e.g., with thrombotic material. As the thrombotic material can be hypoechoic, analysis should always be performed using the combination of B-mode and color-mode ultrasound. The latter considerably facilitates the detection of the residual perfused lumen and helps to avoid overlooking, for example, a fresh, hypoechoic, or small floating thrombus. Care should be taken to adjust pulse repetition frequency (PRF) and color gain to prevent color overlapping beyond the perfused lumen. Plaque calcification, which may lead to pronounced acoustic shadowing, is a limitation for direct morphologic ultrasound assessment. This phenomenon may be observed in up to 7% of patients (Polak et al 1989). Geometric lumen reduction on B-mode and color-coded flow imaging can be assessed in two ways: by calculating the reduction in the cross-sectional diameter or the cross-sectional area.

Diameter: Diameter assessments can be performed in a longitudinal plane, such as with DSA. Color-mode imaging facilitates the recognition of echolucent material and its use is therefore mandatory in addition to B-mode imaging. Because of oversteering artifacts this method is usually used in local ICA or CCA stenoses of ~50% and less. Care should be taken that the anterior and posterior walls are simultaneously visible. Measurements are performed in the region with maximal lumen reduction from the inner border zone of the wall. The grade of stenosis is calculated from the relation of the total vessel diameter (D_total) and the minimal stenosis diameter:

Fig. A5.34 Mathematical flow model in ICA stenosis (Spencer’s curve). Theoretical relationship of stenosis grade (assessed by diameter and area measurements) and systolic blood flow velocity (blue line) within the stenosis. In very high-grade stenosis velocity values drop, and they further decrease in near-occlusion. Note the blood volume flow (red line) values remain constant until a 70–80% diameter reduction, defining the start of a hemodynamic relevant stenosis. (Adapted from Spencer and Reid 1979.)

Area: Unlike DSA but similar to CTA and (with some restrictions) to MRI, duplex ultrasound allows measurement and calculation of the grade of stenosis from the vessel’s cross-sectional area, a parameter which correlates best with results derived from postoperative histologic planimetric analysis (Alexandrov et al 1993, Eckstein et al 2001). It is the true relevant anatomic parameter for the measured flow velocities according to the physics of flow. Area measurement is also independent of the morphological configuration of the stenosis, while the diameter approach only measures correctly in the case of a cylindrical stenosis. The grade of stenosis is calculated from the relation of the total vessel area (A_total) and the minimal stenosis diameter (A_stenosis):

Examples of both diameter and area measurements are given in Fig. A5.32. Although assessed in exactly the same vessel segment, the two methods yield different results. The diameter calculation (D_total = 9 mm, D_stenosis = 3.4 mm) results in a 62% stenosis, the area calculation (A_total = 52.4 mm², A_stenosis = 10.1 mm²) in 81% stenosis. This phenomenon can also be described mathematically (Fig. A5.33). Depending on the type of stenosis (axisymmetric or asymmetric) the nonlinear relation between area and diameter varies in favor of diameter or area (Spencer and Reid 1979). Up to now all major clinical trials with catheter angiography as the main method have used the diameter approach, so it will currently continue to be the preferred method of assessment. In the future, however, the area method recommended for duplex sonography in the 1990s (de Bray and Glatt 1995) will probably gain importance, particularly considering the increasing use of the CTA technique, with which it is also possible to perform exact planimetric cross-sectional measurements (Bartlett et al 2007).

Direct Hemodynamic Assessment

Hemodynamic effects can be observed using the color mode of the ultrasound system. The color signal not only reveals the regions with preserved flow but also gives information about flow direction (antegrade or retrograde flow). Furthermore, a color-aliasing phenomenon may indicate the presence of raised flow velocities, such as those caused by a stenosis of at least medium grade. However, the main source of hemodynamic information is provided by the Doppler spectrum analysis, from which several parameters can be derived.

Blood Flow Velocity: Blood flow velocity values, i.e., the maximal systolic velocity (also referred to as peak velocity), maximal end-diastolic flow velocity, or the mean flow velocity, are derived from the Doppler spectrum (for further details see Chapter 3, “Cerebral Blood Flow Velocity”). Their assessment may reveal normal, raised, or reduced values. Vessel narrowing is directly correlated with increased flow velocities but this relation is not linear over the whole range of stenosis grades. In very high-grade stenosis and near-occlusion, flow velocity drops to normal or below normal values, as demonstrated in the Spencer’s curve (Fig. A5.34). According to this curve, which was developed to describe flow properties in a straight vessel without bifurcation and an axisymmetric stenosis, the initial increase in flow velocity compensates the lumen reduction to maintain a constant blood flow volume. With a stenosis of 70% (which correspond to an area reduction of 90%) or more, the blood flow volume drops despite a further increase of flow velocity. When the stenosis reaches ~85% the blood flow velocity also starts to fall. The same velocity may therefore be seen in a 60% stenosis as in a 90% stenosis (rising or falling shoulder of the curve). To assess the grade of stenosis correctly it is crucial to consider the pre- and poststenotic waveforms as well as the presence of collateral vessel activation. These indirect signs are of paramount importance for confident grading of stenoses. Note that the definition of a hemodynamically relevant stenosis refers to the decreased blood volume flow and not to an increased blood flow velocity; this term should therefore be used only for a stenosis of at least 70–80%.

The Spencer’s curve was mainly developed to predict arterial hemodynamics in a proximal ICA stenosis, and has several limitations. For instance, it applies to short and axisymmetric concentric stenoses, whereas patients with macroangiopathic lesions usually reveal irregular, eccentric stenoses of differing lengths. Atherosclerotic stenoses are usually short compared with stenoses caused by vessel dissection or arteritis. Long-segmented stenoses will be underestimated as high flow velocities are usually missed in such conditions. Also, Spencer’s model implies straight vessel walls resembling a channel: However, the major site of stenotic lesions in extracranial brain-supplying vessels is the proximal ICA which has a physiologic proximal widening, the carotid bulb. The carotid bulb is rarely absent but has a highly variable diameter ranging from normal to pseudo-aneurysmatic dilation doubling the normal ICA diameter (see also Chapter 2, “Extracranial Arterial Anatomy” under “General Arterial Anatomy”). Because of this physiologic dilation a nonlaminar flow, even with partly retrograde flow components, is the usual duplex ultrasound appearance and should not be misinterpreted. A bulb stenosis of ~30% may therefore not reduce the intraluminal diameter compared with that of the distal ICA segment and may even “normalize” the flow signal by showing flow that is now laminar. A 50% stenosis would also not alter waveform and velocities and a local 80% stenosis might not lead to any hemodynamic compromise. The carotid bulb therefore makes interpretation of velocity data somehow difficult. The good news is that all the other brain-supplying vessels follow the prediction of the Spencer’s curve better, as they normally do not have a bulb equivalent. According to the model a 50% stenosis should double blood flow velocities and a 70% stenosis should lead to a fourfold increase in velocity. Despite these and other shortcomings, sonographers should use the Spencer’s curve as the basic tool to understand hemodynamics and the consequences of brain-supplying artery stenosis.

