α_vβ₃ integrin PET imaging in atherosclerosis

Imaging α_vβ₃ receptors in vessels could help to identify high-risk atherosclerotic plaques. In a pilot study of the α_vβ₃ PET tracer ¹⁸F-Galacto-RGD, which included 10 patients scheduled for carotid endarterectomy surgery, increased PET signals were observed in stenotic, compared with non-stenotic areas of carotid arteries, which were significantly correlated with ex vivo α_vβ₃ plaque staining and autoradiographic signal intensity.2 Localisation of α_vβ₃ signals within atherosclerotic plaques has also been demonstrated using similar nuclear imaging probes in preclinical studies. However, despite meaningful attempts to correlate imaging with histology, none of these previous studies could determine conclusively whether detectable vascular α_vβ₃ signals originated from the activated endothelium, macrophages or both.

The data from 46 patients presented by Jenkins et al represent the largest prospective cohort of patients with atherosclerosis imaged using an α_vβ₃ tracer.1 In a related study, the group previously reported that ¹⁸F-fluciclatide could accurately identify areas of acute myocardial infarction (MI), and predict the likelihood of functional recovery in impaired, but viable, myocardium.3 In the current study, aortic ¹⁸F-fluciclatide signals were higher in patients with recent MI, compared with stable CVD. Aortic ¹⁸F-fluciclatide signals were also moderately correlated with both the total plaque volume and calcific burden in the aorta. To further validate the clinical findings of these two studies, ex vivo analyses of α_vβ₃ receptor expression and PET ligand binding within myocardial biopsies and excised carotid arterial specimens were performed.

The authors should be congratulated on these important additions to the literature. However, as they acknowledge, imaging patients prior to surgery would have permitted more direct comparisons with histological markers of plaque instability. As most experimental PET ligands tested in atherosclerosis, including ¹⁸F-fluciclatide, were developed initially for cancer imaging, re-purposing these PET tracers outside their intended clinical indication requires careful consideration. While ligand-binding experiments are often used to infer accuracy of PET tracers for marking an imaging target, histological validation of clinical imaging is challenging in atherosclerosis as no ex vivo technique can accurately replicate the local plaque environment in patients. Moreover, many commonly used histological markers, including CD68 for macrophages and CD31 for endothelial cells lack a high degree of cell specificity. Thus, identifying correlations between ex vivo ligand binding and selected histological markers of interest, while important, might not alone give the full answer.

To more accurately resolve the question of what lies beneath clinical PET signals generated by novel ligands in atherosclerosis, and better understand biological mechanisms that define imaging targets within the heterogeneous range of cells comprising advanced atherosclerotic plaques, where feasible, future studies should consider integration of genomic data. For example, a highly instructive plaque imaging study identified differentially activated macrophage pathways, including interferon signalling, in symptomatic carotid plaques with high lipid content measured by T2 mapping MRI, using microarray data from macrophages procured by laser capture microdissection.4 Advances in methods for single-cell RNA sequencing and spatial transcriptomics will likely present further opportunities for this type of molecular imaging research.

Precision imaging versus signal intensity

When selecting a PET ligand for atherosclerosis imaging, a trade-off exists between tracers with highly precise imaging targets, and those with greater signal intensity. While fludeoxyglucose F 18 (¹⁸F-FDG) is the most widely used PET tracer, providing a strong, highly reproducible signal across multiple vascular beds in patients with atherosclerosis, as a glucose analogue it is less specific for individual inflammatory cell types or pathological processes than other more targeted experimental tracers.5 However, these more narrowly targeted tracers often provide lower signal intensity than ¹⁸F-FDG. In the study by Jenkins et al , relatively low arterial ¹⁸F-fluciclatide signals compared with blood pool activity precluded accurate assessment of tracer uptake in the coronary arteries.

Untangling the intertwined pathological processes of inflammation, hypoxia and neoangiogenesis at play within high-risk atherosclerotic plaques with PET imaging might not be possible, or even beneficial. Robust signal intensity is crucial for identification of high-risk plaques, where a blunt approach might better encompass the overall disease activity. In contrast, PET tracers that report on more precise molecular pathways might be favourable for tracking subtle longitudinal changes in the natural history of the disease, or monitoring treatment response. Indeed, ¹⁸F-fluciclatide PET could open new avenues for testing the efficacy of novel anti-angiogenic therapies in CVD.

