18F-fluoride positron emission tomography for identification of ruptured and high-risk coronary atherosclerotic plaques: a prospective clinical trial

BACKGROUND
The use of non-invasive imaging to identify ruptured or high-risk coronary atherosclerotic plaques would represent a major clinical advance for prevention and treatment of coronary artery disease. We used combined PET and CT to identify ruptured and high-risk atherosclerotic plaques using the radioactive tracers (18)F-sodium fluoride ((18)F-NaF) and (18)F-fluorodeoxyglucose ((18)F-FDG).


METHODS
In this prospective clinical trial, patients with myocardial infarction (n=40) and stable angina (n=40) underwent (18)F-NaF and (18)F-FDG PET-CT, and invasive coronary angiography. (18)F-NaF uptake was compared with histology in carotid endarterectomy specimens from patients with symptomatic carotid disease, and with intravascular ultrasound in patients with stable angina. The primary endpoint was the comparison of (18)F-fluoride tissue-to-background ratios of culprit and non-culprit coronary plaques of patients with acute myocardial infarction.


FINDINGS
In 37 (93%) patients with myocardial infarction, the highest coronary (18)F-NaF uptake was seen in the culprit plaque (median maximum tissue-to-background ratio: culprit 1·66 [IQR 1·40-2·25] vs highest non-culprit 1·24 [1·06-1·38], p<0·0001). By contrast, coronary (18)F-FDG uptake was commonly obscured by myocardial uptake and where discernible, there were no differences between culprit and non-culprit plaques (1·71 [1·40-2·13] vs 1·58 [1·28-2·01], p=0·34). Marked (18)F-NaF uptake occurred at the site of all carotid plaque ruptures and was associated with histological evidence of active calcification, macrophage infiltration, apoptosis, and necrosis. 18 (45%) patients with stable angina had plaques with focal (18)F-NaF uptake (maximum tissue-to-background ratio 1·90 [IQR 1·61-2·17]) that were associated with more high-risk features on intravascular ultrasound than those without uptake: positive remodelling (remodelling index 1·12 [1·09-1·19] vs 1·01 [0·94-1·06]; p=0·0004), microcalcification (73% vs 21%, p=0·002), and necrotic core (25% [21-29] vs 18% [14-22], p=0·001).


INTERPRETATION
(18)F-NaF PET-CT is the first non-invasive imaging method to identify and localise ruptured and high-risk coronary plaque. Future studies are needed to establish whether this method can improve the management and treatment of patients with coronary artery disease.


FUNDING
Chief Scientist Office Scotland and British Heart Foundation.


Introduction
Coronary atherosclerotic plaque rupture is the principal precipitant of acute myocardial infarction and an important cause of sudden cardiac death. Rupture is challenging to predict because most plaques are non-obstructive and are not identifi ed by stress testing or coronary angiography. 1,2 Atherosclerotic lesions at risk of rupture have certain histopathological characteristics that include positive remodelling, microcalcifi cation, and a large necrotic core. [1][2][3] The development of modern molecular imaging techniques targeted at these features could lead to the identifi cation of such high-risk plaques in vivo and guide the development of novel treatment strategies. [4][5][6][7] Combined PET and CT is a non-invasive imaging technique that brings functional molecular imaging together with precise anatomical information. We have recently reported preliminary PET-CT data using the tracer ¹⁸F-sodium fl uoride (¹⁸F-NaF) as a marker of valvular and vascular calcifi cation activity in patients with aortic stenosis. [7][8][9] Other studies have shown the usefulness of ¹⁸F-fl urodeoxyglucose (¹⁸F-FDG) as a surrogate of vascular infl ammation and macrophage burden. 6,[10][11][12][13] We therefore investigated whether, compared with the current non-invasive gold standard of ¹⁸F-FDG, ¹⁸F-NaF uptake could identify ruptured and high-risk athero sclerotic plaques in patients with symptomatic coronary and carotid artery disease. undergoing carotid endartectomy for symptomatic carotid artery disease. 15 Exclusion criteria were age younger than 50 years, insulin-dependent diabetes mellitus, women of childbearing age not receiving contraception, severe renal failure (serum creatinine >250 μmol/L), known contrast allergy, and inability to provide informed consent. Only patients older than 50 years were recruited in the study to reduce any long-term risks associated with radiation exposure. Uncontrolled diabetes and high blood glucose concentrations (>11 mmol/L) interfere with the quality of ¹⁸F-FDG PET imaging because of the competition between glucose and ¹⁸F-FDG for cellular entry. The convention is therefore to exclude such patients from vascular ¹⁸F-FDG PET studies. 7,10,12,13 All patients underwent a comprehensive baseline clinical assessment including evaluation of their cardiovascular risk factor profi le. Plasma troponin I concentrations were measured in patients with stable angina using the ARCHITECT STAT high-sensitivity troponin I assay (Abbott Laboratories, Abbott Park, IL, USA; lower limit of detection 1·2 ng/L; 99th percentile diagnostic threshold 26 ng/L). Studies were done with the approval of the local research ethics committee, in accordance with the Declaration of Helsinki, and with the written informed consent of each participant.

