Activity Transcript

Bruce A. Wasserman, MD: Hello everyone. My name is Bruce Wasserman, and I want to welcome you to this educational program. I'd like to start by thanking Medscape for supporting this program. I've organized this talk as a case-based approach to vessel wall imaging, and we'll start by reviewing the role of vessel wall imaging for identifying and characterizing atherosclerotic plaque; in particular, vulnerable plaque, that is, plaque vulnerable to disruption leading to the clinical events. We'll start by talking about carotid plaque imaging, which the techniques were initially developed for, and then we'll extend these techniques to intracranial vessels and review applications and commonly encountered pitfalls specific to each vascular system.

So, we'll start with why vessel imaging, why not angiography? And what value does wall imaging add to angiography? Well, we know that when plaque forms early on, it forms eccentrically as shown here, and this starts to narrow the lumen and blood velocity starts to increase. Vessels can respond to this by vasodilating. This vasodilatation or vascular remodeling normalizes the lumen. If we were to look at the luminal projection or angiogram of a vessel with early disease, it'll be identical to that of a normal vessel until we get to a critical mass of plaque formation beyond which the vessel can no longer accommodate its formation and we start to see stenosis that is angiographically detectable. This highlights the importance of having a technique like vessel imaging that can identify these angiographically occult lesions.

How do we detect plaque using vessel wall imaging? The most common approach to vessel wall imaging is black-blood magnetic resonance imaging (MRI), which is what I will focus on in this talk. With black-blood MRI we suppress signal in the lumen, and it highlights plaque building up in the lumen wall. Now watch what happens when we give intravenous gadolinium. We get much better conspicuity of the central lipid core because of preferential enhancement of the surrounding fibrous tissue improving its delineation.

Once the plaque is detected, how do we characterize its composition? This is one of the earlier examples that we published of a carotid plaque that was imaged using a contrast-enhanced black-blood MRI technique in a patient who went on to have surgery. Here's the endarterectomy specimen. We see the lumen is nicely suppressed with this black-blood technique. We see, adjacent to the lumen, this triangular enhancing tissue, which corresponds with this triangular fibrous cap on the histology. This crescent of gray tissue corresponds with this crescent of lipid core, and this dark signal corresponds with blue-staining calcification.

There are several features of carotid plaque that can be identified by vessel wall MRI and have been implicated in stroke risk. These include a thin or ruptured fibrous cap, a large lipid core, the presence of intraplaque hemorrhage, ulceration, having evidence of a prior rupture or the so-called culprit plaque, and enhancement that is indicative of inflammation.

With this background, here is our first carotid case. This is a 57-year-old man who came in with left hemisphere ischemic events that were ipsilateral to this carotid artery. We see that he has a low-grade stenosis, certainly not enough to warrant surgery by North American Symptomatic Carotid Endarterectomy Trial (NASCET) criteria, and yet he continued to have events while on medical management. So we were asked to acquire vessel wall imaging. We see on our long axis view that he has this very large plaque that's along the outer wall of the bulb. Here's the internal carotid artery (ICA) lumen in black and the common carotid artery. This plaque is certainly larger than we would've anticipated based on the degree of stenosis. On short axis imaging, we see on the first precontrast T1-weighted images a very hyperintense plaque against this black lumen. This is specific for intraplaque hemorrhage, which is one of the vulnerable features that we discussed. On our postcontrast imaging, we see this linear enhancement along the luminal surface of the plaque that appears to be disrupted. So, this looks like a disrupted fibrous cap. This patient continued to have ischemic events despite being on maximum medical management, and having identified these vulnerable features, the decision was made to refer him for endarterectomy.

Here the specimen shows high concentrations of intraplaque hemorrhage, as we suspected on the MRI. We see here a disruption of the fibrous cap with the cap shown by these short green arrows and the disruption site shown by the long blue arrow. But the most compelling feature of this specimen is the thrombus sitting at the cap disruption site. This was a stroke that was waiting to happen, and it highlights how we can use this technique to prevent strokes by guiding management based on vulnerable feature identification without relying exclusively on stenosis.

Our next case is a 61-year-old man who presented with a 2-year history of left hemisphere strokes. His left carotid artery shows some irregularity along the outer wall of his carotid bulb but certainly nothing that looks worrisome. We were asked to acquire wall imaging because of the continued events on this side. We see that he's got this very large plaque along the outer wall that's causing this irregularity. On our short axis imaging, we see a very big plaque growing around this common carotid artery lumen. As we're coming up into the bifurcation, here's the ICA lumen and the external carotid artery (ECA) lumen with a very big plaque that's almost exophytically compressing adjacent structures such as the jugular vein. But there's relative preservation of the ICA lumen because of this exaggerated vascular remodeling. I had acquired wall imaging before and after one of his strokes, and we see in the poststroke study, he's got this piece of tissue projecting into the lumen that wasn't apparent on the prestroke study. I acquired this with contrast, so we can see that he's got this linear enhancement along the luminal surface of the plaque that appears to be disrupted. This looks like another disrupted fibrous cap.

