Carotid ultrasound measurements of artery wall thickness and plaque are widely used to predict heart attack and stroke risk, but current measurement guidelines may be outdated and inconsistent. This comprehensive review by an international team of researchers examines the forces that influence these guidelines, including disease types, conventional risk factors, measurement tools, and rapid advances in artificial intelligence. The authors argue that newer full-length measurement techniques and machine learning approaches could significantly improve cardiovascular risk prediction compared to the current 10-millimeter segment measurements recommended by existing guidelines.
Cardiovascular diseases (CVD), including myocardial infarction (heart attack) and stroke, remain the leading causes of death both regionally and globally. The underlying culprit in most cases is atherosclerosis, a condition where fatty deposits, calcium, fibrin, fibrosis, and macrophages build up inside artery walls, causing them to thicken and harden.
For the past four decades, doctors have used a non-invasive imaging technique called B-mode carotid ultrasound (CUS) to measure two important markers: carotid intima-media thickness (cIMT) and carotid plaque (CP). The cIMT measures the thickness of the inner two layers of the carotid artery wall, while carotid plaque refers to focal areas of thickening that protrude into the artery. These measurements serve as early warning signs for cardiovascular problems.
More than 2,000 scientific articles have been published on this topic, yet the results have been remarkably inconsistent. This variability likely stems from a lack of standardization in measurement tools, risk stratification methods, and risk assessment approaches. The authors of this review set out to understand why these inconsistencies exist and whether current guidelines need to be updated.
The stakes are high. If doctors can accurately measure carotid artery changes, they can identify patients at risk for heart attacks and strokes before these devastating events occur. Early intervention could save countless lives. But if the measurements are inconsistent or incomplete, patients may receive inaccurate risk assessments, leading to either unnecessary worry or dangerous complacency.
The research team conducted a systematic search of PubMed, the world's largest biomedical literature database, using an extensive list of keywords. These included terms like "carotid intima-media thickness," "intima-media thickness variability," "carotid wall motion," "carotid plaque," "cardiovascular risk prediction," "total carotid plaque area," "automated cIMT," "machine learning and carotid artery," and "deep learning and carotid artery," among many others.
The initial search returned 36,940 articles. The team then applied four exclusion criteria to narrow down the results:
This review ultimately synthesized findings from 107 peer-reviewed articles. The authors focused exclusively on 2D carotid longitudinal scans, noting that approximately 90% of atherosclerotic risk assessment uses this approach. They deliberately excluded 3D imaging techniques from their analysis, as those represent a separate and broader topic.
Two major sets of guidelines currently govern how doctors measure cIMT and carotid plaque: the American Society of Echocardiography (ASE) guidelines, published by Stein et al., and the Mannheim guidelines, published by Touboul et al. Both were developed in the last decade and provide specific recommendations for measurement protocols.
According to the ASE guidelines, doctors should follow a two-step process for risk assessment. First, they must thoroughly scan the carotid artery to identify any plaque. Second, they measure the cIMT within a 10-millimeter plaque-free segment of the far wall of the common carotid artery (CCA). The guidelines note that most plaque is found at the distal end of the common carotid artery, in the carotid bulb (the widened area where the artery branches), and in the internal carotid artery.
The Mannheim guidelines similarly recommend measuring cIMT in a 10-millimeter plaque-free segment, but they allow measurement in the common carotid artery, internal carotid artery, or carotid bifurcation. The idea behind using plaque-free regions was to ensure reproducible measurements.
Both guidelines define carotid plaque similarly but with slight differences. The ASE guidelines define plaque as focal wall thickening that is 50% greater than the surrounding cIMT, or a focal region with cIMT greater than 1.5 millimeters protruding into the lumen (the open channel of the artery). The Mannheim guidelines add one additional criterion: plaque can also be defined as focal thickening encroaching into the lumen by at least 0.5 millimeters.
Several types of cIMT measurements exist under these guidelines:
Despite these detailed guidelines, the authors note that in routine clinical practice, the protocols are followed only to a limited extent and with great diversity. This inconsistency is a major problem that the review seeks to address.
