Association Between Aortic Vascular Inflammation and Coronary Artery Plaque Characteristics in Psoriasis

Key Points

Question  What is the association between aortic vascular inflammation by fludeoxyglucose F 18 (¹⁸F-FDG) positron emission tomography/computed tomography and coronary artery disease by coronary computed tomography angiography?

Findings  In this cross-sectional cohort study of 215 patients with psoriasis assessed with ¹⁸F-FDG positron emission tomography/computed tomography and coronary computed tomography angiography, aortic vascular inflammation was directly associated with quantitative burden of coronary artery disease, luminal stenosis severity within the coronary arteries, and the prevalence of high-risk coronary plaque beyond traditional cardiovascular risk factors.

Meaning  Aortic vascular inflammation by ¹⁸F-FDG positron emission tomography/computed tomography is associated with presence of coronary artery disease by coronary computed tomography angiography, suggesting that ¹⁸F-FDG PET/CT may provide a potential surrogate of early coronary artery disease.

Importance  Inflammation is critical to atherosclerosis. Psoriasis, a chronic inflammatory disease associated with early cardiovascular events and increased aortic vascular inflammation (VI), provides a model to study the process of early atherogenesis. Fludeoxyglucose F 18 positron emission tomography/computed tomography (¹⁸F-FDG PET/CT) helps quantify aortic VI, and coronary computed tomography angiography provides coronary artery disease (CAD) assessment through evaluation of total plaque burden (TB) and noncalcified coronary plaque burden (NCB), luminal stenosis, and high-risk plaques (HRP). To our knowledge, association between aortic VI and broad CAD indices has not yet been assessed in a chronic inflammatory disease state. Such a study may provide information regarding the utility of aortic VI in capturing early CAD.

Objective  To assess the association between aortic VI and CAD indices, including TB, NCB, luminal stenosis, and HRP prevalence, in psoriasis.

Design, Setting, and Participants  In a cross-sectional cohort study at the National Institutes of Health, 215 consecutive patients with psoriasis were recruited from surrounding outpatient dermatology practices. All patients underwent ¹⁸F-FDG PET/CT for aortic VI assessment, and 190 of 215 patients underwent coronary computed tomography angiography to characterize CAD. The study was conducted between January 1, 2013, and May 31, 2017. Data were analyzed in March 2018.

Exposures  Aortic VI assessed by ¹⁸F-FDG PET/CT.

Main Outcomes and Measures  Primary outcome: TB and NCB. Secondary outcomes: luminal stenosis and HRP.

Results  Among 215 patients with psoriasis (mean [SD] age, 50.4 [12.6] years; 126 men [59%]), patients with increased aortic VI had increased TB (standardized β = 0.48; P < .001), and higher prevalence of luminal stenosis (OR, 3.63; 95% CI, 1.71-7.70; P = .001) and HRP (OR, 3.05; 95% CI, 1.42-6.47; P = .004). The aortic VI and TB association was primarily driven by NCB (β = 0.49; P < .001), whereas the aortic VI and HRP association was driven by low-attenuation plaque (OR, 5.63; 95% CI, 1.96-16.19; P = .001). All associations of aortic VI remained significant after adjustment for cardiovascular risk factors: aortic VI and TB (β = 0.23; P < .001), NCB (β = 0.24; P < .001), luminal stenosis (OR, 3.40; 95% CI, 1.40-8.24; P = .007), and HRP (OR, 2.72; 95% CI, 1.08-6.83; P = .03). No association was found between aortic VI and dense-calcified coronary plaque burden.

Conclusions and Relevance  Aortic VI is associated with broad CAD indices, suggesting that aortic VI may be a surrogate for early CAD. Larger prospective studies need to assess these associations longitudinally and examine treatment effects on these outcomes.

