Coronary Vessel Wall Contrast Enhancement Imaging as a Potential Direct Marker of Coronary Involvement Integration

OBJECTIVES This study investigated the feasibility of visual and quantitative assessment of coronary vessel wall contrast enhancement (CE) for detection of symptomatic atherosclerotic coronary artery disease (CAD) and subclinical coronary vasculitis in autoimmune in ﬂ ammatory disease (systemic lupus erythematosus [SLE]), as well as the association with aortic stiffness, an established marker of risk. BACKGROUND Coronary CE by cardiac magnetic resonance (CMR) is a novel noninvasive approach to visualize gadolinium contrast uptake within the coronary artery vessel wall. METHODS A total of 75 subjects (CAD: n ¼ 25; SLE: n ¼ 27; control: n ¼ 23) underwent CMR imaging using a 3-T clinical scanner. Coronary arteries were visualized by a T2-prepared steady state free precession technique. Coronary wall CE was visualized using inversion-recovery T1 weighted gradient echo sequence 40 min after administration of 0.2 mmol/kg gadobutrol. Proximal coronary segments were visually examined for distribution of CE and quanti ﬁ ed for contrast-to-noise ratio (CNR) and total CE area. RESULTS Coronary CE in patients (93%, n ¼ 42) with a diffuse pattern for SLE and a patchy/regional distribution in CAD patients. Compared with control subjects, CNR values and total CE area in patients with CAD and SLE signi ﬁ cantly higher

C oronary contrast enhancement (CE) by cardiac magnetic resonance (CMR) is a novel, noninvasive approach for visualization of gadolinium contrast uptake within the coronary artery vessel wall. Previous studies demonstrated that CE colocalizes with mixed and calcified plaques in advanced coronary artery disease (CAD) (reviewed in Kuo et al. [1]). We have recently shown that coronary CE is also present in patients with systemic lupus erythematosus (SLE) despite the absence of cardiovascular (CV) symptoms (2). Demonstration of subclinical coronary involvement in subjects with persistent systemic inflammation, accelerated atherosclerosis, and increased and premature mortality and morbidity can be potentially useful for identification of individuals at risk (3)(4)(5).Younger women with SLE appear to be particularly affected; they experience an increased rate of coronary events early in the course of systemic disease (6). As current risk stratification schemes fail to detect these patients, a noninvasive and radiation-free approach using coronary CE may help to facilitate identification of early subclinical coronary vessel wall changes in vulnerable subpopulations (7). Patients with systemic inflammatory diseases, including SLE, are for the first time being acknowledged as subpopulations with increased CV risk in practice guidelines (8). Means of early and noninvasive identification of coronary vascular changes or strategies for improved prevention in these patients remain unknown. The aim of the current study was to assess the feasibility of coronary CE to describe differences between patients with established CAD and SLE, as well as control subjects.
In addition to visualization, we employed 2 methods of CE quantification: contrast-to-noise ratio (CNR) and total CE area. We also determined the association of CE with aortic stiffness, an independent predictor of increased CV risk. Germany) using complete short-axis stack coverage and long-axis views (2-, 3-, and 4-chamber views) (11). In-plane flow acquisitions of ascending and descending aorta and aortic arch were obtained for PWV measurement during shallow free-breathing, using a retrospectively gated gradient echo pulse sequence and signal averaging (for imaging parameters please see the Online Appendix) (12). IMAGE ANALYSIS. All routine CMR analysis was performed using commercially-available software following standardized post-processing recommendations (13). Endocardial left ventricle (LV) borders were manually traced at end-diastole and -systole.

METHODS
The papillary muscles were included as part of the LV cavity volume. LV end-diastolic and -systolic volumes were determined using Simpson's rule. Ejection fraction was computed as end-diastolic volume -(endsystolic volume/end-diastolic volume). All volumetric indexes were normalized to body surface area.
Short-axis LGE images were visually examined for the presence of regional fibrosis, which showed as bright areas within the myocardium in 2 fold-over  (13). Microvascular disease was established by circumferential homogeneous gradient of epicardial to endocardial enhancement (13). A spaceaveraged PWV was measured between ascending and thoracic descending aorta using an in-house developed software, as previously described (14)(15)(16).

