Target identification for the diagnosis and intervention of vulnerable atherosclerotic plaques beyond ¹⁸F-fluorodeoxyglucose positron emission tomography imaging: promising tracers on the horizon

Vascular disease is still the leading cause of morbidity and mortality in the Western world, and the primary cause of myocardial infarction, stroke, and ischaemia. Atherosclerosis is a chronic disease of the large arteries that remains asymptomatic for decades. Arterial wall changes in the context of atherosclerosis start in childhood and vascular lesions enlarge with age and typically become symptomatic at 50–60 years of age. The pathobiology of atherosclerosis is complex and lesions can be generally divided into stable and unstable (also termed vulnerable) plaques. While stable plaques can occlude the lumen of the arterial vessel over time, they can develop plaque erosion. Moreover, plaque rupture is the dominant initiating event, responsible for 60–70% of acute coronary syndromes (ACS). It is therefore important to develop tools that discriminate between plaques that become vulnerable and plaques that remain stable but display erosion. State-of-the-art imaging using novel probes identifying molecular and cellular processes associated with plaque rupture have the potential for better patient stratification and personalized strategies.

The process of atherogenesis starts with activation of vascular endothelial cells, either by vascular shear stress [1] or local processes [2]. Also platelets can deposit so-called footprints that direct leucocytes to the vasculature [3]. At the same time as leucocyte infiltration, a heterogeneous population of vascular smooth muscle cells (VSMCs) is established in the subintimal space [4]. It has been shown that VSMCs are important in health and disease [5]. The pool of intimal VSMCs in the atherosclerotic lesion might arise from the infiltration of haematopoietic stem cells that differentiate locally into VSMCs [6]. However, more recently, VSMC lineage tracing in apoE-null mice has shown that a majority of atherosclerotic lesion cells are derived from VSMCs from the vasculature [7]. The accumulation of low-density lipoproteins (LDL), that become oxidized, promotes macrophages to take up these remnants via their scavenger receptor. These foam cells are the starting point of the so-called fatty streak. In response to these proinflammatory macrophages, more VSMCs migrate into the vascular intima producing a fibrous cap around the xanthomal lesion. In addition, the proinflammatory macrophages attract more macrophages which further fuel the inflammatory vascular process. Due to hypoxia in the core of the atherosclerotic plaque, a necrotic core develops and intraplaque neovascularization is upregulated in response. Finally, cellular debris calcifies and the amount of calcification is a measure of the atherosclerotic burden [8].

Many features have been linked to plaque vulnerability including perivascular inflammation, large necrotic core, thin fibrous cap, microcalcifications, intraplaque haemorrhage, neoangiogenesis and hypoxia [9]. A thin fibrous cap is the result of degradation of collagen by metalloproteinase derived from activated macrophages. The rupture of an atherosclerotic plaque usually takes place at the shoulders of the plaque where the cap is thinnest. A large necrotic core and increased angiogenesis are a consequence of inflammation and hypoxia. Endothelial cells of intraplaque (micro)vessels express adhesion molecules favouring leucocyte recruitment. These angiogenic vessels are immature and fragile and due to extravasation of leucocytes further contribute to atherosclerosis progression. Inflammation has been linked to both apoptosis and calcification [10, 11], and oxidative stress has been shown to be responsible for VSMC calcification [12, 13].

Vascular calcification was long considered a passive process not amenable to intervention. A paradigm shift came with the discovery that so-called Gla proteins are involved in the regulation of mineral deposition in the vasculature [14]. Vascular calcification is now considered an active process with cellular and humoral contributions which plays an important role in cardiovascular morbidity and mortality [15]. The discovery that vitamin K-dependent proteins are inhibitors of vascular calcification has advanced our mechanistic understanding of the calcification process and opened novel avenues for diagnosis and treatment [16].

Vascular calcification comes in different flavours: calcification can occur in the vascular intima as well as in the media and it may have different shapes and sizes. Recent data suggest that microcalcification can have a destabilizing effect on atherosclerotic plaques [2, 17], in contrast to the hypothesis that large calcifications are plaque-stabilizing. Microcalcification arises from dedifferentiated VSMCs that start secreting extracellular vesicles, that in the extracellular space form the nidus for calcification [18]. These microcalcifications, which are undetectable by conventional imaging, increase the local stress inside the thin fibrous cap twofold and thus increase the likelihood of rupture [19]. These microcalcifications will eventually result in macrocalcification as a product of osteogenic action by osteocyte-like and chondrocyte-like cells inside the plaque. Novel imaging tools allow the course of microcalcification to be followed [20]. In a rodent model of atherosclerosis, macrophage infiltration in early atherosclerotic plaques has been shown to colocalize with calcification. Indeed, microcalcifications that are phagocytosed by macrophages polarize towards the proinflammatory M1 phenotype [21].

