The initial application of RawVisionTM plaque (RVP) quantitative technique in evaluation for carotid plaque vulnerability Authors: Siqi Cao, Juan Cheng, Xinliang Xu, Li Wei, Jun Jiang, Ying Wang, Yi Dong, Jacques Souquet Published: 2026 ABSTRACT BackgroundNon-invasive assessment for carotid plaque vulnerability is important for risk evaluation of strokes. The RawVisionTM plaque (RVP) quantitative evaluation technique is a novel imaging method based on raw radio frequency (RF) data with real time quantitative assessment of plaques’ vulnerability. This study aimed to explore the value of RVP quantitative technique in assessing the vulnerability of carotid plaques.MethodsIn this prospective study, hospitalized patients diagnosed with carotid atherosclerotic plaques were recruited and divided into symptomatic and asymptomatic groups. The B-mode ultrasound and RVP technique color-coded imaging was performed to evaluated the thickest plaque of carotid arteries in all patients. Various quantitative parameters were created by RVP quantitative analysis, including contrast (CNT), homogeneity (ENT), non-correlation (NRT), capacity (CPT) and the maximum value of absolute strain rate (MASR)—were measured as both mean values (from three regions of interest) and global values (from the overall plaque). All ultrasound features were compared between two groups. The Inter-observer and intra-observer reliability were assessed using the intraclass correlation coefficient (ICC).ResultsThe 123 patients were divided into the asymptomatic group (48/123, 39.0%) and symptomatic group (75/123, 61.0%). Heterogeneous coloring of RVP color-coded imaging in the target plaque could be observed in the symptomatic group. Among all quantitative parameters, the CNT (mean) value of symptomatic carotid plaques was significantly higher than that of asymptomatic carotid plaques(65.424 vs. 47.121, P=0.015). And the CNT (global) value of symptomatic carotid plaques was also significantly higher than that of asymptomatic carotid plaques(57.720 vs. 37.672, P=0.002). Other RVP quantitative parameters showed no significant differences between two groups. The ICC for intra-observer reliability of the RVP quantitative parameters was 0.791–0.827, and was 0.763–0.794 for the inter-observer reliability respectively.ConclusionsAs a novel non-invasive imaging method, the RVP quantitative technique showed potential value in assessing the vulnerability of carotid plaques. Introduction Stroke is the third most common Global Burden of Diseases, and the fourth most common causes of disability-adjusted life-year (1). The vulnerability of carotid plaques, characterized by the propensity for rupture and subsequent thromboembolic events, represents a critical determinant factor of stroke risks (2-4). To make accurate assessment and stratification of carotid atherosclerotic patients based on plaque vulnerability is of great significance to the prevention of ischemic strokes. Vulnerable plaques are defined according to potential risks of subsequent cardiovascular events, such as plaque rupture or erosion (5). Carotid vulnerable plaques are the most common factors that may lead to carotid occlusion or cause embolism (6,7). In histological examination, vulnerable plaques show specific features such as microcalcification, neovascularization, lipid-rich necrotic cores, intraplaque hemorrhage, thin fibrous caps, plaque surface ulceration, huge dimensions, are suggest stenosis, and plaque ruptures (2,8). That’s high-risk features associated with increased risk of subsequent transient ischemic attack (TIA) or stroke in patients with asymptomatic internal carotid artery stenosis (9,10). High-resolution magnetic resonance imaging (MRI) has been established as an effective method for characterization of plaques (11,12). However, the long scan time, high cost, and potential contraindications are the main limitations of MRI (12,13). Computed tomography angiography (CTA) can provide information on plaque morphology and composition, it also shows the clinical value in evaluating the degree of stenosis of carotid plaques (14-16). However, the potential ionizing radiation limited its clinical applications. Furthermore, although CTA provides excellent anatomical information, its direct ability to identify specific vulnerable plaque features—such as fibrous cap rupture or intraplaque hemorrhage.Ultrasound as a non-invasive and real time imaging method, is currently the first-line imaging method for assessment of carotid plaques. B mode ultrasound could evaluate the Intima-Media thickness (IMT) of the carotid and the size, location, echogenicity of the plaques. Color Doppler flow imaging (CDFI) could show real time hemodynamic changes of carotid artery spectrum (17). By injection of contrast agents, contrast-enhanced ultrasound (CEUS) can assess the microcirculation of neovascularization of carotid plaques (18,19). However, ultrasound examination is highly operator-dependent. The relationship between the CEUS value and vulnerability