Aortic imaging as a barometer of systemic atherosclerotic disease activity

By focusing on aortic imaging as a non-invasive barometer of systemic atherosclerotic disease activity, this approach has the potential to identify high-risk features at the patient level, to complement more focal plaque-based assessments of anatomic, haemodynamic and morphological disease severity. These focal disease measures, while essential for guiding everyday clinical management decisions, may be too heavily influenced by local plaque factors to give a balanced indication of overall disease activity. In fact, previous work has shown the ability of ¹⁸F-FDG to differentiate patients with MI versus stable angina when measured in the aorta and left main stem, but not the left anterior descending artery.6

Systemic inflammation measured in the aorta has been linked with cardiovascular risk in patients with psychological stress and increased ¹⁸F-FDG activity in the amygdala and bone marrow.7 In patients with chronic systemic inflammation due to psoriasis, aortic ¹⁸F-FDG uptake is highest in patients with non-calcified coronary disease and low attenuation plaques indicating the presence of lipid-rich necrotic cores.8 In patients with MI, aortic ¹⁸F-FDG uptake is associated with an increased number of lipid-rich coronary plaques identified by optical coherence tomography.9

However, the bigger question is whether sufficient data exist to advocate changes in clinical management from experimental measures of aortic atherosclerosis? Although the comparisons between ¹⁸F-fluciclatide uptake and aortic wall thickness or calcification are interesting, these aortic imaging endpoints are not currently recognised as clinical markers that would themselves prompt treatment for atherosclerosis. Increased aortic wall thickness could reflect structural changes in the vascular wall arising from chronic hypertension rather than a direct measure of atherosclerosis per se, given the close correlation between systolic blood pressure and intima medial thickness. Dense circumferential calcification of the aortic medial layer arising in patients with diabetes mellitus and chronic kidney disease also often coincides with patchy neointimal calcification of typical atherosclerosis.10

Moreover, while thoracic aortic calcification has been identified as an independent risk factor for stroke and all-cause mortality at the population level, measurement of aortic calcification has not reliably demonstrated incremental benefit for prediction of coronary events over traditional risk factors. While molecular imaging of aortic disease activity could be superior for predicting future coronary events to conventional assessments of disease severity, including calcification, in this study there was only a marginal statistical relationship observed between aortic ¹⁸F-fluciclatide uptake and presence of coronary disease in the stable cohort. This marginal result might in part be explained by differing angiogenic responses or macrophage content in aortic compared with coronary plaques.

It is also not clear whether having a recent MI would lead to significant increases in aortic microvessels, or indeed whether these changes are likely to occur within 1–3 weeks post-infarct, although it does seem probable that α_vβ₃ is acting as an activation antigen with its expression induced. In this clinical scenario, these uncertainties suggest that ¹⁸F-fluciclatide might be reporting predominately on systemic inflammation rather than endothelial cell proliferation. Comparison of ¹⁸F-fluciclatide to MRI-based markers of inflammation and neoangiogenesis, for example, using combined PET/dynamic contrast-enhanced or ultrasmall supraparamagnetic iron oxide-MRI, could help to further examine these relationships in future studies.

Future role (and challenges) of PET imaging for CVD risk prediction

The major hurdle for any new imaging method when predicting cardiovascular risk is to demonstrate significant incremental benefit over established measures of disease severity that are often easier and quicker to perform, more widely accessible to patients, and more economical for healthcare systems. However, unlike conventional imaging targets in CVD, PET imaging can provide unique markers of disease activity, such and inflammation and neoangiogenesis, which are central to the pathogenesis of disease. The article by Jenkins et al eloquently demonstrates the use of aortic ¹⁸F-fluciclatide imaging for this purpose. The next translational step to validate novel PET tracers such as ¹⁸F-fluciclatide for cardiovascular risk prediction is to perform larger, collaborative clinical outcome studies in selected high-risk cohorts where there is a realistic expectation that PET imaging could have a true clinical benefit by guiding advanced therapies to alter the course of the disease.