Edinburgh Research Explorer Non-invasive in vivo coronary artery thrombus imaging

Background The diagnosis and management of myocardial infarction is increasingly complex and establishing the presence of intracoronary thrombosis has major implications for both the classification and treatment of myocardial infarction. Objectives To investigate whether positron emission tomography (PET) and computed tomography (CT) could non-invasively detect in vivo thrombus formation in human coronary arteries using a novel glycoprotein IIb/IIIa receptor antagonist-based radiotracer, 18 F-GP1. Methods In a single centre observational case-control study, patients with or without acute myocardial infarction underwent coronary 18 F-GP1 PET-CT angiography. Coronary artery 18 F-GP1 uptake was assessed visually and quantified using maximum target-to-background ratios. Results 18 F-GP1 PET-CT angiography was performed in 49 patients with, and 50 patients without, acute myocardial infarction (61±9 years, 75% male). Coronary 18 F-GP1 uptake was apparent in 39 (80%) of the 49 culprit lesions in patients with acute myocardial infarction. False negative scans appeared to relate to time delays to scan conduct and low thrombus burden in small calibre distal arteries. On multivariable regression analysis, culprit vessel status was the only independent variable associated with higher 18 F-GP1 uptake. Extra-coronary cardiac 18 F-GP1 findings included a high frequency of infarct-related intramyocardial uptake (35%) as well as left ventricular (8%) or left atrial (2%) thrombus.


Introduction
High-sensitivity cardiac troponin testing has transformed cardiological practice and has ensured that at-risk patients and those with myocardial infarction can be effectively identified.(1)(2)(3) However, this greater sensitivity has come at the cost of identifying many more patients with acute myocardial injury who may or may not have myocardial infarction.There has also been the increasing complexity of the diagnosis of myocardial infarction typified by type 2 myocardial infarction and myocardial infarction with non-obstructive coronary arteries.(4,5) Clinicians are often left with the challenge of determining whether an atherothrombotic event has occurred and establishing the underlying cause.(6) This is critically important if patients are to be correctly diagnosed and treated.
At present, invasive coronary angiography is the primary modality used to detect the presence of intracoronary thrombus in patients with acute myocardial infarction.However, it lacks sensitivity and may fail to identify, or misidentify, the culprit artery in over a third of patients with non-ST elevation myocardial infarction.(6) Adjunctive invasive techniques, such as intravascular ultrasound and optical coherence tomography, are valuable tools to clarify angiographic ambiguity.(7,8) However, these catheter-based techniques require instrumentation of the coronary artery and are limited in their ability to assess smaller or stenosed vessels due to the size and physical limitations of the intravascular probes.Like other non-invasive imaging approaches, intravascular imaging also only provides circumstantial evidence of the presence of thrombus based on the characteristic appearances of the images.(9) We have previously described the use of 18 F-sodium fluoride positron emission tomography and computed tomography (PET-CT) imaging in patients with acute myocardial infarction and stable coronary artery disease.(10)(11)(12)(13)(14) We have shown that increased coronary 18 F-sodium fluoride uptake correlates with high-risk plaque characteristics, disease progression and subsequent myocardial infarction.However, assessments of plaque characteristics or disease activity do not directly inform about the presence of plaque rupture and active atherothrombosis.Prior radiotracer imaging studies have investigated thrombus imaging in the heart but have been limited to the assessment of atrial or ventricular thrombus rather than coronary thrombosis.(15)(16)(17)  18 F-GP1 is a novel and selective elarofiban-derived radiotracer that binds with high specificity to the glycoprotein IIb/IIIa receptor on activated platelets.(18) Previous phase 1 in vivo studies (19,20) and case reports (21,22) have demonstrated that 18 F-GP1 PET-CT can detect in vivo venous and arterial thrombi.We hypothesized that 18 F-GP1 PET-CT could provide the first non-invasive technique to identify in vivo coronary thrombosis in patients presenting with acute myocardial infarction.

F-GP1 Binding to Human Thrombus
Fresh human thrombus formation was generated using an ex vivo model of deep arterial injury as described previously (23,24) and performed during infusion of 18 F-GP1 in the presence or absence of tirofiban co-infusion. 18F-GP1 autoradiography was performed in the presence or absence of tirofiban on thrombectomy specimens obtained from patients undergoing percutaneous coronary intervention for acute myocardial infarction.Further details of the methodologies are provided in the Supplementary Material.

