Thoracic Aortic 18F-Sodium Fluoride Activity and Ischemic Stroke in Patients With Established Cardiovascular Disease

Background Aortic atherosclerosis represents an important contributor to ischemic stroke risk. Identifying patients with high-risk aortic atheroma could improve preventative treatment strategies for future ischemic stroke. Objectives The purpose of this study was to investigate whether thoracic 18F-sodium fluoride positron emission tomography (PET) could improve the identification of patients at the highest risk of ischemic stroke. Methods In a post hoc observational cohort study, we quantified thoracic aortic and coronary 18F-sodium fluoride activity in 461 patients with stable cardiovascular disease undergoing PET combined with computed tomography (CT). Progression of atherosclerosis was assessed by change in aortic and coronary CT calcium volume. Clinical outcomes were determined by the occurrence of ischemic stroke and myocardial infarction. We compared the prognostic utility of 18F-sodium fluoride activity for predicting stroke to clinical risk scores and CT calcium quantification using survival analysis and multivariable Cox regression. Results After 12.7 ± 2.7 months, progression of thoracic aortic calcium volume correlated with baseline thoracic aortic 18F-sodium fluoride activity (n = 140; r = 0.31; P = 0.00016). In 461 patients, 23 (5%) patients experienced an ischemic stroke and 32 (7%) a myocardial infarction after 6.1 ± 2.3 years of follow-up. High thoracic aortic 18F-sodium fluoride activity was strongly associated with ischemic stroke (HR: 10.3 [95% CI: 3.1-34.8]; P = 0.00017), but not myocardial infarction (P = 0.40). Conversely, high coronary 18F-sodium fluoride activity was associated with myocardial infarction (HR: 4.8 [95% CI: 1.9-12.2]; P = 0.00095) but not ischemic stroke (P = 0.39). In a multivariable Cox regression model including imaging and clinical risk factors, thoracic aortic 18F-sodium fluoride activity was the only variable associated with ischemic stroke (HR: 8.19 [95% CI: 2.33-28.7], P = 0.0010). Conclusions In patients with established cardiovascular disease, thoracic aortic 18F-sodium fluoride activity is associated with the progression of atherosclerosis and future ischemic stroke. Arterial 18F-sodium fluoride activity identifies localized areas of atherosclerotic disease activity that are directly linked to disease progression and downstream regional clinical atherothrombotic events. (DIAMOND–Dual Antiplatelet Therapy to Reduce Myocardial Injury [DIAMOND], NCT02110303; Study Investigating the Effect of Drugs Used to Treat Osteoporosis on the Progression of Calcific Aortic Stenosis [SALTIRE II], NCT02132026; Novel Imaging Approaches To Identify Unstable Coronary Plaques, NCT01749254; and Role of Active Valvular Calcification and Inflammation in Patients With Aortic Stenosis, NCT01358513)

I schemic stroke remains a leading cause of serious long-term disability and mortality across the world. 1 Current preventative strategies focus on addressing the underlying causes and modifiable risk factors for stroke. 2 Comprehensive analysis of multiple large community data sets has allowed the optimization of clinical risk scores providing generalized estimates of stroke risk. 3,4 Although these well-validated estimates provide a guide to risk at the epidemiological level, noninvasive imaging has the potential to detect and to quantify disease in a more precise and patient-specific manner. Such information can provide a personalized approach to risk stratification and preventative treatment.
Imaging of the thoracic aorta can directly visualize atheromatous lesions, which have been consistently associated with the risk of ischemic stroke. 5,6 Calcified atheromatous lesions of the ascending and arch of the aorta are readily detected and quantified on conventional computed tomography (CT), with calcium scores providing some incremental value for the prediction of future stroke risk. 7,8 However, these overtly calcified vascular lesions are thought to represent a later and more stable stage in the disease process.
Conversely, arterial 18 F-sodium fluoride positron emission tomography (PET) identifies an earlier and more active stage of atheromatous disease that is associated with plaque vulnerability and the culprit lesions underlying atherothrombotic events. [9][10][11] In other cardiovascular conditions, arterial 18 F-sodium fluoride activity provides an assessment of disease activity that is associated with disease progression and clinical events. 12 The potential of aortic 18 F-sodium fluoride activity to assess thoracic aortic atherosclerotic disease progression and to predict downstream clinical outcomes is unknown. 13,14 We here assess whether thoracic aortic 18 Figure 1 for CONSORT diagram). The principal findings of these studies have been reported previously, with both randomized controlled trials reporting no difference in the primary outcome between treatment and placebo groups. 9,[15][16][17] Demographics, clinical risk factors, and history of cardiovascular disease were recorded, and 10-year revised Framingham stroke risk score was calculated for each patient. 3 This study complies with the Declaration of Helsinki, with each of the studies approved by regional ethical committees  Calcium scores, calcium volume and calcium mass were calculated across the ascending aorta and aortic arch on the attenuation correction CT scans for each patient using OsiriX version 12.0.0 (Bernex) as described previously. 22 To allow for direct comparison to thoracic aortic 18   Ascending calcium mass Ascending calcium score, AU

RESULTS
The final study cohort comprised 461 patients with advanced stable coronary artery disease or aortic stenosis followed up for a mean of 6.1 AE 2.3 years (Supplemental Figure 2  Values are n (%) or mean AE SD.

