The predictive value of Plaque-RADS classification in two-dimensional carotid ultrasound for stroke risk stratification

Introduction

Stroke is a major challenge globally. Due to its high mortality and disability rates, it imposes a substantial economic burden on society (1). Carotid artery plaques serve as an indicator of atherosclerosis throughout the body. Approximately 7–18% of ischemic strokes are associated with carotid artery plaques (2). Ulcerated plaques, with intraplaque hemorrhage, and plaques with large necrotic or lipid core areas are referred to as vulnerable plaques. Vulnerable plaques are more likely to cause ischemic stroke (3). Thus, early detection of vulnerable plaques, accurate risk assessment, and appropriate classification management are of vital importance in reducing the incidence of stroke (4).

Saba et al. (5) jointly released the Plaque Reporting and Data System (Plaque-RADS) for stroke risk prediction. It characterizes vulnerable plaques using multimodal imaging techniques, including ultrasound and magnetic resonance imaging (MRI). Employing standardized vocabulary and structured reporting enables risk stratification and prognosis assessment. It also enhances the consistency of results among different examiners and facilitates communication among doctors from various disciplines (6).

At present, the predictive efficacy of each Plaque-RADS classification lacks sufficient clinical validation, and the accuracy of its application in clinical practice requires further validation. This study aimed to verify the clinical value of Plaque-RADS classification in two-dimensional (2D) ultrasound and to provide an evidence-based approach for optimizing stroke risk assessment. We present this article in accordance with the STROBE reporting checklist (available at https://qims.amegroups.com/article/view/10.21037/qims-2025-1152/rc).

Methods

Study cohort

In total, 2,023 patients who underwent carotid artery ultrasound examination at the Sanming Second Hospital from January 2024 to January 2025 were included in this prospective study. The inclusion criteria were as follows: (I) clear carotid artery ultrasound images; (II) complete clinical data; and (III) irregular or no use of statin drugs. The exclusion criteria were as follows: (I) thrombosis of the vertebral artery or subclavian artery; and/or (II) incomplete clinical data or unclear ultrasound images.

This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. This study was approved by the Ethics Committee of Sanming Second Hospital (No. sm2ykj20231006), which waived the requirement of informed consent because the study involved the analysis of existing clinical data and images obtained during routine care, and the research posed no more than minimal risk to participants. All the patient data were anonymized and de-identified prior to analysis.

Clinical information

Clinical data were extracted from the electronic medical records of the patients during carotid ultrasound examination, including demographic information (gender, age, height, and weight), lifestyle factors (e.g., smoking status), and vital signs (e.g., systolic blood pressure). Relevant medical history information was also collected (e.g., the presence or absence of diabetes).

Carotid ultrasound examination and Plaque-RADS classification criteria

Using Philips 7c and GE E11/E95 color Doppler ultrasound diagnostic equipment, with linear array probes and a frequency range of 3–12 Hz, a comprehensive scan of bilateral common carotid arteries and internal and external carotid arteries was performed. The examination included the assessment of the carotid artery intima-media thicknesses and presence of plaques, including the size, internal echogenicity, fibrous cap status, location, and number of any detected plaques, with all parameters recorded. According to the Plaque-RADS classification (5), plaques were categorized as follows: category 1, no plaque; category 2, plaque with a maximum vessel wall thickness <3 mm; category 3, plaque with a maximum vessel wall thickness ≥3 mm; and category 4, plaque with intraplaque hemorrhage, fibrous cap rupture, or intraluminal thrombus. When multiple plaques were present, the highest classification among them was recorded (Figure 1).

