ww.sciencedirect.com

J o u rn a l o f C a r d i o v a s c u l a r C om p u t e d T omog r a p h y 7 ( 2 0 1 3 ) 1 2 5e1 3 2

Available online at w

journal homepage: www.JournalofCardiovascularCT.com

Original Research Article

Relationship of epicardial fat volume to coronary plaque,severe coronary stenosis, and high-risk coronary plaquefeatures assessed by coronary CT angiography

Ronak Rajani DM, MRCPa,b, Haim Shmilovich MDa,b, Ryo Nakazato PhD, MDa,b,Rine Nakanishi PhD, MDa,b, Yuka Otaki PhD, MDa,b, Victor Y. Cheng MDa,b,Sean W. Hayes MDa,b, Louise E.J. Thomson MBChBa,b, John D. Friedman MDa,b,Piotr J. Slomka PhDa,b, James K. Min MDa,b, Daniel S. Berman MDa,b, Damini Dey PhDc,*aDepartment of Imaging, Cedars-Sinai Medical Center, Cedars-Sinai Heart Institute, 8700 Beverly Blvd., Los Angeles, CA 90048, USAbDepartment of Medicine, Cedars-Sinai Medical Center, Cedars-Sinai Heart Institute, 8700 Beverly Blvd., Los Angeles, CA 90048, USAcBiomedical Imaging Research Institute, Department of Biomedical Sciences, Cedars-Sinai Medical Center, 8700 Beverly Blvd., Los Angeles,

CA 90048, USA

a r t i c l e i n f o

Article history:

Received 8 November 2012

Received in revised form

15 January 2013

Accepted 21 February 2013

Keywords:

Epicardial fat volume

Coronary computed tomography

angiography

Coronary artery stenosis

High-risk plaque features

Conflict of interests: The authors report no* Corresponding author.E-mail address: damini.dey@cshs.org (D.

1934-5925/$ e see front matter ª 2013 Sociehttp://dx.doi.org/10.1016/j.jcct.2013.02.003

a b s t r a c t

Background: Associations of epicardial fat volume (EFV) measured on noncontrast cardiac

CT (NCT) include coronary plaque, myocardial ischemia, and adverse cardiac events.

Objectives: This study aimed to define the relationship of EFV to coronary plaque type,

severe coronary stenosis, and the presence of high-risk plaque features (HRPFs).

Methods: We retrospectively evaluated 402 consecutive patients, with no prior history of

coronary artery disease, who underwent same day NCT and coronary CT angiography

(CTA). EFV was measured on NCT with the use of validated, semiautomated software. The

coronary arteries were evaluated for coronary plaque type (calcified [CP], noncalcified

[NCP], or partially calcified [PCP]) and coronary stenosis severity �70% with the use of

coronary CTA. For patients with NCP and PCP, 2 high-risk plaque features were evaluated:

low-attenuation plaque and positive remodeling.

Results: Therewere402patientswithamedianageof66years (range,23e92years)ofwhom226

(56%)weremen.TheEFVwasgreater inpatientswithCP (112� 55 cm3vs89� 39 cm3), PCP (110

� 57 cm3 vs 98� 45 cm3), and NCP (115� 44 cm3 vs EFV 100� 52 cm3). In the 192 patients with

PCP or NCP, on multivariable analysis, after adjusting for conventional cardiovascular risk

factors, EFV was an independent predictor of �70% coronary artery stenosis (odds ratio [OR],

3.0; 95% CI, 1.3e6.6; P¼ 0.008), any high-risk plaque features (OR, 1.7; 95% CI, 0.9e3.4; P¼ 0.04),

and low attention plaque (OR, 2.4; 95% CI, 1.1e5.1; P ¼ 0.02) but not of positive remodeling.

Conclusions: EFV is greater in patients with CP, PCP, and NCP. In patients with NCP and PCP,

EFV is significantly associated with severe coronary stenosis, high-risk plaque features,

and low attenuation plaque.

ª 2013 Society of Cardiovascular Computed Tomography. All rights reserved.

conflict of interest.

Dey).ty of Cardiovascular Computed Tomography. All rights reserved.

