Seeing the Hidden Danger: How MRI Scans of Vulnerable Carotid Plaque Are Changing Stroke Risk Prediction

Table of Contents

• Key Points
• Background: Why Artery Narrowing Alone Isn't Enough
• What Is a "Vulnerable Plaque"?
• How the Research Was Conducted
• Plaque Shape: Thickness, Volume, and Ulceration
• Plaque Composition: The Lipid-Rich Necrotic Core
• Plaque Composition: The Fibrous Cap
• Plaque Composition: Intraplaque Hemorrhage
• Newer MRI Techniques That Speed Up Detection
• Parametric Mapping: Measuring Plaque Tissue by Numbers
• Risk Stratification: What the Studies Show
• Modern Medical Therapy and Leftover Risk
• Treatment Implications
• Limitations and Unanswered Questions
• Recommendations for Patients
• Frequently Asked Questions
• Source Information

Key Points

• MRI can detect three vulnerable plaque features — lipid-rich necrotic core, thin or ruptured fibrous cap, and intraplaque hemorrhage — that predict stroke better than percentage stenosis alone.
• In recently symptomatic patients with more than 50% carotid stenosis, those with intraplaque hemorrhage had an estimated annual stroke risk of 23.2% versus 0.6% without it.
• Most vulnerable plaque features can be detected with standard MRI equipment, adding only 4 to 5 minutes to a routine carotid MRA.
• Statins shrink the lipid-rich necrotic core, but no prospective drug trial has tested whether lipid-lowering therapy can reduce intraplaque hemorrhage or total plaque volume.
• Ask your care team whether plaque composition imaging is appropriate for your situation, especially if narrowing is below 50% with symptoms or significant without symptoms.

Background: Why Artery Narrowing Alone Isn't Enough

Stroke remains a leading cause of illness and death worldwide. Yet doctors still decide how to treat the carotid arteries mostly by measuring one thing: the percentage of the artery blocked by plaque.

These are the two large neck arteries that supply blood to the brain. Randomized clinical trials are studies in which patients are assigned by chance to different treatments. These trials have shown that carotid endarterectomy (CEA) lowers the risk of future stroke. CEA is surgery to remove plaque from the carotid artery. This benefit applies to patients who recently had symptoms and whose narrowed artery is 70% or more blocked on the same side. The benefit is smaller in patients with no symptoms at all.

The problem is that a simple stenosis measurement — stenosis means the degree of narrowing — tells a doctor very little about an individual patient's actual stroke risk. This gap matters most in two groups:

• Symptomatic patients whose narrowing is not severe
• Patients with narrowing but no symptoms

There is a second issue. Since those landmark trials were published, medical therapy has improved so much that stroke risk has fallen across the board. In some patients with carotid stenosis, surgery may no longer add any benefit at all.

This creates a need for better tools. The authors argue that MRI of unstable carotid plaque could provide extra risk information. This MRI is performed after patients are already on guideline-based medical therapy. The extra risk information could help choose the right treatment for each person.

What Is a "Vulnerable Plaque"?

Histology is the microscopic study of tissue. Histology shows that coronary artery plaques with a large lipid-rich necrotic core and a thin fibrous cap overlying it are linked to sudden cardiac death. Coronary artery plaques are plaques in the heart's arteries. This observation led researchers to coin the term "vulnerable plaque."

Two multidisciplinary consensus articles defined the key features of vulnerable plaque. Their authors stressed one important point: rupture-prone plaques are not the only dangerous ones. Any plaque with a high likelihood of causing a blood clot or of progressing rapidly should be considered vulnerable.

The original concept, developed for the heart's arteries, was then extended to the carotid arteries. Researchers found that many of these features could be detected and measured with MRI. These include:

• Carotid plaque volume
• Maximal wall thickness
• A large lipid-rich necrotic core with a thin or ruptured fibrous cap
• Intraplaque hemorrhage

Each feature has its own specific imaging signature on MRI.

How the Research Was Conducted

This article is a review, meaning the authors gathered and interpreted findings from many prior studies rather than running a single new experiment. The authors are radiologists from the Walter Reed National Military Medical Center and the Uniformed Services University of the Health Sciences in Bethesda, Maryland.