The review identifies three major categories of diseases that impact the complexity of atherosclerosis and therefore influence how carotid measurements should be interpreted.
Conventional systemic diseases include diabetes, hypertension (high blood pressure), hyperlipidemia (high cholesterol), obesity, and autoimmune diseases such as rheumatoid arthritis and systemic lupus erythematosus. All of these conditions significantly increase a patient's risk of cardiovascular disease and stroke.
Local arterial vascular diseases affect specific blood vessels throughout the body. These include carotid artery disease, coronary artery disease (affecting the heart's blood vessels), renal artery disease (affecting the kidneys), aortic disease, peripheral artery disease (affecting the legs and arms), and brachial artery disease. Each of these conditions accelerates the atherosclerotic process.
Neurological diseases that are accelerated by atherosclerosis include small vessel disease, large vessel disease, leukoaraiosis (changes in the brain's white matter), and Alzheimer's disease. The review also mentions thyroid disease and erectile dysfunction as conditions accelerated by the same atherosclerotic process that leads to more calcium, fibrin, fibrosis, and macrophages in artery walls.
This broad range of affected conditions underscores why carotid measurements matter beyond just heart attack and stroke prediction. The thickness of the carotid artery wall can serve as a window into the health of blood vessels throughout the entire body.
Conventional risk factors (CRF) include age, height, weight, serum biomarkers, and lipid (fat) biomarkers. The review emphasizes that both conventional risk factors and carotid ultrasound image-based phenotypes (CUSIP) are strongly associated with each other. However, current guidelines place little emphasis on this important connection.
This gap in the guidelines is significant because nearly 90% of patients who die from cardiovascular disease or stroke have at least one conventional risk factor. Understanding how carotid measurements behave in conjunction with these risk factors is essential for accurate risk determination.
Annual progression of carotid wall thickness is influenced by several traditional risk factors:
Tracking the increase in carotid wall thickness over time is extremely important because it directly increases the risk of stroke. Furthermore, increases in cIMT are associated with abnormalities in other vascular beds, resulting in endothelial dysfunction (damage to the inner lining of blood vessels), renal (kidney) diseases, coronary heart disease, and erectile dysfunction.
The authors argue that recommendations for carotid measurements provide vital information about the morphology (shape and structure) and growth of atherosclerotic plaque. This information can support physicians in recommending treatment plans to prevent cardiovascular events. Additionally, understanding how carotid measurements relate to conventional risk factors enables researchers to predict long-term changes in cIMT and plaque, allowing for more reliable prevention planning.
The current guidelines recommend measuring average cIMT within a 10-millimeter segment of the common carotid artery. However, in the past decade, researchers have identified five distinct types of carotid ultrasound image-based phenotypes (CUSIP) that capture plaque variations more completely.
These five phenotypes include four different types of full-length cIMT measurements plus total plaque area:
"Full-length" measurement means measuring cIMT all along the carotid ultrasound scan of the common carotid artery, from the edge of the bulb or flow dividers. Researchers collected 100 sample equidistant points (or vertices) throughout the length of the scan for measurement purposes.
This full-length approach represents a significant departure from the current guidelines, which limit measurement to a 10-millimeter plaque-free segment. The authors note that atherosclerotic plaques are uncommon in the common carotid artery and practically start at the carotid bifurcation — the point where the common carotid artery splits into the internal and external carotid arteries. This anatomical fact may explain why adding cIMT to traditional risk scores showed minimal improvement in risk prediction.
Khanna et al. computed 10-year predictions for all five of these current CUSIP measurements. Additionally, a composite risk score (CRS) has been proposed by Godia et al. that combines the effect of all five current CUSIP measurements to provide a real-valued percentage risk for patients.
The review also highlights a newer phenotype called IMT variability (IMTV), which measures variation in carotid wall thickness in both directions — toward the lumen and toward the adventitia. This is important because atherosclerotic plaque can cause inflammation in both directions, forming balloon-like structures that protrude both inward and outward from the artery wall.