J o u r n a l o f C a r d i o v a s c u l a r C om p u t e d T omog r a p h y 7 ( 2 0 1 3 ) 1 2 5e1 3 2126

1. Introduction

nonfasting glucose of >200 mg/dL. Family history of coronary

Epicardial fat volume (EFV), measured on noncontrast

enhanced CT (NCT), has emerged as an important parameter

in understanding the pathophysiology of coronary athero-

sclerosis. Recent studies have shown a relationship between

EFV and the presence and severity of coronary plaque as

assessed by coronary artery calcification.1,2 Furthermore,

emerging data now suggest that this parameter may be

related to myocardial ischemia,3 acute coronary syndrome,4

and also prognosis.4,5 However, despite these observations,

uncertainty remains as to how epicardial fat may exert these

detrimental effects. One potential hypothesis that has

emerged is that EFV may exert a local paracrine effect on

adjacent coronary artery segments and lead to local inflam-

mation and changes in plaque architecture.6 Supporting this

hypothesis are data showing that specific structural plaque

characteristics identified on coronary CT angiography (CTA)

are associated with culprit coronary lesions and an increased

risk of future cardiac events.7e10

Although EFV has been shown to be related to coronary

plaque11,12 and coronary stenosis,13e15 whether EFV is related

to high-risk plaque features (HRPFs) is a subject of ongoing

investigation. Oka et al16 showed an independent association

between EFV and HRPFs in patients with noncalcified coro-

nary plaques, and Schlett et al17 have shown EFV to be asso-

ciated to HRPFs in 13 patients with a high-risk coronary lesion

presenting with chest pain. The relationship between EFV and

thin-capped fibroatheroma has also been recently confirmed

on invasive coronary angiography and optical coherence

tomography.18 One area of current uncertainty is the associ-

ation of EFV to severe coronary stenosis, plaque type, and

HRPFs in patients with stable symptoms being referred for

coronary CTA. The principal aims of the present study were to

investigate the relationship of EFV to coronary plaque type,

severe coronary stenosis, and HRPFs in stable patients with

either noncalcified or partially calcified coronary plaques.

2. Methods

2.1. Patients

We retrospectively studied 402 consecutive patients who

underwent same-day coronary CTA and NCT at Cedars-Sinai

Medical Center from January 2009 through December 2009.

Patients were excluded if they had a prior history of coronary

artery disease (myocardial infarction, coronary stenting, and

prior bypass surgery), if their bodymass index (BMI; calculated

as weight divided by height; kg/m2) was beyond limits set by

mean BMI � 2 SDs, and if their image quality was not

considered good or excellent by an expert reader. Cardiovas-

cular risk factors were determined by preset criteria. Hyper-

tension was defined as a systolic blood pressure of >140 mm

Hg, a diastolic blood pressure of >90 mm Hg, or antihyper-

tensive drug use. Smoking was defined as a current smoker or

past heavy smoker (>20 package-years). Diabetes mellitus

was defined as a previously established diagnosis, insulin or

oral hypoglycemic therapy, fasting glucose of >126 mg/dL, or

artery disease was defined as myocardial infarction, coronary

revascularization, or sudden cardiac death in a first-degree

male relative <55 years or female relative aged <65 years.

EFV of all patients was measured on NCT with the use of

semiautomated quantitative software. Detailed assessment of

coronary plaque severity, type, and structural characteristics

was performed on coronary CTA images.

2.2. Nonecontrast-enhanced CT scan

All subjects underwent NCT on a on a dual-source CT scanner

(SOMATOM Definition; Siemens Medical Solutions, For-

chheim, Germany). The scan extended from the aortic arch to

the diaphragm and was obtained in a single breathhold. Heart

rateedependent electrocardiogram (ECG)etriggering was

performed, typically at 45%e60% of the R-R interval. The field

of view was 35 cm. Matrix size was 512 � 512. Tube voltage

was 120 kVpwithmultislice scanning. The slice thickness was

3 mm for electron-beam CT and 2.5 mm for multislice CT.

