Journal of Nuclear Cardiology

, 16:860

Who gets the heart attack: noninvasive imaging markers of plaque instability

ASNC 2008 Keynote Address

DOI: 10.1007/s12350-009-9141-6

Cite this article as:
Narula, J. J. Nucl. Cardiol. (2009) 16: 860. doi:10.1007/s12350-009-9141-6

Acute coronary events result from thrombotic occlusion of the coronary artery.1-5 The occlusion is secondary to rupture of an atherosclerotic plaque in up to three-fourths of subjects; plaque erosion is seen in most of the remaining subjects who have died of an acute coronary event.2 Plaque rupture is associated with traditional risk factors, whereas erosion is generally associated with smoking and is commonly observed in women or younger subjects. Upon histopathological examination, the plaques that are prone to rupture and result in an acute event are almost always large.1-5 Such plaques also demonstrate large necrotic cores that occupy a large proportion of the plaque area. These necrotic cores are often associated with intraplaque neovascularization and hemorrhage, and adventitial vasa vasorum proliferation. 6,7 The necrotic cores are covered by rather attenuated fibrous cap, which are intensely inflamed. Therefore, an imaging strategy designed to identify rupture-prone plaques would target the enormity of plaque and necrotic core volumes, positive remodeling, and plaque inflammation.4,5 The morphological characteristics of such plaques can be identified by CT angiography of coronary arteries.4,8 Magnetic resonance has been employed for morphologic characterization of carotids and can also identify intraplaque hemorrhage.9,10 Contrast-enhanced ultrasound examination has revealed plaque neovascularization and vasa vasorum proliferation.11,12 Assessment of the fibrous cap thickness needs intravascular imaging techniques such as the optical coherence tomography.13 Plaque inflammation has been successfully assessed by PET imaging using fluorodeoxyglucose (FDG).14,15 Newer molecular imaging strategies have targeted upregulation of receptors on infiltrating monocyte or cytokine production.5

Assessment of Morphologic Characteristics of Plaque Instability

Almost all plaques that rupture occupy at least half of the total vascular area in cross section; up to 40% of them may occupy more than three-fourths of the vascular area.4 The necrotic cores are invariably voluminous as they usually occupy one-fourths of the plaque area, one-thirds of the plaque circumference, and extend up to a centimeter in longitudinal dimension. In spite of such an extensive plaque morphology, only about one half of the plaques result in critical luminal narrowing. Such plaques with luminal obstruction (even in the absence of an acute coronary event) should result in classic symptoms of inducible ischemia, stress tests should be positive, and patients are expected to seek medical attention and appropriately treated. Such are the patients which provide credibility to the stress tests as prognostic measures. On the other hand, approximately half the subjects, which have large plaques similar to the extent described earlier would induce <50% occlusion of the lumen, may not be symptomatic, defy a diagnosis on stress test, and may be erroneously assured of a better outcome. Regardless of the degree of luminal encroachment, the presence of large necrotic cores is associated with positive remodeling of the involved vascular segment. Such asymptomatic patients may present with acute coronary syndromes including unstable angina, acute infarction, or sudden death, as the first manifestation of their undiagnosed (or subclinical) ailment. An imaging technique is needed which may be able to precisely identify large, not-critically stenotic plaques associated with substantial lipid-rich enclosures and expansive remodeling (Figures 1, 2 and 3).
Figure 1