Inflammation is increasingly recognized as a driver of atherosclerosis.¹ From early endothelial cell dysfunction to late-stage coronary plaque rupture, inflammatory mediators have a critical role in the atherosclerotic cascade.² Psoriasis, a chronic inflammatory skin disorder, is associated with heightened cardiovascular risk, increased vascular inflammation (VI), and increased incidence of early cardiovascular events including myocardial infarction and stroke.^3-6 Psoriasis being a systemic inflammatory state may therefore help understand the role of chronic inflammation in atherosclerosis in vivo.

In 2017 and 2018, findings from the Canakinumab Anti-inflammatory Thrombosis Outcomes Study (CANTOS),^7,8 inflammation reduction was demonstrated to reduce cardiovascular events in direct proportion with the extent of high-sensitivity C-reactive protein reduction, despite null effects on lipids. While these findings highlight the role of residual inflammatory risk in atherosclerosis, emerging imaging techniques provide means to assess inflammation within the vasculature and effect on atherosclerotic plaque development and progression. Aortic vascular inflammation by fluorodeoxyglucose F 18 positron emission tomography computed tomography (¹⁸F-FDG PET/CT) has been used as a surrogate marker of cardiovascular risk and immune-mediated vascular disease during the last decade.^9,10 Histologic studies corroborate that aortic VI by ¹⁸F-FDG PET/CT is associated with the activity of proinflammatory CD-68 macrophages within the arterial wall.¹¹ Aortic VI is associated with hard cardiovascular events, improves cardiovascular risk prediction,¹² and is also amenable to modulation, with known risk-modifying agents including statins, anti-inflammatory therapies, and therapeutic lifestyle changes.^13-16 However, its context in treatment trials as a quantifiable primary end point is only now gaining momentum. Furthermore, how aortic VI is associated with vascular diseases in other territories, including the coronary arteries, is understudied. A prior study characterized the association between aortic VI and high-risk coronary plaque (HRP) morphology in patients with HIV¹⁷ and found that aortic VI was associated with coronary artery disease. To our knowledge, whether aortic VI is associated with broad coronary plaque characteristics has not been evaluated.

Coronary computed tomography angiography (CCTA) is an important imaging tool for the assessment of coronary plaque.¹⁸ Coronary computed tomography angiography provides qualitative assessment of coronary plaque composition and allows for the volumetric quantification of the burden of coronary artery disease (CAD)^19-21 while sparing the risks of invasive assessment tools.²² Noncalcified coronary plaque burden (NCB) by CCTA has been shown to associate with thin cap fibroatheroma by histological evaluation²³ and is elevated in psoriasis.²¹ Furthermore, noncalcified plaques are known to be the culprit lesion in most ST-segment elevation myocardial infarction.^23-25 Total coronary plaque burden (TB) and NCB by CCTA are beneficial in the prediction of prospective cardiovascular outcomes.^26,27 Additionally, CCTA also provides a reliable tool to characterize the lipid-rich, rupture-prone HRPs, which associate with an increased risk of prospective cardiovascular events.²⁴ We have previously demonstrated that psoriasis is associated with a higher prevalence of such HRPs compared with age- and sex-matched healthy volunteers, to a similar extent as seen in hyperlipidemic patients 10 years older.²¹ Here, we used an ongoing cohort study of patients with psoriasis and leveraged multimodality imaging using ¹⁸F-FDG PET/CT and CCTA. Our goal was to analyze the association between aortic VI (Figure 1A) and CCTA-based quantitative and qualitative CAD indices including TB and NCB, luminal stenosis, and HRP characteristics (Figure 1B). We hypothesized that there would be a direct association between aortic VI and CAD indices.