Study Design and Study Population
This was a single centre cross-sectional case-control study (In-vivo Thrombus Imaging With 18 F-GP1, a Novel Platelet PET Radiotracer (iThrombus); NCT03943966).We recruited patients over 40 years of age who presented with type 1 myocardial infarction defined by the fourth universal definition of myocardial infarction (4) and cardiac troponin concentration greater than 50 times the upper reference limit within seven days of presentation to the Edinburgh Heart Centre.All patients underwent invasive coronary angiography.Control subjects consisted of patients with stable coronary artery disease and previous aortic valve replacement who underwent 18 F-GP1 cardiac PET-CT as part of a concurrent study (NCT04073875).(25) These patients had no angina or recent myocardial infarction (Supplementary Figure i).Exclusion criteria included inability to provide informed consent, pregnancy or breastfeeding, contraindications to iodinated contrast, use of anticoagulant therapies, estimated glomerular filtration rate <30 mL/min/1.73m 2 , metastatic malignancy and an inability to tolerate the supine position.
The study was conducted in accordance with the Declaration of Helsinki and approved by the Southeast Scotland Regional Ethics Committee (18/SS/0163).Written informed consent was obtained from all subjects.
Cardiac Imaging and analysis 18 F-GP1 cardiac PET-CT was performed on a 128-multislice scanner (Biograph mCT, Siemens, Germany) with prospective electrocardiogram-gating.PET list mode acquisition was performed 60 min after intravenous injection of 250 MBq 18 F-GP1 with a single bed position centred on the heart, preceded by an attenuation correction CT scan.Further details of the cardiac imaging and analysis are provided in the Supplementary Material.

Definition of Culprit Lesion
The clinical team ascribed the location of the culprit lesion at the time of invasive angiography independent of the research team or knowledge of the 18 F-GP1 PET-CT findings.The culprit lesion was defined as the lesion that led to the index admission and upon which the interventional procedure was performed.Further details of the culprit assessment are provided in the Supplementary Material. 18

F-GP1 uptake assessment
Qualitative and quantitative 18 F-GP1 assessment was performed on the co-registered PET and CT coronary angiogram images. 18F-GP1 PET uptake was assessed in the left anterior descending artery, left circumflex and right coronary arteries where the vessel diameter was ≥2 mm.Further details of the tracer uptake are provided in the Supplementary Material.

Statistical analysis
We assessed the distribution of data with the Shapiro-Wilk test and quantile-quantile plots.
Categorial variables were reported as frequencies (percentages).Continuous data are presented as mean ± standard deviation or median [interquartile interval].Statistical significance was assessed using chi-squared, Wilcoxon, Kruskal-Wallis, or Mann-Whitney U tests as appropriate.Correlations were assessed using Spearman coefficient.A multivariable linear regression model was constructed with log2-transformed TBRmax as the independent variable and age, sex, time after myocardial infarction, presence of culprit lesion and presence of ST elevation myocardial infarction as the independent variables.Model residuals were checked against fitted values and distributions confirmed with quantile-quantile plots.Two-sided pvalues <0.05 were considered statistically significant.Analysis was performed using R version 4.0.3(R Foundation for Statistical Computing, Vienna, Austria).

Results
18 F-GP1 Binding to Human Thrombus 18 F-GP1 uptake occurred in the presence of freshly forming human thrombus in the ex vivo model of deep arterial injury (Figure 1).(23) This uptake was inhibited by co-administration of tirofiban, a glycoprotein IIb/IIIa receptor antagonist.In thrombectomy specimens obtained from patients undergoing percutaneous coronary intervention for acute myocardial infarction, 18 F-GP1 autoradiography demonstrated uptake that co-localized to thrombus on histology and platelets on CD41 immunohistochemistry (Figure 1).

Study Populations
Fifty-nine patients with recent myocardial infarction (time from myocardial infarction onset to scan: 9 [5 to 21] days) were recruited.Eight patients either withdrew or were unable to attend due to government public health restrictions, and two patients were excluded because they were subsequently diagnosed with takotsubo syndrome (Supplementary Figure i).Forty-nine patients (mean age 61±9 years and 75% male) underwent 18 F-GP1 PET-CT and were included in the analysis (Table 1).There were 35 patients with ST-segment elevation myocardial infarction and 14 with non-ST-segment elevation myocardial infarction.Of the latter, 10 had a culprit lesion identified angiographically by the clinical team.All patients presenting with STsegment elevation myocardial infarction received primary percutaneous coronary intervention within 4 hours of symptom onset and all patients with non-ST-segment elevation myocardial infarction underwent invasive coronary angiography within 48 hours of symptom onset.Ten patients were managed with guideline-directed medical therapy only and thirty-nine patients underwent percutaneous coronary intervention.In the control group, we included 50 patients without recent myocardial infarction (Table 1) who had undergone 18 F-GP1 PET-CT as described previously.(25)

Invasive Coronary Angiography
Most patients with myocardial infarction (28/49, 57%; all presenting with ST-segment elevation) had complete vessel occlusion (TIMI 0 flow) at the time of invasive coronary angiography with the remainder having TIMI 2 or 3 flow.Sixteen patients received an intravenous tirofiban infusion for 12 hours after the index event.In the absence of percutaneous coronary intervention, intracoronary thrombus was visible as a filling defect in four cases (8%) which was visible on subsequent computed tomography angiography performed at 5 and 7 days (Supplementary Figure ii).