Histopathological characteristics of a ruptured plaque (A, B) and the plaque which is prone to disruption (C-E). Unlike the stable atherosclerotic plaques, which are rich in collagen and smooth muscle, ruptured plaques are usually large and are formed predominantly by a large necrotic core that is covered by a thin and inflamed fibrous cap. The site of plaque rupture exposes the thrombogenic necrotic core to luminal blood. The culprit plaques observed in the victims of acute coronary syndrome are usually significantly voluminous, even though they may not always be accompanied by significant luminal obstruction (A). The lumen and cross-sectional vessel area are outlined in B; the plaque area has been colored blue and the lumen in green. The disrupted plaques almost always occupy >50% cross-sectional area of the vessel (blue area divided by blue + green area); in up to half of the cases they may occupy >75% cross-sectional area. Disrupted plaques usually demonstrate large necrotic cores (A); necrotic core is outlined and colored red (B). The necrotic cores usually occupy >25% of the plaque area (red area divided by blue area) and show >120° circumferential involvement of the vessel arc in up to 75% of the culprit plaques. The necrotic cores are usually 9-mm long (range 2 to 22 mm). The larger the plaque area and the larger the necrotic core size, higher is the likelihood of plaque vulnerability. Although these plaques are large and reveal large necrotic contents, they usually expand outwards and may not always impinge on the lumen. As such, ante-mortem they may not have caused angina symptoms or produced stress-related myocardial ischemia on noninvasive testing, and may not even have demonstrated luminal involvement by invasive coronary angiography. Positive or outward remodeling is usually not seen in stable plaques which have small or no lipid cores, and their predominant fibrotic composition may result in negative remodeling or segmental shrinkage. Before the occurrence of an acute event, or the disruption of the fibrous cap, presence of the histopathologic signatures described above (C, D) should indicate the vulnerability of plaque to rupture. These lesions include a large plaque (blue) area and a large necrotic core (red, contains abundant cholesterol crystals seen as residual clefts), covered by a significantly attenuated fibrous cap. The clinical detection of unstable plaques would therefore need demonstration of large plaque and necrotic core volumes and the expansive remodeling of the plaque carrying vascular segment. These morphological characteristics of plaque vulnerability are shown in the schematic representation (E). Additional characteristics of vulnerability shown in the scheme are described below (modified from references1,4)

Figure 2

Necrotic core formation in the vulnerable plaques. Worsening hypoxia in the enlarging plaques promotes vasa vasorum proliferation (A) and intraplaque neovascularization (B). A manifold increase in vasa vasorum proliferation has been reported with culprit compared to stable and occlusive lesions. Microvessels that perforate from the adventitial layer to the neointima are leaky (C) and appear to allow extravasation of RBCs into the plaque (D); the plaque hemorrhages are very common in the culprit plaques. The extent of iron deposits (E) in the plaque, and that of RBC membrane-associated protein, glycophorin A (F), is directly proportional to the size of the necrotic core. The cholesterol content of RBC membranes is high, contributes to the necrotic core and perpetuates plaque inflammation and macrophage infiltration (G). Thus, it may be important to identify intraplaque hemorrhage and adventitial vasa vasorum for clinical detection of unstable plaques (modified from references4,19)

Figure 3

The thin fibrous caps of the unstable plaques are markedly inflamed with monocyte-macrophage infiltration. The area enclosed by the yellow rectangle (A) at the site of plaque rupture (light orange colored disrupted fibrous cap has exposed the thrombogenic necrotic core or NC, leading to thrombotic formation (Th) in the coronary lumen) is magnified in B. Immunohistochemical characterization of this area demonstrates abundant macrophages (B, brown and D, red). Analysis of attenuated fibrous caps demonstrates that macrophages are the most dominant cellular population in ruptured and vulnerable plaques, whereas smooth cells are dominant in stable atherosclerotic lesions (C). A large proportion of these macrophages (D, F) are in the process cell death by apoptosis (brown nuclei represent DNA fragmentation by TUNEL staining in D); these macrophages also demonstrate upregulation and activation of caspase 1 (E, brown, CASP1). Identification of inflammatory component and apoptosis can be best exploited by molecular imaging (modified from reference4)

CT angiography, which has been predominantly investigated for the lumen narrowing by the plaque impingement in comparison to the invasive coronary angiography, has a distinct advantage of simultaneous demonstration of the plaque and necrotic core extent and the type of segmental remodeling.8 A comparison of disrupted plaques in patients who had experienced an acute coronary syndrome (ACS) with plaques from patients undergoing coronary intervention for stable angina by CT angiography showed positive vascular remodeling (PR; external vessel wall diameter of >110% compared to a normal proximal or distal segment) and low attenuation plaque (LAP; <30 Hounsefield Units [HU]). These two features demonstrate a high accuracy for identifying culprit lesions. The interpretation of such features on CT angiographic investigation, however, is not without limitations. Suboptimal resolution does not allow precise definition of the vascular boundary, and the assessment of the extent of PR may be over or underestimated. Similarly, the LAP is based on the assessment of the HU and various imaging/technical parameters may seriously influence the soft plaque measurements. As such, various investigators have suggested different cut-off points to define soft plaques. Our comparison of IVUS and CT angiography had demonstrated that a majority of echo-lucent plaque cores could be identified by keeping the upper limit of 30 HU.16 Of interest, spotty calcification was more commonly associated with culprit lesions and large calcific plates with the stable plaques.