This study consisted of a cohort of consecutive patients with psoriasis (n = 215), all of whom were recruited between January 1, 2013, and May 31, 2017, and underwent ¹⁸F-FDG PET/CT. Of 215 patients, 25 patients did not undergo CCTA scans owing to the unavailability of scanners on their specific visit dates; thus, 190 patients had CCTA scans (Figure 2). Of 190 patients with CCTA, 182 scans (546 arteries) were interpretable and were included in the final analysis for quantitative plaque burdens. Furthermore, for analyses of luminal stenoses, scan reads were available for 171 patients (513 arteries), while HRP characterization was available in 169 patients (507 arteries). Coronary plaque burden was evaluated across 3 major epicardial coronary arteries using the dedicated software QAngio CT (Medis) by previously described methods.¹⁹ Semiautomated segmentation of each major coronary artery was performed. The need for manual adjustment was reviewed by assessing transverse reconstructed cross-sections at 0.5-mm increments. Coronary artery disease indices, such as TB and NCB volume indices, assessed as millimeters squared, were calculated by dividing total vessel plaque volume by total vessel length for standardization and were attenuated for luminal intensity for accuracy. Clinical reads were used for analysis of luminal stenosis. Quantitative stenosis grading recommended in Society of Cardiovascular Computed Tomography guidelines was used for categorization of stenosis severity.²⁸ All coronary segments were analyzed for the presence of HRP, defined as a plaque with positive remodeling (PRP) (remodeling index ≥1.1) or presence of low-attenuation plaque (LAP) (<30 Hounsfield units). Low-attenuation plaque and PRP scores were calculated for each artery by calculating the number of LAPs or PRPs present in that artery. High-risk plaque score was calculated by addition of LAP and PRP scores. All scores were confirmed by 2 separate, trained, and blinded readers (intraclass correlation coefficient, 94% for intraexaminer and intraclass correlation coefficient, 93% for interexaminer analysis). The primary outcomes of our study were TB and NCB, whereas secondary outcomes were presence of luminal stenoses with its various severities, HRP assessed by CCTA, and various HRP scores. The main exposures were aortic VI by ¹⁸F-FDG PET/CT assessed as target-to-background ratio (TBR) and other cardiovascular risk factors. All participants provided written informed consent. The study protocols were approved by the institutional review board at the National Institutes of Health. All study protocols are in compliance with the Declaration of Helsinki.

Multivariable linear regressions were performed to examine the association between aortic VI and TB, NCB, and dense-calcified plaque. Standardized β values from these analyses were reported, which indicate number of SD change in the outcome variable (TB or NCB) per SD change in the exposure (aortic VI). Multivariable logistic regressions assessed the aortic VI–luminal stenosis and aortic VI–HRP associations, whereas ordered logistic regressions were used for studying the luminal stenosis severity–aortic VI as well as HRP score–aortic VI associations. Odds ratios (ORs) with 95% confidence intervals were reported for all logistic regression analyses. Detailed methods, including the comprehensive imaging protocols, image analyses protocols, covariate determinations, and statistical analyses, are provided in the eMethods in the Supplement. A 2-sided P value of less than .05 was considered statistically significant.

We summarize the characteristics of our study population in Table 1. The psoriasis cohort included 215 middle-aged patients (mean [SD] age, 50.4 [12.6] years), was male predominant (n = 126 [59%]), and had moderate skin disease severity by psoriasis area severity index score (median, 5.7; interqartile range [IQR], 3-10). Median psoriasis duration was 20 years. Patients with psoriasis had a mean (SD) body mass index (BMI; calculated as weight in kilograms divided by height in meters squared) suggestive of overweight profiles (29.6 [6.2]) and a high prevalence of dyslipidemia (48%) despite near-normal lipid profiles, with mean (SD) total cholesterol of 183 (37.9) mg/dL (to convert to millimoles per liter, multiply by 0.0259) and low-density lipoprotein cholesterol of 102.2 (31.1) mg/dL. Sixty-nine of 215 patients with psoriasis (32%) were receiving lipid-lowering therapy at baseline. We observed increased insulin resistance in psoriasis (homeostatic model assessment insulin resistance, 2.8; IQR, 1.6-5.0), although the prevalence of diabetes mellitus was 11%. Moreover, patients with psoriasis were at low risk for cardiovascular disease by atherosclerotic cardiovascular disease 10-year risk (median, 2.2; IQR, 0.7-6.1). Aortic VI was high in patients with psoriasis at baseline (mean [SD] TBR, 1.70 [0.26]). The quantified TB was 1.14 (0.44) mm², which was predominantly composed of NCB (mean [SD], 1.11 [0.46] mm²). A total of 82 patients (163 arteries) of 171 (513 arteries) evaluated had greater than 25% luminal stenosis. Finally, 57 patients of 169 evaluated had prevalent HRP in at least 1 major epicardial coronary artery.