of plaques still requires further validation through additional clinical studies (20,21). Recent studies indicates that ultrasound localization microscopy (ULM) could successfully visualize IPN in carotid plaques at a significantly higher resolution than conventional CEUS images (22,23). However, further clinical multicenter studies are required to confirm the utility of ULM in distinguishing between symptomatic and asymptomatic carotid artery plaques. The RawvisionTM plaque (RVP) quantitative evaluation technique is a novel imaging method based on raw radio frequency (RF) data with real time quantitative assessment of plaques’ vulnerability. The RVP technique employs a two-step optical flow algorithm applied to RF data, which serves a dual purpose: it first tracks tissue motion between consecutive frames to compensate for respiratory and cardiac movements, and then analyzes the axial strain rate from the motion-corrected data. The frame with the highest axial strain rate is selected for texture feature analysis of the strain rate image. Four standard texture features—contrast (CNT), homogeneity (ENT), non-correlation (NRT), capacity (CPT) are quantified, and the maximum value of absolute strain rate (MASR) is calculated. CNT reflects the depth of textural grooves, deeper grooves indicate higher contrast and greater intraplaque compositional complexity. ENT measures textural uniformity, with higher values indicating minimal local variation and highly uniform regions. NRT assesses the linear dependence of gray levels, measuring the similarity of gray-level co-occurrence matrix (GLCM) elements along rows or columns. CPT represents the sum of squares of GLCM element values, providing information on gray-level orderliness and uniformity. The aim of this study was to explore the initial value of RVP quantitative technique in assessing the vulnerability of carotid plaques. We present this article in accordance with the STROBE reporting checklist (available at https://qims.amegroups.com/article/view/10.21037/qims-2026-0726/rc). Methods Patients This was a prospective, single-center, case-control study designed to evaluate the initial value of RVP quantitative technique in assessing the vulnerability of carotid plaques. No individual matching was performed between the symptomatic and asymptomatic groups. This was an exploratory study without a priori sample size calculation. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. This study was approved bythe Ethics Committee of Xinhua Hospital Affiliated to Shanghai Jiao Tong University (No. XHEC-C-2024-037-1). Informed consent was obtained from all patients. Patients were recruited at Xinhua Hospital Affiliated to Shanghai Jiao Tong University between August 2024 and December 2024. Ultrasound examinations and clinical data collection were performed during hospitalization. No longitudinal follow-up was conducted. The inclusion criteria were as follows: (I) patients over 50 years old; (II) hospitalized patients diagnosed with carotid atherosclerotic plaques with thickness >1.5 mm between August 2024 and December 2024; (III) intima-media layer of the carotid artery can be clearly visualized on B mode ultrasound. The exclusion criteria were: (I) patients with history of carotid endarterectomy surgery or other interventional therapies; (II) patients who could not cooperate during ultrasound examinations. According to clinical symptoms, all patients were divided into symptomatic and asymptomatic groups. Symptomatic group was defined as those with cerebrovascular events (CVEs) within 6 months before the ultrasound examination with at least one atherosclerotic carotid plaque in the ipsilateral region (24). The CVE was defined as carotid territory ischemic event (TIA or stroke) with typical symptoms such as hemiparesis, dysarthria, dysphasia, or monocular blindness (25). Ultrasound examination All ultrasound examination was performed using an eSonic eHertz ultrasound equipment (Beijing eSonic Medical Technology Co., Beijing, China) with a L16-4Ep linear array probe (Bandwidth =4–16 MHz). A novel RVP quantitative evaluation technique was performed with the ultrasound system. Patients laid on a supine position with the entire anterior cervical region fully exposed. First transversal and longitudinal B mode ultrasound scans (BMUS) were performed on the common carotid artery, internal carotid artery, and external carotid artery. After detecting the carotid atherosclerotic plaques, the size, location, echogenicity and degree of carotid stenosis were observed and recorded. If the patient had more than one plaque, the thickest plaque would be observed and analyzed. On the longitudinal view of the target carotid plaques, the “RVP” mode was switched on. A sampling box was set to delineate the target plaque region, ensuring the entire plaque is covered. After the image quality was stabilized, a clip at least 3 seconds was recorded and stored without movement of transducer. The RVP color-coded imaging of the target