Fig. A5.35 Increased flow in the VAs in bilateral ICA occlusion. Left: Ce-MRA, anterior MIP, showing bilateral ICA occlusion (arrowheads) and a bilaterally pronounced VA signal (arrows). Right: Duplex sonography, longitudinal plane, color-mode image and corresponding Doppler spectrum analysis: Turbulent flow and increased flow velocities in the right V2-VA, diameter 3.9 mm, flow velocity 93/44 cm/s (top) and in the left V2-VA, diameter 4.1 mm, flow velocity 108/48 cm/s (bottom). Identical flow patterns were observed in the intracranial VAs and BA. Note, the mild poststenotic and turbulent flow aspect results from the increased VA flow.

Fig. A5.36 Doppler spectra, obtained from follow-up transcranial insonation of the MCA in a patient with severe hypoxia. (A) Normal blood flow and flow velocity. (B) Hyperperfusion—note the increased flow velocities as well as the reduced pulsatility. (C) Hypoperfusion—note the reduced flow velocity as well as the increased pulsatility resembling an extracranial flow signal (“no brain, no flow”).

It is important to know that raised flow velocities are not limited to lumen reductions. For example, a global velocity increase may be observed in hemodilution and anemia to compensate for the loss of oxygen transporters, or in increased blood volume flow (hyperperfusion) which may be seen in the early phase following severe head trauma, in the subacute phase after subarachnoid hemorrhage, or in general hypoxia. In all these conditions flow velocities are globally increased. Segmental hyperperfusion and concomitant raised flow velocities are observed in extra- and intracranial collateral flow pathways in steno-occlusive disorders, e.g., in the anterior (ACoA) or posterior (PCoA) communicating arteries or extracranially in the neck vessels in occlusions of other major vessels. For example, in bilateral VA occlusion an increased blood flow and a corresponding increase in flow velocity is expected in one or both ICAs (see also Case 41). Contrarily and more evidently, in bilateral ICA occlusion raised blood flow and flow velocity may be observed in the extra- and intracranial segments of the VAs and the BA (Fig. A5.35; see also Case 12). Similar findings will be encountered in feeding vessels of arteriovenous angiomas or dural fistulas (see also Case 34 and Case 40). In all these cases the diastolic flow velocity will be disproportionately high, indicating a loss of peripheral resistance and dilation of the low-resistance vessels. Conversely, global low flow velocities can be observed if the hematocrit is high (Brass et al 1988), in severe cardiac output failure, in generalized dilated intracranial vessels—usually in long-standing arterial hypertension—or in the chronic state after severe head trauma or hypoxia. Regional flow velocity reductions may indicate hemodynamically relevant occlusive processes proximal or distal to the measurement site. All the above-mentioned conditions obviously hinder a simple interpretation of velocity ratios. A sophisticated analysis therefore needs a critical overview of all accessible direct and indirect ultrasound findings. Furthermore, the optimal time for insonation must also be considered. In the first hours of severe global cerebral hypoxia, a distinct reduction of cerebral blood flow and consequently of blood flow velocity occurs. Subsequently a reactive hyperemic phase can be observed, comprising generally increased flow velocities as well as reduced pulsatility. If the hypoxia leads to massive brain tissue necrosis, such as in persistent vegetative state, the chronic phase may reveal low flow velocities and a high pulsatility similar to the profiles seen in the ECA. A similar phenomenon occurs after large territorial infarctions in the involved arteries (Fig. A5.36, Fig. A5.37, Fig. A5.38).

Fig. A5.37 CCT and TCCS in chronic MCA infarction. Left: CCT scan, axial planes showing a residual almost complete territorial MCA infarction. Right: TCCS (transtemporal approach), right-sided insonation, midbrain plane, color-mode image and Doppler spectrum analysis of the M1-MCA: Reduced flow velocity of 58/15 and increased PI of 1.84 in the affected side (top) compared with the contralateral M1-MCA (insonated via the same side) with normal flow velocity of 68/28 mm/s and PI of 1.07 (bottom).

Fig. A5.38 MRI and TCCS in complete MCA infarction. (A) 3D TOFMRA, coronal view, indicating proximal MCA occlusion (arrow). (B) MRI, T2-weighted image, axial view. M1-MCA flow void with reduced diameter compared with the contralateral side (arrow) indicating residual flow. (C–E) TCCS, transtemporal approach, axial midbrain plane, color-mode and corresponding Doppler spectra. (C) Normal PRF and color gain TCCS settings easily depicting the main stems of the contralateral MCA, both ACAs, and PCAs. Ipsilateral proximal MCA vessel sheath is visible within the lateral fissure but M1-MCA is missed (arrow), wrongly suggesting occlusion. (D,E) Reduced PRF and increased color gain result in aliasing of color signals (“dirty image”) now showing a proximal MCA color signal (arrow) and a flow signal with low flow velocities (15/6 cm/s).

Blood flow velocity measurement in the brain-supplying arteries crucially depends on the angle of correction as this greatly affects the final velocity measurement (for further details see Chapter 1, “Doppler Shift and Flow Velocity” under “Ultrasound Principles,” and Chapter 3, “Cerebral Blood Flow Velocity” under “Parameters of Cerebral Hemodynamics”). In nonpathologic vessels angle correction should preferentially be done parallel to the vessel walls, which usually corresponds to the flow stream visualized in the color mode. In a stenosed vessel the flow may differ from the anatomic course of the vessel due to eccentric stenoses. Here, the angle should be corrected according to the flow stream (also called the “jet”). A Doppler angle greater than 60° should be avoided as it may lead to velocity overestimation. This problem particularly occurs in V2-VA insonation as the vessel course is practically at a 90° angle in relation to the ultrasound beam, which already causes problems in the color-mode visualization of flow. We recommend improving the insonation angle by mild angulation of the probe into the soft neck tissue. Intracranially, nonoptimal insonation angles occur during the insonation of the A2-ACA, PCoA, and the transitional zone between the proximal and the distal P2-PCA, as well as during coronal transtemporal insonation of distal ICA and BA.

Velocity Ratios: As discussed earlier, flow velocity measurements give only surrogate information about blood flow. Under physiologic conditions cerebral blood flow and flow velocities are well correlated. In pathologic conditions, high velocities may still reflect normal blood flow (e.g., in nonhemodynamically relevant stenosis), decreased blood flow (e.g., in hemodynamically relevant stenosis), and even increased blood flow (e.g., in hyperperfusion). High velocities might even be caused by a mixture of stenosis and hyperperfusion which may be observed in the subacute phase after subarachnoid hemorrhage (SAH) (simultaneous presence of hyperperfusion and vasospasm). In addition to pure velocity measurements, velocity ratios comparing homologous vessels on both sides or intra- and prestenotic or intra- and post-stenotic signals on the ipsilateral side may be of help.

ICA/CCA Index: The index is calculated from the maximal systolic flow velocity within the ICA stenosis (V_{ICA syst stenosis}) and the maximal systolic flow velocity within the ipsilateral CCA (V_{CCA syst}). It is a parameter that is independent of general blood flow alterations, but it only works if the CCA is itself not affected by atherosclerotic vessel wall changes.

ICA/ICA Index: The index is calculated from the maximal systolic flow velocity within the ICA stenosis (V_{ICA syst stenosis}) and the maximal systolic flow velocity of the contralateral (unaffected) ICA (V_{ICA syst contralateral}).