Procedures
Patients with myocardial infarction and stable angina underwent ¹⁸F-NaF and ¹⁸F-FDG PET-CT, CT coronary angiography, and CT calcium scoring (appendix). 7 To minimise myocardial uptake, patients were instructed to adhere to a low-carbohydrate, high-protein, and high-fat diet for at least 24 h before undergoing ¹⁸F-FDG PET-CT.
Electrocardiograph-gated PET images were reconstructed in diastole (50-75% of the R-R interval, Ultra-HD) using the Siemens Ultra-HD algorithm, fused with the CT coronary angiogram, and analysed by experienced observers blinded to the clinical diagnosis (NJ, MD, FC) using an OsiriX workstation (OsiriX version 5·5·1 64-bit; OsiriX Imaging Software, Geneva, Switzerland). Twodimensional regions of interest were drawn around all major (diameter >2 mm) epicardial vessels on 3 mm axial slices just beyond the discernible adventitial border. The maximum standard uptake value (the decay corrected tissue concentration of the tracer divided by the injected dose per bodyweight) was measured and corrected for blood pool activity in the superior vena cava to provide tissue-to-background ratio (TBRs) measurements. Using this method, we have previously shown excellent reproduci bility for ¹⁸F-NaF TBR measurements in the coronary arteries with an intraclass correlation coeffi cient of 0·99. 7 We used a previously established 95% lower reference limit to categorise coronary plaques into ¹⁸F-NaF positive lesions (focal uptake with a TBR more than 25% higher than a proximal reference lesion) and negative plaques if these criteria were not achieved. This limit was based on our previous study, where plaques with high ¹⁸F-NaF uptake had maximum TBRs that were 44% (95% CI 26-62) higher than a proximal quiescent reference lesion. 7 In patients with acute myocardial infarction, ¹⁸F-NaF uptake in the culprit plaque was compared with the highest value in any of the non-culprit vessels.
Quantifi cation of ¹⁸F-FDG uptake was performed as for ¹⁸F-NaF uptake but restricted to the proximal and midportions of the coronary arteries, and to regions where myocardial uptake and spillover could be confi dently excluded. 7 Again, ¹⁸F-FDG positive plaques were defi ned using the 25% threshold as described for ¹⁸F-NaF. Eff ective myocardial suppression of ¹⁸F-FDG was predefi ned as a standard uptake value of 5·0 or less in the basal ventricular septum (appendix) as per published data. 12 In patients with stable angina, PET-CT imaging was prospectively used to direct greyscale and radiofrequency intravascular ultrasound (20 MHz Eagle Eye Platinum Catheters [Volcano Corp, San Diago, CA, USA], motorised pull-back 0·5 mm/s) to the ¹⁸F-NaF positive and negative plaques. The interventional cardiologist acquiring the intravascular ultrasound data was blinded to the PET-CT status of the plaque.
Intravascular ultrasound analysis was done as described previously 16 using dedicated VIAS software (Volcano Image Analysis Software version 3.0) by operators blinded to the PET data. Regions of interest were drawn around the external elastic membrane and luminal borders, and plaque area and composition (dense calcium, necrotic core, fi bro-fatty tissue, and fi brous tissue) calculated. [16][17][18] The presence of microcalcifi cation (spotty calcifi cation in the absence of acoustic shadowing on three or more consecutive frames) and the maximum frame necrotic core (the highest percentage of necrotic core on a single frame) were recorded. 19 The remodelling index was defi ned as the ratio between the external elastic membrane cross-sectional area of the lesion and a proximal reference region in the same vessel. 20 Plaques were classifi ed as thin-cap fi broatheroma, thick-cap fi broatheroma, pathological intimal thickening, or fi brocalcifi c plaque as defi ned previously. 18,21 CT analysis was done on a dedicated cardiovascular workstation (Vital Images, Minnetonka, MN, USA). Vessel-specifi c and total Agatston calcium scores were calculated as described previously. 7 An independent experienced and blinded observer (MW) determined the stenosis severity, plaque composition (calcifi ed, noncalcifi ed, mixed plaque), and presence of high-risk CT features (positive remodelling, microcalcifi cation, necrotic core) according to standard defi nitions in plaques with and without increased ¹⁸F-NaF activity. 22 Intact atherosclerotic plaques were retrieved at the time of carotid endarterectomy and scanned using exvivo PET-CT to allow precise anatomical co localisation of ¹⁸F-NaF activity with pathological evidence of plaque rupture. Plaques were divided into ¹⁸F-NaF positive and negative areas, and histological sections were assessed See Online for appendix using Movat's pentachrome and immunohistochemistry to investigate calcifi cation activity (tissue non-specifi c alkaline phosphatase and osteocalcin), macrophage infi ltration (CD68), and cell death (apoptosis, cleaved caspase 3; presence of necrotic core; appendix).