These two CT angiographic features (i.e., PR and LAP) in the subjects who have not experienced an acute coronary event predict high likelihood of plaque rupture over time.17 In a study of more than 10,000 coronary arterial segments from >1000 patients, presence of the 2-feature positive plaques (PR + LAP) demonstrated a 22% likelihood of an ACS over a 2-year follow-up period compared to <0.5% in the patients with 2-feature negative plaques. Greater the remodeling and larger the plaque area, higher was the likelihood of plaque rupture and earlier occurred the acute event. Patients with no plaques did not suffer from any adverse events. Interventions with statins reduce the plaque and LAP volumes considerably, without significantly influencing the luminal volume.18 The plaques in fact demonstrate negative remodeling after statin treatment and the reduction in the plaque volume was related to the reduction in LAP area (Figure 4).
Figure 4

The morphologic characteristics of a culprit lesion as defined by computed tomography angiographic (CTA) imaging. Multislice CT angiogram in a culprit coronary lesion (D) shows positive remodeling (E, yellow arrows) and low attenuation plaque (E, red arrowheads). Invasive and CT coronary angiograms confirm that the lesion in this 57-year-old patient with unstable angina may not be significantly flow-limiting. Stable lesion (left), which is severely obstructive by invasive (A) and CT (B) angiogram, is not remodeled and shows intermediate attenuation (green arrow) plaque (B). Comparison of CTA with intravascular ultrasound examination has revealed that low attenuation (<30 HU) corresponds to soft plaques and intermediate attenuation (30 to 150 HU) to fibrous plaques. It can be proposed that presence of positive remodeling and low attenuation plaques in an asymptomatic person should be predictive of higher likelihood of major cardiovascular event (compare with Figure 7A, B) (modified from reference8)

Although not yet ready for the primetime, I believe that when reporting a CT angiogram, we should not limit ourselves only to the extent of luminal stenosis or calcified vs. noncalcified plaques, but insist to define the vessel wall characteristics in terms of plaque magnitude and consistency. It requires a cultural change in the way we deal with the coronary disease. However, it should also be realized that heavy calcification in many subjects may preclude such as a judgment and that plaque erosions are not amenable to CT characterization. Nonetheless, the subjects harboring high-risk plaques are candidates of intense and aggressive risk factor reduction including pharmacological intervention with currently available strategies.

Adventitial Vasa Vasorum Proliferation, Intraplaque Neovascularization, and Hemorrhage

Neovascularization of the atherosclerotic plaques is closely associated with the necrotic cores. These nascent vessels are fragile and allow convenient extravasation of erythrocytes and macromolecules.7 RBC membrane is one of the richest sources of free cholesterol, and leaking RBCs or intraplaque hemorrhage contributes substantially to the necrotic core size. Greater the deposition of RBC membrane or the iron deposits in the plaque, larger is the necrotic core size. As such, culprit lesions demonstrate substantially greater density of neovascularization and RBC membrane deposits.19 The magnetic resonance images of the carotid plaques have been compared with the endarterectomy specimens obtained during surgical procedures. High T1-weighted densities and low T2* values closely correlate with the extent of intraplaque hemorrhage.9,10 It has been observed that greater the hemorrhage, larger are the plaque volumes and that the symptomatic carotid disease is almost always associated with hemorrhage.9 It is also observed that the statin administration helps reduce plaque volumes only before the plaques are complicated by hemorrhage. Although magnetic resonance imaging of coronary plaques is not yet feasible, the RBC membrane cholesterol has been shown to be higher in patients undergoing coronary interventions for an acute coronary event than the stable disease, even though these subjects may have similar circulating cholesterol levels.20

The intraplaque hemorrhage and neovascularization are accompanied by greater proliferation of adventitial vasa vasorum.6 The evolving plaque and the increasing vessel wall hypoxia necessitate vasa vasorum proliferation. In the experimental settings, vasa vasorum abundance was seen in hypercholesterolemic state that resolved with statin therapy. Contrast-enhanced ultrasound inquiry allows identification of adventitial vascularity and intimal demarcation in symptomatic carotid disease, which has been pathologically verified in endarterectomy specimens (Figure 5).
Figure 5