CORONARY IMAGING TECHNIQUE AND ANALYSIS.
All coronary imaging studies were performed as free breathing scans with respiratory navigation (details on sequence parameters are provided in the Online Appendix). First, coronary arteries were localized by a 3-dimensional T2-prepared balanced steady-state free precession CMR angiography sequence (17).
Double-oblique imaging planes parallel to the proximal RCA and LCA were then defined using a 3-point plan-scan tool for free-breathing targeted volume CMR angiography with a balanced steady-state free precession sequence with navigator-gating and correction (1,2,17). Coronary CE imaging was performed last using a T1-weighted 3-dimensional gradient echo inversion recovery sequence, typically 40 min after bolus administration, as previously Intraobserver and interobserver reproducibility and agreement were performed using Bland-Altman methods. Associations were explored by simple linear and binary logistic regression analyses. Cut-off values for discrimination between health and disease were derived using receiver-operating characteristic curve analysis using the point that maximized the trade-off between specificity and sensitivity. All tests were 2-tailed, and a p value <0.05 was considered significant.

RESULTS
Patients with CAD were older than SLE patients and control subjects (p < 0.01) ( Table 1)
Compared with control subjects and SLE patients, CAD patients had enlarged cardiac volumes, impaired global systolic function, and increased LV mass ( Table 2). None of the patients had evidence of significant valvular pathology. Twenty-two CAD patients had inducible ischemia on adenosine myocardial perfusion testing, whereas 8 patients with SLE showed a pattern of microvascular disease (Fig. 2).
Eleven SLE patients (41%) showed diffuse perimyocardial LGE, most commonly noted in midbasal inferolateral segments (20). Twenty CAD patients revealed an ischemic type of LGE (13). PWV was similarly significantly increased in both patient groups.  (Fig. 3A). Coronary CE was prevalent in both patient groups ( Table 2). SLE patients showed a diffuse enhancement pattern (Fig. 3B), whereas in patients with CAD, CE was patchy and concentrated to the areas of minimal lumen diameter (Fig. 3C).
Both patient groups had higher CNR and total CE area, with no difference between the patient groups.    and CT coronary angiography, respectively) or detection of myocardial ischemia. Even though these methods provide prognostic information and guide therapy in symptomatic CV disease (24), their use as screening tools in young, vulnerable populations is limited. Repetitive radiation exposure in young patients is likely not justified despite considerable efforts to reduce radiation dose (25,26). High calcium score correlates with more aggressive CV disease and reflects a greater burden in the general population (27). However, increased calcium score has poor correlation with disease activity or with the propensity of atherosclerotic plaques to rupture, making it difficult to predict acute coronary events (28). A previous study using calcium imaging in SLE patients showed that although age and body mass index predicted increased calcium score, SLE disease activity was not associated, challenging the utility of calcium scoring in this particular population (29). Whether calcium scoring is a useful marker of increased CV risk in young SLE patients has not been shown.
Several studies reported abnormal myocardial perfusion in SLE patients; however, perfusion defects were not regional, as seen in CAD patients, but were more diffuse as typically seen in microvascular disease, adding to the pathophysiological distinction between the 2 conditions (30). Gold-standard