Figure 1 Plaque-RADS ultrasound classification diagram (A-H correspond to the characteristics of 1–4 categories of plaques). (A) Plaque-RADS 1 (no plaque); (B) Plaque-RADS 2 (thickness <3 mm); (C) Plaque-RADS 3a (thickness ≥3 mm, thick fibrous cap); (D) Plaque-RADS 3b (thickness ≥3 mm, thin fibrous cap); (E) Plaque-RADS 3c (ulcer plaque); (F) Plaque-RADS 4a (the areas indicated by red circle represent intraplaque hemorrhage); (G) Plaque-RADS 4b (ruptured fibrous cap); (H) Plaque-RADS 4c (the areas indicated by red circle represent intraluminal thrombus). Plaque-RADS, Plaque Reporting and Data System.

Quality control

Two senior attending physicians independently conducted the Plaque-RADS assessment of the most severe plaques in the carotid arteries. If any inconsistencies arose, a third evaluation was performed by a board-certified physician (or a specialist of equivalent or higher qualification) to reach a final consensus. The inter-group consistency test showed that the Kappa value was 0.82 (Figure 2).

Figure 2 Study flowchart (with a total of 2,023 patients ultimately included). Plaque-RADS, Plaque Reporting and Data System.

Follow-up and outcome assessment

The participants who underwent carotid ultrasound examination were followed up. Outcomes were assessed through a combination of medical record reviews and telephone interviews. The primary endpoint was the occurrence or recurrence of ischemic stroke. Stroke recurrence was defined as the emergence of new neurological deficits lasting more than 24 hours, consistent with the territory of the previously affected carotid artery, or the presence of a new acute ipsilateral infarction confirmed by MRI. In cases of multiple stroke events, only the first recurrence was recorded. The follow-up period extended retrospectively from the date of examination to 1 year prior, or until the study conclusion in May 2025, whichever occurred first.

Statistical analysis

The collected data were statistically processed using SPSS 25.0 statistical software. The measurement data that followed a normal distribution were expressed as median (interquartile range), while the categorical variables were described by the frequency (%). Comparisons between groups were conducted using the chi-square test, Fisher’s exact test, or a non-parametric test. A univariate analysis was conducted to identify potential risk factors. A multivariate logistic regression analysis was then performed on the statistically significant factors to evaluate the independent risk factors. A P value <0.05 was considered statistically significant.

Results

Baseline characteristics

A total of 2,023 patients were included in the study (212 in the stroke group and 1,811 in the non-stroke group). The proportion of males (65.1% vs. 56.5%), median age (70.5 vs. 63 years), and prevalence of diabetes (42.6% vs. 29.3%) were significantly higher in the stroke group than the non-stroke group (all P<0.05) (Table 1).

Plaque-RADS classification and risk of stroke

The proportions of Plaque-RADS 1–4 categories in the stroke group were 3.1% (13/412), 10.1% (121/1,198), 17.2% (65/379), and 38.2% (13/34), respectively. The differences in distribution among the groups were statistically significant (P<0.001). The proportions of Plaque-RADS categories 3 and 4 in the stroke group were significantly higher than those in the non-stroke group (30.7% vs. 17.3%, 6.1% vs. 1.2%, both P<0.001) (Table 2). The multivariate analysis revealed that after adjusting for gender, age, and diabetes, Plaque-RADS category 4 was associated with the highest risk of stroke [odds ratio (OR) =8.13, 95% confidence interval (CI): 3.18–20.78], and the risk increased linearly as the classification increased (trend test P<0.001) (Table 3).

Plaque-RADS stratification and short-term risk of stroke recurrence

Over a 1-year follow-up period, stroke recurred in 33 (16.6%) of the 199 study participants. The recurrence rate varied significantly with the Plaque-RADS classification (P=0.008), exhibiting a graded increase from the lower to higher categories. A detailed time-stratified analysis revealed that this association was confined to the acute phase and subacute phase (0–6 months). During this period, a clear hierarchical risk was evident: category 2 lesions had the lowest recurrence (4.9%, 6/121), category 3 had an intermediate recurrence (12.3%, 8/65), and category 4 had the highest recurrence (30.8%, 4/13). Conversely, during the 6–12-month period, the recurrence risks converged and no significant inter-category differences were observed (Table 4).

