Imaging of Vascular Inflammation and Unstable Plaque
Imaging of Vascular Inflammation and Unstable Plaque
The serendipitous observation that some arteries can take up FDG administered for tumour visualization has spawned much interest in using this convenient and well-established agent to visualize atherosclerosis. The first clinical study of PET-FDG imaging of human atherosclerosis was published in 2002, and it demonstrated increased FDG uptake in carotid plaques associated with recent cerebrovascular ischaemic events. When tissue from human atherosclerotic plaques, obtained following endarterectomy, was incubated ex vivo with tritiated deoxyglucose, the tracer was taken up by plaque macrophages, but not by surrounding vascular smooth-muscle cells or by normal vessels. Subsequent studies have shown that the uptake of FDG correlates significantly with plaque macrophage content, and is ~20 times higher in the most inflamed plaques than in control arteries. The use of PET-FDG imaging has been extended to other large arteries of the body (aorta, femoral, and iliac arteries) since then. The FDG signal in atherosclerotic plaques correlates well with circulating inflammatory biomarkers, older age, male sex, and the presence of traditional cardiovascular risk factors. Additionally, some studies have documented a good short-term inter-scan repeatability (2 weeks to 3 months), an important prerequisite for studies using FDG uptake as a surrogate endpoint for plaque-stabilizing treatments. Early pilot studies using a high-fat dietary preparation to prevent spillover artefacts from myocardial FDG uptake have also demonstrated that coronary and aortic FDG uptake often associates with symptoms of acute myocardial ischaemia.
These observations led to the conclusion that the FDG signal in atheromata reports on inflammation. Several groups have shown that interventions—notably, treatment with HMG-CoA reductase inhibitors (statins)—can reduce FDG signals. Effective therapeutic intervention to reduce FDG uptake has raised hope that FDG signals can allow evaluation of novel therapeutics that may alter inflammation in atherosclerotic plaques. Indeed, several dozen clinical trials currently underway use FDG uptake as a biomarker of inflammation within plaques.
Despite these attractions, some issues require resolution before embracing FDG uptake in this regard. Firstly, only limited prospective data correlate FDG uptake, or changes in FDG uptake, with cardiovascular events or altered rates of such complications, and we eagerly await the results of larger prospective cohort studies, such as the High Risk Plaque Initiative and BioImage studies. We lack sufficient validation of the ability of multi-centre studies to provide standardized data on FDG uptake. Information regarding false positives, false negatives, sensitivity, and reproducibility on longitudinal follow-up remains rudimentary. Although FDG signals that co-localize with carotid atheromata present little difficulty with identifying the region of interest, and show modest background signal, considerable barriers to using FDG uptake to visualize coronary atheromata exist—because of avid glucose uptake by myocytes, and substantial myocardial background even under conditions of high-dietary fat intake to suppress cardiac myocyte FDG uptake.
In addition, mechanisms other than inflammation may generate FDG signal associated with atheromata. Microvessels in the plaque may increase delivery of the radiolabelled glucose analogue, enriching its local-specific radioactivity in the glucose pool and giving rise to greater accumulation within cells and signal generation that may not reflect absolute increases in glucose transport. Other biological processes, including hypoxia, may drive increased glucose utilization by mononuclear phagocytes, a putative major source of the FDG signal in atheromata. In particular, studies of human mononuclear phagocytes have not consistently shown increased glucose uptake in response to pro-inflammatory mediators. In contrast, hypoxia readily stimulates glucose uptake by these cells. Hence hypoxia, rather than inflammation, may increase glucose and hence FDG uptake by inflammatory cells in atheromata. In addition, statins can decrease hypoxia-induced augmentation in glucose uptake in human mononuclear phagocytes in vitro. Thus, the attenuation of the FDG signal by statins may not necessarily reflect an anti-inflammatory action. In this regard, nuclear agents that image hypoxia, already in use in oncology, may be useful in imaging atherosclerosis, as hypoxic conditions may prevail in the core of lesions.
Moreover, some have questioned the specificity of FDG for activated macrophages. A micro-autoradiography study of aortic sections of ApoE mice showed a poor correlation of C-FDG uptake with fat content and selective macrophage staining with anti-CD68 in atherosclerotic plaques. Davies et al. reported that, at variance with ex vivo findings, in vivo FDG micro-PET uptake in atherosclerotic lesions of rabbit aorta did not correlate with macrophage accumulation (r= 0.16, P= 0.57), and that FDG uptake in rabbits with highly inflamed aortic walls, those with low levels of inflammation, or controls did not differ significantly. In summary, although FDG has many attractive aspects for clinical use in characterizing atheromata, its underlying biological mechanisms and clinical significance remain incompletely understood and validated. As a result, other compounds have emerged for interrogating vascular inflammation with PET.