Coronary Wall Contrast Enhancement
A U G U S T 2 0 1 4 : 7 6 2 -7 0 methods for assessing coronary wall remodeling such as intravascular ultrasound or optical coherence tomography (19) are invasive and unsuitable for screening of young patients with recognized high risk.
As a noninvasive imaging strategy that is radiation-free and acceptable in most subjects, coronary CE imaging thus has the potential to develop into a novel marker to detect vulnerable individuals with subclinical coronary involvement. The physiological mechanism of gadolinium tissue uptake is well documented and corresponds with the extracellular space (31). CE imaging of the myocardium marks histological substrates such as scar, fibrosis, or extracellular edema (32,33). Previous studies have shown that vessel wall CE colocalizes with areas of mixed and calcified plaque on CT. Coronary CE was also shown to be associated with acute or chronic systemic inflammation (1,34). Validation studies performed in human vascular, mainly carotid, tissue suggest that gadolinium uptake in coronary vessels may bear similarities; however, direct histology, gadolinium quantification, and colocalization in the coronary vessel wall tissue is not available (35). In the absence of a direct histological correlate, we assume that findings in SLE patients correspond to subclinical coronary wall remodeling due to persistent systemic inflammation. A study in patients after cardiac transplantation lends support in showing that increased coronary CE corresponds to increased plaque burden, as demonstrated by intravascular ultrasound (19). Therefore, it might be possible to We demonstrate that coronary CE allows for discrimination between healthy subjects and patients with overt CAD and subclinical disease in SLE. We found that control subjects show relatively little contrast uptake and lower CNR values, whereas coronary CE is substantial (visually and quantifiably) in patients with CAD and SLE. The values of CNR and total CE area in both patient groups are comparable, despite the differences in age and distribution of the risk factors and clinical condition. We further show that CE pattern in SLE is predominantly diffuse, consistent with accelerated vessel remodeling, contrasting the patchy and regional involvement in CAD, which is more limited to the areas of focal plaque formation. Thus, qualitative assessment of coronary CE as present/absent and diffuse/patchy may allow for rapid discrimination between health/disease and type of disease, and is potentially of great clinical utility. Quantification of coronary CE may add a means to grade severity of disease.
The associations with increased aortic stiffness for the total cohort as well as for separate groups further support the concept that coronary CE concords with adverse vessel wall remodeling and may potentially provide a new and direct marker for individualized CV risk assessment in vulnerable subpopulations (21).
Compared with previous studies reporting on coronary CE in patients with systemic inflammatory diseases (1,34,35), the prevalence of findings is higher in our study, and several reasons may explain this observation. Patient selection in the present study likely contributes most to this difference: patients with SLE have been recruited from a highlyspecialized rheumatology service for complex systemic inflammatory diseases. It is conceivable that we included patients with more aggressive systemic disease, which is also reflected in their coronary involvement. Patients with CAD were similarly preselected as those with known and significant CAD.
We also employed an optimized sequence with high spatial resolution at a higher field strength with greater signal-to-noise ratio (1). A sufficient time delay between contrast administration and post-contrast acquisition is crucial for maximized background tissue and blood suppression by inversion-recovery pre-pulse to allow for visualization of gadolinium uptake within the enhanced vessel wall, as previously reported (between 30 to 45 min after contrast administration) (18). Finally, we used gadobutrol, a gadolinium contrast agent with the highest magnetic "relaxivity," allowing for the greatest shortening of T1 relaxation (and contrast) within accumulated tissues (36). STUDY LIMITATIONS. This was an exploratory and hypothesis-generating pilot study with a limited sample size to test the feasibility of coronary CE imaging in pre-selected patient populations with a high CV risk. Thus, a few limitations apply to this study.
We were not able to assess the true interstudy reproducibility (a repeated study on a different occasion). In a small number of patients, we repeated coronary acquisitions within the same study for  Correlations were performed using Pearson or Spearman tests, as appropriate for the type of the data.
Abbreviations as in Tables 1 and 2. presentation to CMR (upon which they were recruited to this study) for a test of myocardial ischemia to guide subsequent coronary intervention.
Because we intend coronary CE as a means to identify the subclinical coronary injury early in the course of disease, we believe that it is unlikely that stents would be present in such patients and they would not pose a limitation to its use in practice. Associations with soluble biomarkers of disease and atherosclerotic plaque activity and the correlation with novel imaging techniques for characterization of atherosclerotic coronary plaque composition, such as coronary CT angiography, need to be examined in larger studies (19,27,35). Finally, the relation with poorer prognosis and ability to inform and guide the individualized risk assessment requires confirmation in future studies (37).