The evidence they synthesize comes from several types of research:

1. Repeatability studies measuring how consistent plaque measurements are. One multi-institution study compared measurement error at multiple locations against a single location. The multiple-location measure was total plaque volume (TPV, the overall volume of plaque in the artery). The single-location measure was maximal wall thickness. Error was generally lower for total plaque volume, making it an attractive measure for future drug trials guided by imaging.
2. Histology validation studies comparing MRI findings in living patients against the actual plaque tissue removed during surgery.
3. Prospective cohort studies that followed patients over time to see who developed stroke or transient ischemic attack (TIA, a "mini-stroke" that resolves).
4. Meta-analyses that pool results from multiple single-center studies to increase statistical power.
5. A randomized controlled trial of 2,323 patients with atherosclerotic disease (hardening and narrowing of the arteries).

A key practical point: most vulnerable plaque features can be detected using commercially available coils and sequences already used in routine clinical practice. This takes as little as 4 minutes of scanner time. A more detailed, comprehensive evaluation using dedicated carotid surface coils and specialized MR sequences is also available, but it requires about 40 minutes of scanner table time.

Plaque Shape: Thickness, Volume, and Ulceration

Carotid plaque is best found at or near the carotid bifurcation — the point where the main neck artery splits into two branches. A plaque there may or may not cause significant narrowing.

In the North American Symptomatic Carotid Endarterectomy Trial (NASCET), researchers defined a plaque as ulcerated if an ulcer niche appeared in profile as a crater. The crater extends from the lumen into a narrowed plaque. The lumen is the open channel inside the artery. A plaque was also defined as ulcerated if it appeared as a double density when viewed face-on. A plaque was labeled irregular if multiple small craters were possible.

A separate study of 128 symptomatic patients with severe narrowing examined this question more closely. It found a very strong association between detailed histology from surgical specimens and preoperative intra-arterial angiography (an X-ray technique that injects dye into the artery). The angiographic ulcerations were linked to intraplaque hemorrhage and a large lipid-rich necrotic core.

Researchers then extended these definitions to MRI. They use both the luminal view on bright-blood MR angiograms and direct visualization of an ulceration niche larger than 1 millimeter extending into the plaque on cross-sectional black-blood MR images.

Plaque Composition: The Lipid-Rich Necrotic Core

The lipid-rich necrotic core (LRNC) is a mix of cholesterol crystals, debris from dying cells, and calcium particles. A large LRNC occupying more than 40% of the plaque, covered by a thin fibrous cap infiltrated by inflammatory macrophages (immune cells), is prone to rupture.

Multicontrast carotid artery MRI using black-blood T2 weighted (T2W), T1 weighted (T1W), and bright-blood time-of-flight (TOF) MR angiography has been validated against histology. It accurately identifies and quantifies LRNC. Adding contrast-enhanced T1W (CE-T1W) images improves the ability to tell a non-enhancing LRNC apart from the surrounding fibrous plaque tissue.

On the images themselves, LRNC appears:

• Hypointense (dark) on T2W images
• Non-enhancing on CE-T1W images

An LRNC may or may not be accompanied by intraplaque hemorrhage.

Plaque Composition: The Fibrous Cap

The fibrous cap (FC) is the protective layer separating the plaque's contents from the bloodstream. It is characterized most accurately with 3D TOF MR angiography and CE-T1W images.

Multiple research teams have used dedicated carotid coils, research MR sequences, and either noncontrast 3D TOF MRA or CE-T1W MRI. These teams have shown they can distinguish a thick, intact fibrous cap from a thin or ruptured one.

Two validation studies are worth noting. Using 3D TOF MRA, Hatsukami and colleagues found a high level of agreement — 89% — between MRI and histological findings. Cai and colleagues instead used CE-T1W and found a moderate-to-good correlation between carotid MRI findings and the excised tissue specimens.

Here is how radiologists read the images:

• Intact cap: an enhancing band next to the dark lumen on CE-T1W, with a smooth luminal surface on TOF and CE-T1W images.
• Thin but intact cap: a smooth luminal surface, but the enhancing band is lost on CE-T1W images.
• Ruptured cap: a disrupted, dark band on CE-T1W and an irregular luminal surface on all images.

Ultimately, the distinction between "thin" and "ruptured" matters less than the distinction between "thick" and "not thick." Both thin and ruptured fibrous caps raise the risk of future stroke or TIA.