Each set of NCT images was evaluated for Agatston score by

an experienced reader blinded to results of epicardial fat

measurements, using semi-automatic, commercially avail-

able software (ScImage, Los Altos, CA, USA). The total Agat-

ston score was the sum of calcified plaque (CP) scores of all

coronary arteries.19

2.3. Epicardial fat quantification

Epicardial fat quantification was performed on the NCT scan

with the use of validated software (QFAT) developed at the

Cedars-Sinai Medical Center.20 Scans were presented to blin-

ded experienced readers in random order. Readers identified

the superior epicardial fat boundary at the takeoff of the right

pulmonary artery, and the inferior boundary at the first slice

where the posterior descending artery was first visualized.

Five to 10 contour points were then traced on the pericardium

at each slice from the upper slice limit to the lower slice in the

axial view (total number of slices ranged from 20 to 40 for

typical subjects). From these control points, piecewise cubic

Catmull-Rom spline functions were automatically generated

to form a smooth, closed pericardial contour (Fig. 1). The EFV

was then automatically calculated (reported in cm3), using

contiguous 3-dimensional voxels with Hounsfield units

between �190 and �30 as the range of attenuation values

defining fat.21e23 The reproducibility of QFAT software

measurement of EFV has previously been shown to be high for

both intrascanner (multidetector CT [MDCT]eMDCT) and

interscanner (electron beam CTeMDCT) data (correlation

coefficient � 0.98). The reproducibility coefficient values are

lowest (4.3% for EFV) for intrascanner same-observer

measurement. For intrascanner cross-observer measure-

ment, reproducibility values are 10.7% for EFV.24

2.4. Coronary CTA image acquisition and reconstruction

Coronary CTA was performed on a dual-source CT scanner

(SOMATOM Definition; Siemens Medical Solutions). The

imaging protocol has been previously described in detail.25,26

Figure 1 e The measurement of EFV in a patient with low (AeC) and high (DeF) EFV. The user identifies the pericardium and

places approximately 10 contour points manually on the pericardium at each slice from the upper slice limit to the lower

slice in the axial view. From these control points, the software (Q-FAT) automatically quantifies all fat voxels within the

pericardial contour (blue line) and generates the EFV (red ). EFV, epicardial fat volume.

J o u rn a l o f C a r d i o v a s c u l a r C om p u t e d T omog r a p h y 7 ( 2 0 1 3 ) 1 2 5e1 3 2 127

b-Blockadewithmetoprolol was used to achieve a heart rate of

<70 beats/min, and 0.4 mg of nitroglycerin spray (Sciele

Pharma, Alpharetta, GA, USA) was administered 3e5 minutes

before the scan. Eighty milliliters of intravenous contrast

(Omnipaque; GE Healthcare, Princeton, NJ, USA) followed by

50e80mL of saline at a rate of 5mL/swere power-injected into

the antecubital vein. Ascending aorta contrast-triggered (100

HU), ECG-gated helical scanning was then performed in

a single breathhold. Scanning parameters included heart

rateedependent pitch (0.2e0.45), gantry rotation time of 330

milliseconds, tube voltage of 100 or 120 kVp, depending on

patient BMI,26 and reference tube current of 330e350mAs. The

acquired coronary CTA datawas reconstructed inmid diastole

and at end systole with the use of 0.6-mm slice thickness

(0.75 mm if BMI was >35), 0.3-mm slice increment, 250 mm

field of view, 512 � 512 matrix, and B26f “medium-smooth”

J o u r n a l o f C a r d i o v a s c u l a r C om p u t e d T omog r a p h y 7 ( 2 0 1 3 ) 1 2 5e1 3 2128

kernel. If reconstruction from standard phases of the cardiac

cycle resulted in uninterpretable segments, additional phases

were reconstructed and analyzed.

2.5. Coronary CTA analysis

For the purposes of this research, plaque assessment on

coronary CTAwas independently performed by 2 experienced

readers with the use of axial images, oblique multiplanar

reformations, and oblique maximum intensity projections.27

Plaque type was classified into CP (plaque consisting of only

calcium), noncalcified plaque (NCP; plaque that was calcium

free), and partially calcified plaque (PCP; plaque that had

calcified and noncalcified components). Stenosis severity was

manually quantified with the luminal diameter ratio between

the sites of maximal stenosis and proximal healthy reference,

as previously reported.28 Severe coronary stenosis was

defined as �70% reduction in luminal diameter. For each NCP

and PCP, readers determined the presence of 2 HRPFs, shown

to predict subsequent adverse cardiovascular events (Fig. 2):