Discussion

Reilly et al. (7) first introduced a classification system for carotid plaques in 1983, establishing a systematic correlation between ultrasonographic appearances and histopathological features. Based on echogenicity, plaques were categorized as homogeneous or heterogeneous: homogeneous plaques, which are hyperechoic, correspond to fibrous plaques; while heterogeneous plaques, which exhibit mixed echogenicity, are associated with intraplaque hemorrhage, lipid cores, or ulceration (8). Steffen et al. (9) later refined this classification and correlated it with stroke risk, defining four plaque types: Type 1, uniformly hypoechoic; Type 2, predominantly hypoechoic with small hyperechoic areas; Type 3, predominantly hyperechoic with small hypoechoic areas; and Type 4, uniformly hyperechoic. Building on this, Geroulakos et al. (10) added a Type 5 category for plaques that were unclassifiable due to acoustic shadowing from calcification, and further validated the strong association between plaque echographic features and symptomatic status, underscoring the importance of “hypoechoic plaque” as a marker of stroke risk. Although the relationship between plaque morphology and stroke risk is well established, a standardized, cross-modality reporting system has been lacking. To address this, Saba et al. (5) proposed the Plaque-RADS classification, which aims to integrate plaque morphology with quantified stroke risk, thereby facilitating clinical communication and standardizing research data.

This study verified the clinical applicability of Plaque-RADS in 2D ultrasound. The risk of stroke was significantly increased for Plaque-RADS categories 3 and 4 (with high-risk features), which is consistent with the findings of previous studies (5,11). Reporting and Data Systems (RADSs) are imaging guidelines designed to assess disease risk and guide classification and management. Like the Breast Imaging Reporting and Data System (BI-RADS), Thyroid Imaging Reporting and Data System (TI-RADS), and Liver Imaging Reporting and Data System (LI-RADS) (12-14), Plaque-RADS specifies the characteristics of plaques and classifies them to predict the risk of stroke occurrence. However, the specific ranges for each type of prediction of stroke risk are still unclear.

Plaque-RADS describes plaque-related terms, including maximum vessel wall thickness, fibrous membrane thickness and morphology, ulcerated plaques, intraplaque hemorrhage, and intraluminal thrombus. The research suggests that the maximum thickness of plaque is a risk factor for unstable plaques. When the vessel wall maximum thickness is ≥2.5 mm, the risk of cardiovascular disease is high (11). Song et al. (15) reported that in cases of mild to moderate stenosis, a vessel wall maximum thickness ≥3 mm was consistent with the ipsilateral side of stroke. The arc-shaped linear hyperechoic line located between the plaque and the lumen is the fibrous cap, which is another characteristic for evaluating the vulnerability of the plaque.

A thickness <100 µm is defined as a thin fibrous cap. Evidence shows that a thin fibrous cap is a precursor lesion for plaque rupture (16). Ultrasound cannot accurately measure this thickness. Conversely, MRI can accurately determine the thickness and integrity of the fibrous cap of carotid atherosclerotic plaques (17). Despite its limitations in evaluating thin fibrous caps, the convenience and wide availability of ultrasound make it the preferred tool for screening.

Ulcerated plaques are defined as plaques with a surface depression depth of ≥2 mm (18). After ulcer formation, the high shear stress at the front end of the plaque and the circumferential stress at the shoulder exceed the surface strength limit of the plaque, which may lead to plaque rupture. Studies have shown that the incidence of ulcerated plaques is relatively low and the natural healing rate is relatively high (19). Studies have also shown that intraplaque hemorrhage is caused by the rupture of immature neovascularization (20).