Statistical analysis
The primary endpoint of the study was the comparison of ¹⁸F-fl uoride tissue-to-background ratios of culprit and nonculprit coronary plaques of patients with acute myocardial infarction. The main secondary endpoints were comparative imaging and histol ogical characterisation of ¹⁸F-fl uoride positive and negative atherosclerotic plaques in patients with coronary and carotid artery disease. Based on our previous data, 7 we required 36 patients with myocardial infarction to detect a diff erence of 0·23 in the tissue-to-background ratio between culprit and non-culprit plaques at 90% power and two-sided p<0·05. We recruited 40 patients to account for incomplete data and recruited a similar sized (n=40) comparator group of patients with stable angina.
Continuous data were tested for normality with the D'Agostino-Pearson omnibus test. Continuous parametric variables were expressed as mean (SD) and

Role of the funding source
The funding source had no role in the study design (except through its external peer review process), data collection, data analysis, data interpretation, or writing of the report. All authors had access to the primary data and have fi nal responsibility to submit for publication.
The median duration between clinical symptoms and carotid endarterectomy was 17 [IQR [10][11][12][13][14][15][16][17][18][19][20][21][22][23][24][25][26][27] days (appendix). Carotid endarterectomy specimens (fi gure 3; appendix) were obtained for 12 patients, although three specimens could not be excised intact and were discarded. Ex-vivo ¹⁸F-NaF PET-CT was undertaken in nine specimens and uptake was localised to the site of macroscopic plaque rupture in all patients (fi gure 3). Compared with sections of tissue without uptake (n=15), those with increased ¹⁸F-NaF uptake (n=24) (table 1). Focal ¹⁸F-NaF uptake was noted in 18 patients (45%), which did not seem to be related to percutaneous coronary intervention and stent angiography that were both stented during the index admission. Only the culprit lesion had increased ¹⁸F-NaF uptake (¹⁸F-NaF, tissue-to-background ratios, culprit 2·03 versus reference segment 1·08 [88% increase]) on PET-CT (E) after percutaneous coronary intervention. Corresponding ¹⁸F-fl uorodeoxyglucose PET-CT showing no uptake either at the culprit (¹⁸F-FDG, tissue-to-background ratios, culprit 1·62 versus reference segment 1·49 [9% increase]) or the bystander stented lesion. Note intense uptake within the ascending aorta. In a patient with stable angina with previous coronary artery bypass grafting, invasive coronary angiography (G) showed non-obstructive disease in the right coronary artery. Corresponding PET-CT scan (H) showed a region of increased ¹⁸F-NaF activity (positive lesion, red line) in the mid-right coronary artery (tissue-to-background ratio, 3·13) and a region without increased uptake in the proximal vessel (negative lesion, yellow line). Radiofrequency intravascular ultrasound shows that the ¹⁸F-NaF negative plaque (I) is principally composed of fi brous and fi brofatty tissue (green) with confl uent calcium (white with acoustic shadow) but little evidence of necrosis. On the contrary, the ¹⁸F-NaF positive plaque (J) shows high-risk features such as a large necrotic core (red) and microcalcifi cation (white). Although predefi ned myocardial suppression of ¹⁸F-FDG uptake was achieved in 34 (85%) patients (median myocardial standard uptake value 2·60 [IQR 1·83-3·83]), coronary ¹⁸F-FDG uptake could not be confi dently interpreted in 45% of vessel territories. Increased focal ¹⁸F-FDG uptake was noted in just four patients: three at the site of recent coronary stenting and one at the ostium of a saphenous vein graft.