Plaque Composition: Intraplaque Hemorrhage

Intraplaque hemorrhage (IPH) occurs when the proinflammatory, lipid-rich membranes of red blood cells and iron leak into the artery wall. This process destabilizes the plaque.

IPH shows a high signal on all T1W imaging sequences. These include magnetization-prepared rapid acquisition gradient-echo (MPRAGE), TOF, and fast spin-echo (FSE). It is typically seen inside the LRNC, but it can appear elsewhere in the plaque.

Detection depends on the strength of the MRI scanner. A 3 Tesla (3T) scanner outperforms a 1.5 Tesla (1.5T) scanner, but IPH is still readily detectable at 1.5T. In a study by Saam and colleagues, researchers detected IPH using T1W and TOF MRA images at 1.5T with a sensitivity of 82% and a specificity of 77%. Sensitivity is the ability to correctly identify those who have the condition; specificity is the ability to correctly rule out those who do not.

At 3T, the MPRAGE sequence depicted IPH with a similar sensitivity of 80% but a much higher specificity of 97%. Compared with FSE and TOF, MPRAGE showed higher diagnostic capability for detecting and quantifying IPH. MPRAGE is clinically available on most MR manufacturers' machines at either 1.5T or 3T.

A concrete example from the research illustrates the differences between sequences. In a 75-year-old man with a moderately calcified IPH in the left internal carotid artery, the high-signal area measured 15.8 mm² on the TOF image. The high-signal area measured 8.6 mm² on the FSE image. The high-signal area measured 24.8 mm² on the MPRAGE image. The matching histologic specimen showed a large necrotic core filled with IPH measuring 34.9 mm².

Researchers further classify IPH as type I or type II based on its appearance on T2W images:

• Type I IPH is dark on T2W with a short T2, and it correlates with a history of recent stroke or TIA on the same side.
• Type II IPH has a different signal pattern.

Newer MRI Techniques That Speed Up Detection

One difficulty with multicontrast carotid plaque MRI is the need to align, or co-register, multiple sequences. Two newer approaches address this.

The first is the Sequence for Hemorrhage assessment using INversion recovery and multiple Echoes (3D SHINE). It modifies the original MPRAGE sequence to include multiple echoes. This allows automatic detection of IPH and characterization of type I versus type II IPH in a single 4-minute scan, replacing the more time-consuming multicontrast approach.

The second is Simultaneous Noncontrast angiography and intraPlaque hemorrhage (SNAP). This recently developed sequence uses a phase sensitive inversion recovery technique. The phase sensitive reconstruction separates the high T1 signal of IPH from the negative signal of flowing blood inside the lumen. Displaying only the negative signals produces a noncontrast MRA. Displaying only the high signals produces a highly T1 weighted image suitable for IPH detection. Both are produced in the same 4- to 6-minute acquisition.

Important caveat: neither 3D SHINE nor SNAP is currently clinically available.

Simpler Options Already in Use

In 2015, Moody and Singh made a compelling case for adding carotid plaque imaging to routine clinical carotid MRA. They showed that IPH can be reliably detected at either 1.5T or 3T using a variety of vessel wall sequences, including the widely available MPRAGE. The added scanner time is only 4 to 5 minutes.

Separately, a head-to-head comparison tested dedicated carotid research coils and sequences against large field-of-view, clinically available plaque sequences using a standard clinical neurovascular coil. The clinically available setup achieved high sensitivity, specificity, and accuracy for detecting LRNC and IPH, though its assessment of the fibrous cap was limited. Even so, the ability to identify, characterize, and quantify most vulnerable plaque features with standard equipment removes a major barrier to adopting this approach.

A recent consensus statement from the Vessel Wall Imaging study group of the American Society of Neuroradiology confirmed that current carotid vessel wall imaging techniques can be informative. The group made specific recommendations about imaging parameters for both noncontrast and contrast-enhanced carotid plaque MRI.

Single-Acquisition and Machine Learning Approaches

A newly developed 3D sequence can obtain three different contrast weightings during a single 5-minute acquisition. Called the Multicontrast Atherosclerosis CHarterization (MATCH) sequence, it was tested in 53 consecutive patients undergoing conventional multicontrast carotid plaque MRI. MATCH was comparable — and possibly superior — to conventional multicontrast imaging in identifying and quantifying major carotid plaque components.