This study validated Plaque-RADS as a robust tool for stroke risk stratification. Category 4 plaques (intraplaque hemorrhage/thrombus) showed an eight-fold stroke risk, aligning with prior MRI-based studies. While ultrasound has limitations in detecting thin fibrous caps (<100 µm) and intraplaque hemorrhage, its accessibility supports its widespread clinical adoption. Complementary MRI or contrast-enhanced ultrasound may enhance accuracy in high-risk cases.

The pathogenesis of stroke is complex. The Plaque-RADS classification for predicting the risk of stroke is only a relative concept. It should be noted that even patients in the low-risk category (categories 1–2) may still have a stroke, suggesting that a comprehensive assessment should be conducted by combining other risk factors. This study demonstrated that the risk of Plaque-RADS category 4 was significantly elevated (OR =8.13), which is consistent with the conclusion of McNally et al. (21) that intraluminal thrombus is an independent risk factor. The incidences of stroke in Plaque-RADS categories 1, 2, 3, and 4 in this study were 3.1%, 10.1%, 17.2%, and 30.2%, respectively, which were consistent with the low, medium, and high risks defined by the Plaque-RADS criteria.

Conclusions

The Plaque-RADS classification, as applied in 2D carotid ultrasound, effectively stratifies both initial and short-term recurrent stroke risk, supporting its integration into routine clinical practice for risk assessment.

Acknowledgments

The authors thank Guorong Lyu, MD, PhD, for editing the manuscript and editorial assistance.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://qims.amegroups.com/article/view/10.21037/qims-2025-1152/rc

Data Sharing Statement: Available at https://qims.amegroups.com/article/view/10.21037/qims-2025-1152/dss