Imaging for intraplaque hemorrhage and adventitial vasa vasorum: Magnetic resonance imaging of carotid arteries has successfully identified iron load associated with intraplaque hemorrhage by high T1-weighted images (A). These findings are closely correlated with histological evidence of plaque hemorrhage in resected carotid endarterectomy specimens (B). MR imaging of coronary vessels for identification of plaque hemorrhage has not been reported. On the other hand, ultrasound microbubble-enhanced visualization of intimal-medial thickness of the anterior and posterior carotid artery walls (C) suggests vasa vasorum proliferation and correlates with the evidence of neovascularization in the endarterectomy specimens (D). The contrast enhancement which correlates with histological density of neovessels and ultrasonic plaque echolucency (also a marker of high-risk lesions) is unrelated to the degree of stenosis. Successful visualization of coronary vasa vasorum has not yet been reported (modified from references9,11)

Molecular Imaging for Plaque Inflammation

Plaque inflammation is an important constituent of the plaque instability.5 The cells of monocyte-macrophage origin are seen abundantly in the fibrous caps as also around the necrotic core. These cells develop various receptors for integrins and cytokines when they interact with the injured endothelial cells and negotiate through to the subendothelial space. The macrophages in neointima develop scavenger receptors to remove modified lipoprotein cholesterol particles. Macrophages eventually succumb to necrotic and apoptotic cell death process and add to the expanding necrotic core.21 Molecular imaging has been successfully employed for targeting the receptor upregulation, macrophage metabolism, and cell death as a marker of plaque inflammation. Fluorine-18-labeled FDG and annexin A5 (AA5) have been used clinically.14,15

FDG imaging has been commonly employed for the differentiation of malignant from benign tumors on the premise that the cells with high respiratory burst will preferentially retain radiolabeled glucose. Coincidental uptake of FDG in large and medium-sized arteries (including coronary arteries) has been anecdotally reported, which has been traced to activated macrophages. A prospective study14 was recently reported wherein myocardial FDG uptake was prevented by administration of low carbohydrate diet and beta blockers, and coronary FDG was localized by concomitant PET imaging and CT angiography after coronary stent placement for either acute coronary events or stable symptomatic coronary artery disease. FDG uptake was only seen in stented segments after an acute event. This study demonstrated the feasibility of targeting coronary inflammation; FDG uptake was also observed in presumably inflamed nonstented coronary lesions and aortic roots (Figure 6).
Figure 6

Imaging of inflammation by targeting of activated macrophages. Imaging with F-18-labeled FDG, which has been commonly used to differentiate malignant and benign tumors, often shows coincidental FDG uptake in various large vessels. One such example of a patient undergoing FDG imaging for the exclusion of lung malignancy is shown here (A, B). The thoracic CT image (A) shows calcific coronary vessel (A), but FDG uptake in the noncalcific part of the vessel (B). FDG uptake in coronary arteries has also been shown in a prospectively designed study that took advantage of recently implanted stents for the localization of radiotracer uptake in a coronary vessel. While stented segments in stable coronary disease did not accumulate FDG, stents (C) in acute coronary syndromes showed intense FDG uptake (D). Commonly observed FDG uptake in the myocardium, which is likely to interfere with the interpretation of coronary vascular uptake, was cleverly obviated by effective beta blockade. The successful demonstration of molecular imaging of coronary arteries has instilled substantial enthusiasm in imaging community and has highlighted the feasibility of characterization of hitherto forbidden target (Figure 6A, B: courtesy H. William Strauss, MD, New York; Figure 7A, B: Ahmad Tawakol, MD, Boston)

Since apoptosis of macrophages is commonly seen in the high-risk plaques and radiolabeled AA5 can selectively bind to apoptotic cells, annexin positivity has been proposed as a marker of instability.15 Positive Tc-99m-AA5 uptake in a patient with carotid artery disease is traced to macrophages in the endarterectomy specimen, whereas the patient with negative scan had smooth muscle-rich lesion. In experimental settings, AA5 uptake correlates closely with macrophage density and the magnitude of apoptosis in atherosclerotic plaques. Diet modification and statin treatment reduce AA5 uptake.22,23 The experimental studies have demonstrated AA5 localization in aortic lesions in rabbits22 and mice,24 as also in coronary vessels in pigs.25