We now have compelling evidence that this plaque is temporally linked to his stroke events, suggesting that it is a culprit lesion. It's important to understand that the risk of stroke in the territory of a plaque that has previously ruptured is greatly increased. If you can identify a culprit plaque, you've identified a plaque at risk for future events. So, we had recommended surgery for this patient. Here's the post-endarterectomy result; a nice, wide open carotid bulb. Here's the specimen, and on our serous red stain, we see a disrupted fibrous cap as we saw on the vessel wall imaging. This area of hyperintense signal on the MRI corresponds with this area of blood products with inflammatory cells. This technique is really changing how we assess stroke risk, especially for lesions that have little hemodynamic impact on the lumen.

It's important to understand that the reason plaque typically forms initially along the outer wall of the carotid bulb is because of low shear stresses created by the recirculating blood flow currents in the bulb. This causes the endothelial cells to be randomly organized with more endothelial dysfunction leading to plaque formation, first along the outer wall, but eventually it will progress to involve other parts of the vessel. It's also important to understand that the same recirculating currents that lead to plaque formation here make it difficult to suppress the signal of luminal blood flow and lead to flow artifacts. When we're imaging the carotid bulb and looking for plaque, we want to focus on the outer wall. We might see something that looks like this. But we could also see a carotid artery that looks like this, and what we see on the carotid on the left is actually flow artifact, whereas this is a true plaque. How do I know this?

Well, if we rely on time-of-flight or noncontrast magnetic resonance angiography (MRA) for this left carotid artery, we'll be misled because the time-of-flight MRA is susceptible to the same artifacts that we see on vessel wall imaging. So, we will erroneously think that this luminal signal is a real plaque. However, if we acquire a contrast-enhanced MRA, the gadolinium nicely delineates the luminal boundary and confirms this to be an artifact.

So now we are going to extend these techniques to the intracranial circulation, which can be very challenging to image, because intracranial vessels are inherently small and torturous. So, we shift from 2D imaging, which I had shown you for the carotids, to 3D techniques, which allow for higher resolutions, so we can resolve these small vessels and also more efficiently capture these inherently torturous structures. For example, here is a high-grade atherosclerotic plaque involving the mid basilar artery on magnetic resonance (MR) angiography. We can acquire a 3D black-blood volume, and this is acquired at 400 μm isotropic resolution. Because it's isotropic, we can reconstruct a 2D plane through the long axis of the basilar artery precontrast. We can repeat this process postcontrast and reconstruct a short axis view. This 3D isotropic technique gives us very robust postprocessing capabilities and allows us to capture a large volume of the intracranial circulation. We don't need to know where the lesions are at the time of the exam, which can be very helpful.

There are a number of reports on using vessel imaging to characterize intracranial atherosclerotic disease or ICAD, and more recently vasospasm or RCVS and vasculitis. They all rely on varying degrees of eccentricity and enhancement of the vessel wall. However, we don't want to rely exclusively on wall imaging features to come to a diagnosis. We need to interpret our vessel wall images in the context of the vessel involved and its angiographic appearance. For example, we wouldn't diagnose an atherosclerotic plaque in a small arterial perforator or in a vein. We need to integrate the clinical and laboratory data with our imaging, because without this, we lack the precision to distinguish these vascular diseases.

I'm going to go through the use of wall imaging for diagnosing these vasculopathies, beginning with ICAD. These lesions are eccentric, and because these lesions are typically too small to characterize their internal components, we see heterogeneous enhancement, which represents a mixture of its components imaged at inadequate spatial resolution. The enhancement can range anywhere from none at all to high-grade enhancement. We've shown in the past that the degree of enhancement correlates quite strongly with the likelihood that the plaque is a culprit lesion or is responsible for a recent ischemic event, which can be very useful clinically.

Here's a patient who presented with multifocal narrowing on cerebral angiography. Clinically, he had a headache. We see the vessel wall imaging shows thin concentric thickening with little to no enhancement on these postcontrast reconstructed images. This is typical for RCVS. More recently, enhancement has been reported as being moderate to even marked with vasospasm. So, this feature lacks specificity, and the enhancement really depends on the etiology that underlies the vasospasm. So, the lack of enhancement is really more useful for ruling in RCVS since it's often difficult to rule it out when you do see enhancement.

