Shear wave elastography imaging of carotid plaques: feasible, reproducible and of clinical potential

Carotid artery plaque is a major risk factor for stroke, a leading cause of death and of adult disability worldwide. Although landmark European and North American clinical trials have shown carotid endarterectomy to be beneficial for symptomatic patients with stenosis over 70% [1], the benefits of carotid endarterectomy for moderate stenosis patients and for asymptomatic patients continues to be debated. For a severe stenosis (≥70%), the number needed to treat (i.e. the number of operations required to save one life at 5 years) has been estimated as 6, which increases for moderate stenosis (50-69%) to 15 [2]. As carotid endarterectomy has a 5% risk of an intra-operative stroke, there is considerable interest in developing improved methods of identifying the unstable plaque in order to improve clinical risk stratification for predicting stroke.

Shear Wave Elastography (SWE) is a novel imaging technique developed to assess tissue elasticity and to differentiate diseased tissue from normal tissue. Shear Wave Elastography exploits acoustic radiation force to generate shear wave propagation in tissue [3]. Application of a theoretical model of wave propagation enables the quantification of the Young’s Modulus (YM) from the measurement of shear wave velocity. Studies have shown potential clinical applications in a wide range of tissues including the breast, thyroid, liver, prostate, spleen, salivary glands, and cervical lymph nodes [4] but there are very few vascular studies [5, 6]. Application of SWE for imaging carotid plaques is particularly challenging as the carotid plaque is often small, of variable morphology, subject to arterial pulsatile motion, with heterogeneous surrounding tissue. Despite these challenges, we have previously shown that SWE estimates of YM using in-vitro vessel stenosis phantom models are feasible, and able to distinguish soft and hard regions with good reproducibility [7]. Also, our case study demonstrated the clinical application of SWE for imaging carotid plaques in comparison with the greyscale median (GSM) and histological assessment [6].

The purpose of this clinical study was to investigate the feasibility of using SWE imaging for measuring the YM of carotid plaques and the potential clinical value of SWE imaging to help identify the unstable carotid plaque. The underlying hypothesis behind this study is that unstable carotid plaques can be identified by characterisation of the carotid plaque stiffness.

The study was approved by the National Research Ethics Service (NRES) Committee East Midlands - Northampton (reference 11/EM/0249), followed institutional guidelines, and each patient gave informed consent before participating in the study. Eighty-one patients undergoing a clinical carotid ultrasound scan were recruited from the rapid-access TIA Clinic and the Vascular Studies Unit. Patients were defined as atherosclerotic if they had carotid artery stenosis of ≥30% (North American Symptomatic Carotid Endarterectomy Trial (NASCET) criteria) measured using blood flow velocities in addition to B-mode diameter measurements and color Doppler findings. Plaques were classified as either having caused focal neurological symptoms relating to the ipsilateral brain hemisphere within the past six-month period (i.e. symptomatic) or as asymptomatic following specialist review of patient symptoms and clinical or radiological (computed tomography and/or magnetic resonance imaging) findings by the stroke physician. This classification based on symptoms was used as a surrogate marker of plaque vulnerability in this study.

Forty-seven patients had atherosclerosis and 54 plaques of ≥30% stenosis were analysed, due to the occurrence of bilateral stenosis in seven patients. Six patients had unilateral internal carotid artery (ICA) occlusion, and were excluded from this study. Successful analysis was carried out on 52/54 (96%) of plaques. Data collection was unsuccessful in two plaques, both in obese patients, owing to attenuation of the ultrasound by superficial tissue.

Table 1 summaries patient demographics, grouped into patients with no or minor atherosclerosis (0-30%) stenosis, and patients with >30% atherosclerosis. Table 1 shows age, a history of smoking and a positive vascular diagnosis (transient monocular blindness, transient ischaemic attack and/or stroke) were associated with the presence of (>30%) atherosclerosis (p < 0.05). Only smoking remained significant when comparing the occurrence of symptoms in those patients with atherosclerosis (>30%). Other risk factors including diabetes mellitus, hypertension, hypercholesterolaemia and atrial fibrillation (AF) were not significantly different between groups. The demographics for symptomatic and asymptomatic patients are summarised in Table 2.

Figure 2 shows scatter plots of age compared to YM in vessel wall and carotid plaque, and GSM of plaque. There was no significant relationship between age and YM with respect to the vessel wall (r² = 0.002; p = 0.91) or YM within plaques (r² = 0.004; p = 0.66), and no significant relationship between age and GSM of plaques (r² = 0.00005; p = 0.96).