Funding: This study was supported by the 2023 Sanming City Health Science and Technology Innovation Joint Project (No. 2023-S-95).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://qims.amegroups.com/article/view/10.21037/qims-2025-1152/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. This study was approved by the Ethics Committee of Sanming Second Hospital (No. sm2ykj20231006). The ethics committee waived the requirement for informed consent due to the retrospective analysis of anonymized routine clinical data, which posed minimal risk.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Global burden of 369 diseases and injuries in 204 countries and territories, 1990-2019: a systematic analysis for the Global Burden of Disease Study 2019. Lancet 2020;396:1204-22. [Crossref] [PubMed]
2. White H, Boden-Albala B, Wang C, Elkind MS, Rundek T, Wright CB, Sacco RL. Ischemic stroke subtype incidence among whites, blacks, and Hispanics: the Northern Manhattan Study. Circulation 2005;111:1327-31. [Crossref] [PubMed]
3. Wang Y, Wang T, Luo Y, Jiao L. Identification Markers of Carotid Vulnerable Plaques: An Update. Biomolecules 2022;12:1192. [Crossref] [PubMed]
4. Prati F, Arbustini E, Alfonso F. Potential of an Approach Based on the Identification and Treatment of Vulnerable Coronary Plaques. JACC Cardiovasc Interv 2021;14:468-73. [Crossref] [PubMed]
5. Saba L, Cau R, Murgia A, Nicolaides AN, Wintermark M, Castillo M, et al. Carotid Plaque-RADS: A Novel Stroke Risk Classification System. JACC Cardiovasc Imaging 2024;17:62-75. [Crossref] [PubMed]
6. Saba L, Scicolone R, Chabert GL, et al. Carotid Plaque-RADS: Inter- and Intrareader Agreement and Learning Curve Analysis in CTA. AJNR Am J Neuroradiol 2025;46:2009-17. [Crossref] [PubMed]
7. Reilly LM, Lusby RJ, Hughes L, Ferrell LD, Stoney RJ, Ehrenfeld WK. Carotid plaque histology using real-time ultrasonography. Clinical and therapeutic implications. Am J Surg 1983;146:188-93. [Crossref] [PubMed]
8. Gray-Weale AC, Graham JC, Burnett JR, Byrne K, Lusby RJ. Carotid artery atheroma: comparison of preoperative B-mode ultrasound appearance with carotid endarterectomy specimen pathology. J Cardiovasc Surg (Torino) 1988;29:676-81.
9. Steffen CM, Gray-Weale AC, Byrne KE, Lusby RJ. Carotid artery atheroma: ultrasound appearance in symptomatic and asymptomatic vessels. Aust N Z J Surg 1989;59:529-34. [Crossref] [PubMed]
10. Geroulakos G, Ramaswami G, Nicolaides A, James K, Labropoulos N, Belcaro G, Holloway M. Characterization of symptomatic and asymptomatic carotid plaques using high-resolution real-time ultrasonography. Br J Surg 1993;80:1274-7. [Crossref] [PubMed]
11. Johri AM, Nambi V, Naqvi TZ, Feinstein SB, Kim ESH, Park MM, Becher H, Sillesen H. Recommendations for the Assessment of Carotid Arterial Plaque by Ultrasound for the Characterization of Atherosclerosis and Evaluation of Cardiovascular Risk: From the American Society of Echocardiography. J Am Soc Echocardiogr 2020;33:917-33. [Crossref] [PubMed]
12. Spak DA, Plaxco JS, Santiago L, Dryden MJ, Dogan BE. BI-RADS(®) fifth edition: A summary of changes. Diagn Interv Imaging 2017;98:179-90.
13. Tessler FN, Middleton WD, Grant EG. Thyroid Imaging Reporting and Data System (TI-RADS): A User's Guide. Radiology 2018;287:1082. [Crossref] [PubMed]
14. Bartolotta TV, Terranova MC, Gagliardo C, Taibbi A. CEUS LI-RADS: a pictorial review. Insights Imaging 2020;11:9. [Crossref] [PubMed]
15. Song JW, Phi HQ, Koneru M, Cao Q, Rubin J, Sakai Y, Ibrahim L, Zhou SE, Woo JH, Kasner SE, Saba L, Cucchiara BL. Prevalence of High-Risk CTA-Based Carotid Plaque-RADS Subtypes in Patients With Embolic Stroke of Undetermined Source. Stroke 2025;56:737-40. [Crossref] [PubMed]
16. Funayama H, Yoshioka T, Ishikawa SE, Momomura SI, Kario K. Close Association of Matrix Metalloproteinase-9 Levels With the Presence of Thin-Cap Fibroatheroma in Acute Coronary Syndrome Patients: Assessment by Optical Coherence Tomography and Intravascular Ultrasonography. Cardiovasc Revasc Med 2021;32:5-10. [Crossref] [PubMed]
17. Torres G, Czernuszewicz TJ, Homeister JW, Farber MA, Caughey MC, Gallippi CM. Carotid Plaque Fibrous Cap Thickness Measurement by ARFI Variance of Acceleration: In Vivo Human Results. IEEE Trans Med Imaging 2020;39:4383-90. [Crossref] [PubMed]
18. de Bray JM, Baud JM, Dauzat M. Consensus Concerning the Morphology and the Risk of Carotid Plaques. CEREBROVASC DIS 1997;7:289-96.
19. Simonetto M, Dharmadhikari S, Bennett A, Campo N, Asdaghi N, Romano J, Koch S. Do Carotid Plaque Ulcers Heal? Potential Detection of Carotid Artery Plaque Healing by Carotid Ultrasound Imaging. J Ultrasound Med 2021;40:973-80. [Crossref] [PubMed]
20. Parma L, Baganha F, Quax PHA, de Vries MR. Plaque angiogenesis and intraplaque hemorrhage in atherosclerosis. Eur J Pharmacol 2017;816:107-15. [Crossref] [PubMed]
21. McNally JS, McLaughlin MS, Hinckley PJ, Treiman SM, Stoddard GJ, Parker DL, Treiman GS. Intraluminal thrombus, intraplaque hemorrhage, plaque thickness, and current smoking optimally predict carotid stroke. Stroke 2015;46:84-90. [Crossref] [PubMed]