Fig. A5.39 Top: Schematic of flow pattern in a normal and in an increasingly stenosed blood vessel. Bottom: Corresponding Doppler spectra: (A) Preserved systolic window in the unaffected vessel (arrow); (B) Turbulent flow with spectral broadening and disappearing systolic window but still normal velocity; (C) Turbulent flow with increased velocity indicating a higher grade of stenosis; (D) Severe turbulent flow with retrograde flow components and high flow velocity.

Fig. A5.40 Musical murmurs. (A) TCD Doppler spectrum of a high-grade M1-MCA stenosis (peak systolic flow velocity >300 cm/s). Note the mirror-image parallel strings as the visual correlate of a musical murmur. (B–D) TCCS Color-mode images and Doppler spectra: (B) Transtemporal approach, midbrain plane showing a high-grade P1-PCA stenosis; (C) Transtemporal approach, midbrain plane demonstrates a functional stenosis at the ACoA; (D) Transtemporal approach, posterior coronal plane showing a high-grade basilar artery stenosis.

It only works if the contralateral ICA shows normal flow profiles. Flow velocity measurement should not be performed in the carotid bulb but in a straight segment of the unaltered ICA. Because of this limitation the ICA/ICA index is not commonly used. The angle-corrected intrastenotic flow velocities of the ICA may be compared with the flow velocities of the distal ICA. To get reliable values the poststenotic flow signals should not be disturbed and a sufficient length of the distal ICA must be visualized to use angle correction. This index resembles the Lindegaard Index (LI) which is widely used after SAH (see also Case 33). As with the LI, a high ICA/ICA index indicates a more severe stenosis (Alexandrov 2013).

Flow Profile Alterations—Spectral Broadening: Doppler spectrum analysis of normal blood flow classically reveals a laminar flow characterized by a systolic window, which means that the highest velocity is in the center of the vessel and the lowest at its wall, best seen in the systolic phase as an “empty” area below the systolic peak. Turbulent flow is observed when blood starts to form eddy currents. Turbulence starts early but is usually seen as a progressive disappearance of the systolic window at a local diameter reduction of ~50%. Turbulence may also be present without pathologic meaning near sharp changes of flow direction, e.g., in vessel bifurcations and loops. Because of their short lengths and smaller sizes this is more often observed in the intracranial arteries and especially in the carotid siphon (Fig. A5.39). Experienced ultrasonographers can also identify turbulence by hearing a disturbed audio signal. In very high-grade stenosis a harmonic phenomenon, the so-called musical murmur, can be observed. Acoustically it resembles a bird call and is therefore also frequently called the “seagull cry” or “goose cry.” In the Doppler spectrum, mirror-image parallel strings or bands can be observed (see Fig. A5.40 and Video A5.7). The phenomenon presumably results from harmonic frequencies, generated from regular vibrations of the vessel walls caused by the increased blood flow velocities. Musical murmurs are usually observed in intracranial stenosis. A recent study reporting on 66 musical murmurs found 94% of murmurs occurring in intracranial vessels and 6% in extracranial vessels (Lin et al 2006). In 88% of cases a high-grade stenosis was detected. In the remaining cases, the musical murmur was found mainly in the communicating intracranial arteries. Here the musical murmur indicates a “functional stenosis,” when blood flow is too high for the size of the ACoA or PCoA. As a rule of thumb it can be postulated that whenever a musical murmur is detected in one of the communicating arteries, even if the maximal flow velocities are not clearly increased, a proximal steno-occlusive process has to be present.

Spectral broadening and musical murmurs, however, are additional and not exclusive criteria for stenosis. They depend on the grade as well as the configuration of the stenosis and are not mandatory.

Indirect Hemodynamic Assessment

In any case of a suspected or known stenosis, not only the intrastenotic flow signal but also the waveforms from vessel segments proximal and more importantly distal to a stenosis (prestenotic and poststenotic flow pattern) must be analyzed. This distinguishes between stenoses with or without a hemodynamic effect (Fig. A5.41 and Fig. A5.42). If the poststenotic flow of an ICA stenosis is not visualized extracranially, the C6-ICA and OA may be analyzed for validation of its hemodynamic relevance (Fig. A5.43). By definition, stenoses are hemodynamically relevant if they cause a reduced blood volume flow and poststenotic pressure drop. According to the Spencer’s curve discussed earlier, this occurs if the diameter is decreased by more than 70–80% or the cross-sectional area is reduced by 90–95%. Archie and Feldtman (1981) found similar results, suggesting that relevant blood flow reduction of 40% begins at 75% diameter stenosis or 94% area stenosis. Considering waveform appearance together with flow velocities will therefore help to recognize a hemodynamically relevant stenosis.

Fig. A5.41 Left: Contrast-enhanced MRA, coronal MIP showing a severe extracranial hemodynamically relevant high-grade 90% ICA stenosis. Note the signal void at the stenosis and the collapse of the poststenotic segment of the ICA due to the drop of pressure and reduced volume flow. Right: Extracranial duplex, longitudinal view. Doppler spectrum analysis in post-stenotic ICA, flow velocity 26/5 cm/s (top), intrastenotic ICA, flow velocity 360/105 cm/s (middle) and prestenotic CCA, flow velocity 38/12 cm/s (bottom). Note the increased pulsatility in the prestenotic CCA and the distinctly delayed systolic flow acceleration in the poststenotic ICA.

Fig. A5.42 Extracranial duplex of a hemodynamically relevant ICA stenosis. Longitudinal view, color-mode image, and Doppler spectrum analysis. Top: Shortly after the stenosis (arrow) a serrated flow signal indicates relevance of the stenosis. The acceleration time (AT) is normal: 50 milliseconds. Bottom: Insonation further distal reveals an obvious poststenotic flow pattern with prolonged AT (120 milliseconds). For further reading on the AT see below.

Fig. A5.43 (A) Extracranial duplex, longitudinal view. Doppler spectrum analysis showing a proximal ICA stenosis (flow velocity: 414/181 cm/s). (B) TCCS, color-mode image and corresponding Doppler spectrum analysis, transtemporal axial insonation, lower pontine plane. Poststenotic flow pattern and reduced flow velocity in the C6-ICA (24/13 cm/s). (C) TCCS, transorbital approach, color-mode image and corresponding Doppler spectrum analysis revealing an OA flow signal similar to C6-ICA (15/6 cm/s).

Fig. A5.44 Flow pattern distal to a hemodynamically relevant VA stenosis. (A) Increased flow velocity in a distal V1-VA stenosis (257/46 cm/s). (B) Turbulent flow in the proximal V2-VA between C5 and C6 (69/17 cm/s). (C) Mild poststenotic flow pattern in the mid V2-VA between C3 and C4 (108/34 cm/s, AT 136 milliseconds). (D) Apparent poststenotic flow pattern in V3-VA (53/23 cm/s, AT 200 milliseconds).

Fig. A5.45 TCCS, Doppler spectrum analysis of normal M1-MCA (1) and moderate (2,3) to severe (4) and pseudo-venous (5) poststenotic flow pattern.

