Carotid plaque dynamic contrast-enhanced magnetic resonance imaging normalised signal intensity reproducibly differs between plaque and vessel wall

This study demonstrates that a simplified, muscle-normalized dynamic contrast-enhanced MRI method reliably distinguishes carotid plaque cores from remote vessel walls with high reproducibility over six months, although the metric remained unaffected by low-dose colchicine treatment.

The Gist

Imagine your arteries are like old garden hoses. Over time, they can get clogged with "gunk" (plaque), which makes them stiff and prone to bursting. If a piece of this gunk breaks loose, it can cause a stroke. Doctors want to know not just how much gunk is there, but how dangerous it is. Is it a solid, stable rock, or is it a fragile, spongy sponge full of tiny, leaky pipes?

This study is about a new way to take a "leak test" on these arteries without cutting anyone open.

The Problem: The "Black Box" of Scans

Doctors use a special camera called an MRI to look at these arteries. To see the "leakiness" (which indicates inflammation and danger), they inject a glowing dye (contrast agent) into the patient's bloodstream.

Usually, analyzing this dye is like trying to solve a complex math equation while juggling. You have to guess how fast the dye moves, how much space it takes up, and how the body processes it. It's complicated, expensive, and hard to do the same way every time.

The Solution: The "Muscle Ruler"

The researchers in this paper asked a simple question: What if we just compare the glow of the plaque to the glow of something nearby that we know is healthy?

They decided to use the neck muscles as a "ruler."

• The Plaque: The dangerous, leaky gunk.
• The Remote Vessel: A healthy part of the artery far away from the gunk.
• The Muscle: The "ruler" (skeletal muscle in the neck).

They took the brightness of the plaque and divided it by the brightness of the muscle. This gave them a "Normalized Score." It's like saying, "The plaque is glowing 3 times brighter than the muscle," instead of trying to calculate the exact chemical concentration of the dye.

What They Did

They looked at 28 patients with mild-to-moderate artery clogging.

1. They scanned them at the start.
2. They gave half the patients a cheap anti-inflammatory pill (colchicine) and the other half a fake pill (placebo).
3. They scanned them again 6 months later.

The Findings (The "Aha!" Moments)

1. The Plaque Glows Brighter (And that's bad news for the plaque)
Just like a sponge soaked in water glows brighter than a dry rock, the dangerous plaque core glowed significantly brighter than the healthy artery wall.

• Analogy: Imagine the healthy artery is a dry sponge, and the plaque is a sponge soaked in glowing water. The study confirmed that the "soaked" plaque lights up much more than the "dry" healthy parts. This proves the method works to spot the trouble spots.

2. The "Ruler" is Reliable
The most important finding was reproducibility. If you scan the same person twice, six months apart, the "Normalized Score" stays almost exactly the same.

• Analogy: It's like using a ruler that doesn't stretch or shrink. No matter when you measure, the number is consistent. This means doctors could use this method to track patients over time without the numbers jumping around due to technical errors.

3. The Pill Didn't Change the Glow
The researchers hoped the anti-inflammatory pill (colchicine) would make the plaque "dry out" (reduce the glow) by calming the inflammation.

• The Result: The glow didn't change. The plaque looked just as "wet" and leaky after 6 months, whether the patient took the real pill or the fake one.
• Takeaway: This specific type of scan didn't show that the pill worked on the blood vessel walls in this short time, or perhaps the pill works in a different way that this specific "leak test" can't see yet.

Why This Matters

This study suggests a simpler, cheaper, and more reliable way to check if an artery plaque is "angry" and unstable.

• Before: You needed a super-computer and a PhD in math to analyze the dye.
• Now: You can just compare the plaque to the neck muscle. It's like using a simple ruler instead of a laser interferometer.

The Catch

The study had a few hiccups:

• Motion Sickness: If the patient moved their head or if their heart beat too hard, the images got blurry, and they had to throw those scans away (about 1 in 3 scans were unusable).
• Small Group: They only looked at 28 people. It's a great start, but they need to test it on thousands to be sure.

The Bottom Line

The researchers found a simple "glow-ratio" trick that consistently spots dangerous, leaky plaque in neck arteries. It's a promising new tool that could help doctors decide who needs urgent treatment, without needing complex math. While the specific pill they tested didn't show a change in this short trial, the method of measuring the plaque is now proven to be a reliable "thermometer" for artery health.

Technical Summary

1. Problem Statement

Carotid atherosclerotic plaque neovascularization and endothelial permeability are critical indicators of plaque instability and the risk of major adverse cerebrovascular events (MACCE). While Dynamic Contrast-Enhanced Magnetic Resonance Imaging (DCE-MRI) can non-invasively quantify these processes, current clinical translation is hindered by:

• Complexity: Standard quantitative approaches rely on complex pharmacokinetic modeling (e.g., Tofts, Patlak models) to derive parameters like the volume transfer constant ($K_{trans}$) and plasma volume fraction ($v_p$).
• Data Requirements: These models often require an Arterial Input Function (AIF), which is difficult to derive from "black-blood" MRI sequences commonly used for plaque imaging.
• Debate: There is no consensus on which kinetic model best reflects plaque biology, limiting reproducibility and standardization.
• Variability: Non-linear relationships between MR signal intensity and gadolinium concentration create inter-scan and inter-patient variability.

The study addresses the need for a simplified, reproducible, and accessible DCE-MRI biomarker that avoids complex modeling while retaining physiological relevance to plaque vulnerability.

