Ultrasound Assessment in Ischemic and Hemorrhagic Stroke

10.1055/b-0034-80434

Ultrasound Assessment in Ischemic and Hemorrhagic Stroke

Romero, Javier M., Meader, Anna G.

Pitfall

Although peak systolic velocity is the most common criterion for grading stenosis in duplex ultrasonography (DUS), it is not always entirely reliable in its estimation. Cases in point are the differentiation between occlusion and pseudo-occlusion of the carotid artery and subclavian steal physiology, as well as pseudonormalization, otherwise known as “velocities falling off.”

♦ Carotid Duplex Ultrasonography

Justification for Use

Although many causes of anterior circulation stroke exist, emboli from carotid occlusive disease account for a significant percentage as well as for the majority of stroke morbidity and mortality.1 The North American Carotid Endarterectomy Trial (NASCET) (1991) demonstrated strong evidence that, for symptomatic patients with greater than 70% stenosis, surgical intervention provides a greater reduction in ischemic stroke risk than medical treatment alone.2 Furthermore, the Asymptomatic Carotid Atherosclerosis Studies (ACAS) as well as the Asymptomatic Carotid Surgery Trial (ACST) showed a clear benefit, albeit less dramatic than the NASCET, for intervention in asymptomatic patients with greater than 60% stenosis.3 ^, 4 In addition to carotid endarterectomy (CEA), surgical options for carotid revascularization include carotid angioplasty and stent placement.5 ^, 6 Because the efficacy of intervention in reducing ischemic events directly relates to the severity of stenosis of the internal carotid artery (ICA) in symptomatic patients, noninvasive imaging determining the degree of stenosis plays a pivotal role in defining which patients may benefit from carotid artery revascularization.

Noninvasive imaging of the ICAs is indicated in several clinical settings ( Table 10.1 ), including as a preoperative evaluation prior to coronary artery bypass graft (CABG) surgery,7 in patients who have experienced transient ischemic attacks (TIAs), patients with ischemic stroke, and asymptomatic patients who have carotid bruits. Carotid bruits are neither a specific nor a sensitive indicator for severe carotid disease, however, as many as one third of patients with bruits do have severe ICA stenosis.8 Patients with cervical bruits should undergo evaluation for carotid disease. All patients who have TIAs should be evaluated for carotid stenosis as soon as possible due to a significantly increased risk of stroke in the months following a TIA.6 Patients undergoing evaluation for acute stroke should undergo carotid imaging due to the large proportion of embolic strokes originating from the carotid artery.1 Although carotid screening is not recommended for every CABG candidate, all patients over the age of 60 years or presenting with a minimum of two risk factors, such as elevated cholesterol or known coronary artery disease, should undergo carotid evaluation to reduce the risk of perioperative stroke.9 ^, 10 Sequential follow-up status post-carotid revascularization should also be performed to search for restenosis and intimal hyperplasia.2 Intimal hyperplasia results secondary to operative trauma and is characterized by redundant granulation or scar tissue. Intimal hyperplasia with smooth muscle cell and matrix accumulation is the prominent feature in all of these situations, with evidence of intense cell proliferation and cell death.11 Intimal hyperplasia usually peaks at 9 months postsurgery and generally does not progress after the first year.

Technique

In a standard carotid duplex ultrasonography (DUS), a bilateral examination of the common carotid arteries (CCAs), ICA, external carotid arteries (ECAs), and vertebral arteries (VAs) is performed. Gray-scale images in both transverse and longitudinal planes are obtained throughout the course of the CCA and ICA, whereas in the ECA and VA, longitudinal images only are obtained. B-mode imaging is conducted as a separate segment of the exam from the spectral Doppler evaluation. Utilizing gray-scale B-mode imaging, color and Doppler flow provide the necessary imaging tools to determine the severity of ICA stenosis. The presence of hemodynamically significant disease in the ICA is determined by an increase in the Doppler-derived blood flow velocity, B-mode findings, and color flow images. Spectral Doppler waveforms are obtained from eight standardized sites, with a defined technique (Table 10.2): the proximal, middle, and distal CCA; the proximal, middle, and distal ICA; the proximal ECA; and the proximal/middle VA. Although many guidelines have been published for the detection of significant stenosis (>70%), we favor the internal correlation with angiograms (computed tomography [CT] or conventional catheter) for a practiceoriented guideline ( Table 10.3 ).

**Table 10.1** Clinical Indications for Carotid Duplex Examination
Carotid bruit
Transient ischemic attack
Ischemic stroke
Preoperative coronary artery bypass graft
Post-carotid endarterectomy or poststenting follow-up

Each vessel has a characteristic normal waveform pattern based on the vascular bed distal to the artery ( Table 10.4 ).

