Plaque biomarkers from high-frame rate vector flow imaging and shear wave elastography in mild carotid stenosis

Compliance with Ethical Standards

The study protocol was approved by the relevant institutional ethics committee of Shanghai General Hospital ([2022]028), and informed consent was obtained from all participants.

Study Population and Patient Selection

This retrospective case-control study initially included 616 patients (aged 37-86 years) diagnosed with carotid plaques who underwent B-mode US, high-frame rate VFI, and SWE examinations between September 2023 and April 2024 (Fig. 1).

Fig. 1.

Flowchart of patient inclusion and exclusion.

n, number of patients.

The inclusion criteria were as follows: (1) the presence of a single plaque located at the CB and (2) mild carotid stenosis (<50%) on B-mode US, as defined by the North American Symptomatic Carotid Endarterectomy Trial criteria [20]. The exclusion criteria comprised: (1) other thromboembolic sources, such as cardiac or intracranial embolism; (2) transient ischemic attack without definitive imaging evidence of an anterior circulation lesion ipsilateral to the CB plaque; (3) poor image quality caused by issues such as severe plaque calcification (Gray-Weale–Nicolaides Type V [21]); (4) multiple plaques, moderate to severe carotid stenosis, or plaques located outside the CB; (5) a history of neck irradiation or prior carotid surgery/stenting; and (6) the absence of head computed tomography (CT) or magnetic resonance imaging (MRI) examinations, or other incomplete clinical data.

Patients were classified as symptomatic or asymptomatic based on head CT or MRI findings. The symptomatic group comprised individuals who experienced their first anterior circulation ischemic stroke ipsilateral to the carotid plaque within the previous 30 days [22,23]. The asymptomatic group included patients without a history of stroke or intracranial lesion.

Clinical Variables and US Parameters

The following variables were collected: (1) demographic data, including age, sex, and body mass index (BMI); (2) laboratory values, including total cholesterol (TC), triglycerides, low-density lipoprotein (LDL), high-density lipoprotein, serum creatinine, and uric acid; (3) medical history, including hypertension, diabetes mellitus, and hyperlipidemia; (4) medication history, including statins, antihypertensives, and hypoglycemics; (5) personal history, including smoking and alcohol consumption; (6) B-mode US parameters, including plaque echogenicity, plaque length, and plaque thickness; (7) high-frame rate VFI parameters, namely wall shear stress (WSS), oscillatory shear index (OSI), and turbulence (Tur) index; and (8) plaque elasticity on SWE, specifically the Young modulus (YM).

US Data Acquisition

All standard B-mode US, high-frame rate VFI, and SWE examinations were performed using a Mindray Resona R9 system (Mindray Bio-Medical Electronic Co., Ltd., Shenzhen, China), as previously described [12,24]. A linear-array transducer with a 3-11 MHz frequency range was employed for all scans. Examinations were conducted by radiologists with more than 5 years of carotid US experience; in the event of disagreement, a senior radiologist with over 10 years of experience made the consensus decision.

B-Mode US Analyses

The carotid arteries were examined bilaterally in both longitudinal and transverse planes, and B-mode US parameters—including plaque echogenicity, plaque length, and plaque thickness—were documented [12,17,24]. Color Doppler flow imaging (CDFI) was then performed in the longitudinal plane along the plaque surface, followed by Doppler spectral analysis. Peak systolic velocity, end-diastolic velocity, and resistance index were recorded for further analysis.

High-Frame Rate VFI Protocol

For VFI video recordings, the following settings were used: arrow life cycle of 25 ms; frequency of 5.0 MHz; depth range of 2-4 cm; and arrow density of 10%. The frame rate was set to 498 Hz with a pulse repetition frequency of 4.0 kHz. Each plaque was recorded in the longitudinal plane for 36 seconds. A representative image illustrating blood flow velocity vectors, streamlines, and vortices over one carotid plaque is shown in Fig. 2.

Fig. 2.