We compared the patients with psoriasis with TBR at the median or higher (1.66) with those with TBR less than the median to assess whether coronary findings tracked alongside TBR. Patients with psoriasis with higher TBR (mean [SD], 1.89 [0.22]) had similar age compared with those with lower-than-median TBR (mean [SD], 1.51 [0.10]) but were predominantly composed of men (73 of 109 [67%] vs 53 of 106 [50%], P = .001) (Table 1). They also had a higher BMI (mean [SD], 32.7 [6.6] vs 26.6 [3.9]; P < .001), despite similar prevalence of diabetes, hyperlipidemia, and comparable traditional lipid profiles except high-density lipoprotein cholesterol, which was lower (mean [SD], 49.5 [13.7] mg/dL vs 62.5 [19.1] mg/dL; P < .001). Alhough both subgroups had low atherosclerotic cardiovascular disease 10-year risk, patients with higher TBR had increased atherosclerotic cardiovascular disease risk (3.5; IQR, 1.6-7.9 vs 1.5; IQR, 0.4-4.4; P < .001) compared with the other subgroup. Finally, despite similar psoriasis duration, patients with higher TBR had elevated insulin resistance and increased levels of high-sensitivity C-reactive protein concomitant with increased psoriasis severity.

On comparing the CCTA-based characteristics, quantitative TB was increased in patients with higher TBR (mean [SD], 1.33 [0.47] vs 0.97 [0.39]; P < .001), which was predominantly secondary to increased NCB (mean [SD], 1.30 [0.47] vs 0.93 [0.37]; P < .001) (Table 2). Dense-calcified burden did not differ between these groups (mean [SD], 0.04 [0.12] vs 0.04 [0.08]; P = .64). Concurrent with quantitative analyses, more patients with higher TBR had presence of luminal stenosis in at least 1 major epicardial coronary artery (49 of 82 [60%] vs 37 of 89 [37%]; P < .001) compared with patients with TBR lower than median. Number of arteries with luminal stenosis was also greater in this group (98 [40%] vs 65 [24.5%]; P < .001). Furthermore, severity of luminal stenosis was significantly worse in patients with higher TBR (P = .002). Finally, HRP prevalence was higher in those with higher TBR (42% vs 26%, P = .02). The distribution of HRP score differed significantly between 2 groups and was driven by the difference in LAP score, while PRP scores were essentially similar (Table 2). Similar analyses were performed to test whether the reverse was true for TBR, which demonstrated increased TBR in patients with luminal stenosis as well as in patients with presence of HRP compared with those without (eTables 1 and 2 in the Supplement).