Qualitative Coronary 18 F-GP1 Uptake in the Culprit Plaque
Most patients with myocardial infarction (39/49, 80%) had focal coronary 18 F-GP1 uptake corresponding to the culprit vessel (Table 2).There was no qualitative coronary 18 F-GP1 uptake in the non-culprit vessels of patients with myocardial infarction or in any vessel of patients without myocardial infarction.Coronary 18 F-GP1 uptake was present in 30 (86%) patients with ST-segment elevation myocardial infarction, all of which co-localized to the culprit lesion (Figure 2).Coronary 18 F-GP1 uptake was present in 9 out of 14 (64%) patients with non-STsegment elevation myocardial infarction, and this co-localized to the culprit lesion in each case (Supplementary Figure iii).Detection of intracoronary thrombus formation appeared to be less frequent in patients with non-ST-segment elevation myocardial infarction compared to those with ST-segment elevation myocardial infarction, but this was not statistically significant (64% vs 86% respectively, p=0.197).
Overall, the identification of intracoronary thrombus in the culprit artery with 18 F-GP1 uptake had a sensitivity of 80%, specificity of 100%, positive predictive value of 100%, negative predictive value 92% and accuracy of 96% in this population.

Quantitative Coronary 18 F-GP1 Uptake in the Culprit Vessel
Per-vessel coronary 18 F-GP1 uptake (SUVmax and TBRmax) was higher in culprit vessels compared to non-culprit vessels as well as all coronary arteries in control subjects without myocardial infarction (Figure 3 and Supplementary Table i).Linear regression models demonstrated univariable associations between coronary 18 F-GP1 TBRmax and the timing of the scan, as well as culprit vessel status.On multivariable analysis, only culprit vessel status was associated with TBRmax (r 2 =0.246, p=0.006;Table 3).Based on the Youden's index, TBRmax >1.16 was the optimal threshold for identifying the culprit artery with high sensitivity (83%), specificity (86%) and accuracy (84%).

Sub-types of Myocardial Infarction
One patient had an apparent spontaneous coronary artery dissection involving the distal left anterior descending artery and demonstrated focal 18 F-GP1 uptake at the site of the dissection (Supplementary Figure iv).Another patient presented in sinus rhythm with an infero-posterior ST-segment elevation myocardial infarction.They had diffuse coronary atherosclerosis and a heavy thrombus burden in the culprit artery on invasive coronary angiography and underwent primary percutaneous coronary intervention to the left circumflex artery.The patient had one brief episode of paroxysmal atrial fibrillation on day 2. Focal 18 F-GP1 uptake was seem both in the left circumflex artery and the left atrial appendage suggesting a thromboembolic event and reclassification from type 1 to a type 2 myocardial infarction (Supplementary Figure v).(21)

Identification of the Culprit Lesion
Coronary 18 F-GP1 uptake identified the culprit lesion (25/27; 93%) in patients with multivessel disease (Figure 4).In a patient with triple vessel coronary artery disease and bypass grafting, 18 F-GP1 uptake only occurred in the culprit lesion of a degenerated saphenous vein bypass graft and no uptake was seen in the other severely diseased native vessels, bypass grafts or indeed the prior stented segment of the bypass graft upstream of the culprit lesion (Supplementary Figure vi).In all five patients with triple vessel disease undergoing multivessel percutaneous coronary intervention, coronary 18 F-GP1 uptake was only seen within the stented segment of the culprit lesion but not the stented segments of non-culprit lesions.Bystander disease was treated at the index procedure, but in each of these cases, the non-culprit stented vessel demonstrated no 18 F-GP1 uptake (Supplementary Figure vii).