Occlusion Versus Pseudo-Occlusion

Vessel occlusion, for instance in the ICA, is typically indicated by a lack of signal or Doppler shift within the vessel. However, even with a critical stenosis very slow flow may be present, and routine ultrasonography may be insensitive to such low velocities. Therefore, the possibility of a pseudoocclusion, otherwise known as a hairline residual lumen, must be considered. The technique should be optimized by placing the Doppler signal in the ghost vessel while increasing the color gain to the maximum level and decreasing filters and pulse repetition frequency (PRF). With these parameters carotid ultrasound shows an 80 to 90% sensitivity in detecting hairline lumina.12

**Table 10.2** Standard Spectral Carotid Duplex Ultrasound Technique
Obtain waveforms from a longitudinal axis view of the artery, using a small sample volume placed in the center stream of flow or center to the flow “jet” Align the cursor parallel to the vessel wall/flow jet Standardize the Doppler angle, with the recommended Doppler angle being 40–60 degrees; if the angle is correct, the cursor will be parallel to the vessel wall; a correct Doppler angle is required to calculate the peak systolic velocity (PSV)

When clinical suspicion is high for pseudo-occlusion, meaning a lack of flow with ultrasound or magnetic resonance angiography (MRA) two-dimensional (2D) time of flight (TOF), CT angiography may be required for a definitive diagnosis.13 Perhaps the most definitive technique may be conventional catheter arteriography, still considered the standard of reference for distinguishing between pseudo- and true occlusions.14

Velocities Falling Off

Typically, as the percent stenosis in a vessel increases, peak systolic velocities increase correspondingly due to the increase of the pressure gradient. However, when stenosis approaches 100%, leaving only a residual hairline lumen open, velocities may suddenly decrease in a process referred to as pseudonormalization ( Fig. 10.1 ). This phenomenon, otherwise known as “velocities falling off,” results in a downward recapitulation of the velocity values into ranges associated with less severe stenosis or absence of stenosis.2 Pseudonormalization is problematic due to the fact that it may lead to an inaccurate description of vessel stenosis, whereas in fact the vasculature may be critically stenotic ( Fig. 10.2 ), and peak systolic velocities indicate a low degree of stenosis ( Fig. 10.1 ), if any, with the result that severe carotid disease may be overlooked.

**Table 10.3** Carotid Noninvasive Ultrasound
Degree of Stenosis	Peak Systolic Velocity (cm/s)	Approx. % Degree of Stenosis	Luminal Diameter (mm)	Other Indirect Signs (Periorbital Doppler TCD, Assumes NL ECAs)
NL	<150		>3
Mild	150–200	50–60%	2.5–3
Mod	200–300	60–70%	2–2.5
Severe	300–400	70–80%	1–2 mm	Reversal flow ipsilateral ACA (TCD) Supraorbital Doppler reversal
Critical	>400	80–90%	0.7–1	Reversal flow Ophthalmic artery (TCD) Supratrochlear Doppler reversal
	Falling off	>90%	<0.7	MRA shows absent distal ICA signal (slow flow versus occlusion), consider CTA
Abbreviations: ACA, anterior cerebral artery; CTA, computed tomography angiography; ECA, external carotid artery; ICA, internal carotid artery; MRA, magnetic resonance angiography; NL, normal; TCD, Transcranial Doppler ultrasonography.
Source: Velocity guidelines for determining degree of vessel stenosis in carotid DUS adapted from the Neurovascular Laboratory, Massachusetts General Hospital, Boston.

**Fig. 10.1** Carotid duplex ultrasound demonstrating dampened waveforms and low peak systolic velocities in the left internal carotid artery, a result of a hairline lumen. (Courtesy of Massachusetts General Hospital Imaging.)

**Table 10.4** Typical Vascular Waveform Characteristics for Carotid Duplex Ultrasonography
Vessel	Normal Carotid DUS Waveform Patterns
ICA	Low resistance waveform with continuous forward flow throughout the diastole
ECA	High-resistance waveform with a rapid sharp systolic upstroke and diminished diastolic flow
CCA	A mixture of the ICA and ECA waveforms with a low resistance waveform and continuous flow throughout the diastole
VA	Low-resistance waveform with continuous forward flow throughout the diastole
Abbreviations: CCA, common carotid arteries; DUS, duplex ultrasonography; ECA, external carotid artery; ICA, internal carotid artery; VA, vertebral artery.

Fortunately, turbulent flow ( Fig. 10.3 ), decreases in the resistance index, spectral broadening ( Fig. 10.4 ), dampening of the distal vessel waveform (curved waveforms Fig. 10.5 ), and a pinpoint residual lumen on color flow images in either the longitudinal or transverse plane may be used to correctly identify a “velocities falling off” situation. Additionally, a retrograde flow direction in the ophthalmic artery detected via transcranial Doppler ultrasonography may be used to identify pseudonormalization. Evidence of reversal of flow begins at approximately the same time velocities start to fall off as the internal carotid artery narrows. Also a dissociation among the degree of stenosis being evaluated in b-mode, the color flow, and the measured velocities should always be interrogated and further investigated.

**Fig. 10.2** Computed tomography angiogram: curved reformat centered in the left internal carotid artery demonstrating a hairline lumen of the proximal segment *(arrow)*. (Courtesy of Massachusetts General Hospital Imaging.)

**Fig. 10.3** Turbulent color flow demonstrated in the left internal carotid artery. (Courtesy of Massachusetts General Hospital Imaging.)