The review examines various tools and techniques used for carotid plaque burden measurement and 10-year risk calculation. These include different distance measurement methods such as centerline, polyline, shortest distance, and Mahalanobis distances.
Several risk calculators are currently used in clinical practice to predict cardiovascular risk:
A pivotal moment in this field came in 2012 when Ruijter et al. published a meta-analysis with 11 years of follow-up data. This study measured cIMT within the 10-millimeter region of the far wall of the common carotid artery, free of plaque, and added it to the Framingham Risk Score. The results showed only minimal improvement in 10-year cardiovascular risk prediction, and the authors concluded the improvement was of little clinical importance.
This study had a strong impact on modifying guidelines. The American College of Cardiology (ACC) and American Heart Association (AHA) subsequently recommended against the routine use of cIMT for risk assessment. However, the authors of this review point out a critical flaw: atherosclerotic plaques are uncommon in the common carotid artery and typically start at the carotid bifurcation. By measuring only the plaque-free region of the common carotid artery, the Ruijter study may have missed the most important indicator of disease.
In contrast, studies using full-length measurements of all three arterial segments (common carotid artery, carotid bulb, and internal carotid artery) have demonstrated that plaque thickness has a stronger association with future cardiovascular events. The European Society of Cardiology and European Atherosclerosis Society have recently stated that the presence of carotid plaque on ultrasound unequivocally identifies individuals at high risk of myocardial infarction or stroke.
The review discusses how advances in computer technology, particularly machine learning (ML) and deep learning (DL), are transforming cardiovascular risk assessment. These artificial intelligence approaches can analyze carotid ultrasound images automatically, potentially reducing human error and improving consistency.
Automated tools like AtheroEdge™ 2.0 can process carotid ultrasound images and measure cIMT along the full length of the artery. Figures in the original article demonstrate how these tools identify the lumen-intima and media-adventitia interfaces in the common carotid artery, carotid bulb, and internal carotid artery on both sides of the neck.
The authors suggest that machine learning and deep learning methods may play a significant future role in cardiovascular disease and stroke risk assessment. These technologies can integrate image-based phenotypes with conventional risk factors to improve prediction accuracy.
However, the review notes that all interacting factors — including disease types, conventional risk factors, measurement tools, and risk calculators — are affected by the evolution of technology. This creates both opportunities and challenges for standardizing measurement guidelines.
For patients, this review has several important implications. First, if you have had a carotid ultrasound that measured only the 10-millimeter plaque-free segment of your common carotid artery, your doctor may not have received the most accurate picture of your cardiovascular risk. Full-length measurements that include the carotid bulb and internal carotid artery may provide better risk prediction.
Second, the presence of carotid plaque on ultrasound is now recognized as a strong indicator of high risk for heart attack and stroke. The European Society of Cardiology and European Atherosclerosis Society have stated that carotid plaque on ultrasound unequivocally identifies individuals at high risk. This means that if plaque is found, aggressive risk factor management is warranted.
Third, the combination of conventional risk factors (age, smoking, blood pressure, cholesterol, diabetes, and BMI) with carotid ultrasound measurements provides a more complete picture of cardiovascular health than either approach alone. Patients should expect their doctors to consider both when assessing risk.
Fourth, the review suggests that variability in carotid wall thickness (cIMTV) is associated with cerebrovascular events. This means that even if your average cIMT is normal, significant variation in wall thickness along the artery could indicate increased risk.
Finally, the development of automated tools and artificial intelligence approaches may lead to more consistent and accurate measurements in the future. This could reduce the variability that has plagued carotid ultrasound measurements and improve risk prediction for all patients.
This review has several limitations that should be acknowledged. First, it focuses exclusively on 2D carotid longitudinal scans and does not address 3D imaging techniques, which represent a separate and broader topic in cardiovascular imaging.
Second, the review is based on published literature, which may be subject to publication bias — studies with positive findings are more likely to be published than those with negative results.
