How I Do It: Using 3D MRI to Identify High-risk Vascular Disease
Stroke and heart attack are major killers, not only in the Western world but, increasingly, throughout the developing industrial nations. The cost of these diseases is counted not only in lives lost, but in the health care costs incurred by those patients suffering the chronic manifestations of stroke and cardiac failure. While the acute and more chronic aspects of these diseases appear quite disconnected (for example, brain disease versus heart disease), and we are used to thinking of these conditions in terms of these end-organ manifestations, they do, in fact, share a common causality: vascular disease within the blood vessel supplying the end-organ tissue. In the case of stroke, this is in the carotid artery, and for heart attack, the coronary artery, so coronary thrombosis is synonymous with heart attack. While much medical attention has been focused on the acute diagnosis and treatment of both of these conditions to avoid or restrict end-organ damage, knowing that the acute disease is preceded by a protracted period of developing vascular disease provides the opportunity to detect and treat the vascular disease before it has the opportunity to cause downstream end-organ disease. The diagnostic challenge is, therefore, to identify biomarkers that discriminate between those people with and without high-risk vascular disease. Until recently, the most common and reliably used measure of vascular disease has been the degree of narrowing of the blood-vessel lumen (stenosis). It is known that as vascular disease progresses, the degree of stenosis will also increase. This has been readily measurable using techniques such as conventional angiography, ultrasound, and (more recently) CT and MR angiography. The diagnostic usefulness of stenosis, however, only becomes manifest once the disease is severe. Regrettably, symptom-provoking vascular disease is not confined to this advanced-disease group and often occurs in patients with lesser degrees of stenosis. In recent years, we have gained a better understanding of the biology of vascular disease, which accounts for this disconnection between stenosis and symptoms, but also gives us clues about newer markers of vulnerability. It is now clear that much of the early phase of vessel-wall disease, which causes wall thickening, results in an outward expansion with no significant luminal encroachment (positive remodeling); hence the hidden advance of significant and potentially dangerous disease. The pathobiology of vessel-wall disease has also been better defined. It is now apparent that atherosclerotic disease is fundamentally an inflammatory disease that drives a number of discrete processes that result in progression and destabilization of the atherosclerotic plaque, with eventual complication and rupture. One major manifestation of this chronic inflammatory process is neoangiogenic intraplaque hemorrhage (IPH). The plaque itself causes the growth of disorganized and fragile neovessels, much as a malignant tumor would, in order to sustain its increase in size. These vessels, however, are leaky and prone to rupture, thus allowing the egress of red blood cells into the substance of the vessel wall itself. This introduces two important elements directly into the vessel wall. First, as the red blood cells break down, their surface membranes are deposited deep within the plaque. These membranes are rich in cholesterol—higher in it than any other cell membrane in the body—and the add to the fatty lipid core at the heart of the atherosclerotic plaque, fueling the inflammatory process. Second, the hemoglobin within the red cell is released from its protecting membrane. Extracellular hemoglobin is highly inflammatory, and in the intraplaque environment, it is relatively hidden from the normally protective mechanisms present within the circulation. Intraplaque hemorrhage therefore provides two separate but additive elements that have the ability to drive, and even accelerate the atherosclerotic process. Within the plaque’s inflammatory milieu, therefore, there exist a number of potential biomarkers that signal increased risk. One of these is the presence of hemoglobin. Technique MRI has the ability to differentiate between tissues based on their molecular environment and their influence on local protons. High signal intensity on T1-weighted image acquisitions occurs when there is shortening of proton relaxation within the static field. This is most commonly brought about by the presence of an injected exogenous contrast agent, such as gadolinium. Other molecular species can also cause the same T1-shortening effects. One such species is generated during the breakdown process of hemoglobin. During this process, methemoglobin is produced, which causes endogenous T1 shortening and, therefore, a bright signal on a T1-weighted imaging sequence. Methemoglobin thus provides a natural endogenous contrast agent that identifies the presence of intraplaque hemorrhage. We have developed an MRI technique that exploits this contrast-generating property of the blood within high-risk atherosclerotic plaques. First, the sequence uses 3D acquisition; the voxels are isotropic, with 500-micron spatial resolution in all planes. This provides the opportunity to analyze the data in any plane, which is highly relevant in the investigation of disease of the carotid bifurcation and demonstrates complex morphology. Furthermore, the provision of 3D data allows accurate quantitative measurements not only of the total plaque volume, but also of the intraplaque hemorrhage. 3D imaging gives unique insights into the distribution of disease and occurrence of IPH, which cannot be achieved easily using the gold standard of histology. This will shed light on the location of elements such as IPH not only in the axial plane (deep or superficial), but also in the longitudinal plane (proximal or distal within the plaque). As these techniques gain greater acceptance in the clinical arena, comparisons between studies to assess the progression of disease will become a prerequisite. The 3D technique, again, is ideally suited for this, allowing ease of multivolume registration and overcoming the need for exact image repositioning during follow-up scanning.
A sagittal view (left) of the carotid bifurcation shows high signal intraplaque hemorrhage within atheromatous plaque (arrow). From the same 3D dataset, the axial view (right) demonstrates the distribution of intraplaque hemorrhage within the plaque (arrowheads).
A sagittal view (left) of the carotid bifurcation shows high signal intraplaque hemorrhage within atheromatous plaque (arrow). From the same 3D dataset, the axial view (right) demonstrates the distribution of intraplaque hemorrhage within the plaque (arrowheads). Two further technical features have been incorporated into the sequence in order to increase the conspicuity of the high signal generated by methemoglobin. The fat of the surrounding tissues within the neck is suppressed by applying specific prepulses to the 3D sequence that selectively remove the fat signal. This same prepulse and the multiple repetitive pulses that are required for the acquisition also result in the blood within the lumen of the vessel losing its signal, so that it, too, appears dark on the resulting image. The combination of these effects means that the only residual high signal is generated by the IPH, making detection easier. Recently published data have shown that this imaging technique, compared with histological sections, is highly accurate at detecting IPH, not only on a slice-by-slice basis, but within each slice. This has been achieved by acquiring 3D data at 500-micron resolution. Using a similar, though lower-resolution, technique, it has been possible to show, in a number of outcome trials, that the presence of MRI-detected IPH is a good predictor of patients who will undergo future neuroischemic events. Perhaps more important, the absence of MRI-detected IPH identifies patients at low risk for such events. This technique, therefore, has the potential to characterize atherosclerotic plaque better and to identify the high-risk patient. Evidence so far has come from small, single-center trials. The true value of this technique will only be demonstrated by large, multicenter trials. Through the Canadian Atherosclerosis Imaging Network, a large, multicenter trial is underway, with the aim of defining the utility of this newly recognized and MRI-visible biomarker for characterizing high-risk carotid-artery disease. Alan R. Moody, FRCR, is radiologist in chief, Sunnybrook Health Sciences Centre, Toronto, and professor, Department of Medical Imaging, University of Toronto. He and his coauthors described this technique in detail in an article published in Radiology (2008;249:259-267).