plaque was then performed. Three regions of interest (ROIs) in the hyperechoic, hypoechoic, and isoechoic region of the target plaque were selected. The obtained RVP quantitative parameters include the following parameters for the plaques: CNT, ENT, NRT, CPT and MASR as well as the CNT (mean), ENT (mean), NRT (mean), CPT (mean) and MASR (mean). In addition, the entire boundary of the target plaque was manually delineated to define the overall plaque area. The machine then automatically identified plaque and non-plaque regions using GLCM algorithms and calculated the RVP quantitative parameters for the entire plaque. These were recorded as the CNT (global), ENT (global), NRT (global) and CPT (global) for the plaques. The measurements were repeated three times by two radiologists with 10 years’ experience in carotid ultrasound examination, and the mean values were calculated for analysis. Clinical data Data of the following clinical parameters were recorded: (I) age, sex and body mass index (BMI); (II) potential clinical risk factors, including hypertension, hyperlipidemia, diabetes mellitus, coronary artery disease and smoking history; (III) serological test biomarkers were performed such as triglyceride (TG), cholesterol (TC), low-density lipoprotein (LDL) and high-density lipoprotein (HDL) levels. Statistical analysis Statistical analysis was performed using SPSS Statistics (version 26.0; IBM Corp., Armonk, NY, USA). The Kolmogorov-Smirnov test was used to test for the normality of the metrology data. For continuous variables, data of normally distributed variables were descripted as mean ± standard deviation, while the data of non-normally distributed variables were descripted as median and interquartile range (IQR). Mann-Whitney U and t-tests were used to compare continuous variables between two groups. Frequencies and percentages (%) were used to express categorical variables. The Chi-squared test and Fisher exact test was used to compare categorical variables between two groups. The Inter-observer and intra-observer reliability were assessed using the intraclass correlation coefficient (ICC). P<0.05 was considered as statistically significant. Results General characteristics From August 2024 to December 2024, 123 patients (mean age 72.6±10.6 years) diagnosed with carotid plaques were finally enrolled, including 43 females (35.0%) and 80 males (65.0%). According to clinical symptoms, all patients were divided into the symptomatic group (75/123, 61.0%) and asymptomatic group (48/123, 39.0%). The general characteristics of the patients are shown in Table 1. The number of patients with diabetes mellitus and smokinghistory in the ‘symptomatic group’ was higher than that in the ‘asymptomatic group’ (P=0.006 and P=0.044, respectively). Table 1Summary characteristics of the patientsVariablesAsymptomatic (n=48)Symptomatic (n=75)P valueAge (years)72.22±11.4272.76±10.05 0.180BMI (kg/m2)22.60±2.6423.64±3.12 0.037*Hypertension29 (60.4)57 (76.0)0.073Hyperlipidemia26 (54.2)48 (64.0)0.346Diabetes mellitus15 (31.3)43 (57.3)0.006*Coronary artery disease2 (4.2)12 (16.0)0.077Smoking history19 (39.6)44 (58.7)0.044*Serological indicators Triglyceride level (mmol/L)1.23±0.711.46±0.840.078 Cholesterol level (mmol/L)3.53±0.843.72±0.870.594 LDL level (mmol/L)2.03±0.832.35±0.830.121 HDL level (mmol/L)1.07±0.701.02±0.650.321Data are presented as mean ± standard deviation or n (%). *, statistically significant difference. BMI, body mass index; HDL, high-density lipoprotein; LDL, low-density lipoprotein. Summary characteristics of the patients Data are presented as mean ± standard deviation or n (%). *, statistically significant difference. BMI, body mass index; HDL, high-density lipoprotein; LDL, low-density lipoprotein. B mode ultrasound imaging features On BMUS, 203 carotid plaques were detected in 123 patients. Most of carotid plaques were multiple ones (67/123, 54.5%). The thickest plaque was analyzed in this study. The mean thickness measured by B mode ultrasound of all carotid plaques was 2.29±0.56 mm. Most carotid plaques in both groups showed mixed echogenic (65/123, 52.8%, P>0.05). In addition, 113 mild stenosis cases, 8 moderate stenosis, and 2 severe stenosis were diagnosed on BMUS. There were no significant differences in terms of BMUS imaging features between two groups (P>0.05). RVP technique imaging features Comparing the RVP color-coded imaging features between the two groups, heterogeneous coloring of RVP color-coded imaging in the target plaque could be observed in the symptomatic group (Figures 1,2). While homogeneous coloring of RVP color-coded imaging in the target plaque could be observed in the asymptomatic group. RVP measurement of a carotid plaque in the asymptomatic group. A 69-year-old man without known history of stroke. The B mode ultrasound scan detected a mixed echogenic plaque in the anterior wall of the left common carotid artery, approximately 19.0 mm in size and 2.3 mm in thickness (A). Color Doppler flow imaging showed blood flow filling