Third, the authors note that the 2012 Ruijter meta-analysis, which strongly influenced current guidelines against routine cIMT use, measured cIMT only in the 10-millimeter plaque-free region of the common carotid artery. This may have underestimated the value of cIMT measurement, as atherosclerotic plaques are uncommon in this region and typically begin at the carotid bifurcation.
Fourth, while the review identifies multiple factors that influence measurement guidelines, it does not provide a new set of guidelines. Instead, it calls for revisiting the current guidelines and highlights areas where improvements are needed.
Finally, the review acknowledges that the use of carotid ultrasound phenotypes for routine cardiovascular risk assessment remains a debatable topic due to variations in measurement protocols, types of measurement tools, different ways of computing 10-year risk, specific carotid segments analyzed, scientific validation protocols, and the type of study utilized.
Based on this review, patients and healthcare providers should consider the following recommendations:
The authors conclude that both machine learning and non-machine learning strategies will flourish for current and 10-year cardiovascular disease and stroke risk prediction, as long as they integrate image-based phenotypes with conventional risk factors. This integrated approach represents the future of cardiovascular risk assessment.
cIMT is the thickness of the inner two layers of the carotid artery wall, measured by ultrasound. It serves as an early warning sign for cardiovascular problems, helping doctors assess risk for heart attack and stroke. However, current guidelines may not capture the full picture, as they often measure only a small segment.
Current guidelines recommend measuring a 10-millimeter plaque-free segment of the common carotid artery. But plaques typically start at the carotid bifurcation, so this method may miss important disease. Full-length measurements including the bulb and internal carotid artery may provide better risk prediction.
The presence of carotid plaque on ultrasound is a strong indicator of high risk for heart attack and stroke. The European Society of Cardiology and European Atherosclerosis Society state that carotid plaque unequivocally identifies individuals at high risk, so aggressive risk factor management is warranted.
Conventional risk factors (age, smoking, blood pressure, cholesterol, diabetes, BMI) are strongly associated with carotid ultrasound measurements. Nearly 90% of patients who die from cardiovascular disease have at least one risk factor. Combining both risk factors and carotid measurements gives a more complete picture of cardiovascular health.
The 10-mm segment is a small plaque-free area in the common carotid artery, while full-length measurements cover the entire scan, including the bulb and internal carotid artery. Full-length measurements capture plaque variations better and may improve risk prediction, as plaques are uncommon in the common carotid artery.
Yes, machine learning and deep learning tools can automatically analyze carotid ultrasound images, potentially reducing human error and improving consistency. Automated tools like AtheroEdge™ 2.0 can measure cIMT along the full length of the artery, which may lead to more accurate risk assessment.
Ask whether the measurements included the carotid bulb and internal carotid artery, not just the common carotid artery. Also ask about your cIMT number, plaque presence, and how these combine with your blood pressure, cholesterol, and other risk factors. Discuss if automated tools were used for measurement.
Original Article Title: Many carotid intima-media thickness and plaque- should the current measurement guidelines be revisited
DOI: 10.23736/S0392-9590.19.04267-6
Authors: Luca Saba, Ankush Jamthikar, Deep Gupta, Narendra N. Khanna, Klaudija Viskovic, Harman S. Suri, Ajay Gupta, Sophie Mavrogeni, Monika Turk, John R. Laird, Gyan Pareek, Martin Miner, Petros P. Sfikakis, Athanasios Protogerou, George D. Kitas, Vijay Viswanathan, Andrew Nicolaides, Deepak L. Bhatt, and Jasjit S. Suri
Journal: International Angiology, December 2019, Volume 38, No. 6, Pages 451-465
DOI: 10.23736/S0392-9590.19.04267-6
Publisher: Edizioni Minerva Medica
Key Words: Ultrasonography; Carotid arteries; Cardiovascular diseases; Stroke
Note: This patient-friendly article is based on peer-reviewed research. It is intended for educational purposes and should not replace professional medical advice. Always consult your healthcare provider about your individual cardiovascular risk and treatment options.
Use collapsible tabs for more detailed information that will help customers make a purchasing decision.
Ex: Shipping and return policies, size guides, and other common questions.