DISCUSSION
This is the largest cohort of patients with cardiovascular disease undergoing prospective thoracic 18 F-sodium fluoride PET-CT. We demonstrate, for the first time, that thoracic aortic 18 F-sodium fluoride activity is associated with both thoracic aortic atheromatous plaque disease progression and a 10-fold increased future risk of ischemic stroke. Importantly, we found that regional 18  We have previously explored total coronary 18 F-sodium fluoride activity using coronary 18 F-sodium fluoride activity as a marker of global tracer activity across the coronary circulation. We demonstrated that a higher coronary 18 F-sodium fluoride activity was associated with faster coronary disease progression and served as a powerful predictor of myocardial infarction, outperforming clinical risk scores and coronary artery calcium scores. 12,18 This work led us to hypothesize that similar associations might apply to thoracic aortic atheroma and the risk of ischemic stroke. Given that the thoracic aorta is in the field of view of all of our previous prospective cardiovascular 18 F-sodium fluoride PET-CT studies, we aimed to assess the relationship between aortic 18 F-sodium fluoride activity, the progression of aortic calcification, and the risk of ischemic stroke. We were able to demonstrate that thoracic aortic 18 F-sodium fluoride activity was associated with both progression of aortic atherosclerosis and subsequent ischemic stroke. This is consistent with prior work in other disease states that have found 18 F-sodium fluoride activity to be indicative of disease activity, providing powerful prediction of disease progression and clinical events. 9,18,24 We explored whether the relationship between 18 F-sodium fluoride activity and cardiovascular events was specific to the vascular territory under evaluation. Would thoracic aortic 18 F-sodium fluoride activity predict myocardial infarction, and would coronary 18 F-sodium fluoride activity predict ischemic stroke? We observed that atherosclerotic 18 F-sodium fluoride activity was specific to the circulation being assessed, with coronary 18 F-sodium fluoride activity being able to identify those at risk of myocardial infarction but not stroke, whereas thoracic aortic 18 F-sodium fluoride activity identified ischemic stroke risk but not the risk of myocardial infarction (Central Illustration). These findings are intuitive and plausible given that coronary plaque will not cause stroke and aortic plaque will not cause myocardial infarction. How might this relationship be used to improve risk stratification and targeted treatment strategies? Current strategies to prevent future stroke involve identifying those at risk and targeting modifiable risk factors. 25 Clinical risk scores, such as revised Framingham stroke risk, can identify those at highest risk but suffer from a lack of patient-level specificity: like many risk scores, most events occur in low-risk patients. 26 The ability to detect the activity of thoracic aortic atheroma may help adjust future stroke risk profiles and potentially lead to better targeted treatment for patients at the highest risk (for an example see Figure 5). 27 Previous work has identified that thoracic aortic calcium scores are associated with stroke risk independent of clinical factors. 7,8,28,29 This is consistent with aortic atheroma proximal to the origins of the carotid and vertebral arteries representing a major source of stroke-related atherothrombotic emboli. Our work builds on this concept by incorporating biological disease activity and demonstrates that thoracic aortic 18  We set out to explore whether thoracic aortic atherosclerosis disease activity was associated with the future risk of atherothromboembolic clinical events. We therefore specifically examined for ischemic stroke events but excluded lacunar strokes, because of its presumed differing pathophysiology, as well as transient ischemic attacks, because there is often clinical uncertainty regarding these events and we did not wish to introduce unnecessary noise from the misclassification of events. Consequently, we restricted our analysis to patients with clinical presentations of ischemic stroke that are likely to be of atherothromboembolic origin. Although we were able to demonstrate a strong association with thoracic aortic 18 F-sodium fluoride activity, we cannot be certain that these events were attributable to thromboembolism from aortic atheroma. Indeed, we accept that thoracic aortic 18 F-sodium fluoride activity may also be indicative of disease activity within the head and neck vessels. Unfortunately, the head and neck STUDY LIMITATIONS. It is important to highlight some further limitations of our work. We acknowledge that the overall number of stroke events is relatively small although the overall incidence rate of 9.2 per 1,000 patient-years is almost double that reported in a cohort of similar age and ethnicity, likely reflecting the enrichment and inclusion criteria of cardiovascular disease in our study cohort. 30 Although the current work represent largest study assessing of thoracic aortic 18 F-sodium fluoride PET, the small number of stroke events limits the robustness of our conclusions and requires further validation in bigger cohorts with larger numbers of events.
We have combined 4 cohorts of patients with a combination of coronary artery disease and aortic stenosis representing a relatively heterogeneous cohort.
Overall, combining these groups reflects a cohort of patients with prevalent cardiovascular risk factors, but the results may not be applicable to those with lower overall cardiovascular risk. Finally, while demonstrating an association between aortic 18 Fsodium fluoride activity and stroke, the mechanism by which risk is conferred can only be notional, with some patients having competing risks for stroke.
Large prospective studies assessing the relationship between thoracic aortic 18 F-sodium fluoride activity and stroke, as well as coronary 18 F-sodium fluoride activity and myocardial infarction, are now required to validate our findings in further external patient cohorts.

CONCLUSIONS
We have found that high thoracic aortic 18