Relationship Between 18F-fluorodeoxyglucose Uptake on Positron Emission Tomography and Aortic Calcification.

Introduction: Although 18F-fluorodeoxyglucose (FDG) positron emission tomography (PET) has been widely utilized to assess the extent of inflammation, the association between the extent and severity of atherosclerosis and 18F-FDG uptake on PET remains unexamined. The current study aimed to investigate whether aortic calcium (AC) scores were associated with increased aortic uptake of 18F-FDG on PET. Methods: A total of 167 consecutive patients with suspected lung cancer but unproven malignancy who underwent non-contrast-enhanced computed tomography (CT) and 18F-FDG PET/CT were enrolled. The average standardized uptake values in the ascending aorta were used to calculate the target-to-background ratio (Mean TBR). The total (thoracic and abdominal) AC scores were measured on non-contrast-enhanced chest and abdominal CT using the Agatston method, and were categorized into three groups (0, 1-399, and ≥400). The relationship between total AC scores and 18F-FDG uptake in the ascending aorta was assessed using multivariate linear regression analysis. Results: In total, 68.26% were male, and a mean age was 67.10±14.70 years. Mean TBR values increased progressively with total AC score 0, 1-399, and ≥400 (1.01±0.07, 1.08±0.09, and 1.11±0.11, respectively; p<0.00001). Multivariate linear regression analysis revealed that increased total AC scores of 1-399 (β=0.06, 95% CI: 0.01-0.11, p=0.02) and ≥400 (β=0.11, 95% CI: 0.06-0.16, p<0.001) were significantly associated with higher Mean TBR. Conclusions: The current study demonstrated that total AC scores were associated with Mean TBR. Patients with a greater extent and severity of aortic calcifications may possess increased atherosclerotic inflammatory activity as measured by 18F-FDG PET/CT.


Introduction
Atherosclerosis, a chronic disease of the arterial wall, remains one of the leading causes of mortality worldwide. Estimates show that coronary heart disease will become the largest cause of disability and death globally in the future [1]. Various studies are currently being conducted to provide clear evidence regarding the importance of processes such as lipoprotein oxidation, in ammation, and immunity in human atherosclerosis. In ammation has been long known as a risk factor for developing atherosclerosis, with recent studies highlighting it as a target marker for the treatment of atherosclerosis apart from cholesterol control [2][3][4] given the approximately 15-20% additional risk reduction [3,5,6].
Coronary artery and aortic calci cation have been surrogate markers of atherosclerosis, the severities of which have been associated with the risk of cardiovascular disease [7][8][9]. However, recent evidence has emerged suggesting that calci cation occurs during the late stage of atherosclerosis and may not capture the early stages of the disease. Therefore, no de nite conclusions have been established regarding the association between calci cation and in ammatory activity.
The current study aimed to investigate whether the calcium score of arteries measured via thoracoabdominal plain computed tomography (CT) was associated with arterial accumulation of 18 Fuorodeoxyglucose (FDG) on positron emission tomography (PET)/CT among patients without cancer.

Study population
Between February 2015 and September 2017, 1157 patients with suspected lung cancer underwent noncontrast-enhanced chest and abdominal CT and 18 F-FDG PET examination within 6 months at our institute (Toho University Omori Medical Center, Tokyo, Japan). Among such patients, the following were sequentially excluded: patients with diagnosed lung cancer (n = 926), any other malignancies (n = 53), strong lymph node accumulation (n = 9), and a history of thoracic or abdominal endovascular aortic repair (n = 2). Ultimately, the current study enrolled 167 patients (Fig.1), the medical records of whom were then retrospectively reviewed. Our study protocol was approved by the ethics committee of Toho The slice thickness was 3 mm. The CT-based attenuation correction technique was used on PET data. A model-based scatter correction for PET was performed and then reoriented in axial, sagittal, and coronal slices.
Imaging analysis of non-contrast-enhanced chest and abdominal CT Image analysis of AC scores was performed on a dedicated workstation (IntelliSpace Portal, PHILIPS).
Ascending AC, descending AC, and coronary artery calcium (CAC) scores were measured using the Agatston method [10]. AC scoring was performed using methods described by Agatston  surrounding blood vessels, and measure SUV (Fig.2). Mean 18 F-FDG uptake on PET/CT using the SUV (SUV mean ) was measured at the ascending aorta in every three slices (9 mm) and as blood-pool SUV at Comparison of SUV values measured over the entire ascending aorta and limited areas thereof This study calculated 530 slices of the entire aorta from 5 randomly selected patients in order to compare the SUV mean at the entire aorta to that measured at the ascending or descending aorta in every three slices. Strong correlation among SUV mean values was also observed (correlation coe cient 0.98, 95% con dence interval 0.87-0.99, p < 0.001). Considering the high concordance, our analyses therefore included the SUV values at the ascending aorta in every three slices.