Fig. A5.46 TCCS, Doppler spectrum analysis of normal M1-MCA and M1-MCA of a patient with an ipsilateral ICA occlusion revealing typical signs of a poststenotic flow pattern. Top: Reduced blood flow velocity 42/24 cm/s versus 137/44 cm/s. Middle: Prolonged acceleration time (AT) 120 milliseconds versus 56 milliseconds. Bottom: Reduced pulsatility index (PI 0.59 versus 0.88).

Note that the terminology for pre- and poststenotic flow signals is not well standardized. In this book, we use the terms pre- and poststenotic flow pattern for detectable proximal or distal flow signal alterations in stenoses and occlusions.

The prestenotic flow pattern comprises a mostly normal systolic flow velocity and always a normal rise in systolic flow. However, distal flow obstruction leads to raised peripheral resistance, reduced prestenotic diastolic flow, and subsequently raised pulsatility. In cases with an unclear cause of the distal flow obstruction (occlusion or stenosis) but highly pulsatile flow signals, the term “high-resistance flow pattern with increased pulsatility” might be more appropriate.

A poststenotic flow pattern requires a relevant proximal obstruction of at least 70–80%, which then leads to the phenomenon of delayed systolic acceleration. For assessment of a poststenotic flow pattern, the insonation has to be performed at a certain distance from the stenosis to avoid signal artifacts caused by severe turbulence. The poststenotic flow may also become more obvious in the run of a vessel with a long course. This can also be observed in proximal VA stenoses when the V2 segment reveals no obvious delayed systolic increase, but this becomes evident in its V3 or V4 segments (Fig. A5.44). A reduced blood flow results in a compensatory dilatation of the resistance vessels to avoid downstream hypoperfusion. As a consequence the total arterial cross-sectional area increases, with subsequent reduction of the peripheral resistance causing a raised diastolic flow component and reduced pulsatility. For these flow patterns, terms such as “blunted flow,” “low-resistance flow,” “tardus parvus waveform” or, in case of distinct alterations, “venous-like flow” can be used (Fig. A5.45). The delayed systolic acceleration (delayed upstroke/upslope) as integral part of the poststenotic flow pattern can be quantified by the acceleration time (AT), measured from the onset of systole to the first peak systole. The AT is widely used in renal artery stenosis to diagnose hemodynamically significant diseases, with 70 ms generally considered as the cut-off. An AT of >100 ms has a sensitivity of 32% and a specificity of 100% in diagnosing a hemodynamically relevant renal stenosis (Motew et al 2000). AT analysis may also be of interest in neurosonology, but here its use is not yet well established. In clinical practice a visual comparison of homologous vessel segments appears to be adequate.

Besides delayed systolic flow acceleration and a concomitantly increased diastolic flow, reduced flow velocities are usually present as signs of a hemodynamically compromised poststenotic flow (Fig. A5.46). Reduced or normal flow velocities help to distinguish real poststenotic flow from similar patterns which may be seen in hyperperfused vessels like feeders of dural arteriovenous fistulas and arteriovenous malformations (AVMs) or otherwise normal vessels that serve as collaterals in major vessel occlusions (see Fig. A5.35). General low-resistance flow patterns in all brain-supplying arteries may be also seen in severe aortic valve stenosis.

To obtain the greatest diagnostic certainty in everyday clinical practice, we recommend that signals in an assumed stenosis should be obtained from all three vessel segments (prestenotic, intrastenotic, and poststenotic) whenever possible. The above criteria are of special significance when performing serial measurements over long periods to detect disease progression. For instance, if initially absent indirect hemodynamic criteria develop over time, an increase of the stenosis to a range of at least 80% is likely. Assessment of only intrastenotic flow velocities alone is problematic, as velocities may even decrease with increasing grade of stenosis (see the Spencer’s curve, Fig. A5.34).

Fig. A5.47 Left: Contrast-enhanced MRA, coronal MIP showing an ICA stenosis. The ICA was considered to reveal only wall irregularities. Top right: Extracranial duplex, color-mode image, cross-sectional plane: Area reduction according to a 65% stenosis (left), diameter reduction according to a 50% stenosis (right). Bottom right: Color-mode image and corresponding Doppler spectrum analysis revealing a nonturbulent flow with normal velocities of 72/22 cm/s.

Fig. A5.48 Distal ICA stenosis. Color-mode, longitudinal plane, composed image. Aliasing and lumen reduction is visible ~2 cm distal of the ICA bifurcation (arrow) indicating ICA stenosis. Doppler spectrum reveals increased velocities (top left: 265/87 cm/s). The proximal ICA itself showed low velocities (bottom right: 18/6 cm/s).

Ultrasound Criteria of Occlusions

Direct Morphologic Assessment

Extracranial B-mode duplex ultrasound may reveal complete filling of the vessel lumen with thrombotic material of varying echogenicity (see Video A5.8). In chronic occlusion, precise vessel identification and differentiation of the vessel lumen might be difficult. A fresh thrombotic occlusion usually presents with hypoechoic thrombotic material. In contrast to a recent occlusion, a chronic occlusion may demonstrate a reduction or loss of vessel distensibility which defines variation of diameter during the systolic and diastolic phase (Alexandrov 2013). B-mode insonation alone, however, is not sufficient for diagnosis of occlusion. It should always be combined with color-mode and Doppler spectrum analysis.

Direct Hemodynamic Assessment

Occlusions result in a complete absence of color-flow signal, even after adjustment for very low flow signals (lowest PRF and increased color gain settings). Doppler spectrum analysis reveals no flow signal. In cases with a proximal vessel stump, a distinctly reduced, alternating flow pattern with a short systolic peak and a small retrograde flow component (“stump signal” or “to-and-fro signal”) can be found. Diagnostic certainty may be increased by using intravenously administered ultrasound contrast agents. On transcranial insonation, a missing flow signal does not necessarily imply occlusion. For example, the P1-PCA segment might be absent due to P1 hypo- or aplasia in case of a fetal-type PCA. The A1-ACA segment might also be missing in distinct hypo- or aplasia. In these circumstances, indirect hemodynamic criteria might help to distinguish normal anatomic variants from pathologic findings.

Indirect Hemodynamic Assessment

In occlusion, the same criteria as for hemodynamically relevant high-grade stenoses can be applied for the analysis of flow profiles proximal and distal to the occlusion (see also “Ultrasound Criteria of Stenoses” above). Because of the ability of collateral vessels to bypass occlusions, their detection and consideration of their capacity is of paramount importance in the acute and chronic state after stroke (for further details see also “Collateral Pathways” below).

Extracranial Pathology

Extracranial Anterior Circulation

ICA Stenosis

Auscultation with the stethoscope can be considered as an inadequate screening method as it detects only 25% of ICA stenoses and has a high number of false-positive findings: 10% are not confirmed by conventional angiography (Ziegler et al 1971).

Since a variety of alternative invasive, less invasive, and noninvasive imaging methods are available that permit visualization of vessel pathology in vivo, the evaluation of ICA pathology and grading of ICA stenosis has been and remains a matter of extensive debate. A special anatomic variant of the ICA which hinders a simple interpretation of stenoses in each modality, unlike other brain-supplying arteries, is the variable widening of the carotid bulb. ICA stenoses of atherosclerotic origin are mainly located just at this site, which is explained by the nonlinear flow at the bifurcation. In methods primarily outlining the intraluminal patency a relevant bulb stenosis may be underestimated. In particular, homogeneous noncalcified plaques may have a smooth surface and may be completely overlooked in catheter or MR angiography (Fig. A5.47). Precise analysis of vessel lumen reduction within this region is therefore a challenge for all currently available imaging methods. ICA dissections usually alter the ICA in its more distal parts but atherosclerotic lesions may also affect the ICA at more atypical distal sites (Fig. A5.48).