LAP, defined as visually distinct intraplaque hypoattenuation

that contained �30 HU, and PR, defined as maximal outer

arterial wall diameter along the plaque exceeding proximal

reference by �5% (Fig. 2). Patients were classified as having

none, 1, or 2 HRPFs. All readers were blinded to the results of

Figure 2 e The assessment of HRPFs. (A) The presence of 3

HRPFs are shown within the proximal right coronary

artery: a >70% stenosis, LAP (closed arrow), and PR at the

site of noncalcified plaque (open arrow). (B) The presence of

a proximal partially calcified plaque is shown within the

right coronary artery (open arrow) and a more distal

noncalcified plaque with a LAP component (closed arrow).

HRPFs, high-risk plaque features; LAP, low attenuation

plaque; PR, positive remodeling.

the EFV measurements. Consensus was used to resolve

discrepancies. Study quality was graded on a 5-point quali-

tative scale that ranged from uninterpretable to excellent.

Only studies graded as being good (some minor artifact

present but all coronary segments evaluable) and excellent

(no artifact present, all coronary segments evaluable) were

included in the study.

2.6. Statistical analysis

All continuous variables included in the analysis are pre-

sented as mean � SD. Variables with non-normal distribu-

tions are presented asmedianwith range. Univariate analyses

were performed on continuous variables with the use of the

2-sample t test for normally distributed variables and the

Mann-Whitney U test for non-normally distributed data. The

distribution of EFV was positively skewed; therefore, natural-

log transformed values were used throughout, unless other-

wise specified. Spearman correlation coefficient was used to

assess the relationship between continuous variables. Multi-

variable logistic regression was used to determine the

predictors of severe coronary stenosis and HRPFs with the use

of age, sex, BMI, hypertension, hypercholesterolemia, dia-

betes mellitus, active smoking, a family history of premature

coronary disease, and EFV as covariates. Statistical signifi-

cance for all analyses was set at the 5% level. All data were

collected and analyzed with SPSS for MAC version 17 (IBM,

Somers, NY, USA).

3. Results

There were 402 patients with a median age of 66 years (range,

23e92 years) of whom 226 were men (56%). Coronary CTAwas

predominantly performed for symptoms of chest pain in 227

patients (56.4%), shortness of breath in 107 (26.6%), as

a preoperative assessment in 15 (4%), and for an abnormal

resting ECG in 15 (4%). Other indications included equivocal

stress test results and prior positive Agatston scores. Table 1

shows the baseline demographics of the population. EFV

was weakly correlated to increasing age (r ¼ 0.29, P < 0.001),

BMI (r ¼ 0.42, P < 0.001), and also body surface area (r ¼ 0.39,

P< 0.001). On univariate analysis, EFVwas also greater inmen

[median (range): 102 cm3 (16e476 cm3) vs 85 cm3 (19e310 cm3);

P < 0.001], in patients with hypertension [107 cm3

(16e310 cm3) vs 81 (19e476 cm3); P < 0.001], hypercholester-

olemia [101 cm3 (30e310 cm3) vs 84 cm3 (16e476 cm3); P ¼0.001], and a prior smoking history [127 cm3 (42e266 cm3) vs

83 cm3 (16e476 cm3); P ¼ 0.007], but not in patients with dia-

betesmellitus or a significant family history of coronary artery

disease.

3.1. EFV and type of coronary artery plaque

Coronary CTA detected coronary plaques in 294 patients (73%)

and no plaques in 108 patients (27%). Calcified plaques were

present in 237 patients (59%) with a median plaque number of

3 (range, 1e12). Partially calcified coronary plaques were

present in 166 patients (41%) with a median plaque number of

2 (range, 1e8). NCPs were present in 76 patients (19%) with

Table 1 e Baseline demographics.

Total

Number 402

Age, y, median (range) 66 (23e92)

Sex ratio, M:F, % 56:44

BMI, mean � SD 26.7 � 4.4

Hypercholesterolemia, n (%) 253 (63)

Smoking, n (%) 41 (10)

Diabetes mellitus, n (%) 55 (13.7)

Family history, n (%) 148 (37)

Hypertension, n (%) 215 (54)

EFV, cm3, mean � SD 103 � 51

Log EFV, cm3, mean � SD 4.52 � 0.49

Agatston score, AU, median (range) 57 (0e5272)

AU, Agatston unit; BMI, body mass index (calculated as weight

divided by height; kg/m2); EFV, epicardial fat volume.