Vector flow imaging (VFI) measurements of the upstream and downstream carotid plaque surfaces.

A. Wall shear stress (WSS), oscillatory shear index (OSI), and turbulence (Tur) index were measured at the upstream surface. B. WSS, OSI, and Tur index were measured at the downstream surface. Blue dots mark locations for WSS and OSI measurements; green squares denote regions for Tur index calculation. Vel, velocity; ROI, region of interest; T, time; TA, time average.

For each plaque, the optimal frame at peak systole was extracted from the VFI video for analysis. Measurement points (blue dots) were placed on the upstream and downstream plaque surfaces to derive WSS and OSI values (Fig. 2). Then, measurement boxes (green 2 mm×2 mm squares) were positioned on those same surfaces to calculate the Tur index in both regions (Fig. 2).

SWE Examination

SWE was performed on the plaque’s longitudinal section as previously described [17]. A sampling box was positioned to encompass the entire plaque, with real-time quality control. Builtin software was used to quantify the YM values, a process that involved manually delineating several regions of interest: the entire plaque margin (delimited by lines) (Fig. 3A, B) and the downstream and upstream plaque areas (denoted by 2 mm-diameter circular region of interests) (Fig. 3C). Each measurement was repeated three times consecutively, and the mean value was recorded. The images were saved for analysis.

Fig. 3.

Shear wave elastography (SWE) measurements of the carotid plaques.

A. B-mode ultrasound image of the plaque is shown. B. SWE measurement of the entire plaque is presented. C. SWE measurements at the downstream (a) and upstream (b) regions are shown: white solid lines delineate the plaque margins, while white circles indicate regions of interest for mean Young modulus calculation.

Assessment of Inter- and Intra-observer Agreement

Inter-observer agreement for VFI and SWE parameters was evaluated in a blinded fashion by two independent radiologists using images from 30 randomly selected patients. To assess intra-observer agreement, one radiologist re-analyzed the data from 30 randomly selected patients after a 1-month interval, without access to the initial results.

Statistical Analyses

All analyses were performed using SPSS Statistics for Windows, version 26.0 (IBM Corp., Armonk, NY, USA). Categorical variables were presented as frequency (%) and analyzed using the Pearson chi-square test. Normally distributed continuous variables were expressed as mean±standard deviation and compared using one-way analysis of variance, whereas non-normally distributed data were reported as median (interquartile range [IQR]) and evaluated with the non-parametric rank-sum test. Multivariate logistic regression models were constructed to examine associations between VFI parameters, SWE parameters, B-mode US parameters, and the occurrence of ischemic stroke, with odds ratios and 95% confidence intervals calculated. Receiver operating characteristic (ROC) curves were generated to assess each model’s diagnostic performance, and optimal cutoff values were determined using the maximum Youden index (sensitivity+specificity-1). The Spearman rank correlation coefficient was used to explore relationships between VFI parameters and SWE values. Intra- and inter-observer reliability were quantified by the Kendall W coefficient of concordance. A two-sided P-value of less than 0.05 was considered to indicate statistical significance.

Clinical Characteristics

Of the 616 patients initially identified, 494 were excluded for various reasons (Fig. 1), leaving 122 patients for analysis. Among these 122 patients, 82 (67.2%) were male. The median age was 67 years (IQR, 61.75 to 71 years), and the median BMI was 23.81 kg/m² (IQR, 21.73 to 26.20 kg/m²). Clinically, 75 patients (61.5%) had hypertension, 51 (41.8%) had diabetes mellitus, and 67 (54.9%) had hyperlipidemia. As shown in Table 1, the symptomatic group (n=60) had significantly higher TC and LDL levels than the asymptomatic group (n=62) (P=0.011 and P=0.001, respectively), while other clinical characteristics did not differ significantly between groups (all P>0.05).

Table 1.

Patient clinical characteristics

Comparisons of US-Based Parameters between Symptomatic and Asymptomatic Groups

As shown in Table 2, plaque length and thickness were greater in the symptomatic group (P<0.05), whereas plaque echogenicity and all CDFI parameters did not differ significantly between groups (P>0.05).