Grading of ICA Stenosis by Digital Subtraction Angiography Evaluation of stenoses and occlusions has so far been dominated by the first available method: conventional angiography. Several important clinical trials have been based on angiographic data, the results of which form the basis for current treatment decisions in carotid stenosis. The European Carotid Surgery Trial (ECST Collaborative Group 1991) and the North American Symptomatic Carotid Endarterectomy Trial (NASCET Collaborators 1991) compared medical treatment and carotid endarterectomy (CEA) in patients with different grades of symptomatic ICA stenosis. They found that patients with stenoses between 70% and 95% significantly benefit from the surgical intervention. However, the two studies used different approaches to determine the grade of stenosis. The NASCET study used the diameter of the unaffected distal ICA and the narrowest stenosis diameter for calculation of stenosis (distal grade of stenosis). The ECST used the stenosis diameter and the estimated diameter of the nonvisualized outer walls of the ICA stenosis (local grade of stenosis). A third method, defining the grade of stenosis between the stenosis diameter and the proximal unaffected CCA (CC method) has not yet been used in a large clinical trial (de Bray and Glatt 1995) (Fig. A5.49). Currently treatment decisions on whether or not to perform CEA rely on the NASCET and ECST data. Although numerically identical, a 70% NASCET ICA stenosis is not equal to a 70% ECST ICA stenosis. Rothwell and co workers compared NASCET and ECST grades of stenosis and found a linear correlation which allows an estimated conversion between the two approaches (Rothwell et al 1994):

Fig. A5.49 Schematic of an ICA stenosis, similar to the images derived from DSA. Illustration of three angiographic methods to determine the grade of ICA stenosis. Note that the ECST method uses an “eyeball” estimate of the nonvisible outer wall of the ICA at the carotid bulb (dashed gray line). Comparative analysis of the three methods yielded a linear relationship, allowing an estimated conversion.

Fig. A5.50 DSA, selective CCA injection, lateral view. (A–C) Potential error in ICA stenosis estimation using the ECST method. Different estimates of the presumed carotid sinus will result in a calculated grade of stenosis increasing from A to C. Diameter of stenosis: 3.3 mm. Diameter of the carotid bulb, assessed by duplex ultrasound: 9.2 mm. Resulting degree of stenosis: 64%, best corresponding to the estimate in C.

The same relationship can be applied to the CC criteria:

Following this conversion, a 70% ECST ICA stenosis is equivalent to a 50% NASCET stenosis. Despite this correlation and the positive findings in the above two clinical studies, both approaches have considerable methodological problems (Alexandrov et al 1993). The NASCET approach is unable to account for low-grade stenosis as a 40% ECST stenosis equals 0% NASCET, and, for example, a 30% ECST stenosis translates to an absurd –17% NASCET stenosis. The ECST approach, on the other hand, relies on an “eyeball” estimation of the presumed carotid bulb diameter, which has the potential for considerable error (Fig. A5.50 and Fig. A5.51). To reduce this error some authors use the carotid stenosis index, which is largely based on a publication by Williams and Nicolaides, who found a fixed carotid bulb to proximal CCA ratio of 1.2 in 96% of 61 angiograms of presumably normal carotid bifurcations (Williams and Nicolaides 1987). Use of this ratio certainly improved the comparability between both angio-graphic methods but did not improve their diagnostic accuracy. In fact other studies found an ICA/CCA ratio ranging from 0.7 to 1.4 (Rothwell et al 1994). A CTA approach has questioned a fixed ICA/CCA ratio (Bartlett et al 2007).

Considering these shortcomings, it is surprising that DSA has so far remained the diagnostic gold standard with which all other methods have to compete. The relevance of extended meta-analyses which try to analyze sensitivity and specificity values for the less invasive methods (duplex ultrasound, MRA, CTA) in comparison to DSA (Patel et al 2002, Wardlaw et al 2006a) is also questionable. As with all other nonangiographic methods, the duplex ultrasound findings have to be “imported” into the angio graphic scales. Continuous-wave Doppler sonography, in contrast, can no longer be recommended because of its low diagnostic accuracy. For instance, in a comparative Doppler and angiographic study that aimed to identify patients with ICA stenoses >60% (NASCET criteria), the Doppler technique yielded 41% false-positive results (Qureshi et al 2001). For further details, see Case 1.

Grading of ICA Stenosis by Duplex Ultrasonography
Duplex ultrasonography allows the grading of ICA stenosis according to the ECST or NASCET criteria, so ultrasound reports should always specify the classification system used. The ultrasound technique works at its best if the local grade of stenosis is assessed, i.e., if the ECST grades are measured by direct analysis of cross-sectional area or lumen reduction. The most confident results can be achieved in low and moderate stenoses with up to 50% lumen reduction (Fig. A5.52). The distal ICA lumen, required for NASCET grading, is usually difficult to assess. However, the NASCET criteria have become the current base for clinical decision-making worldwide. The Neurosonology Research Group of the World Federation of Neurology (NSRG) has therefore proposed adapted duplex criteria, which are able to define stenoses according to the NASCET system by using a multiparametric approach (von Reutern et al 2012). With increasing grade of stenosis there is a shift of importance from B-mode imaging to velocity measurements and finally hemodynamic parameters. To achieve this goal, all morphologic as well as direct (flow velocities) and indirect hemodynamic information (collateral flow, poststenotic flow pattern) needs to be considered by also including orbital and intracranial flow parameters. In the following, we present our recommendations for the grading of ICA stenoses which are mainly based on the NSRG criteria.

Fig. A5.51 Extracranial duplex corresponding to the DSA in Fig. A5.50. (A) B-mode image, longitudinal plane: Large, mildly hypoechoic structure in the carotid bulb as well as at the ECA origin. (B) Color-mode, cross-sectional plane: Following the ECST criteria, the diameter of the vessel (9.2 mm) and the residual lumen (3.3 mm) are assessed, resulting in a 64% stenosis. (C) Color-mode, longitudinal plane: Clear delineation of the remaining perfused lumen and confirmation of the plaque extension. (D) Doppler spectrum analysis. Flow velocity 185/102 cm/s, indicating local stenosis, grade 60–70%.

Fig. A5.52 Extracranial duplex, color-mode image, cross-sectional plane, ICA bulb stenosis. (A,B) ICA bulb with a diameter reduction from 10.8 mm to 4.6 mm leading to a 57% local stenosis (A) and an area reduction from 32.7 mm² to 15.9 mm² which results in a 51% stenosis (B). (C,D) Distal ICA on the same side with a diameter of 6 mm and an area of 28.9 mm² resulting in a stenosis according to NASCET of 23%.