Discussion
We have shown that intense ¹⁸F-NaF uptake localises to recent plaque rupture in patients with acute myocardial infarction and in those with symptomatic carotid disease. Moreover, in patients with stable coronary artery disease, ¹⁸F-NaF uptake seems to identify coronary plaques with high-risk features on intravascular ultrasound. This technique holds major promise as a means of identifying high-risk and ruptured plaque, and potentially informing the future management and treatment of patients with stable and unstable coronary artery disease. Over 90% of our patients with myocardial infarction had increased ¹⁸F-NaF uptake at the site of their culprit ruptured plaque, with TBR values that were a third higher than the maximum activity anywhere else in the coronary vasculature. These fi ndings were not unique to the coronary circulation since we also noted increased focal ¹⁸F-NaF uptake at the site of plaque rupture in all excised carotid endarterectomy specimens from patients with symptomatic carotid disease. However, we do acknowledge that this was not a universal fi nding. Of the three patients with myocardial infarction who had no uptake, two were younger smokers with only mild underlying irregularities on coronary angiography, implicating plaque erosion and thrombosis as the mechanism of their infarction rather than plaque rupture. 23 The third patient sustained an inferolateral non-ST segment elevation myocardial infarction and had a lesion stented in the right coronary artery. Increased ¹⁸F-NaF activity was seen in the co-dominant circumfl ex artery that could have equally explained the clinical presentation, raising the intriguing possibility that ¹⁸F-NaF might have a clinical role for patients in whom the culprit lesion is not readily apparent.
Focal regions of increased ¹⁸F-NaF activity were seen in almost a half of our patients with stable coronary artery disease. To understand the mechanism of uptake in these patients, we sought to compare plaque characteristics of lesions with and without increased ¹⁸F-NaF uptake. Because histology of the coronary arteries in this population is not feasible, we undertook greyscale and radiofrequency intravascular ultrasound, a widely used and validated process that provides detailed characterisation of plaque composition. 21 This method showed that lesions with increased ¹⁸F-NaF uptake were associated with greater positive remodelling, more microcalcifi cation, and a larger necrotic core. These fi ndings were corroborated by, and consistent with, the fi ndings of plaque analysis done with CT coronary angiography. Plasma troponin concentrations measured by a novel high-sensitivity assay were also higher in those patients with ¹⁸F-NaF positive plaques than in patients with ¹⁸F-NaF negative plaques, perhaps implicating subclinical plaque rupture with embolisation and microinfarction.
Why does ¹⁸F-NaF bind to ruptured or high-risk plaque? Similar to the caseating granulomata of tuberculosis, atherosclerotic vascular calcifi cation is a controlled cellular response to an intense, necrotic, and chronic infl ammatory stimulus. Indeed, direct links between infl ammatory cells and osteoblastic metaplasia in the vasculature are well described. 24,25 Hydroxyapatite is the central structural component of vascular calcification and is laid down during the earliest and most active stages of mineralisation: 24 hydroxyapatite nanocrystals nucleate, propagate, and mineralise the extracellular matrix. Fluoride ions are incorporated into the hydroxyapatite by ion exchange with hydroxyl groups at the crystal surface. This process is dependent on the crystal surface area that will be greatest in the earliest and most active nanocrystalline stages of mineralisation associated with plaque infl ammation and necrosis. We believe that these processes are responsible for the observed ¹⁸F-NaF uptake and is consistent with our data showing ¹⁸F-NaF uptake in regions of necrosis, macrophage infi ltration, apoptosis, microcalcifi cation, and alkaline phosphatase and osteocalcin staining. Moreover, mathematical modelling indicates that microcalcifi cation at the surface of thin-capped atheroma (fi gure 1) can intensify and double incident stresses. 26 Microcalcifi cation is therefore not only a marker of acute plaque rupture but is implicated in its precipitation.
Coronary arterial calcifi cation is considered pathognomonic of atherosclerosis and is a powerful independent risk predictor for cardiovascular events that can be further refi ned by the rapidity of its progression. 27,28 Why then not rely on CT coronary calcium scoring alone as a biomarker? Microcalcifi cation cannot be detected on CT and confl uent coronary macrocalcifi cation develops slowly, taking many months or years to become apparent on CT, and can become dormant once infl ammation in Minimal diameter (mm) Remodelling index 1·12 (1·09-1·19) 1·01 (0·94-1·06) 0·0004