defects (B). The measured peak systolic velocity at the plaque was 48.77 cm/s with RI 0.76 (C). The RVP color-coded imaging showed homogeneous colors within the plaque. The measured RVP parameters including contrast (CNT global) =26.598, homogeneity (ENT global) =0.580, non-correlation (NRT global) =99.920%, capacity (CPT global) =0.003 (D). CNT, contrast; CPT, capacity; ENT, homogeneity; NRT, non-correlation; RI, Resistance Index; RVP, RawVisionTM plaque. RVP measurement of a carotid plaque in the symptomatic group. A 75-year-old man diagnosed with stroke 3 months before the ultrasound examination. The B mode ultrasound scan detected a hypoechoic plaque in the anterior wall of the right internal carotid artery, approximately 14.6 mm in size and 2.5 mm in thickness (A). Color Doppler flow imaging showed the narrowed blood flow signals (B). The measured peak systolic velocity at the place of plaque was 46.13 cm/s, with RI 0.51 (C). The RVP color-coded imaging showed heterogeneous coloring within the plaque. The measured RVP parameters including contrast (CNT global) =68.400, homogeneity (ENT global) =0.401, non-correlation (NRT global) =99.885%, capacity (CPT global) =0.001 (D). CNT, contrast; CPT, capacity; ENT, homogeneity; NRT, non-correlation; RI, Resistance Index; RVP, RawVisionTM plaque. The success rate of RVP technique measurements was 100% (123/123) in both two groups. The CNT (mean) value of symptomatic carotid plaques was significantly higher than that of asymptomatic carotid plaques (65.424 vs. 47.121, P=0.015). The CNT (global) value of symptomatic carotid plaques was also significantly higher than that of asymptomatic carotid plaques (57.720 vs. 37.672, P=0.002) (Figure 3). Other RVP quantitative parameters showed no significant differences between two groups (Table 2). Comparison of the contrast (CNT mean) and CNT (global) values of carotid plaques between the asymptomatic and symptomatic groups. The CNT (mean) value of symptomatic carotid plaques was significantly higher than that of asymptomatic carotid plaques (A). The CNT (global) value of symptomatic carotid plaques was also significantly higher than that of asymptomatic carotid plaques (B). *, P<0.05. CNT, contrast. Table 2Comparisons between full name (RVP) parameters between symptomatic and asymptomatic groupsVariablesAsymptomatic (n=48)Symptomatic (n=75)P valueMann-Whitney UMedianIQRMedianIQRCNT (mean)47.12149.62965.42471.4460.015*2,270.5ENT (mean)0.3700.2310.3330.2410.4071,640NRT (mean) (%)99.9270.11599.9120.1170.1691,535CPT (mean)0.2300.0130.0200.0110.1941,550MASR (mean)−1.2660.437−1.1410.5780.1352,088CNT (global)37.67227.93757.72064.9790.002*2,393ENT (global)0.5230.1060.4900.1420.2841,593.5NRT (global) (%)99.8880.10899.8730.1110.5611,688CPT (global)0.0020.0010.0020.0010.6561,880.5*, P<0.05. CNT, contrast; CPT, capacity; ENT, homogeneity; IQR, interquartile range; MASR, maximum value of absolute strain rate; NRT, non-correlation; RVP, RawVisionTM plaque. Comparisons between full name (RVP) parameters between symptomatic and asymptomatic groups *, P<0.05. CNT, contrast; CPT, capacity; ENT, homogeneity; IQR, interquartile range; MASR, maximum value of absolute strain rate; NRT, non-correlation; RVP, RawVisionTM plaque. Reproducibility of RVP technique To establish the reproducibility of RVP quantitative evaluation Technique, intra-observer and inter-observer reliability analyses were conducted. The ICC for intra-observer reliability of the RVP quantitative parameters was 0.791–0.827, and the inter-observer reliability was 0.763–0.794 respectively, showed good reliability. Discussion Non-invasive assessment for carotid plaque vulnerability is of importance for risk evaluation of strokes. Ultrasound imaging as a safe, non-invasive, and real time imaging method, is the current first-line imaging method for evaluating carotid plaques. Comparing to conventional imaging methods, RVP technique is a new quantitative evaluation technology which could provide both intuitive color-coded imaging features as well as various quantitative parameters of the carotid plaques. It would quantify four common textural features of plaques (CNT, ENT, NRT and CPT) and directly extracts the average value of MASR and the tension fluctuation curve based on raw strain rate images, which might be related to the vulnerability of carotid plaques. CNT reflects the intraplaque heterogeneity of multiple features, which is directly correlated with stability of carotid plaques. ENT represents the homogeneity of subtle strain distribution within the carotid plaque. NRT evaluates the texture distribution relationships among multiple features in the tissue. CPT indicates variance variability during vascular pulsation, reflecting tissue compliance. In our study, the success rate of RVP technique measurements was 100% (123/123) in both symptomatic and asymptomatic groups. Moreover, the RVP technique enables rapid acquisition of at least 3 seconds of clip, followed by data analysis during idle time. This demonstrates