In another effort to streamline analysis, one study used machine learning to identify LRNC, IPH, calcification, and fibrous tissue from a single SNAP acquisition. The initial results were promising.

Parametric Mapping: Measuring Plaque Tissue by Numbers

Multicontrast sequences have been the mainstay of vulnerable plaque imaging. But recently developed apparent diffusion coefficient (ADC), T1, and T2 parametric maps have pushed the field toward more quantitative analysis.

The main benefit of quantitative ADC, T1, and T2 imaging is improved reproducibility. This makes it easier to monitor changes in carotid plaque composition over time, with little if any learning curve for interpreting the images.

Early results show that combining ADC with the longitudinal relaxation rate allowed researchers to tell apart LRNC, IPH, and fibrous tissue inside surgical specimens of carotid plaque. The longitudinal relaxation rate is R1, defined as the reciprocal of the T1 value.

Risk Stratification: What the Studies Show

Recent studies demonstrate that characterizing carotid plaque features offers far more information — and better risk stratification — than percentage stenosis and other cardiovascular risk factors alone.

The Lipid-Rich Necrotic Core and Thin Fibrous Cap

A sentinel (landmark) paper by Takaya and colleagues found that four features all predicted future stroke or TIA. These features were a thin or ruptured fibrous cap, a larger maximum percentage of LRNC, the presence of IPH, and larger maximal wall thickness. The study looked at patients with moderate carotid stenosis who had no symptoms.

A prospective study of 120 patients with carotid stenosis found that increasing LRNC size was associated with the development of new ulceration or fibrous cap rupture and with increasing plaque burden. Notably, percentage carotid stenosis showed no association with either outcome at follow-up.

A meta-analysis by Gupta and colleagues pooled multiple single-center studies. It confirmed that the presence of LRNC, a thin or ruptured fibrous cap, and IPH is associated with increased risk of future stroke or TIA. There was no statistically significant difference in hazard ratios between the different plaque components. Hazard ratios express how much more likely an event is in one group versus another. In this analysis, LRNC showed the lowest hazard ratio, IPH an intermediate one, and thin or ruptured fibrous cap the highest. That ranking matches the original vulnerable plaque classification scheme.

Perhaps the most striking finding comes from a prospective randomized controlled trial of 2,323 patients with atherosclerotic disease. Increasing LRNC size and the presence of a thin or ruptured fibrous cap in carotid plaques were strongly associated with:

• Fatal and nonfatal myocardial infarction (heart attack)
• Ischemic stroke (stroke caused by a blocked blood vessel)
• Hospitalization for acute coronary syndrome (ACS, a sudden reduction in blood flow to the heart)
• Symptom-driven revascularization (procedures to restore blood flow)

This finding is important because it means MR biomarkers of carotid plaque vulnerability reflect systemic athero-thrombotic risk — risk throughout the body's arteries — and not just stroke or TIA risk.

Intraplaque Hemorrhage as a Stroke Predictor

IPH is a known predictor of future stroke on the same side as the affected artery. Carotid plaques with IPH represent a more advanced stage of atherosclerotic disease than plaques showing only LRNC with a thick, intact fibrous cap.

In a meta-analysis, Schindler and colleagues showed that the presence of IPH increased the risk of future stroke in both 560 symptomatic patients and 136 asymptomatic patients. Multivariate analysis — a statistical method that accounts for multiple variables at once — identified IPH as a predictor of stroke independent of percent stenosis. There was no statistical difference between men and women. This means simple carotid stenosis measurements and traditional risk factor analysis may be inadequate for identifying the patients at highest risk of adverse cerebrovascular events.

Consider the numbers for recurrent stroke in recently symptomatic patients with more than 50% carotid stenosis:

• IPH-positive patients: estimated annual stroke risk of 23.2% — roughly 1 in 4 patients per year
• IPH-negative patients: estimated annual stroke risk of only 0.6% — roughly 6 in 1,000 patients per year

This gap calls into question the current risk-benefit assessment for carotid endarterectomy.

Additionally, a recent meta-analysis suggests that IPH-positive plaque may be the cause of embolic strokes in patients with symptomatic narrowing below 50%. These strokes were previously labeled "embolic stroke of undetermined source" (ESUS). ESUS means strokes with no clear explanation found after standard testing.