Figure 3 e Comparison of EFV (mean ± SD) in patients with

and without ‡70% stenosis (A), LAP (B), and PR (C). EFV,

epicardial fat volume; LAP, low attenuation plaque; PR,

positive remodeling.

J o u rn a l o f C a r d i o v a s c u l a r C om p u t e d T omog r a p h y 7 ( 2 0 1 3 ) 1 2 5e1 3 2 129

a median plaque number of 1 (range, 1e6). EFV was greater in

patients with coronary plaques (108 � 53 cm3) than patients

without (89 � 41 cm3; P < 0.001). EFV was greater in patients

with CP [EFV with CP 112 � 55 cm3 [n ¼ 165] vs EFV without CP

89 � 39 cm3; P < 0.001), PCP (EFV with PCP 110 � 57 cm3 [n ¼236] vs EFV without PCP 98 � 45 cm3; P ¼ 0.02), and NCP (EFV

with NCP 115 � 44 cm3 [n ¼ 326] vs EFV without NCP 100 � 52

cm3; P ¼ 0.03).

3.2. EFV and number of coronary artery plaques

EFV was weakly correlated to the total number of plaques

(P < 0.001, r ¼ 0.28), the number of CPs (P < 0.001, r ¼ 0.28), the

number of PCPs (P ¼ 0.009, r ¼ 0.13), the number of NCPs (P ¼0.003, r ¼ 0.15), and the Agatston score (P < 0.001, r ¼ 0.27).

3.3. EFV and presence of severe coronary stenosis

Seventy patients (36%) had coronary stenosis �70%. In these

patients EFV was greater (EFV with �70% stenosis, 126 � 54

cm3 vs 100 � 54 cm3 without; P < 0.001) (Fig. 3A) as was

Agatston score [Agatston score with �70% stenosis, 28 Agat-

ston unit [AU] (0e3681 AU) vs 544 AU (6e4537 AU) without;

P< 0.001]. Onmultivariable logistic regression analysis for the

determinants of coronary stenosis �70%, age (odds ratio [OR],

1.0; 95% CI, 1.0e1.1; P ¼ 0.009), diabetes mellitus (OR, 2.1; 95%

CI, 1.3e3.4; P ¼ 0.004), smoking (OR, 3.1; 95% CI, 1.4e6.9; P ¼0.005), and EFV (OR, 3.0; 95% CI, 1.3e6.6; P ¼ 0.008) were

significant independent determinants (Table 2).

3.4. EFV and HRPFs

At least 1 HRPF was present in 113 (59%) of the 192 patients.

Table 3 outlines the demographics of the patients with and

without HRPFs. Category of coronary artery Agatston score

was higher in patients with HRPFs than in patients without

HRPFs (P< 0.001). At least 1 plaque exhibited LAP in 67 patients

(35%) and PR in 93 patients (48%). EFV was greater in patients

with at least 1 plaque exhibiting LAP (EFV with LAP, 120 � 48

cm3 vs 103 � 58 cm3 without; P ¼ 0.008) (Fig. 3B), PR (EFV with

PR, 115 � 54 cm3 vs 104 � 56 cm3 without; P ¼ 0.07) (Fig. 3C), in

patients who had either LAP or PR (EFV with HRPFs, 116 � 53

cm3 vs 99 � 57 cm3 without; P ¼ 0.009), and in patients with

both LAP and PR (EFV with LAP and PR, 119 � 47 cm3 vs 101 �51 cm3; P ¼ 0.006).

Table 2 e Multivariable logistic regression analysis forthe prediction of severe coronary stenosis (‡70%).

Odds ratio 95% CI P value

Age 1.0 1.0e1.1 0.009

BMI 1.0 0.9e1.0 0.3

Diabetes mellitus 2.1 1.3e3.4 0.004

Hypercholesterolemia 1.9 0.9e3.6 0.1

Smoking 3.1 1.4e6.9 0.005

Family history 1.0 0.6e1.9 0.9

Hypertension 1.3 0.7e2.4 0.4

Log EFV 3.0 1.3e6.6 0.008

BMI, body mass index (calculated as weight divided by height;

kg/m2); EFV, epicardial fat volume.