Plaque classifi cation, n (%)
Thin-cap fi broatheroma 7 (47%) 4 (16%) 0·068 Thick-cap fi broatheroma 5 (33%) 9 (38%) 1·0 Pathological intimal thickening 0 7 (29%) 0·003  the plaque has subsided. By identifying areas of nascent and ongoing calcifi cation activity, ¹⁸F-NaF uptake allows us to detect regions of metabolically active plaque, thus providing complementary information to CT. [29][30][31][32] Indeed, we noted large areas of coronary CT calcium in the absence of increased ¹⁸F-NaF uptake (fi gure 1) whereas other regions with minimal or no CT calcium had intense ¹⁸F-NaF uptake (appendix) in keeping with previous observations in the aorta by Derlin and colleagues (panel). 29,32 Moreover, given that ¹⁸F-NaF seems more closely aligned with the process of necrotic infl ammation and plaque metabolic activity, we believe that it potentially off ers major improvements to the prediction of cardiovascular risk compared with calcium scoring. Our data have already established that ¹⁸F-NaF identifi es plaque with multiple high-risk features, but prospective studies are now needed in a broad range of patients to assess whether increased coronary ¹⁸F-NaF activity will ultimately translate into future adverse events. If the results prove confi rmatory then this technique has the potential to fundamentally alter the way we treat coronary artery disease: moving us away from the current framework based on lesion severity and ischaemia to one focused on plaque metabolism and infl ammation. It could, for example, permit the identifi cation of the vulnerable patient with single or multiple high-risk or silently ruptured plaques, providing an opportunity to treat and modify their risk to prevent future adverse cardiovascular events.
By contrast with ¹⁸F-NaF, ¹⁸F-FDG imaging was hampered by problems related to tracer uptake in the myocardium. Our stringent dietary recommendations resulted in suppression of myocardial activity in 70-85% of patients: a rate that compares favourably with previous studies (57-84%). 6,12,13 However, this suppression resulted in a patchy distribution of myocardial uptake that frequently obscured activity in one or more coronary vessels. Increased ¹⁸F-FDG uptake might possibly occur in the culprit plaque and we failed to show this because of incomplete data or the delay in scanning. However, given its limitations, we believe that ¹⁸F-FDG is unlikely to become suffi ciently robust to permit its clinical application to the coronary circulation. Nevertheless, ¹⁸F-FDG uptake remains an important measure of general vascular infl ammation in the aorta and carotid arteries, providing complementary and distinct metabolic information to that of ¹⁸F-NaF uptake.
We acknowledge that there are limitations of our study that include a lack of respiratory gating, potential partial volume artefacts, and the use of surrogate measures for coronary histology. 21 However, we believe that the totality of our comprehensive evidence using multiple approaches and imaging modalities provides a robust and cogent argument to support our contention that ¹⁸F-fl uoride uptake identifi es vulnerable and high-risk plaques in patients with stable and unstable coronary heart disease. Further work is now needed to establish whether ¹⁸F-NaF PET-CT will provide a clinically useful technique capable of improving risk stratifi cation, monitoring disease progression, guiding therapeutic interventions, and assessing novel anti-atherosclerotic therapies.