2. Methodology

Study Design:

• Type: Retrospective longitudinal cohort study (sub-study of the CAPRI clinical trial).
• Population: 28 patients with mild-to-moderate carotid atherosclerosis (confirmed by Doppler ultrasound).
• Intervention Context: Patients were originally randomized to receive low-dose colchicine (0.5mg daily) or placebo for 6 months. This sub-study analyzed the DCE-MRI data from this cohort.
• Ethics: Approved by local IRBs; retrospective analysis permitted via amendment.

Image Acquisition:

• Scanner: 3T MRI (Philips Ingenia).
• Protocol: T1-weighted black-blood DCE-MRI performed at baseline and 6-month follow-up.
• Contrast: 20mL Gadobutrol (Gadavist) administered via hand injection.
• Timing: 21 dynamic frames acquired over 256 seconds (12-second intervals) post-injection.

Image Analysis & Quantification:

• Software: Osirix MD.
• Regions of Interest (ROIs): Manually delineated on the most-diseased slice (identified via multi-planar reconstruction):
  1. Plaque Core: The atherosclerotic lesion.
  2. Remote Vessel Wall: Atherosclerosis-free vessel wall.
  3. Reference Tissue: Adjacent posterior neck skeletal muscle (splenius capitis).
• Normalization Strategy: To eliminate variability, Signal Intensity (SI) was normalized against the skeletal muscle SI in the same slice at every time point:
  $$SInorm = \frac{SI_{tissue}}{SI_{muscle}}$$
• Derived Metrics:
  • PEAK $SInorm$: Maximum normalized signal intensity.
  • AUC $SInorm$: Area under the time-intensity curve.
  • $\Delta SInorm$: Absolute difference between baseline and 6-month follow-up.
• Quality Control: Cases with severe motion artifacts or inadequate image quality were excluded (37% exclusion rate from the original pool).

Statistical Analysis:

• Comparison of plaque core vs. remote vessel wall.
• Comparison of baseline vs. 6-month follow-up (reproducibility).
• Comparison between Colchicine vs. Placebo groups.
• Correlation analysis with systemic inflammatory biomarkers (hs-CRP, IL-6, IL-18, VCAM-1).

3. Key Contributions

• Simplified Biomarker Validation: Demonstrated that a muscle-normalized signal intensity approach ($SInorm$) is a viable, simplified alternative to complex pharmacokinetic modeling for carotid plaque assessment.
• Reproducibility: Established that $SInorm$ measurements are highly reproducible over a 6-month period in the same patients.
• Discriminatory Power: Confirmed that $SInorm$ can reliably distinguish between the pathological plaque core and healthy remote vessel wall without complex modeling.
• Feasibility for AI: Proposed that $SInorm$ is a more pragmatic target for future Artificial Intelligence automation compared to $K_{trans}$, as it does not require arterial input function estimation.

4. Results

Demographics:

• 28 lesions analyzed (13 Colchicine, 15 Placebo).
• Median age: 72 years; 36% female.

Primary Findings:

• Plaque vs. Vessel Wall: Normalized peak signal intensity was significantly higher in the plaque core compared to the remote vessel wall at both baseline and follow-up ($p < 0.001$).
  • Baseline: Plaque (3.5) vs. Remote (2.1).
  • Follow-up: Plaque (3.2) vs. Remote (2.0).
• Reproducibility (Time): There was no significant difference in $SInorm$ measurements between baseline and 6-month follow-up for either the plaque core or remote vessel wall ($p > 0.80$).
  • Mean variance ($\Delta SInorm$) for plaque core: -0.1 ± 0.96.
  • Mean variance for remote vessel wall: 0 ± 0.55.
• Treatment Effect: There were no significant differences in the change of $SInorm$ between the Colchicine and Placebo groups ($p > 0.35$ for all metrics). Colchicine did not alter the contrast enhancement patterns over 6 months.
• Biomarker Correlation: No significant correlations were found between plaque core $SInorm$ and circulating inflammatory markers (hs-CRP, VCAM-1, IL-6, IL-18).

5. Significance and Limitations

Significance:

• Clinical Utility: The study validates a "black-box" free, simplified method to assess plaque neovascularization. This could facilitate broader clinical adoption of DCE-MRI for monitoring plaque vulnerability without the need for complex kinetic modeling software.
• Longitudinal Monitoring: The high reproducibility suggests $SInorm$ is suitable for tracking plaque evolution or response to therapy over time, provided the imaging protocol remains consistent.
• Future Directions: The authors suggest this metric is a stepping stone for AI-driven automated segmentation, as it bypasses the computational complexity of deriving arterial input functions.

Limitations:

• Lack of Validation: The metric is physiologically plausible but not yet validated against histological gold standards (neovascularization density).
• Sample Size & Exclusion: The study had a small sample size (n=28) and a high exclusion rate (37%) due to motion artifacts, which limits generalizability.
• No T1 Mapping: The absence of T1 mapping prevented absolute quantification of contrast concentration, relying instead on relative signal intensity.
• Inter-reader Reliability: While multiple readers reviewed the data, formal inter-reader reliability statistics were not calculated in this specific analysis.
• Negative Treatment Result: The lack of effect from colchicine may be due to the short duration (6 months), the specific dose, or the fact that $SInorm$ may not be sensitive enough to detect subtle anti-inflammatory changes in this specific cohort.

Conclusion:
The study concludes that muscle-normalized DCE-MRI signal intensity is a reproducible and distinct marker for carotid plaque neovascularization compared to healthy vessel walls. While it did not detect changes induced by colchicine in this cohort, the method offers a robust, simplified framework for future longitudinal studies and AI integration in atherosclerosis imaging.