Statistical analysis
Continuous variables were expressed as mean ± SD, whereas categorical variables were expressed as frequencies or percentages. Total AC score and CAC score were categorized into three groups (0, 1-399, and ≥400). The Mann-Whitney U test was performed to determine whether the total AC score was associated with clinical risk factors. SUV and TBR values were compared between total AC score and CAC score groups using one-way analysis of variance or the Kruskal-Wallis test. Mean TBR values were compared between the total AC score or CAC score groups strati ed according to C-reactive protein (CRP) and low-density lipoprotein cholesterol. Multivariate linear regression analysis was used to assess whether the total AC score, including thoracic and abdominal AC or CAC scores, was associated with 18 F-FDG uptake on PET in the ascending aorta after adjusting for age, gender, body mass index (BMI), history of coronary artery disease (CAD), diabetes mellitus (DM), dyslipidemia (DL), hypertension (HT), and CRP value. All analyses were conducted using STATA (Version 11, Stata Corp LP, College Station, Texas, USA), with a p value of <0.05 indicating statistical signi cance.

Results
The baseline characteristics of the study population are listed in Table 1. Accordingly, the included patients, 68.26% of whom were male, had a mean age of 67.10 ± 14.70 years and a mean BMI of 21.90 ± 3.70 kg/m 2 . Clinical CAD risk factors, such as HT, DL, and DM, were present in 28.74%, 11.98%, and 11.98% of the patients, respectively. Approximately half of patients were smokers, while 17.37% were current smokers. Compared to patients with a total AC score of 0, those with the total AC score of 1-399 and ≥400 were older and had greater rates of HT and history of CAD. Regarding the laboratory data, CRP and blood glucose levels were signi cantly increased in patients with a higher total AC score (<0.001 for all), whereas no signi cant differences in other laboratory data were noted between the three groups. Max TBR and Mean and Mean SUV values between the total AC groups and CAC score groups. Compared to those with a total AC score of 0, those with a total AC score of 1-399 and ≥400 had progressively greater Mean TBR values. However, such a relationship was not observed between patients with a CAC score of 0, 1-399, and ≥400.  Table 3 compares the Mean TBR values between the total AC score groups strati ed according to lowdensity lipoprotein cholesterol (LDL-C) and CRP levels. Accordingly, those with a total AC score of 1-399 had higher Mean TBR values compared to those with a total AC score of 0, with those having a total AC score of ≥400 showing even greater Mean TBR values, regardless of CRP values (<0.5 or ≥0.5).
Regarding the lower LDL-C values with <120, Mean TBR gradually increased with total AC 1-399 and ≥400 compared to total AC score=0, and higher Mean TBR was likely to be increased in total AC 1-399 and ≥400 compared to total AC score=0 when strati ed by LDL-C ≥120. In terms of CAC score, signi cant increase in Mean TBR was not shown across the CAC groups regardless of CRP and LDL-C levels. Table 3 The comparison of Mean TBR values between the total AC score groups strati ed according to C-reactive protein and low-density lipoprotein cholesterol. Abbreviations: TBR, target-to-background ratio; AC, aortic calcium; CRP, C-reactive protein; CAC, coronary artery calcium; LDL-C, low-density lipoprotein cholesterol The relationship between Mean TBR and total AC or CAC score is presented in Table 4. After adjusting for age, sex, BMI, CAD, DM, DL, HT, and CRP value, multivariate linear regression analysis revealed that increased total AC scores (1-399 and ≥400) were signi cantly associated with higher Mean TBR compared to total AC score 0. In contrast, after adjusting for age, sex, BMI, CAD, DM, DL, HT, and CRP value, multivariate linear regression analysis revealed that CAC scores of 1-399, and ≥400 were not signi cantly associated with Mean TBR.