The C-labelled PET tracer PK11195 is a selective ligand of the translocator protein (TSPO, 18 kDa), formerly known as peripheral benzodiazepine receptor. Translocator proteins were discovered in the 1970s as benzodiazepine-binding sites outside the central nervous system. Subsequent studies have shown a high TSPO density in circulating human phagocyte populations, particularly in monocytes and neutrophils, with up to 750 000 binding sites per cell. Mature monocytic cell lines have higher TSPO expression than do pro-monocytic or pro-myelocytic lines. Levels of TSPO increase two- to three-fold after monocyte activation in vitro with interferon-gamma or phorbol 12-myristate 13-acetate. Stimulated human monocytes can express >2 000 000 binding sites for PK11195. This increase accompanies enhanced expression of CD11a and CD11b surface antigens and increased production of interleukin-1, interleukin-8, and tumour necrosis factor, indicating that TSPO over-expression associates with activation of phagocytes.C-PK11195 has been used extensively for the non-invasive imaging of neuroinflammatory and neurodegenerative conditions due to its high uptake in activated microglia and low uptake in neurons. Subsequent studies have demonstrated uptake of C-PK11195 by synovial macrophages in patients with rheumatoid arthritis.
Fujimura et al. showed specific binding of H-PK11195 to macrophages in samples of human carotid atherosclerotic plaque. In a small proof-of-concept study by Pugliese et al. in patients with large-vessel vasculitis (predominantly giant-cell arteritis and Takayasu arteritis), C-PK11195 PET/CT allowed detection and quantification of aortic inflammation. Of 15 patients, all 6 patients with signs and symptoms of active vasculitis (visual disturbance, headache, bruit or vascular pain/tenderness, new claudication, fever, night sweats, and/or arthralgia) had markedly increased vascular uptake of C-PK11195 [target-to-background ratio (TBR), 2.41], whereas patients with quiescent disease had background uptake (TBR, 0.98). Standardized uptake values for C-PK11195 correlated well with quantitative total intra-plaque volumes of distribution, calculated from dynamic tissue kinetic modelling using a 1-tissue compartment model. In a subsequent study from the same group, 32 patients with carotid stenoses (of which 9 had recently suffered an acute cerebrovascular ischaemic event) underwent C-PK11195 PET/CT angiographic (CTA) imaging. Carotid plaques associated with ipsilateral symptoms [stroke or transient ischaemic attack (TIA)] had higher TBR (1.06 ± 0.20 vs. 0.86 ± 0.11, P= 0.001) and lower CT attenuation [(median, inter-quartile range) 37, 24–40 vs. 71, 56–125 Hounsfield units, P= 0.01] than those without ipsilateral symptoms (Figure 2). Eight patients underwent carotid endarterectomy, and plaques were harvested for ex vivo analysis: on immunohistochemistry and confocal fluorescence microscopy, CD68 and TSPO co-localized with H-PK11195 uptake at autoradiography. C-PK11195 TBR correlated significantly with percentage specific binding of H-PK11195 determined by autoradiography (r= 0.77, P= 0.025).
(Enlarge Image)
Figure 2.
Computed tomography angiography (A), [C]-PK11195 positron emission tomography (B), and positron emission tomography/computed tomography fusion (C) in a 66-year-old right-handed male patient with a 90% left internal carotid artery stenosis (solid arrows) who developed a facial droop and dysphasia 3 weeks before the positron emission tomography study. Note focal [C]-PK11195 uptake along the convexity of the plaque (B, C, solid arrows). In contrast, images in a 78-year-old asymptomatic female patient with an 80% right ICA stenosis (D–F). There is no visible [C]-PK11195 uptake in the region of the plaque (open arrows). The black arrowhead denotes high [C]-PK11195 uptake in the submandibular gland, and the asterisk denotes high uptake in bone marrow. Reproduced with permission from Pugliese et al.
These preliminary results indicate that ligands of the TSPO hold promise as molecular-targeted imaging compounds to interrogate the presence of intra-plaque inflammation in patients with atherosclerotic disease, and should encourage further prospective studies to assess the predictive value of C-PK11195 or similar TSPO ligands for cardiovascular events. Extending C-PK11195 PET/CT imaging to the coronary arteries represents a great challenge due to the small vessel calibre, the relatively low spatial resolution of PET, and cardiac and respiratory motion. Nonetheless, a preliminary report of imaging temporal arteritis with C-PK11195 PET/CT demonstrates its feasibility in smaller arteries (with a diameter of ~2 mm). Moreover, myocardial uptake of C-PK11195 should be lower than for FDG, thus causing less confounding effects from background signal for imaging coronary arteries. The short physical half-life (20 min) of C-PK11195, which mandates an on-site cyclotron facility, may limit wide clinical applicability. But newly introduced F-labelled TSPO ligands, currently undergoing preclinical investigation, have shown applicability across species in the brain and may overcome some of these barriers.