B-mode and color-mode imaging, especially in the cross-sectional plane, are the most important parameters in low-grade bulb stenoses <50% as usually no obvious velocity increase is present (see Fig. A5.11). Here we recommend describing the lumen reduction in local grades and to start diagnosing a stenosis from a 30% narrowing onwards, stepwise in 10% steps up to a 50% stenosis. The calculated stenosis according to the NASCET criteria may additionally be reported. In a local stenosis between 50% and 60% (NASCET 15–35%), the main criterion is a velocity increase without indirect signs of collateral flow or of poststenotic flow pattern. In local stenoses of more than 80% (NASCET 70%), indirect hemodynamic criteria are the dominant parameters demonstrating a collateral flow via alternative pathways (retrograde ophthalmic flow, cross-flow via ACoA or PCoA or leptomeningeal collateral flow), and/or a poststenotic flow pattern characterized by a prolonged AT and/or decreased pulsatility index (PI) in the distal ICA, and/or a prestenotic flow pattern in the CCA with increased pulsatility. A comparison with the unaffected side by eyeball estimate is necessary to prove a pre- or poststenotic flow pattern. The ICA/CCA, ICA_ipsilateral/ICA_{contralateral} or ICA_{intrastenotic}/ICA_poststenotic indices may be useful as second-line criteria. However, these are less frequently used in our daily routine and their importance should not be overestimated. This means, in case of conflicting findings, first-line criteria (one or more signs of a hemodynamic compromise, such as collaterals or a poststenotic flow pattern) overrule second-line criteria such as the above-mentioned ratios. If a collateral flow is clearly documented, a hemodynamically relevant stenosis is proven irrespective of the measured intrastenotic flow velocity.

For exact grading, the highest velocity within a stenosis should be identified. Depending on the configuration of the stenosis this point can be at its origin (most frequent location) but also more distal, as often seen in case of dissections. In most cases a direct assessment of the highest velocities is possible; however, evaluation may be hindered or may even be impossible if severe plaque calcification and subsequent acoustic shadowing are present. The latter is found in up to 7% of cases (Polak et al 1989) (Fig. A5.53). A deeply located vessel or an angulated vessel course may also impair the visualization of a stenosis. Velocity measurements should be performed with the lowest possible angle of correction (see also Chapter 1, “Doppler Shift and Flow Velocity” under “Ultrasound Principles”), following the jet of the flow and not the anatomic course of the affected vessel (Fig. A5.54). If present, prestenotic flow alterations in the distal CCA are easily detectable (Fig. A5.55). Extracranial measurements distal to the stenosis are often hindered if the bifurcation is near the mandible or if the stenosis extends over a long segment distance. In some of these cases, the distal extra cranial ICA can be assessed in the axial plane. Unfortunately in this plane the angle of insonation is not well defined and reliable velocity measurement cannot be performed. However, spectrum analysis may still be sufficient to identify a poststenotic flow pattern and sometimes there are also perivascular color artifacts, mainly during systole, surrounding the stenotic vessel (the “confetti effect;” see also Fig. A1.45). Alternatively, the flow pattern of the intracranial ICA below the communicating arteries, preferably at the C6-ICA segment (Fig. A5.56, top), or the OA if anterograde and not activated as a collateral vessel itself (Fig. A5.57) may show a poststenotic flow pattern, illustrating the hemo dynamic relevance of an extracranial ICA stenosis. The ipsilateral MCA and ACA may also be analyzed. However, in their assessment it has to be considered that collateral filling might have already occurred via ACoA, PCoA, and/or OA, and the observed MCA and ACA waveforms might no longer reveal the poststenotic ICA flow pattern (Fig. A5.56, bottom). In the case of a collateral flow via the PCoA, the ACA might show a more distinct poststenotic flow pattern compared with the MCA which can be best explained by a functioning ACoA and a patent contralateral A1-ACA ensuring flow to both ACA territories (Fig. A5.58). A synopsis of recommended duplex ultrasound criteria for grading of a proximal ICA stenosis considering the multipara-metric approach of morphology, direct and indirect flow parameters based on the NSRG criteria (von Reutern et al 2012) is given in Fig. A5.59. Particularly important are the threshold values which define an ICA stenosis >50% according to the NASCET criteria. The NSRG recommended using the relatively low USA threshold values (see below). However, they also added an analysis of data from five DSA-correlated studies including 977 stenoses, which together revealed higher average peak systolic flow velocities. Interestingly, the latter are closely equivalent to the threshold values usually used in European countries. It seems therefore more reasonable to use these average velocities as threshold values for grading of ICA stenosis in addition to the indispensable indirect hemodynamic marker. The 90% grade according to NASCET (95% according to ECST) represents the specific findings in near-occlusion. Ultrasound reports should always state the classification system used. This approach, including a grading in 10% steps, is used in many European centers and is well accepted as it has been shown to be a reliable approach if applied by experienced sonographers (Dippel et al 1997). In North America the intracranial and intraorbital flow parameters are normally not analyzed and therefore the grading of ICA stenosis is mainly based on intrastenotic velocity parameters which have been adapted to the rough NASCET grading system (Alexandrov and Needle-man 2012). According to a consensus report the following criteria were proposed (Grant et al 2003):

Fig. A5.53 Extracranial duplex, color-mode image, longitudinal view. (A,B) Examples of a distinct acoustic shadowing phenomenon caused by calcified ICA plaques impeding assessment of ICA flow. Shortly distally a flow signal can be detected (B).

Fig. A5.54 Extracranial duplex, color-mode image, longitudinal plane. Image of a high-grade ICA stenosis with lumen reduction and color-aliasing at the stenosis. Direction of the flow stream (“jet”) and vessel course is not equivalent, complicating the exact placement of the angle correction. (A) Inappropriate angle correction placement following the vessel course (398/197 cm/s). (B) Preferred placement following the flow jet within the stenosis resulting in a different velocity (270/128 cm/s).

Fig. A5.55 Assessment of pre- and poststenotic blood flow in the CCA, proximal ICA, and distal ICA. Left: Doppler spectra in unaffected vessels. Right: Flow patterns in a >80% ICA stenosis (ECST criteria): Prestenotic flow pattern in the CCA with reduced flow velocity and increased pulsatility (bottom), intrastenotic elevated flow velocity and spectral broadening (middle), poststenotic flow pattern in the distal ICA with reduced flow velocity, pulsatility and delayed systolic flow acceleration (top). Note the distal ICA collapse.

Fig. A5.56 Downstream flow patterns in hemodynamically relevant high-grade ICA stenosis. TCCS, color-mode image and corresponding Doppler spectrum analysis, transtemporal approach. (A) Lower pontine plane, C6-ICA: pronounced poststenotic flow pattern (flow velocity 21/10 cm/s, PI 0.6 and a markedly prolonged AT). (B) Midbrain plane, M1-MCA: mild poststenotic flow pattern only (flow velocity 63/25 cm/s, PI 1.1 and mildly increased AT) because of collateral filling via the ACoA.

Fig. A5.57 Downstream flow patterns in hemodynamically relevant high-grade ICA stenosis. TCCS, color-mode image and corresponding Doppler spectra. (A) Transorbital approach: antegrade OA flow with mild poststenotic flow pattern (flow velocity 24/10 cm/s). (B) Transtemporal approach, lower pontine plane: similar flow pattern in the C6-ICA (flow velocity 20/10 cm/s).