Discussion
Although the current study demonstrated that the extent and severity of aortic calci cations was associated with increased 18 F-FDG uptake on PET, our results showed no association between CAC scores and 18 F -FDG uptake. Regardless of CRP and LDL-C levels, those with higher total AC scores exhibited increased 18 F-FDG uptake on PET. In theory, vascular calci cation and vascular metabolic activity rarely overlap, suggesting that these ndings represent different stages of atheroma evolution [11]. While macro-calci cations are thought to occur at the later stages of the atherosclerosis process, global calci cations have been suggested to re ect overall atherosclerosis, including noncalci ed and calci ed atherosclerosis. Therefore, atherosclerosis in the coronary artery or aorta (i.e., CAC or AC scores) have been associated with higher cardiovascular events or mortality [8,12,13]. Numerous studies have evaluated 18 F-FDG uptake on PET in vascular in ammation [14,15] and atherosclerotic lesions in patients with cancer, psoriasis, rheumatoid arthritis, and chronic kidney disease, as well as those taking antiin ammatory drugs [11, [16][17][18][19]. Moreover, limited studies have reported an association between vascular calci cations and in ammation [20,21]. However, there is still an ongoing debate regarding the association between calci cation, plaque vulnerability, and in ammatory activity in plaque. Our group recently reported that details related to calci ed plaque (i.e., calci ed density) measured by non-contrastenhanced CT in the coronary artery were associated with optical coherence tomography (OCT)-derived calci ed size but not with OCT-derived plaque vulnerability [22]. The aforementioned study emphasized that CT-derived calcium density in local macro-calci cations may not always indicate local plaque vulnerability, although the association between calci cations and plaque activity had not been assessed.
Similarly, a study of 183 patients showed that those with increased local coronary 18 F-Fluoride uptake in at least one coronary artery were likely to have higher overall CAC scores [23]. However, local coronary 18 F-Fluoride uptake was not associated with overall CAC progression. Our ndings expanded these results by showing that Mean TBR of the aorta re ected the overall extent and severity of atherosclerosis (i.e., total AC scores in the current study). The Mean TBR value of 1.08 ± 0.10 in the aorta obtained herein was relatively low compared to that presents in previous studies, which ranged from 1.13 to 1.97 in the carotid and other vascular arteries [20,[24][25][26][27][28]. This may have been due to the lower presence of traditional risk factors for CAD, such as HT, DL, and DM, among our patients. However, no prior study had compared the association between calci cation and 18 F-FDG uptake on PET in the aorta among low-risk patients.
Despite such a lower risk of CAD, the current study observed a signi cant association between Mean TBR and total AC scores among patients with suspected but undiagnosed lung cancer. Additionally, the association between increased total AC scores and higher Mean TBR values was consistently observed regardless of CRP or LDL-C values. Our ndings suggest that 18 F-FDG PET can be an indicator of imperceptible vasculitis and that the association between 18 F -FDG uptake on PET and the extent and severity of calci ed plaque in the aorta may be consistent regardless of coronary vascular risk, potentially suggesting that 18 F-FDG PET can be utilized for the early detection of atherosclerotic activity.
Investigating methods for assessing arterial calci cation in combination with 18 F-FDG uptake on PET/CT may provide additional insights into atherosclerosis and facilitate new clinical applications. Moreover, studies investigating the prognostic utility of combined evaluation will be required.
Some limitations of the current study are worth noting. First, this study was a single-center, retrospective study with a relatively small sample size. Second, as noted earlier, our study population comprised patients who underwent 18 F-FDG PET/CT due to suspicion of malignant disease. Therefore, the association between 18 F-FDG PET and aortic calci cations in patients at higher CAD risk still remains unknown. Third, although several cytokines, such as IL-6, have been associated with 18 F-FDG uptake on PET, such variables had not been assessed herein.

Conclusion
The current study demonstrated that among patients without cancer who underwent 18 F-FDG PET, the total AC scores were associated with Mean TBR. Patients with greater extent and severity of aortic calci cations may exhibit increased atherosclerotic in ammatory activity as measured by 18  Image measuring standardized uptake values (SUV)