Modern Medical Therapy and Leftover Risk

Medical treatment of TIA and minor stroke has improved substantially, and these changes have significantly reduced the risk of future stroke and ACS.

The TIAregistry.com project enrolled 4,789 patients between 2009 and 2011 who had an acute TIA or minor stroke. Researchers tracked their risk of stroke and ACS over one and five years while on modern medical therapy. The results:

• At one year: the composite cardiovascular outcome — stroke, ACS, or death from cardiovascular causes — was 6.2%. Most events were recurrent strokes, with a 5.1% stroke rate at one year.
• At five years: the composite cardiovascular outcome rate was 12.9%, with 50% of events occurring in years 2 through 5. Strokes occurred in 345 patients, a 9.5% occurrence rate, with 43% of strokes happening in the second through fifth years.

The lesson is clear: despite better medical therapy, many patients remain at high risk of future stroke or ACS long after their initial event. This raises an important question — can carotid plaque MRI improve risk stratification in these recently symptomatic patients?

Treatment Implications

What happens to vulnerable plaque features with treatment? The evidence so far points in encouraging directions.

LRNC typically shrinks with appropriate statin therapy — statins are cholesterol-lowering medications. In patients who respond inadequately to statins, PCSK9 inhibitors (a newer class of injectable cholesterol-lowering drugs) may play a role.

IPH is a different story. There are no prospective drug trials testing whether any lipid-lowering therapy can reduce IPH and/or total plaque volume. Evidence is continuously increasing that IPH predicts carotid plaque progression and future adverse vascular events. Given this evidence, the authors argue that trials aimed at targeted therapy for IPH represent a significant need.

The clinical stakes are high. The dramatic difference in stroke risk between IPH-positive and IPH-negative patients with more than 50% stenosis suggests something. Doctors may be operating on some patients who would do fine without surgery. Doctors may also be missing others who would benefit.

Limitations and Unanswered Questions

The authors are careful to acknowledge several gaps in the evidence.

First, some of the most promising newer imaging techniques — 3D SHINE and SNAP — are not yet clinically available. This limits how quickly the research can translate into routine care.

Second, researchers compared dedicated research coils and sequences against clinically available coils and sequences. The clinical setup performed well for LRNC and IPH. The clinical setup had limited ability to assess the fibrous cap. Since the fibrous cap carries the highest hazard ratio for future events in the Gupta meta-analysis, this is a meaningful gap.

Third, and most importantly, there are no prospective drug trials testing whether lipid-lowering therapies can reduce IPH or total plaque volume. This means doctors currently have no proven medication strategy specifically targeting these high-risk plaque features.

Finally, much of the evidence for MRI's predictive value comes from single-center studies and meta-analyses rather than large multicenter randomized trials. More prospective, randomized data would strengthen the case for routine clinical use.

Recommendations for Patients

If you or a loved one has been told you have carotid artery narrowing, here are practical points to keep in mind.

1. Ask what your percentage stenosis means for you personally. Stenosis percentage alone gives limited information about your individual stroke risk. It is one piece of a larger picture.
2. Ask whether plaque imaging is appropriate for your situation. If your narrowing is below 50% and you have had a stroke or TIA, or if you have significant narrowing but no symptoms, plaque MRI may add useful information. Discuss this with your neurologist or vascular specialist.
3. Understand that carotid plaque imaging can be done with standard equipment. Most vulnerable plaque features can be detected with widely available MRI coils and sequences, adding only 4 to 5 minutes to a routine carotid MRA.
4. Stay on your medical therapy. Improved medical management has lowered stroke risk substantially. The TIAregistry.com data show that even with modern treatment, however, some patients remain at high risk — which is exactly why better risk assessment tools are needed.
5. Take lipid-lowering medications as prescribed. Statins have been shown to shrink the lipid-rich necrotic core. If your response is inadequate, ask your doctor about additional options such as PCSK9 inhibitors.
6. Recognize that IPH-positive plaque carries substantially higher risk. If plaque imaging shows intraplaque hemorrhage, discuss the timing of any procedure with your care team. The annual stroke risk difference — 23.2% (about 1 in 4) versus 0.6% (about 6 in 1,000) — is large enough to change treatment decisions.