Fig. A5.58 Downstream flow patterns in hemodynamically relevant high-grade ICA stenosis. TCCS, color-mode image and corresponding Doppler spectra. (A) Transtemporal approach, anterior coronal plane: M1-MCA with almost normal flow pattern. (B) In contrast, a mild poststenotic flow pattern is present in the A1-ACA. The best explanation is that the contralateral A1-ACA takes care of both ACA territories and the affected MCA has more need of the remaining blood flow from the ICA.

<50% stenoses (peak systolic velocity <125 cm/s, ICA/ CCA ratio <2.0, end-diastolic velocity <40 cm/s, <50% diameter reduction).
50–69% stenoses (peak systolic velocity 125–230 cm/s, ICA/CCA ratio 2.0–4.0, end-diastolic velocity 40– 100 cm/s, ≥50% diameter reduction).
Stenoses >70%–near-occlusion (peak systolic velocity >230 cm/s, ICA/CCA ratio >4.0, end-diastolic velocity >100 cm/s, ≥50% diameter reduction).
Near-occlusion with variable velocities and ICA/CCA ratio and occlusion.

An adapted “sonographic NASCET index” has since been proposed, which also took the distal ICA flow information into account and which consequently yielded better correlation with angiographic findings than the conventional peak systolic flow velocity measurements alone (Hathout et al 2005). Nevertheless, the American Heart Association decided to recommend duplex ultrasound as a screening tool only and not as a possible sole examination technique before ICA intervention because of divergent published studies and its low sensitivity and specificity compared with other imaging modalities (Latchaw et al 2009).

Despite the above well-defined grading criteria, they still have to be applied with caution considering other potentially modifying aspects. The given cut-off values for blood flow velocities might not be applicable in patients with generally altered cerebral perfusion, e.g., in severe hyperemia, in young subjects who in general show higher blood flow velocities of the brain-supplying arteries, or in elderly and hypertensive patients who have often generalized dilated vessels and subsequent lower flow velocities. Furthermore, a contralateral high-grade ICA stenosis or ICA occlusion may lead to an ipsilateral compensatory raised ICA flow and flow velocity with subsequent overestimation of the stenosis (Henderson et al 2000). Vice versa, endarterectomy of a contralateral high-grade ICA stenosis often results in an ipsilateral decrease of intrastenotic blood flow velocity. Following stent treatment, the mean observed drop of contralateral peak systolic velocity was reported to be 60.3 cm/s. In 71% of patients with the initial flow velocity-derived diagnosis of significant contralateral stenosis, interventional angiography refuted the presence of any significant contralateral stenosis (Sachar et al 2004). In a CEA study, the intervention led to a contralateral decrease of blood flow velocity in 52% of cases with an average drop of 20% duplex-defined stenosis grade (Busuttil et al 1996). It can be assumed that such effects will only occur if the untreated ICA serves as a collateral vessel supplying blood via the ACoA before intervention. Interestingly, none of these studies analyzed the presence or absence of intracranial collaterals, which probably explains why this phenomenon was not observed in all of the patients.

Fig. A5.59 Ultrasound grading of ICA stenosis at its origin according to the NASCET and ECST criteria. Flow velocities are given in cm/s. Occl. = occlusion; diameter = residual diameter assessed in cross-sectional plane; turbulence = aliasing in color-mode and/or systolic broadening in Doppler spectrum analysis; V_{ICA syst} = peak systolic flow velocity within the stenosis; V_{ICA diast} = end-diastolic flow velocity within the stenosis; V_{ICA syst post} = poststenotic ICA peak systolic flow velocity; ICA/CCA = peak systolic velocity ICA/CCA ratio; prestenot. CCA = prestenotic flow pattern = reduced flow velocity and increased pulsatility in CCA; poststenot. ICA_dist = poststenotic flow pattern = reduced flow velocity, delayed systolic flow acceleration and/or reduced pulsatility; poststenot. C6 = poststenotic flow pattern = reduced flow velocity, delayed systolic flow acceleration and/or reduced pulsatility in C6-ICA; OA collateral = ↑↓: OA flow may be normal, absent, or retrograde; IC collaterals = intracranial collateral activation via communicating arteries: retrograde ipsilateral A1-ACA, raised flow velocities in ipsilateral P1-PCA and PCoA and/or via leptomeningeal arteries: raised flow velocities in ipsilateral P2-PCA and P3-PCA branches in variable combinations. ↑ ↓ = variable.

Another potential pitfall to be considered is the occurrence of tandem stenoses, i.e., the simultaneous presence of an extra- and intracranial ICA stenosis. In such cases flow velocity and therefore also the stenosis grade may be underestimated. A tandem stenosis should be considered if indirect signs of a hemodynamically relevant distal flow obstruction are present, which cannot solely be explained by the findings of the extracranial ICA stenosis. Finally, it has to be mentioned that the stenosis categories are based on the presence of an isolated, short, and circumscribed ICA stenosis. In stenoses more than 2 cm in length, such as are often seen in dissection, flow velocities lower than expected might be observed due to the increased flow resistance.

A near-occlusion may be difficult to be distinguished from a vessel occlusion, especially if the site of the vessel obstruction is slightly distal to an otherwise not severely affected bulb. In these circumstances the bulb may reveal a stump signal, but the following ICA segment may present a turbulent flow pattern with variable flow velocities. A marked poststenotic waveform in the distal extracranial or intracranial petrosal ICA segment assures a near-occlusion of the ICA. Its clinical significance is limited and mainly of interest in patients with enduring insufficient collateral function in whom opening of the near-occlusion may be indicated. In suspected proximal ICA occlusion sonographers should follow the ICA from its origin to the more distal segments, reduce the PRF and the size of the color box, and increase the color gain in both the longitudinal and transverse planes in order not to overlook near-occlusion (Fig. A5.60). For further details, see Case 1 and Case 11. For details of ICA near-occlusion, see Case 15 and Case 18.

Graduation of Stenosis after Stenting
Carotid artery stenting is a well-established alternative to carotid endarterectomy. Postinterventional duplex sonography allows assessing stent placement as well as comorbidities like dissection or residual or reoccurring stenosis (Fig. A5.61). Velocity criteria for in-stent restenosis evaluation are still not well established. Comparing carotid angiograms with duplex ultrasound for luminal stenosis, increasing peak systolic velocities and ICA/CCA ratios correlate with evolving restenosis within the stented carotid artery. However, application of velocity criteria for nonstented arteries leads to an overestimation of the stenosis grade. Analysis of flow after stent placement in nonstenotic vessel segments found a 22% increase of in-stent peak systolic flow velocity but no significant increase of end-diastolic velocity (Hakimi et al 2012). The following ICA in-stent restenosis threshold criteria for peak systolic flow velocity and ICA/CCA ratio have been proposed (Lal et al 2008): ≥20% (PSV ≥150 cm/s and ICA/CCA ratio ≥2.15); ≥50% (PSV ≥220 cm/s and ICA/CCA ratio ≥2.7); ≥80% (PSV 340 cm/s and ICA/CCA ratio ≥4.15). Similar thresholds were reported by Stanziale et al (2005): ≥50% (PSV ≥225 cm/s and ICA/CCA ratio ≥2.5); ≥70% (PSV ≥350 cm/s and ICA/ CCA ratio ≥4.75).