Effective screening to identify the vulnerable plaque is still lacking. This is largely due to the absence of insight into the pathophysiology of vulnerable plaque and thus the lack of cost-effective interventions. Since the genesis of microcalcification is considered one of the earliest precursor processes in the formation of vulnerable plaque [2, 22], adequate early intervention and prevention are possible. Therefore, it is important to understand the factors that drive microcalcification as well as the underlying mechanisms. However, the vast majority of so-called ‘vulnerable’ plaques do not exhibit clinical ‘instability’ and indeed seldom induce ACS [23]. Due to effective treatment, i.e. with statins, lesions became more stable. These lesions underlying superficial erosion do not have thin fibrous caps, have fewer inflammatory cells and lack a large lipid core. Rather they are more collagen-rich lesions and have more calcification, but they are still prone to form occlusive thrombi [23]. These considerations, taken together, indicate the high promise of vascular calcification as a new target for diagnosis and intervention in cardiovascular diseases, with an urgent need to translate experimental findings into adequate diagnostic, preventive and therapeutic solutions.

The strength of nuclear medicine is its ability to provide quantitative information at a functional level, such as the density of a specific receptor or the metabolic activity of a plaque. Nuclear imaging is based on radiolabelled biomarkers with a signal sensitivity in the picomolar range, which compares favourably with both MRI and especially CT, both of which have sensitivities up to a trillion times lower [21, 22]. PET images are derived from the detection of positron-emitting radionuclides labelling biochemical and metabolic substrates. The radionuclide is administered intravenously and circulates within the body. This allows time for the tracer to accumulate at the site of interest. To obtain a favourable target-to-background ratio, the radionuclide should be cleared from the blood quickly. The high sensitivity of PET coupled with the favourable resolution of CT and MRI, as already achieved in PET/CT and PET/MRI hybrid imaging systems provides valuable information about the localization of the acquired signal [24‐26].

To obtain good quality images a high target-to-background ratio is mandatory for vascular PET imaging, for which the radiotracer must show high uptake in the target and preferably also rapid clearance from the bloodstream. Due to the small size of plaques, a high background activity mainly in the blood would severely impair the quality of PET images. Furthermore, coronary artery imaging presents special problems. Cardiac and respiratory movement, myocardial tracer uptake, especially of ¹⁸F-FDG, and the small size (3 to 4 mm) of the coronary arteries may impair and even preclude imaging of the coronary arteries by PET. Whereas motion artefacts can be reduced by using modalities such as cardiac gating, especially with the relatively slow PET data acquisition, suppression of the unwanted myocardial ¹⁸F-FDG uptake might be achieved using a different available dietary protocols [21, 22].

The radioactive tracer most commonly used clinically for PET imaging is ¹⁸F-FDG. ¹⁸F-FDG is mainly used for oncological studies and is well-established for the diagnosis and staging of malignant tumours and the monitoring of therapy response. The tracer is transported into cells by glucose transporters and is phosphorylated by hexokinase to ¹⁸F-FDG-6-phosphate that is not further metabolized [27]. The degree of cellular ¹⁸F-FDG uptake is related to the metabolic rate and the number of glucose transporters [28]. Therefore, increased ¹⁸F-FDG uptake in malignant tumour cells is partly due to an increased number of glucose transporters found in these cells. An increased number of glucose transporters is also found in activated inflammatory cells, such as macrophages. Additionally, the affinity of glucose transporters for deoxyglucose is apparently increased in these cells under inflammatory conditions by various cytokines and growth factors [27, 29]. However, potential competition between ¹⁸F-FDG and nonradioactive stable glucose for uptake in metabolically active cells (for example, inflammatory) might affect the imaging results in patients with high prescan glucose levels. Furthermore, an excess of unlabelled glucose and the action of insulin may increasingly affect the accumulation of ¹⁸F-FDG due to saturation of the glucose transporters [30, 31]. In addition, increases in plasma insulin levels result in translocation of the GLUT-4 transporters from an intracellular pool to the plasma membrane [32, 33]. This may enhance the uptake of ¹⁸F-FDG in myocardial and skeletal muscle, which in turn could significantly impair the quality and interpretability of ¹⁸F-FDG PET scans. This issue needs to be kept in mind particularly in patients with diabetes who are frequently affected by atherosclerosis, as diabetes is one of the main cardiovascular risk factors [34‐39].