Subsequent analyses showed that aortic VI was associated with both TB (β = 0.48; P < .001) and NCB (β = 0.49; P < .001) (Table 3). This association persisted after adjustment for age, sex, BMI, diabetes, hypertension, smoking, hyperlipidemia, lipid-lowering treatment with statins, high-sensitivity C-reactive protein, and psoriasis treatment with systemic/biologic agents (TB, β = 0.23; P < .001 and NCB, β = 0.24; P < .001) (eFigure 1 in the Supplement). No such association was observed for dense-calcified burden (eFigure 2 in the Supplement). For each 1-SD unit increase in aortic VI, we observed a 20% increase in NCB in unadjusted analyses and 10% higher NCB in fully adjusted analyses. Furthermore, aortic VI was associated with luminal stenosis in both unadjusted (OR, 3.63; 95% CI, 1.71-7.70; P = .001) and fully adjusted (OR, 3.40; 95% CI, 1.40-8.24; P = .007) analyses (eTable 3 in the Supplement). Analogous results were found for luminal stenosis severity (unadjusted OR, 4.03; 95% CI, 1.90-8.55; P < .001 and fully adjusted OR, 3.38; 95% CI, 1.42-8.07; P = .006). Per 1-SD increase in VI was associated with significantly increased odds of prevalent luminal stenosis (unadjusted, 1.38 and adjusted, 1.36) and severity of luminal stenosis (unadjusted, 1.42 and adjusted, 1.36). Finally, we also observed an association between aortic VI and HRP prevalence (OR, 3.05; 95% CI, 1.48-6.29; P = .002) that persisted beyond full adjustment (OR, 2.72; 95% CI, 1.08-6.83; P = .03) (eTable 4 in the Supplement). Similarly, aortic VI was also associated with HRP score (unadjusted OR, 3.03; 95% CI, 1.42-6.47; P = .004 and fully adjusted OR, 2.95; 95% CI, 1.21-7.18; P = .02), which was primarily driven by an association with LAP score (unadjusted OR, 5.63; 95% CI, 1.96-16.19; P = .001 and fully adjusted OR, 6.33; 95% CI, 1.88-21.28; P = .003), whereas PRP score did not associate with aortic VI (unadjusted OR, 1.51; 95% CI, 0.64-3.58; P = .70). Ultimately, per 1-SD increase in VI was associated with increased HRP prevalence at 1.34 times increased odds in unadjusted and 1.3 times increased odds in fully adjusted analyses.

Our analysis of aortic VI by ¹⁸F-FDG PET/CT and CAD indices by CCTA demonstrates the following findings: (1) when stratified by high aortic VI, the extent of CAD is greater in those with elevated aortic VI, in line with greater quantitative total CAD burden, increased prevalence of luminal stenoses, more severe luminal stenosis, and a higher prevalence of HRP; (2) the increase in TB is predominantly driven by NCB, whereas dense-calcified burden does not differ by varying degrees of aortic VI; and (3) we found a strong association between aortic VI by ¹⁸F-FDG PET/CT and CAD indices derived from CCTA in unadjusted analyses and beyond adjustment for traditional cardiovascular risk factors. Our study adds to the growing body of literature demonstrating that aortic VI by ¹⁸F-FDG PET/CT may be a potential surrogate for early coronary disease.

Inflammation is critical to atherosclerosis.²⁹ Psoriasis, a chronic inflammatory disease associated with increased systemic inflammation and aortic VI,^5,6 greater burden of lipid-rich NCB,²¹ higher prevalence of HRP,²¹ and elevated risk of myocardial infarction and major adverse cardiovascular events,^3,4 provides an opportunity to study the role of inflammation in early CAD. Within the last decade, aortic VI by ¹⁸F-FDG PET/CT has been validated as a surrogate marker of atherosclerosis and prospective cardiovascular risk.³⁰ Aortic VI by ¹⁸F-FDG PET/CT is associated with prospective risk of cardiovascular events¹² and modulates following cardiovascular risk-modifying therapies such as statins^14,15 and therapeutic lifestyle changes.¹³ In the past, we have demonstrated that aortic VI is increased in psoriasis and is associated with psoriasis severity^5,6 and novel nuclear magnetic resonance–derived markers of inflammation such as glycoprotein acetylation³¹ and is responsive to treatment with anti-inflammatory medications in patients with psoriasis.¹⁶