Discussion
This is the first demonstration that non-invasive imaging can identify in vivo intracoronary thrombus in patients with acute myocardial infarction.We have confirmed the high selectivity and specificity of 18 F-GP1 binding to activated platelets within fresh human thrombus and coronary thrombectomy specimens.We have applied this tracer to populations of patients with and without acute myocardial infarction and have observed 18 F-GP1 uptake only occurs within the culprit coronary arteries of those with acute myocardial infarction.We also demonstrate a high frequency of extra-coronary thrombosis including unrecognized left ventricular and atrial thrombus as well as infarct-related intramyocardial 18 F-GP1 uptake.This technique appears to hold major promise as a method of determining the role and origin of thrombosis in acute myocardial infarction, especially in cases of thromboembolism or myocardial infarction with non-obstructive coronary arteries.It could also assist in cases of multivessel disease where the localization of the culprit lesion or presence of thrombus may influence patient management. 18F-GP1 is a radiolabelled analogue of elarofiban, a glycoprotein IIb/IIIa receptor antagonist.

Unlike other drugs in this class, elarofiban only binds to platelets with activated glycoprotein
IIb/IIIa receptors, rendering it a very specific marker of fresh platelet-containing thrombus.
Indeed, we have previously demonstrated that 18 F-GP1 binding is markedly increased when platelets are activated.(25) In the current study, we have further assessed this tracer for the coronary circulation.First, we used an ex vivo model of deep arterial injury analogous to coronary plaque rupture.(23,26,27) We demonstrated selective binding of 18 F-GP1 to freshly forming human thrombus which could be inhibited by the glycoprotein IIb/IIIa receptor antagonist, tirofiban.We then undertook autoradiography and histology of human coronary thrombi extracted at the time of acute myocardial infarction and confirmed that 18 F-GP1 does indeed bind to activated platelets within human coronary thrombus.
Qualitative assessments of 18 F-GP1 uptake in patients with acute myocardial infarction demonstrated the accurate localization of thrombus formation within the coronary arteries.
Neither the non-culprit arteries of patients with myocardial infarction nor any coronary arteries in patients without myocardial infarction demonstrated coronary 18 F-GP1 uptake.However, some patients with acute myocardial infarction did not have demonstrable 18 F-GP1 uptake in the culprit coronary artery.This imperfect sensitivity appeared to relate to three main factors: delays in performing the scans, thrombosis of small calibre coronary arteries and overspill from hepatic 18 F-GP1 uptake.All false negative cases were scanned more than 8 days after myocardial infarction which perhaps reflects the prompt resolution of intracoronary thrombus with pharmacotherapy.Indeed, in the Clopidogrel as Adjunctive ReperfusIon TherapY (CLARITY) TIMI 28 study, dual antiplatelet therapy was associated with resolution of angiographic evidence of intracoronary thrombus in over a half of patients by a median of 8 days.(28) Previous cases series suggest that up to 92% of patients with ST-segment elevation myocardial infarction have evidence of intracoronary thrombus with 16% presenting with a large thrombus burden.(29,30) In these series, a large thrombus burden was linked to larger vessels (>2 mm in diameter) and led to worse outcomes including distal embolization and microvascular obstruction with no reflow.We found evidence of a large thrombus burden in a similar proportion of patients with ST-segment elevation myocardial infarction on invasive angiography.For those without obvious thrombus on invasive or CT coronary angiography, intense coronary 18 F-GP1 uptake was seen in majority of cases.Cases with no obvious coronary 18 F-GP1 uptake involved culprit lesions in smaller distal vessels (<2 mm) leading to more modest elevations in cardiac troponin indicative of a lower thrombus burden.(31,32) 18 F-GP1 offers thrombus detection beyond the resolution of CT.Even though most culprit arteries were stented, there was demonstrable 18 F-GP1 uptake in most of these segments, consistent with a degree of abluminal thrombus trapped behind the stent which is not visible on invasive or computed tomography angiography.This is supported by cases where severe non-culprit disease was concurrently stented during the index procedure, but only the stented culprit segment demonstrated 18 F-GP1 uptake.The fact that later scans show no thrombus may be a testament to the efficacy of contemporary antiplatelet regimes.Although there are no prospective studies investigating the time course of intracoronary thrombus resolution after myocardial infarction, case reports of serial coronary angiography confirm that the thrombus burden resolves over a few days (2 to 7 days) with administration of modern antiplatelet therapies including glycoprotein IIb/IIIa receptor antagonists.(33,34) In our study, imaging was performed between 5 and 31 days post myocardial infarction and delays in imaging beyond one week from the index event appears to have reduced the sensitivity of GP1 to identify intracoronary thrombus.Therefore, for diagnostic purposes, it should be performed within the first 7-10 days of acute myocardial infarction.
Infarct-related intramyocardial 18 F-GP1 uptake was present in more than a third of cases, and we believe that this likely reflects intramyocardial hemorrhage or thrombotic microvascular obstruction.Intramyocardial hemorrhage occurs because of loss of vascular integrity following infarction and activated platelets will play a role in achieving hemostasis and limiting extravasation of blood.In contrast, microvascular obstruction predominantly occurs due to distal embolization of thrombus with occlusion of the microvasculature leading to no reflow phenomenon and potential expansion of the infarct zone.If confirmed, uptake of 18 F-GP1 within regions of microvascular obstruction would reinforce the role of thromboembolism in the pathogenesis of this condition and provide a biomarker to test the efficacy of potential therapeutic interventions.
There were notable examples of extra-coronary thrombus identified by 18 F-GP1 PET-CT that included undiagnosed left ventricular thrombus and pulmonary thromboembolism requiring treatment with oral anticoagulant therapy.Again, this may well have clinical utility in cases where the presence of thrombus is uncertain or suspected.We are currently exploring the incidence and the natural history of left ventricular thrombus as well assessing the treatment efficacy of oral anticoagulant therapy in patients with large myocardial infarctions (NCT04829825).
It is important to consider the clinical utility and potential application of 18 F-GP1 PET-CT.In our modest-sized case series, we had examples where 18 F-GP1 PET-CT changed both the diagnosis (type 1 reclassified as type 2 myocardial infarction) and the treatment (initiation of anticoagulation) of patients presenting with myocardial infarction.In patients presenting with myocardial infarction and non-obstructive coronary artery disease, 18 F-GP1 could potentially identify those with plaque rupture and intracoronary thrombus who would truly benefit from dual antiplatelet therapy.Similarly, thromboembolic causes of type 2 myocardial infarction can be distinguished from those with in situ thrombosis typical of type 1 myocardial infarction.
We therefore believe that 18 F-GP1 PET-CT will have a potential role for patients where the diagnosis is uncertain or where the role or origin of thrombus is in question.However, this needs further prospective evaluation in future studies.