Here's a 44-year-old man who presented with an anaplastic oligodendroglioma that was resected and treated with chemoradiation. He also had cardiovascular risk factors in his complex history. So, it wasn't clear when he came back with these 3 focal acute infarcts on diffusion imaging what the etiology was. We started with an MR angiogram to look at his intracranial vessels and we see that they're fairly normal. We were then asked to acquire a vessel wall image. This is the top slice of a 3D black-blood postcontrast volume. I want to draw your attention to this ring-enhancing lesion. As we scroll, we see that this enhancement is surrounding these 3 small branch vessels. If I were to characterize this kind of enhancement, I would describe it as concentric, diffuse, and periadventitial, which is a term that I use when the enhancement clearly extends beyond the expected location for the wall of the vessel. In my experience, it's a pretty good marker of inflammation, and there's now literature on this. So, we continue to scroll, and we see that this is actually an insular branch that's dividing into 2 opercular branches, or M3 branches, and the more anterior branch shows severe narrowing.

You can see that tiny black dot represents the residual lumen of this narrowed branch. We continue to scroll and eventually the enhancement goes away. So, this is localized vessel wall enhancement that's indicative of inflammation, presumably radiation-induced vasculitis. The question was why didn't we see the stenosis of this opercular branch on our MRA? Well, we look back and we don't see it again. So I pulled up the vessel wall slice that showed it, and it turns out it was above the level of our MRA. I acquired a higher-level MRA, we then oblique it, subtract out the left side, so we're only focused on the right middle cerebral artery (MCA) branches, we zoom in, and there is the opercular branch that was severely narrowed at its origin. So, we see how vessel wall imaging can identify inflammation and serve as a complimentary tool to MRA. In fact, in this case, it was the vessel wall scan that drew our attention to the abnormality on the MRA.

Here's another case. This is a 33-year-old man who presented with these acute ischemic strokes on diffusion imaging in the posterior fossa or in posterior circulation in the supratentorial brain. He had more chronic appearing infarcts, and it wasn't clear whether he had a vasculitis because he wasn't responding, as we would've expected, to steroids. We were asked to acquire wall imaging, and here we see a lenticular stripe branch that is surrounded by this periadventitial enhancement, which is fairly good for an inflammatory process. As I scroll along the coronal images, here's a cortical vein that appears inflamed, abnormal anterior cerebral artery (ACA) wall enhancement, and another cortical vein here, and a cortical vein here that both appear inflamed. Here is an arterial branch that shows concentric segmental circumferential enhancement, which is fairly typical, again, for a vasculitic process.

We decided to take this a step further and biopsy this vessel, which I thought to be a vein based on imaging that I'm not showing. We sent the patient to the operating room, and using vessel wall imaging for surgical navigation, the vessel was excised. We see here that there was lymphocytic infiltration throughout the wall of this vessel. This was primary angiitis of the central nervous system (CNS), or PACNS. In this case, we also did a brain biopsy, which was normal. This is really an important point to highlight, because as many of you know the yield on biopsies for vasculitis is quite low. The reason for that is that there are often skip lesions with vasculitis. It's very difficult to know if you have a target that is inflamed. But now we have a fairly reliable technique that can identify inflamed targets. We've had great success using this to guide our biopsies, and it's really changed how we approach the diagnosis of CNS vasculitis.

As I've shown you, we rely on eccentricity enhancement for diagnosing these vasculopathies, but these features can also be unreliable predictors of disease. This brings us to pitfalls. Here is a V4 segment of the right vertebral artery, and it shows dense circumferential enhancement. The reason for this is that as vessels enter the intracranial compartment, they drag the extracranial vasa vasorum intracranially to involve the proximal portions of intracranial vessels and this contributes to this enhancement. I point this out because not infrequently this is misdiagnosed as vasculitis or sometimes atherosclerotic plaque, even in journal articles.

Here's an unfortunate woman who presented with bilateral cerebral hemisphere infarcts due to bilateral supraclinoid ICA stenosis. These are consecutive images on a coronal postcontrast black-blood volume that shows the carotid arteries as they emerge from the cavernous sinus. We see that they're moderately to severely narrowed by this eccentric, enhanced wall thickening, continuing to the communicating segments, which appear relatively normal. Tragically, she died, and we were able to study her vessels. We see that the communicating ICA in the right side, which we thought was normal, showed thinning of the media and smooth muscle cell degeneration. So, this was a medial wall fibromuscular dysplasia. The ICA at the site of enhancement shows, again, thinning of the media from this fibromuscular dysplasia (FMD), but within the lumen, we see this organized thrombus that's tightly adherent to this fibrointimal thickening that accounts for the enhancement. What we thought was pathology, this enhancement, was really just the consequence to it. It was a propagating thrombus related to dissection, and the underlying dysplasia was undetectable by vessel wall imaging because it neither enhanced nor did it cause abnormal wall thickening.