that RVP technique offers a novel, quick and easy way to make assessment for carotid plaque vulnerability. Meanwhile, our reproducibility study results showed that the intra-observer and inter-observer ICC of CNT (global) value and CNT (mean) value also showed good reliability. And intra-observer and inter-observer ICC of CNT (global) value is a little higher than that of CNT (mean) value. As a novel non-invasive imaging method, it shows stable and repeatable evaluations in our initial clinical applications. In the future, it might facilitate bedside scanning for hospitalized patients, eliminating the risks associated with unnecessary patient transfers. Using RVP visual parametric imaging, it might be helpful to reflect the texture features of carotid plaques (26). RVP analysis is based on RF data. The advantage of RF data lies in its unprocessed nature, which ensures objectivity by eliminating observer bias and post-processing artifacts, thereby providing more comprehensive tissue characterization (27). By showing the homogeneity of color signal distribution, it would be helpful to make quick preliminary differentiation between vulnerable and stable plaques in clinical practice. In our result, by comparing the RVP technique color-coded imaging features between the symptomatic and asymptomatic groups, heterogeneous coloring of RVP color-coded imaging in the target plaque could be typically observed in the symptomatic group. RVP technique analyzes the variation capability of internal texture indicators or texture images in plaques under regularly changing dynamic conditions, thereby generating visualized parametric imaging (26). Therefore, RVP technique provides a convenient and direct visualization of vulnerability of carotid plaques. The analysis of various quantitative parameters can be achieved based on RVP imaging. While comparing the symptomatic and asymptomatic groups, the CNT (mean) and CNT (global) values of symptomatic carotid plaques were significantly higher. The CNT value was reported to represent the variability of multiple characteristics in the plaques, which is related to plaques’ stability (26). It reflects the depth of texture grooves in the image. The deeper the texture grooves, the higher the CNT value, indicating more complex internal plaque composition. Depending on RF signals, CNT measures the image variation. It was assumed that plaques with larger deformations and higher levels of complexity in deformation patterns are more likely to be vulnerable (26). Therefore, by quantifying the CNT parameter, it would provide critical insights into the heterogeneity of plaques’ composition which is related to the plaque vulnerability. Previously, some studies used compound ultrasound strain imaging (CUSI) to make measurements of stiffness of carotid plaques, that vulnerable plaques exhibited higher strain values than stable plaques (28). According to the design of the machine, the MASR (mean) value could reflect the stiffness of carotid plaques. However, in our results, the MASR (mean) value showed no significant differences between the symptomatic and asymptomatic groups. The potential reason may be that our study only compared with the presence or absence of clinical symptoms. Overall, the analysis of various RVP quantitative parameters is feasible in real clinical practice. Some of the quantitative parameters, such as CNT value may be related to the vulnerability of plaques. Limitations There are several limitations in our study. First, limited cases of patients were included in this single center initial study. Second, only the thickest plaques in the ipsilateral lesion were assessed. Although this strategy reduces the data clustering effect among different plaques within the same patient and simplifies the analysis, it also overlooks potentially vulnerable features that may exist in non-thickest plaques. In the future, we will conduct multicenter studies employing a multi-plaque, vessel-by-vessel analytical approach, combined with sensitivity analyses stratified by plaque thickness, to comprehensively evaluate the clinical application value of the RVP technique. Third, the current gold standard for evaluating the vulnerability of carotid plaques is the histopathological examination of plaques after carotid endarterectomy; however, our patients did not meet the surgical criteria, hence we can only use whether they had clinical symptoms for group classification. Future studies should be designed as prospective, multicenter, imaging-pathology correlation investigations that directly compare RVP parameters with histopathological features of plaques obtained following carotid endarterectomy, in order to definitively establish the diagnostic value of the RVP technique for assessing true tissuelevel vulnerability. Conclusions As a novel non-invasive imaging method, the RVP quantitative evaluation Technique showed potential value in assessing the vulnerability of carotid plaques. Supplementary The article’s supplementary files as