This was the first study to demonstrate the feasibility of SWE imaging for the assessment of tissue elasticity in a range of carotid plaque stenoses and vessel walls. Our findings have demonstrated the potential clinical value of SWE imaging to help identify the unstable carotid plaque by using symptomatic patients as a surrogate measure of the unstable plaque in a retrospective study design.

SWE imaging of carotid plaques was successful in 52 out of 54 plaques giving an excellent technical sucess rate of 96%. The inter-frame reproducibility of YM estimates, represented by the inter-frame CV values was 22% for the vessel wall and 19% for carotid plaques. In comparison, our previous in-vitro phantom study produced CV values ranging from 8-20% [7]. Estimates of YM are affected by the pulsatile motion of the carotid arteries and the restricted temporal resolution of the commercial Aixplorer® scanner (1 Hz). Other clinical studies have quoted CV values of 12-17% in SWE assessment of liver [13] a clinical application, like breast, demonstrating growing clinical potential and good reproducibility [4]. SWE is considered less operator dependent, and with better reproducibility than earlier ultrasound elastography techniques that are based on operator compression of the tissue to induce a transient stress and assess tissue deformation [14]. Whilst the inter-frame reproducibility of YM may be clinically acceptable, the accuracy of YM estimates is dependent on the assumed relationship between shear wave velocity and YM. As discussed in previous studies, shear wave propagation in thin vessel walls and plaque is complex, and therefore the use of a different theoretical model than the commercial implementation adopted in this study is required for an accurate quantification of YM [7, 15]. Nevertheless, our study showing a mean YM across all plaques of 75 kPa (95% CI: 64-85 kPa) is comparable to previous measurements, although these are scarce. A recent study, directly measuring mechanical properties of carotid plaque tissue showed highly non-linear stretch–stress behaviour, with an increased stiffness with stretching due to the inherent microstructure. The YM of fibrous tissue had a large range from 6 to 891 kPa (median 30 kPa) and lipid ranged from 9 to 143 kPa (median 16 kPa) [16].

In our study, the YM of plaques correlated with plaque grading based on echogenicity using the Gray-Weale classification and this relationship might be expected, since previous work suggested that poorly reflective plaques contain more lipids [11]. Moreover, a significant relationship was evident between the YM of plaques and percentage stenosis. This relationship, however, may be the result of a confounding factor as the degree of stenosis of symptomatic plaques was generally higher than for asymptomatic plaques, and the mean YM of symptomatic plaques was significantly lower than asymptomatic plaque YM. Plaque GSM also correlated with the Gray-Weale classification. This association, which was higher than that observed with SWE, was expected as both methods are grades of plaque echogenicity. By contrast, and unlike YM, the GSM was not related to degree of stenosis. Although symptomatic plaques had lower GSM than asymptomatic plaques, this did not reach statistical significance (p = 0.19). This differs from previous studies, including our own, which find a significant association between lower GSM and symptoms [9, 17‐19], and benefitted from larger sample size. Other studies [20] have also found no relation between the GSM and symptoms. Hence the value of GSM in evaluating plaque vulnerability continues to be debated and additional B-mode risk indices investigated [21].

The assessment of clinical potential for the ultrasound techniques used in this study suggested that percentage stenosis was the strongest individual marker for predicting symptomatic patients (AUC = 0.75). SWE estimates of YM were a better marker than GSM, and combining the YM with percentage stenosis yielded improved diagnostic performance (AUC = 0.78). Therefore, whilst quantifying plaque YM will not replace grading of plaque stenosis to determine eligibility for surgery, the additional information provided by SWE imaging in the vascular examination may help improve individual risk stratification. To benefit from the added value of SWE imaging of carotid plaque is straightforward, as implementation in the vascular ultrasound clinic is a simple extension to the current ultrasound examination.

Further work is required to substantiate our initial findings and to extend the study to assess additional parameters such as YM heterogeneity within plaque, a parameter which may further help identify the unstable carotid plaque. Advances in SWE technology, such as improved frame rate should improve reproducibility and enable assessment of temporal variation in YM through the cardiac cycle. Our retrospective study design used clinical symptoms as a surrogate measure of the unstable plaque. A large prospective multicentre study design is required to assess the diagnostic performance and prognostic ability of SWE imaging in identifying the unstable plaque and risk of stroke. This would provide the evidence required for implementation of SWE imaging technology into the routine vascular ultrasound clinic.

Our study showed SWE is able to quantify carotid plaque elasticity and provide additional information that may be of clinical benefit to help identify the unstable carotid plaque.