Arterial ¹⁸F-FDG uptake was first noted in the aorta of patients undergoing ¹⁸F-FDG PET for cancer imaging [66]. It was also shown that the amount of ¹⁸F-FDG uptake increased with age and was higher in patients with cardiovascular risk factors [29, 67, 68]. Supported by cell culture work, ¹⁸F-FDG uptake in the arterial wall can probably be attributed to uptake in plaque macrophages. These studies demonstrated that increased oxidative metabolism and glucose use in response to cellular activating agents is accompanied by a dramatic increase in ¹⁸F-FDG uptake in both leucocytes and macrophages [36, 69].

Following the dedicated preclinical studies of the use of ¹⁸F-FDG PET for imaging atherosclerosis, Rudd et al. performed one of the first clinical studies of atherosclerosis imaging with ¹⁸F-FDG PET in patients with transient ischaemic attack [70]. The patients were scanned shortly after symptom onset and a significantly higher ¹⁸F-FDG uptake (almost 30%) was seen in symptomatic carotid plaques than in the asymptomatic artery [70]. This indicates the feasibility of ¹⁸F-FDG PET for the imaging of inflammation in the aorta, and in peripheral and vertebral arteries [71‐73]. The relationship between uptake of ¹⁸F-FDG and inflammation has been demonstrated using histology linking ¹⁸F-FDG uptake with the number of macrophages in arterial specimens [74, 75]. In addition to findings indicating ¹⁸F-FDG uptake as a marker of arterial inflammation, preliminary studies have indicated that ¹⁸F-FDG has the potential to predict plaque rupture as well as clinical events. Regarding the predictive value of ¹⁸F-FDG PET, in a study including more than 2,000 patients with cancer disease, Paulmier et al. found that, compared with patients with a low arterial ¹⁸F-FDG uptake, patients with the highest ¹⁸F-FDG uptake were more likely to have either a previous vascular event or to experience an event during the 6 months after PET imaging [76]. Furthermore, the short-term interscan reproducibility of ¹⁸F-FDG PET in the imaging of atherosclerosis has been shown to be excellent for the carotid artery, aorta and peripheral arteries [69, 71, 73, 77]; this is an important finding mainly for the use of ¹⁸F-FDG PET as an endpoint parameter in intervention studies.

Despite the widespread use of ¹⁸F-FDG PET/CT in cardiovascular disease, nonspecific uptake of the tracer is undoubtedly the most important drawback of the use of ¹⁸F-FDG for imaging atherosclerosis [54]. ¹⁸F-FDG is nonspecifically taken up by almost all human tissues and organs to a certain degree. ¹⁸F-FDG uptake analysis in arterial and/or venous vessels is therefore hampered by significant spill-over of tracer signal from closely neighbouring structures; for example, ¹⁸F-FDG uptake in the thyroid gland during carotid PET imaging or myocardial ¹⁸F-FDG uptake during PET imaging of the thoracic aorta or coronary arteries. Furthermore, pathologically increased ¹⁸F-FDG uptake in structures close to the target vessel further impair ¹⁸F-FDG uptake analysis in atherosclerotic plaques; for example, inflamed cervical lymph nodes close to the carotid arteries. More importantly, significant ¹⁸F-FDG uptake has been found in atherosclerosis-prone mice in the vicinity of calcified structures in plaques, and this has been confirmed in in vitro studies of atherosclerotic human arteries that showed marked binding of ¹⁸F-FDG to calcifications but not to other structures of the artery wall [61]. On the basis of these findings taken together, studies to evaluate new or already existing PET tracers currently used for indications other than atherosclerosis are warranted. These PET tracers might be more specific and possibly more sensitive for detecting high-risk vulnerable plaques.

Atherosclerosis is a multifactorial process that involves both local and systemic processes. The vulnerable atherosclerotic plaque cannot simply be detected by plaque size, and thus identification of molecular and cellular processes underlying plaque vulnerability is crucial for diagnosis and treatment (Fig. 6). Recent progress in imaging has improved detection of atherosclerotic lesions and has led to the ability to screen for vulnerable atherosclerotic plaques. Combining expertise in biology, chemistry, nuclear medicine and cardiology will advance our understanding of critical determinants of the high-risk plaque and result in tailor-made imaging radiotracers to differentiate the vulnerable plaque. Ultimately, this strategy will open novel avenues for diagnosis and treatment of the patient with high-risk disease.