Limitations
We should acknowledge some further limitations of our work.Although these data represent the first description of the application of coronary 18 F-GP1 PET-CT, the sample size is modest and further confirmatory studies are required in larger populations.Coronary arteries are small structures, and given the resolution of PET, partial volume effects as well as issues relating to sensitivity in smaller distal or tributary vessels must be considered.The latter may well be overcome by more modern PET-CT scanners which continue to evolve with great detector sensitivity combined with wider field of view scanners that can increase the sensitivity by up to 40-fold.(35,36) We also found that 18 F-GP1 liver uptake can preclude analysis of uptake in the distal right coronary artery due to signal overspill.This will be difficult to overcome due to the biliary excretion of 18 F-GP1.Although we believe that infarct-related intramyocardial uptake is likely to represent intramyocardial haemorrhage or thrombotic microvascular obstruction, this is currently speculative and further investigation of this interesting observation is warranted and is the subject of ongoing investigation (NCT04829825).Finally, we acknowledge that there may have been some inherent bias in the qualitative assessment of 18 F-GP1 uptake given the presence of intracoronary stenting will be readily apparent on review of the PET-CT scans.However, the culprit vessel was correctly identified in those without stent implantation and those with multivessel stenting, and our qualitative findings were consistent with the quantitative analysis which is more objective and independent of observer bias.
In conclusion, coronary 18 F-GP1 PET-CT can detect in vivo intracoronary thrombus in patients with acute myocardial infarction.This has the potential to help in the diagnosis, management and treatment of patients presenting with suspected or confirmed acute myocardial infarction.18F-GP1 uptake measured as maximum standardized uptake values and maximum tissue to background ratios.(p value was calculated using pairwise Wilcox test) Coronary 18 F-GP1 uptake was present only in the culprit vessel (right coronary artery; RCA) and there was no uptake in the bystander disease lesions (left anterior descending (LAD) and left circumflex (LCx) coronary arteries).

0.001
Model for log2 coronary maximum 18 F-GP1 target-to-background ratio

COMPETENCY
IN PATIENT CARE AND PROCEDURAL SKILLS: Coronary 18 F-GP1 PET-CT angiography is the first non-invasive selective technique to identify in vivo coronary thrombosis in patients with acute myocardial infarction and can potentially inform the diagnosis, management, and treatment of patients with acute myocardial infarction.TRANSLATIONAL OUTLOOK: Additional prospective trials are necessary to assess the utility of 18F-GP1 in recognising and treating culprit lesions in patients with non-ST segment elevation myocardial infarction and multivessel disease.

Figure 2 .
Figure 2. Three exemplar cases of ST segment elevation myocardial infarction.

Figure 4 .
Figure 4. 18 F-GP1 uptake in a patient with triple vessel disease.

Table 3 .
Linear regression models for coronary maximum18F-GP1 target-to-background ratio