While ¹⁸F-FDG PET/CT provides a reliable tool to assess aortic VI, CCTA remains the ideal noninvasive modality for the comprehensive evaluation of CAD,²² including the assessment of coronary atherosclerotic plaque and its morphological features, beyond the coronary artery lumen.¹⁸ While CCTA-based analysis of luminal stenosis provides important prognostic information,^32-34 quantitative volumetric coronary artery plaque assessment by CCTA has been shown to provide value in assessing future risk of major adverse cardiovascular events.^20,27 Finally, imaging vulnerable coronary plaques with CCTA has also revealed significant prognostic implications in recent studies.^18,24 Given the efficacy of multimodality imaging to detect different phases of atherosclerosis in multiple vascular beds, we used concurrent imaging with ¹⁸F-FDG PET/CT and CCTA in patients with psoriasis to understand the association between aortic VI and CAD indices.

In this study, we demonstrated an association between aortic VI and TB as well as between aortic VI and NCB, both of which persisted beyond adjustment for traditional risk factors, signifying that the presence of early vascular disease may in part be comparable across multiple major vascular beds. Interestingly, dense-calcified coronary plaque burden was not associated with aortic VI, which upheld the notion that aortic VI maybe a marker of early vascular disease.⁹ Furthermore, we demonstrated that aortic VI was directly associated with presence as well as severity of luminal stenosis in major epicardial coronary arteries, denoting aortic VI as a clinically reliable indicator of CAD. Finally, aortic VI was associated with HRP prevalence and HRP score in patients with psoriasis. This was principally driven by the association between aortic VI and LAP, whereas the association between aortic VI and PRP was statistically nonsignificant. While this is a biologically interesting finding, caution needs to be exercised in interpreting this result because the absolute number of PRP was small in our study. Taken together, these findings suggest that aortic VI may be a potential surrogate marker for the evaluation of early CAD. Finally, building on the prior literature, these associations suggest that the inflammatory processes captured by ¹⁸F-FDG PET/CT may be implicated in the development as well as progression of coronary atherosclerosis.³⁵

Despite our significant findings, our study does have notable limitations. We lack hard cardiovascular outcomes. The small sample size, observational nature of our study, and cross-sectional analyses preclude any attempt at establishing causality and also make our analyses prone to unmeasured confounding. When imaging targets of small size, such as atherosclerotic plaques, the spatial resolution limitations of PET imaging create a close association between lesion size and the apparent lesion intensity.^36,37 Thus, an interaction is possible between apparent plaque size measured by CT and standardized tracer activity of the aorta, considering that larger individuals will have larger coronary arteries and aortic wall sizes. To attenuate such an interaction, we included adjustment for BMI within our multivariable model and still observed significance in the association between CAD indices and aortic VI, albeit with attenuated strength. Another limitation of our study is that VI was assessed in the aorta and not in the coronary arteries. There are several technical hurdles in attempting to assess coronary inflammation by PET. Electrocardiogram and respiratory motion gating were not available for ¹⁸F-FDG PET/CT, thus precluding fusion of coronary PET imaging with CCTA; the lesion size expected for coronary plaques, with their low-level activity, is less than the resolution of PET to quantify regional background activity; and we did not have our patients consume a high-fat diet for suppression of intense myocardial activity that would be required for coronary VI quantification. We acknowledge that coronary VI imaging with PET/CT is feasible,³⁵ especially with emerging tracers,³⁸ but these techniques are still evolving. Finally, statin therapy reduces both aortic VI and CAD plaque burden.^14,15,39,40 Although we adjusted for statin use in our regression models, we did not have data on duration of statin therapy in our patients.

Our findings build on previous evidence of an association between VI in the aortic arch by PET and LAP by CCTA in an inflammatory condition (eg, HIV).¹⁷ We confirm this association using VI from the entire aorta and also extend these findings by providing further evidence for an association between aortic VI and coronary plaque burden and composition, including the whole vessel NCB, HRP prevalence, and HRP subtypes. These findings are potentially generalizable to other inflammatory diseases but would need to be confirmed in noninflammatory states using larger generalizable cohorts in longitudinal studies. Additionally, future studies should use multimodality imaging, including CCTA concurrent with ¹⁸F-FDG PET/CT, to investigate the effect of reducing inflammation on atherosclerosis in various vascular beds.