Positron Emission Tomography
The serendipitous observation that some arteries can take up FDG administered for tumour visualization has spawned much interest in using this convenient and well-established agent to visualize atherosclerosis. The first clinical study of PET-FDG imaging of human atherosclerosis was published in 2002, and it demonstrated increased FDG uptake in carotid plaques associated with recent cerebrovascular ischaemic events. When tissue from human atherosclerotic plaques, obtained following endarterectomy, was incubated ex vivo with tritiated deoxyglucose, the tracer was taken up by plaque macrophages, but not by surrounding vascular smooth-muscle cells or by normal vessels. Subsequent studies have shown that the uptake of FDG correlates significantly with plaque macrophage content, and is ~20 times higher in the most inflamed plaques than in control arteries. The use of PET-FDG imaging has been extended to other large arteries of the body (aorta, femoral, and iliac arteries) since then. The FDG signal in atherosclerotic plaques correlates well with circulating inflammatory biomarkers, older age, male sex, and the presence of traditional cardiovascular risk factors. Additionally, some studies have documented a good short-term inter-scan repeatability (2 weeks to 3 months), an important prerequisite for studies using FDG uptake as a surrogate endpoint for plaque-stabilizing treatments. Early pilot studies using a high-fat dietary preparation to prevent spillover artefacts from myocardial FDG uptake have also demonstrated that coronary and aortic FDG uptake often associates with symptoms of acute myocardial ischaemia.
These observations led to the conclusion that the FDG signal in atheromata reports on inflammation. Several groups have shown that interventions—notably, treatment with HMG-CoA reductase inhibitors (statins)—can reduce FDG signals. Effective therapeutic intervention to reduce FDG uptake has raised hope that FDG signals can allow evaluation of novel therapeutics that may alter inflammation in atherosclerotic plaques. Indeed, several dozen clinical trials currently underway use FDG uptake as a biomarker of inflammation within plaques.
Despite these attractions, some issues require resolution before embracing FDG uptake in this regard. Firstly, only limited prospective data correlate FDG uptake, or changes in FDG uptake, with cardiovascular events or altered rates of such complications, and we eagerly await the results of larger prospective cohort studies, such as the High Risk Plaque Initiative and BioImage studies. We lack sufficient validation of the ability of multi-centre studies to provide standardized data on FDG uptake. Information regarding false positives, false negatives, sensitivity, and reproducibility on longitudinal follow-up remains rudimentary. Although FDG signals that co-localize with carotid atheromata present little difficulty with identifying the region of interest, and show modest background signal, considerable barriers to using FDG uptake to visualize coronary atheromata exist—because of avid glucose uptake by myocytes, and substantial myocardial background even under conditions of high-dietary fat intake to suppress cardiac myocyte FDG uptake.
In addition, mechanisms other than inflammation may generate FDG signal associated with atheromata. Microvessels in the plaque may increase delivery of the radiolabelled glucose analogue, enriching its local-specific radioactivity in the glucose pool and giving rise to greater accumulation within cells and signal generation that may not reflect absolute increases in glucose transport. Other biological processes, including hypoxia, may drive increased glucose utilization by mononuclear phagocytes, a putative major source of the FDG signal in atheromata. In particular, studies of human mononuclear phagocytes have not consistently shown increased glucose uptake in response to pro-inflammatory mediators. In contrast, hypoxia readily stimulates glucose uptake by these cells. Hence hypoxia, rather than inflammation, may increase glucose and hence FDG uptake by inflammatory cells in atheromata. In addition, statins can decrease hypoxia-induced augmentation in glucose uptake in human mononuclear phagocytes in vitro. Thus, the attenuation of the FDG signal by statins may not necessarily reflect an anti-inflammatory action. In this regard, nuclear agents that image hypoxia, already in use in oncology, may be useful in imaging atherosclerosis, as hypoxic conditions may prevail in the core of lesions.
Moreover, some have questioned the specificity of FDG for activated macrophages. A micro-autoradiography study of aortic sections of ApoE mice showed a poor correlation of C-FDG uptake with fat content and selective macrophage staining with anti-CD68 in atherosclerotic plaques. Davies et al. reported that, at variance with ex vivo findings, in vivo FDG micro-PET uptake in atherosclerotic lesions of rabbit aorta did not correlate with macrophage accumulation (r= 0.16, P= 0.57), and that FDG uptake in rabbits with highly inflamed aortic walls, those with low levels of inflammation, or controls did not differ significantly. In summary, although FDG has many attractive aspects for clinical use in characterizing atheromata, its underlying biological mechanisms and clinical significance remain incompletely understood and validated. As a result, other compounds have emerged for interrogating vascular inflammation with PET.