Monocyte chemoattractant protein (MCP-1) plays an important role in recruitment of monocytes to the atherosclerotic plaque, and MCP-1 receptors on macrophages have been targeted with radiolabeled MCP-1.26 The increasing severity of atherosclerosis in a rabbit model is associated with increasing uptake of MCP-1. Macrophages infiltration in atherosclerotic plaque in addition to lipid scavenging leads to production of cytokine and proteases. Radiolabeled MMP has been used for the MMP activity in atherosclerotic lesions.27 Increasing severity of atherosclerosis is associated with increased uptake of MPI. Extent of MMP activity in the lesions is correlated to the fibrous cap attenuation and positive remodeling, both of which are obligatory components of plaque instability (Figures 7 and 8).
Figure 7

Future directions: Plaque morphology or plaque inflammation in asymptomatic subjects? Positively remodeled and low attenuation plaques by CTA (A) [similar characteristics as described above after acute coronary syndrome] were observed proximal to the first septal branch in a 57-year-old male who was ruled out for acute coronary syndrome. He returned with acute anterior myocardial infarction 6 months later resulting from the rupture of the same plaque (B) identified during coronary angiography. If positively remodeled soft plaques are identified incidentally, up to one-fourth of them may develop acute coronary syndrome during a 2-year follow-up period. The plaques that are in imminent danger of rupture show large plaque volume, large low attenuation plaque volume with significant expansive remodeling. Plaques with stable characteristics are associated with <0.5% likelihood of an acute cardiac event. Normal coronary vessels exclude the possibility of ensuing major cardiac events. Although no follow-up data on FDG uptake-verified inflamed plaques is available, an anecdotal report of incidental carotid FDG uptake is informative. A 79-year-old male was ruled out for lung malignancy after FDG study but showed marked FDG uptake in the carotid region. The patient was mildly hypertensive and showed modest stenotic lesion upon carotid ultrasound. He was admitted 2 days later with cerebrovascular accident potentially attributed to the inflamed carotid lesion. These two examples demonstrate the predictive usefulness of the assessment of high-risk plaques, which may be conducive to primary prevention (modified from references5,17)

Figure 8

Future directions (Contd.): Novel molecules for targeting plaque inflammation. The thin fibrous caps of the ruptured or vulnerable plaques are markedly inflamed with monocyte-macrophage infiltration and successful imaging of coronary inflammation as proposed in the Figure 6 offers an optimistic note. The molecular imaging techniques have targeted upregulation of surface molecules or secreted products which are uniquely expressed by the inflammatory cells associated with unstable plaques. Most of the molecular imaging studies have been conducted in the experimental models of atherosclerosis or large immobile vessels (such as carotid arteries, abdominal aorta, and iliofemoral vessels). Such studies have only provided the proof of principle. This figure summarizes the results of three studies undertaken in the rabbit model of atherosclerosis. Technetium-99m-labeled monocyte chemoattractant protein (MCP-1, that targets CCR2 receptor expression on infiltrating monocytes, A & B), matrix metalloproteinase inhibitor (MPI, that targets active metalloproteinases released by macrophages and other cells; C & D), and annexin A5 (AA5, that targets macrophages and other cells dying by the process of apoptosis; E & F) were administered intravenously. The percent injected dose per gram uptake of the radiotracers is 6- to 10-fold higher in the foam cell-rich atherosclerotic lesion compared to the normal aortic segments (A, C, E). As expected, MCP-1 uptake correlates closely with pathologically verified prevalence of macrophages in the plaque (B), MPI uptake correlates with MMP-2 and 9, and AA5 correlates with the extent of apoptosis in the plaque (F). As the molecular imaging technology matures, many of these or similar strategies will become applicable for the characterization of various stages of plaque inflammation and their prognostic significance (modified from references22,26,27)


It will be mandatory to develop worthy diagnostic and therapeutic strategies targeted at the prevention of plaque rupture. If identified correctly, such plaques would be treated with aggressive statin and antiplatelet therapy.28 It is possible that newer antiinflammatory agents will be developed. It is also possible that novel stents would become available such as those which are bioabsorbable and targeted at inflammation or neovascularization. Although defining the plaque characteristics by an imaging technique is within the realm of feasibility, it will be important to identify the group of high-risk asymptomatic subjects which will benefit most by an imaging procedure.

Copyright information

© American Society of Nuclear Cardiology 2009

Authors and Affiliations

  1. 1.Memorial Heart & Vascular InstituteLong BeachUSA
  2. 2.University of California, IrvineIrvineUSA

Personalised recommendations