Our findings demonstrate that aortic VI by ¹⁸F-FDG PET/CT is associated with multiple indices of coronary plaque extent and composition derived from CCTA. This provides evidence that aortic VI by ¹⁸F-FDG PET/CT may be a potential surrogate for assessment of early CAD captured by CCTA. Larger studies assessing these associations over longer periods of time are warranted. Concurrent use of multimodality imaging should assess whether targeted therapies to curb inflammation affect early vascular disease in the aorta and coronary arteries.

Corresponding Author: Nehal N. Mehta, MD, MSCE, Section of Inflammation and Cardiometabolic Diseases, National Heart, Lung, and Blood Institute, National Institutes of Health, 10 Center Dr, Clinical Research Center, Room 5-5140, Bethesda, MD 20892 (nehal.mehta@nih.gov).

Accepted for Publication: July 12, 2018.

Published Online: September 12, 2018. doi:10.1001/jamacardio.2018.2769

Author Contributions: Drs Joshi and Mehta had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Joshi, Belur, Aberra, Chung, Dave, Rivers, Mehta.

Acquisition, analysis, or interpretation of data: Joshi, Lerman, Dey, Sajja, Belur, Elnabawi, Rodante, Aberra, Chung, Salahuddin, Natarajan, Dave, Goyal, Groenendyk, Baumer, Teague, Playford, Bluemke, Ahlman, Chen, Gelfand, Mehta.

Drafting of the manuscript: Joshi, Lerman, Sajja, Belur, Elnabawi, Dave, Groenendyk.

Critical revision of the manuscript for important intellectual content: Joshi, Dey, Belur, Elnabawi, Rodante, Aberra, Chung, Salahuddin, Natarajan, Goyal, Groenendyk, Rivers, Baumer, Teague, Playford, Bluemke, Ahlman, Chen, Gelfand, Mehta.

Statistical analysis: Joshi, Dey, Sajja, Belur, Elnabawi, Aberra, Chung, Dave, Goyal, Groenendyk, Rivers, Mehta.

Obtained funding: Mehta.

Administrative, technical, or material support: Joshi, Lerman, Belur, Rodante, Chung, Baumer, Teague, Playford, Bluemke, Ahlman, Chen.

Supervision: Belur, Rodante, Chung, Gelfand, Mehta.

Image analysis: Ahlman.

Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Mehta is a full-time US government employee and receives research grants to the National Heart, Lung, and Blood Institute from AbbVie, Janssen, Celgene, and Novartis. Dr Gelfand in the past 12 months has served as a consultant and received honoraria for Coherus (DSMB), Dermira, Janssen Biologics, Merck (DSMB), Novartis Corp, Regeneron, Dr Reddy’s labs, Sanofi, and Pfizer Inc; receives research grants (to the Trustees of the University of Pennsylvania) from Abbvie, Janssen, Novartis Corp, Regeneron, Sanofi, Celgene, and Pfizer Inc; and received payment for continuing medical education work related to psoriasis that was supported indirectly by Lilly and Abbvie. Dr Gelfand is a co–patent holder of resiquimod for treatment of cutaneous T-cell lymphoma. No other disclosures were reported.

Funding/Support: This study was supported by National Heart, Lung, and Blood Institute Intramural Research Program grant HL006193- 02. Dr Gelfand is supported by National Institute of Arthritis and Musculoskeletal and Skin Diseases grant K24 AR064310.

Role of the Funder/Sponsor: The funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Additional Contributions: We thank the National Institutes of Health Clinical Center outpatient clinic-7 nurses for their invaluable contribution to the process of patient recruitment. They were compensated by the National Institutes of Health for their contributions.