The C-labelled PET tracer PK11195 is a selective ligand of the translocator protein (TSPO, 18 kDa), formerly known as peripheral benzodiazepine receptor. Translocator proteins were discovered in the 1970s as benzodiazepine-binding sites outside the central nervous system. Subsequent studies have shown a high TSPO density in circulating human phagocyte populations, particularly in monocytes and neutrophils, with up to 750 000 binding sites per cell. Mature monocytic cell lines have higher TSPO expression than do pro-monocytic or pro-myelocytic lines. Levels of TSPO increase two- to three-fold after monocyte activation in vitro with interferon-gamma or phorbol 12-myristate 13-acetate. Stimulated human monocytes can express >2 000 000 binding sites for PK11195. This increase accompanies enhanced expression of CD11a and CD11b surface antigens and increased production of interleukin-1, interleukin-8, and tumour necrosis factor, indicating that TSPO over-expression associates with activation of phagocytes.C-PK11195 has been used extensively for the non-invasive imaging of neuroinflammatory and neurodegenerative conditions due to its high uptake in activated microglia and low uptake in neurons. Subsequent studies have demonstrated uptake of C-PK11195 by synovial macrophages in patients with rheumatoid arthritis.
Fujimura et al. showed specific binding of H-PK11195 to macrophages in samples of human carotid atherosclerotic plaque. In a small proof-of-concept study by Pugliese et al. in patients with large-vessel vasculitis (predominantly giant-cell arteritis and Takayasu arteritis), C-PK11195 PET/CT allowed detection and quantification of aortic inflammation. Of 15 patients, all 6 patients with signs and symptoms of active vasculitis (visual disturbance, headache, bruit or vascular pain/tenderness, new claudication, fever, night sweats, and/or arthralgia) had markedly increased vascular uptake of C-PK11195 [target-to-background ratio (TBR), 2.41], whereas patients with quiescent disease had background uptake (TBR, 0.98). Standardized uptake values for C-PK11195 correlated well with quantitative total intra-plaque volumes of distribution, calculated from dynamic tissue kinetic modelling using a 1-tissue compartment model. In a subsequent study from the same group, 32 patients with carotid stenoses (of which 9 had recently suffered an acute cerebrovascular ischaemic event) underwent C-PK11195 PET/CT angiographic (CTA) imaging. Carotid plaques associated with ipsilateral symptoms [stroke or transient ischaemic attack (TIA)] had higher TBR (1.06 ± 0.20 vs. 0.86 ± 0.11, P= 0.001) and lower CT attenuation [(median, inter-quartile range) 37, 24–40 vs. 71, 56–125 Hounsfield units, P= 0.01] than those without ipsilateral symptoms (Figure 2). Eight patients underwent carotid endarterectomy, and plaques were harvested for ex vivo analysis: on immunohistochemistry and confocal fluorescence microscopy, CD68 and TSPO co-localized with H-PK11195 uptake at autoradiography. C-PK11195 TBR correlated significantly with percentage specific binding of H-PK11195 determined by autoradiography (r= 0.77, P= 0.025).
(Enlarge Image)
Figure 2.
Computed tomography angiography (A), [C]-PK11195 positron emission tomography (B), and positron emission tomography/computed tomography fusion (C) in a 66-year-old right-handed male patient with a 90% left internal carotid artery stenosis (solid arrows) who developed a facial droop and dysphasia 3 weeks before the positron emission tomography study. Note focal [C]-PK11195 uptake along the convexity of the plaque (B, C, solid arrows). In contrast, images in a 78-year-old asymptomatic female patient with an 80% right ICA stenosis (D–F). There is no visible [C]-PK11195 uptake in the region of the plaque (open arrows). The black arrowhead denotes high [C]-PK11195 uptake in the submandibular gland, and the asterisk denotes high uptake in bone marrow. Reproduced with permission from Pugliese et al.
These preliminary results indicate that ligands of the TSPO hold promise as molecular-targeted imaging compounds to interrogate the presence of intra-plaque inflammation in patients with atherosclerotic disease, and should encourage further prospective studies to assess the predictive value of C-PK11195 or similar TSPO ligands for cardiovascular events. Extending C-PK11195 PET/CT imaging to the coronary arteries represents a great challenge due to the small vessel calibre, the relatively low spatial resolution of PET, and cardiac and respiratory motion. Nonetheless, a preliminary report of imaging temporal arteritis with C-PK11195 PET/CT demonstrates its feasibility in smaller arteries (with a diameter of ~2 mm). Moreover, myocardial uptake of C-PK11195 should be lower than for FDG, thus causing less confounding effects from background signal for imaging coronary arteries. The short physical half-life (20 min) of C-PK11195, which mandates an on-site cyclotron facility, may limit wide clinical applicability. But newly introduced F-labelled TSPO ligands, currently undergoing preclinical investigation, have shown applicability across species in the brain and may overcome some of these barriers.
Source...