Table 2.

Comparison of US-based parameters of carotid plaques between asymptomatic and symptomatic groups

Regarding VFI-derived biomechanical forces at the plaque periphery (Table 2), the symptomatic group exhibited significantly higher maximum and mean WSS values on both upstream and downstream plaque surfaces (P<0.05). The Tur index for the downstream surface was also significantly elevated in the symptomatic group compared to the asymptomatic group (P<0.05). However, OSI and the Tur index on the upstream surface, as well as downstream OSI, did not differ significantly between groups (all P>0.05). Notably, mean YM for the entire plaque and for its downstream region were significantly lower in the symptomatic group (P=0.015 and P=0.035, respectively). In contrast, mean upstream YM did not differ significantly between the two groups (P>0.05) (Table 2).

Comparisons of VFI and SWE Parameters within Symptomatic and Asymptomatic Groups

Table 3 presents a within-group comparison of VFI parameters between the plaque’s upstream and downstream surfaces in both symptomatic and asymptomatic groups. In both groups, upstream WSS values were significantly greater than downstream values (P<0.001), whereas upstream OSI and Tur index values were significantly lower than their downstream counterparts (P<0.001).

Table 3.

Comparison of US-based parameters of carotid plaques within asymptomatic and symptomatic groups

SWE parameters between the plaque’s upstream and downstream areas were also compared within both asymptomatic and symptomatic groups (Table 3). In both groups, the mean YM in the downstream area was significantly lower than in the upstream area (P<0.001).

Correlations between VFI and SWE Parameters in Symptomatic and Asymptomatic Groups

As shown in Supplementary Tables 1 and 2, no significant correlations were observed between VFI parameters on the plaque’s upstream surface and SWE values in the upstream area, nor between VFI parameters on the downstream surface and downstream SWE values, in either the symptomatic or asymptomatic group (all correlation coefficients r ranged from -0.3 to 0.3; all P>0.05).

Associations of US-Based Parameters in Mild Stenosis with Ischemic Stroke

Multivariable logistic regression models were constructed to explore predictors of ischemic stroke (Table 4). In model 1, mean WSS and Tur index on the downstream plaque surface, along with maximum WSS on the upstream surface, were significantly associated with stroke occurrence. After adjustment for the SWE-derived mean YM and B-mode US parameters (plaque length and thickness), the downstream Tur index (cutoff, 1.4%; area under the curve [AUC], 0.607; sensitivity, 0.774; specificity, 0.450) and the whole-plaque mean YM (cutoff, 16.52 kPa; AUC, 0.627; sensitivity, 0.839; specificity, 0.400) remained significantly correlated with ischemic stroke in patients with mild stenosis (models 2 and 3) (Table 4).

Table 4.

Multivariable logistic regression analyses of US-based parameters in mild stenosis for predicting ischemic stroke

Regarding predictive performance Supplementary Table 3 shows that model 3 achieved the highest area under the ROC curve (AUC, 0.765), with a sensitivity of 0.871 and a specificity of 0.533. This model outperformed the downstream Tur index alone (AUC, 0.607), the plaque mean YM alone (AUC, 0.627), model 1 (AUC, 0.737), and model 2 (AUC, 0.752) (Supplementary Table 3), indicating that the combined use of high-frame rate VFI and SWE examinations improves stroke risk prediction.

Consistency Analyses of High-Frame Rate VFI and SWE Examinations

Assessment of intra- and inter-observer agreement for these parameters demonstrated exceptionally high concordance (Supplementary Table 4). For intra-observer analysis, the Kendall W ranged from 0.826 to 0.953 (all P<0.05) for VFI parameters and from 0.917 to 0.969 (all P<0.05) for SWE measurements. Inter-observer Kendall W coefficients ranged from 0.786 to 0.917 (all P<0.05) for VFI and from 0.906 to 0.973 (all P<0.05) for SWE.