Contributors
NVJ designed the study, undertook experiments, analyse d results, and interpreted the data. ATV undertook experiments, analysed and interpreted the data, and prepared the report. MCW, ASVS, PAC, FHMC, SEY, AMF, EJRvB, and KAAF collected, analysed, and interpreted data, and prepared the report. ADF, NGU, MWHB, NLMC, and NLM collected the data and prepared the report. JHFR, MRD, and DEN contributed to the study design, supervision, and interpretation of data. DEN is the chief

Systematic review
We searched PubMed using variations of the keywords "high-risk plaques", "vulnerable plaques", "ruptured plaques", "¹⁸F-fl uorodeoxyglucose positron emission tomography", "¹⁸F-fl uoride positron emission tomography", and "coronary arteries". The search was restricted to human studies. We assessed the quality of the evidence specifi cally related to cardiovascular disease by reviewing the patient population studied and the methodology for the positron emission and CT imaging.
Non-invasive imaging of carotid plaque infl ammation using ¹⁸F-fl uorodeoxyglucose positron emission tomography was reported by Rudd and colleagues in 2002. 11 Since then, this tracer has been validated and widely used as a surrogate of large vessel infl ammation. 8,10 Increased ¹⁸F-fl uorodeoxyglucose in the coronary arteries has been described in patients with coexisting malignancy. 12,33,34 Since then, three prospective studies have examined the feasibility and reproducibility of assessing uptake of this tracer in the coronary vasculature. 6,7,13 Only two small studies (n=10-20) 6,13 have suggested that ¹⁸F-fl uorodeoxyglucose might identify some infl amed plaques in patients with recent myocardial infarction, although the largest study showed that in 50% of patients with acute myocardial infarction, there was no uptake of ¹⁸F-fl uorodeoxyglucose in the culprit plaque. 13 Four retrospective studies in patients with cancer have recently reported cardiovascular uptake of ¹⁸F-fl uoride. 29,31,32,35 The aortic uptake of ¹⁸F-NaF was fi rst reported by Derlin and colleagues 29 and cardiac ¹⁸F-fl uoride uptake by Beheshti and colleagues. 31 We reported the coronary uptake of ¹⁸F-NaF in a prospective clinical trial involving patients with aortic stenosis, 7,8 and these results were subsequently corroborated by Li and colleagues in their retrospective study of patients with cancer. 30 No study has prospectively assessed this tracer in patients with stable or unstable coronary artery disease or validated its activity against histology or invasive intracoronary imaging, such as intravascular ultrasound. There are no previous reports of ¹⁸F-fl uoride uptake in relation to plaque vulnerability or rupture.

Interpretation
There are currently no non-invasive imaging techniques that can identify high-risk and ruptured coronary atherosclerotic plaques in vivo in patients with coronary heart disease. For the fi rst time, we have shown that ¹⁸F-fl uoride positron emission tomography can identify culprit and ruptured plaques in patients with myocardial infarction and symptomatic carotid disease. Moreover, histological characterisation demonstrates that ¹⁸F-fl uoride activity localises to regions of plaque rupture with evidence of increased infl ammation, calcifi cation activity, necrosis, and cell death. In patients with stable angina, ¹⁸F-fl uoride is associated with coronary plaques that have high-risk features on intravascular ultrasound, including positive remodelling, microcalcifi cation, and necrosis. Given its ability to identify high-risk or ruptured coronary atherosclerotic plaque, this non-invasive imaging technique has the potential to change how we identify, manage, and treat patients with stable and unstable coronary artery disease. Further work is now needed to establish whether ¹⁸F-fl uoride positron emission tomography will provide a means of improving risk stratifi cation, monitoring disease progression, guiding therapeutic interventions, and assessing novel anti-atherosclerotic therapies.
investigator for the study and obtained funding for all studies. All authors participated in data interpretation. NVJ drafted the fi rst and subsequent versions of this report with key input from MRD and DEN, and revisions from all authors, who reviewed and approved the fi nal submitted report.

Confl icts of interest
NLM has received honoraria for Abbott Diagnostics and acted as a consultant for Abbott Diagnostics. The other authors declare that they have no confl icts of interest.