**Fig. 10.4** Spectral broadening and decreased resistance index, demonstrated in the waveform of the distal right internal carotid artery. (Courtesy of Massachusetts General Hospital Imaging.)

**Fig. 10.5** A curved waveform demonstrated in the right distal internal carotid artery. (Courtesy of Massachusetts General Hospital Imaging.)

Tandem Lesions

A tandem lesion in the carotid artery is defined as two separate stenoses at least 3 cm apart that result in severe stenosis along the carotid artery. In such cases, the second lesion, located downstream from the insonated lesion, causes dampening of the peak systolic velocity between the two lesions. It is the stenosis with the greatest narrowing that determines the hemodynamic compromise.13 ^, 15 In this situation, despite evidence of severe stenosis as indicated by B-mode imaging, peak systolic velocities and spectral configuration of flow of the poststenotic segment remain within normal or slightly abnormal limits ( Fig. 10.6 ).13 When tandem lesions are suspected due to a discrepancy between B-mode images and Doppler velocities, CT angiography, gadolinium-enhanced magnetic resonance imaging (MRI), or conventional angiography should be used for a definitive diagnosis ( Fig. 10.6C ).

Subclavian Steal

A subclavian steal phenomenon is clinically characterized by vertigo during arm exertion, as well as by hand or arm pain due to hypoperfusion. Moreover, patients typically exhibit marked differences in blood pressure between the left and right arms. The incidence of subclavian steal is rare, approximately 2%, and the relevance of a steal in asymptomatic patients remains small or unknown.16

**Fig. 10.6** Carotid duplex ultrasound images demonstrating tandem lesions in the right internal carotid artery. **(A)** Gray-scale imaging. **(B)** Waveforms. (Courtesy of Massachusetts General Hospital Imaging.) **(C)** Tandem lesion: coronal curved reformat of the right carotid artery. Severe stenosis of the proximal right internal carotid artery *(long arrow);* severe stenosis of the cavernous segment of the right internal carotid artery *(short arrow)*. (Courtesy of Massachusetts General Hospital Imaging.)

Retrograde flow evident in the VA is the classic hallmark of an advanced subclavian steal phenomenon ( Fig. 10.7 ); however, subtle waveform changes are evident even in the early stages of steal development. In a 957-subject trial, Kliewer et al10 demonstrated four distinctly identifiable waveform changes indicative of advancing subclavian steal physiology. Although all four waveforms share an abrupt decline in flow velocity in early systolic upstroke, each characteristic waveform is defined by the ratio of the flow velocity at the midsystolic notch to the flow velocity at end diastole. The greater the ratio, the more advanced the steal phenomenon, with the class IV steal displaying the greatest degree of retrograde flow ( Fig. 10.8 ). Angiographic correlation supports the conclusion that increasing hemodynamic changes evident in the VA are associated with disease severity.10

♦ Vulnerable Plaque Characterization

Many recent trials have concentrated on the characterization of vulnerable plaque morphology. Seminal papers regarding plaque characterization were based on ultrasound technology; the Cardiovascular Health Physicians Trial in particular demonstrated a close relationship between hypoechoic carotid plaque and neurologic symptoms.17 This was one of the largest trials to date demonstrating a relationship of symptoms different from the one risk based on significant stenosis. In contrast, more recent trials have indicated that a hyperechoic state may represent calcification of the plaque,18 and therefore may confer or represent a certain degree of protection, probably secondary to a more mature stage of this plaque and lack of inflammatory component; in other words, a hyperechoic plaque may represent a stable plaque.

**Fig. 10.7** Depiction of a left subclavian steal phenomenon. The vessel diagram depicts the left vertebral artery (1), the left subclavian artery (2), the aortic arch (3), the brachiocephalic arch (4), and the subclavian region of stenosis (5). (Adapted from Horrow MM, Stassi J. Sonography of the vertebral arteries: a window to disease of the proximal great vessels. AJR Am J Roentgenol 2001;177:53–59.)

Detection of the lipid core as a defined characteristic of vulnerable plaque has also been demonstrated with multiple modalities such as MRI, CT, and ultrasound; a few studies have demonstrated a good correlation of hypoechoic plaque with the presence of a lipid core.19 ^– 21 Although other plaque characteristics such as a thin fibrous cap and plaque enhancement have been shown to possess a strong relationship with symptoms, these findings are difficult to evaluate with ultrasound due to the small dimension of the fibrous capsule, particularly as it is apparently more prone to rupture when the capsule reaches a very thin diameter.22

Recent ultrasound trials have additionally demonstrated the possibility of detecting vasa vasorum with contrast agents.23 ^, 24 This finding has been investigated with computed tomographic angiography (CTA) and demonstrated a larger proportion of vessels with increased vasa vasorum enhancement, hence neovascularization in symptomatic patients compared with asymptomatic patients.25 Neovascularization of the vasa vasorum within the adventitia is a hallmark of inflammation not only of the carotid plaque ( Fig. 10.9 ) but also within the iliac and coronary arteries.26