Fig. A5.60 Left: Ce-MRA, coronal rotated MIP showing an ICA near-occlusion stenosis (arrow). Note the collapsed poststenotic ICA (arrows). Right: Extracranial duplex, color-mode image, longitudinal plane with corresponding Doppler spectrum analysis: Top: Stump signal in the carotid bulb suggestive for ICA occlusion. Bottom: Color-mode and Doppler spectra assessment with adapted PRF and gain reveal a turbulent flow with moderate increase velocities (121/38 cm/s) corresponding to an ICA near-occlusion.

Fig. A5.61 Proximal ICA after stent treatment of a high-grade stenosis. Duplex ultrasound of the CCA and ICA. (A) B-mode image, longitudinal plane revealing the stent in situ, a large shadow of a calcified plaque (arrowhead), and a residual hypoechoic plaque more distally (arrow). (B) Color-mode image, cross-sectional plane showing the CCA circumference. (C,D) Longitudinal color-mode image and simultaneous Doppler spectrum analysis of the proximal ICA revealing a mildly disturbed flow without increase of velocities, indicating a mild residual stenosis. Note also the residual hypoechoic plaque (arrow).

Fig. A5.62 Proximal ICA occlusion. Left: DSA, selective left CCA filling, lateral view. Small blind sack ICA (arrow) followed by a complete absence of ICA contrast filling. Right: Extracranial duplex. Top: B- and color-mode image, longitudinal view with corresponding Doppler spectrum analysis. Isoechoic intraluminal material filling the ICA is seen which could be either an atherothrombotic plaque or subacute thrombus. Note the anechoic proximal area with absent color filling corresponding to a nonperfused but still open vessel segment, the blind sack. Bottom: Doppler spectrum analysis. Alternating “to-and-fro” stump signal.

Fig. A5.63 Proximal ICA occlusion. Left: Ce-MRA coronal rotated MIP. Note the large blind sack (arrow). Right: Extracranial duplex. Color-mode image, longitudinal view (top) with corresponding Doppler spectrum analysis (bottom). Note the anechoic intraluminal material. Here the blind sack ICA signal reveals a retrograde flow. Note the aliasing effect in the ECA, caused by the low PRF setting used for the detection of low residual ICA flow.

In comparison to ultrasound and catheter angiography, MRA is of limited value in the evaluation of intracranial as well as extracranial vessels after stent implantation (Golshani et al 2013). In contrast, CTA was reported to yield similar results to color-coded duplex sonography (Nolz et al 2012). Further improvements in the assessment of intracranial stents might result from applying flat-panel cone beam CT—a high-resolution imaging technique based on projection radiography which lacks the typical beam-hardening CT artifacts (Hu et al 2015, Ott et al 2016). For further reading on stenting in ICA stenosis, see Case 15; for intracranial stenosis, see Case 26.

Fig. A5.64 Distal supraophthalmic ICA occlusion. Left: DSA, selective right CCA filling, lateral view. (A) Early arterial phase: Proximal ICA contrast filling but apparent contrast stops after 4 cm. (B) Late arterial phase demonstrates preserved contrast filling of the whole ICA (arrows). Note: The ICA lumen is smaller than the ECA lumen. Right: Extracranial duplex. Top: Color-mode image, longitudinal plane. Preserved color filling within the proximal ICA. Bottom: Doppler spectrum analysis: Small antegrade flow with reduced flow velocities and increased pulsatility resembling OA flow pattern (flow velocity 40/12 cm/s).

Fig. A5.65 ICA dissection. Left: DSA, selective right CCA filling, lateral view. (A) Proximal ICA dissection with typical flamed-like occlusion. Intracranial collateral flow was assured by a cross-flow via the ACoA. (B) 6 months later, ICA reopened (arrows). Note: The ICA lumen is smaller than the ECA lumen. Angiographically a cross-flow persisted despite the recanalization of the ICA. Right: Extracranial duplex corresponding to B. Top: Color-mode image, longitudinal plane. Preserved color filling within a slim ICA. Bottom: Doppler spectrum analysis: Small antegrade flow with reduced flow velocities and increased pulsatility resembling OA flow pattern (flow velocity 30/6 cm/s).

ICA Occlusion

ICA occlusion should be defined according to its location in relation to the OA offspring. A proximal ICA occlusion below the origin of the OA (infraophthalmic occlusion) and near the carotid bifurcation is usually characterized by a missing color-mode and Doppler flow signal. The most frequently observed cause in near-bifurcation occlusion is atherothrombotic pathology. In cases with a small remaining “blind sack” a stump signal (low-amplitude, alternating “to-and-fro” Doppler signal) with a short anterograde systolic flow and a retrograde diastolic flow or only a retrograde flow can be seen. If the PRF settings are low, a weak color signal may be present (Fig. A5.62 and Fig. A5.63). In distal infraophthalmic occlusions, e.g., near the skull base— as typically seen in dissections—the ICA may initially remain open from its origin up to the occlusion site. However, as no relevant vessel branch is present, a stump signal and weak color signal may be observed over the whole visible course. If the occlusion persists, the remaining ICA usually occludes by thrombus formation starting from the occlusion site and advancing in a stepwise manner toward the carotid bifurcation. The diameter of the CCA subsequently decreases (Kubis et al 2001). In distal ICA occlusion, adaptive diameter reduction is also observed in the ICA itself. In acute and subacute occlusions, the vessel walls are usually well delineated; in chronic occlusion, no clear vessel borders can be distinguished. The latter is caused by the less well delineated or abolished intima-media complex. Flow patterns in the depending distal vessel segments, i.e., the carotid siphon, MCA and ACA depend on the presence and quality of collateral pathways (for further details, see “Collateral Pathways” below). For further details on infraophthalmic ICA occlusion, see Cases 11–13, Case 20, Case 24, Case 28, Case 39, and Case 40.

In supraophthalmic ICA occlusions (distal to the OA origin) the ICA provides blood for the OA only, but therefore remains open. In this constellation the ICA can be considered as an extended OA. Consequently, the ICA lumen collapses to a level below the ECA diameter and the flow pattern resembles that of the OA with low flow velocities and a higher pulsatility but preserved diastolic flow (Fig. A5.64). For further reading on supraophthalmic ICA occlusion, see also Case 37. It is noteworthy that in ICA dissection and ICA occlusion with collateral flow via the ACoA and/or PCoA an “extended OA” and collateral flow via the communicating arteries may persist even after recanalization of the ICA, as we have observed for many years (Fig. A5.65).

CCA Stenosis and Occlusion

In contrast to the ICA, an exact grading system for CCA stenoses does not exist. However, the relation of diameter and area assessments as well as the diameter/flow velocity relation, the criteria for local findings, and the pre- and poststenotic flow alterations apply similarly to the NASCET-adapted criteria for ICA stenoses while a bulb is not present (see “ICA Stenosis” above). Stenoses in the mid- and distal CCA are easily accessible to duplex ultrasound (Fig. A5.66). Proximal low- and medium-grade CCA stenoses are probably often missed as the CCA origin is not directly accessible and a poststenotic flow pattern can only be seen in high-grade stenoses (>70–80%).