Here is a woman who presented with these diffuse white matter infarcts bilaterally, and we were asked to acquire wall imaging because there was concern for vasculopathy. We see this linear configuration of enhancement, which could be the signal of a small-vessel vasculitis. However, when we look back at the diffusion scan, we see that the infarcts, some of them have a linear configuration as well. We know that subacute infarcts can enhance. So, this is not necessarily vascular inflammation. The patient underwent a skin biopsy, and this turned out to be intravascular lymphoma, which can look identical to small-vessel CNS vasculitis on vessel wall imaging.

There are a variety of false positives that we need to be aware of. For example, these are reconstructed images from a 3D contrast-enhanced black-blood volume. Coronal reconstruction shows this M2 branch with enhancement along its inferior and lateral wall. This could easily be an atherosclerotic plaque. However, we acquire a higher-resolution 2D vessel wall scan, and we see that this is, in fact, a small vessel running alongside the M2 branch that's causing this enhancement. On our contrast-enhanced MR venogram, we see this is a small vein. This highlights why it's important to have adequate resolution so that we can resolve these small structures. But even more importantly, veins can enhance normally on 3D vessel wall imaging. In this case, it's because likely those slow flow, it's just inadequately suppressed. But we also see that the wall of veins can also enhance normally on wall imaging. That's because of the low luminal PO2 in veins that cause the veins to have a higher density vasa vasorum than arteries of comparable size, so you will see enhancement in the walls of these veins.

This is a common reason for misinterpretation on vessel wall imaging, to misinterpret this as an enhancing artery. Here is a 3D black-blood volume in a coronal acquisition that is capturing this MCA branch in its short axis. We see circumferential enhancement that looks like vasculitis. I repeat the same exact acquisition, but I acquire it in an axial plane rather than in the coronal plane. I reconstruct a coronal view, so it matches the first view and all the enhancement in this MCA goes away. That's because this is all artifact. The reason for that is that flow suppression is most effective when the vessel orientation is in the readout frequency-encoding direction. If not, you will get these flow artifacts, which is why we acquire our 3D volumes in multiple orientations to maximize flow suppression of vessels running in multiple directions. We should also be able to confirm what we think is pathology here on the angiogram. If these were real, we should be seeing stenosis at the corresponding location on the MRA, or computed tomography angiography (CTA), or whatever it is that we have available.

This is another example of inadequate flow suppression, in this case, in areas of disturbed flow on pre- and postcontrast 3D black-blood volumes. This is the long axis and then the short axis view of the juxtacellular ICA segment showing this presumed filling defect that could be an atherosclerotic plaque along the inferior aspect of the lumen. However, this is a flow artifact. How do I know that for sure? I can do a contrast-enhanced MRA, and the gadolinium fills in the lumen quite nicely confirming this to be artifactual. A word of caution, time-of-flight MRA can mimic these artifacts and may not be helpful to corroborate plaque presence, just as we saw with our carotid case earlier.

So in summary, we've seen that vessel wall imaging can be a powerful tool for imaging neurovascular pathologies. We can identify these angiographically occult lesions because of the vessel's ability to remodel to accommodate plaque formation and preserve the lumen. We can evaluate lesions that are low grade. Carotid lesions that are low grade and might not otherwise be considered high risk by NASCET criteria. We can identify vulnerable features of plaque, which I've listed some here, and we can identify the culprit lesion, which, in my opinion, is the most important clinical application of carotid vessel MRI today. We can also determine the reasons for intracranial narrowing and add specificity to cerebral angiography. We talked about the techniques that we've used, or we can use, for these applications, and the value of contrast for characterizing plaque, and for distinguishing intracranial vasculopathies.

We need to be aware of pitfalls. We need adequate flow- suppression resolution to avoid these, and I went through some of the sequences as well. Using multiple orientations for our 3D black-blood volumes in order to maximize flow suppression for vessels flowing in different directions. It's important to understand that enhancement does not always mean inflammation. There are many reasons for abnormal enhancement. I only touched on a few here, but it is important to recognize this. Also, we need to interpret our images in the context of the angiographic appearance and the vessel that's involved. It's very important to integrate the clinical and laboratory data with our vessel imaging data. Most important is to have adequate training for interpretation, to understand how to recognize pitfalls especially, and for accurate interpretation of these studies.

I want to thank you for participating in this activity. Please continue on to answer the questions that follow and complete the evaluation.

This transcript has been edited for style and clarity.