For now, the strongest action a patient can take is to ensure their care team is aware of plaque composition imaging as an emerging tool. As the authors emphasize, the field urgently needs prospective drug trials targeting intraplaque hemorrhage — a gap that directly affects how patients like you may be treated in the years ahead.

Frequently Asked Questions

What is a vulnerable carotid plaque and why does it matter more than the percentage of narrowing?

A vulnerable plaque is a fatty buildup likely to cause a stroke. MRI can detect three features — a lipid-rich necrotic core, a thin or ruptured fibrous cap, and intraplaque hemorrhage — that predict future stroke better than percentage stenosis alone. This matters because narrowing percentage tells a doctor little about an individual patient's actual stroke risk.

I had a TIA and my carotid narrowing is less than 50%. Could plaque imaging help me?

Possibly. A recent meta-analysis suggests that intraplaque hemorrhage-positive plaque may be the cause of embolic strokes in patients with symptomatic narrowing below 50%. These strokes were previously labeled 'embolic stroke of undetermined source.' These are strokes with no clear explanation found after standard testing. Discuss with your neurologist or vascular specialist whether plaque MRI is appropriate for your situation.

What does it mean if my MRI shows intraplaque hemorrhage (IPH)?

IPH-positive plaque carries substantially higher risk. In recently symptomatic patients with more than 50% carotid stenosis, those with IPH faced an estimated annual stroke risk of 23.2% — about 1 in 4 — versus only 0.6% (about 6 in 1,000) in those without it. If plaque imaging shows IPH, discuss the timing of any procedure with your care team.

Can plaque imaging be done with standard MRI equipment, or do I need special scans?

Most vulnerable plaque features can be detected using commercially available coils and sequences already used in routine clinical practice. This takes as little as 4 minutes of scanner time. A more detailed evaluation using dedicated carotid surface coils and specialized sequences is also available, but it requires about 40 minutes of scanner table time.

What are the treatment options for a vulnerable plaque, and can medication shrink it?

Statins have been shown to shrink the lipid-rich necrotic core. If your response is inadequate, ask your doctor about additional options such as PCSK9 inhibitors. However, no prospective drug trial has yet tested whether lipid-lowering therapies can reduce intraplaque hemorrhage or total plaque volume — a major unmet need identified by researchers.

I have significant carotid narrowing but no symptoms. Should I consider plaque imaging?

Plaque MRI may add useful information if you have significant narrowing but no symptoms. The benefit of surgery is smaller in patients without symptoms, and percentage stenosis alone gives limited information about your individual stroke risk. Discuss with your neurologist or vascular specialist whether plaque imaging is appropriate for your situation.

What is the risk of stroke if I have a thin or ruptured fibrous cap on MRI?

A thin or ruptured fibrous cap predicts future stroke or TIA. In a meta-analysis by Gupta and colleagues, the presence of a thin or ruptured fibrous cap was associated with increased risk. A thin or ruptured fibrous cap showed the highest hazard ratio among plaque components. Both thin and ruptured caps raise the risk of future stroke or TIA.

If I have carotid artery narrowing and my doctor is recommending surgery, when should I get a second opinion?

A second opinion is worth considering when treatment decisions rest on percentage stenosis alone. Plaque features seen on MRI — a lipid-rich necrotic core, a thin or ruptured fibrous cap, and intraplaque hemorrhage — predict future stroke better than narrowing percentage. In recently symptomatic patients with more than 50% narrowing, intraplaque hemorrhage carries an estimated annual stroke risk of 23.2%, versus 0.6% without it. This gap can change whether surgery is appropriate, so an independent review of your imaging and records may clarify your options. Diagnostic Detectives Network provides independent expert second opinions.

Source Information

Original article title: MR imaging of vulnerable carotid plaque

Publication: Cardiovascular Diagnosis and Therapy, Volume 10, Number 4, August 2020, pages 1019–1031. Published as part of a Review Article on Advanced Imaging in the Diagnosis of Cardiovascular Diseases.

DOI: 10.21037/cdt.2020.03.12

Submitted: January 26, 2020. Accepted: March 3, 2020.

This patient-friendly article is based on peer-reviewed research. It is intended for educational purposes and should not replace discussion of your individual medical situation with your healthcare provider.