J o u r n a l o f C a r d i o v a s c u l a r C om p u t e d T omog r a p h y 7 ( 2 0 1 3 ) 1 2 5e1 3 2130

On multivariable logistic regression analysis for the

determinants of having at least 1 HRPF, diabetes mellitus,

hypercholesterolemia, and EFV were significantly associated

with having an HRPF (OR, 1.7; 95% CI, 0.9e3.4; P ¼ 0.038). For

LAP, diabetesmellitus and EFVwere independently associated

to LAP (OR, 2.4; 95% CI, 1.1e5.1; P ¼ 0.02). For PR, EFV was not

associated to the presence of PR. For both LAP and PR, diabetes

mellitus, smoking, hypertension, and EFV were significantly

associated with having both HRPFs (OR, 2.6; 95% CI, 1.1e6.2;

P ¼ 0.03) (Table 4).

4. Discussion

In the present study we show that EFV is greater in patients

with CP, PCP, and also NCP. Furthermore, we show that EFV is

associated with severe coronary stenosis (�70%) and also to

HRPFs in patients with NCP or PCP when clinical characteris-

tics and BMI are accounted for.

Table 3 e Comparison of risk factors in patients with andwithout HRPFs.

HRPFs(e)

HRPFs(þ)

Pvalue

Number 289 113

Age, y, median (range) 63 (23e92) 66 (41e87) 0.01

BMI, mean � SD 26.5 � 4.5 26.9 � 4.1 0.3

Hypercholesterolemia, n (%) 169 (59) 84 (74.3) 0.003

Smoking, n (%) 25 (8.7) 16 (14.2) 0.1

Diabetes mellitus, n (%) 31 (11.1) 23 (20.3) 0.01

Family history, n (%) 108 (37.4) 40 (35.4) 0.7

Hypertension, n (%) 148 (51.2) 67 (59.3) 0.1

EFV, cm3, mean � SD 98 � 50 116 � 53 <0.001

Log EFV, cm3, mean � SD 4.5 � 0.5 4.7 � 0.4 <0.001

Agatston score, AU,

median (range)

18 (0e5272) 268 (0e2600) <0.001

Agatston score category <0.001

0 AU, n (%) 110 (38) 4 (4)

1e99 AU, n (%) 84 (29) 29 (26)

100e399 AU, n (%) 53 (18) 35 (31)

>400 AU, n (%) 42 (15) 45 (30)

AU, Agatston unit; BMI, body mass index (calculated as weight

divided by height; kg/m2); EFV, epicardial fat volume.

4.1. Association of EFV to coronary disease

We found sizeable differences in EFV in patients who showed

any type of coronary plaque and also in patients who had

a significant coronary stenosis �70%. These findings are

consistent with prior studies that have suggested that EFV is

associated with significant coronary disease. Iwasaki et al14

measured EFV in 197 patients who underwent 64-slice coro-

nary CTA and observed on univariate analysis greater EFV in

patients with significant coronary stenosis (�50%) than

patients without significant coronary stenosis. Wang et al15

showed that epicardial fat thickness within the left atrioven-

tricular groove was independently related to the extent and

severity of significant coronary artery disease (� 50%), and

Ueno et al29 have shown that EFV is related independently to

chronic total occlusions. This finding supports the hypothesis

that EFV is associated to not only the presence of coronary

plaque disease but also its severity.

4.2. Association of EFV to HRPFs

Evidence is now emerging that EFV, as well as being associ-

ated with atherosclerosis and plaque composition,30 is asso-

ciated with adverse outcome. Studies have shown that EFV is

associated with acute coronary syndrome,4 myocardial

ischemia,3,13 and also cardiac events.5,14 Despite this, uncer-

tainty remains as to how exactly EFV exerts its detrimental

effect. A potential explanation is that EFV causes a local

inflammatory response on adjacent coronary arterial

segments that may in turn mediate atherosclerosis. In the

present study we therefore examined whether EFV was

related to the plaque characteristics of positive remodeling

and low attenuation plaque. These 2 features on coronary CTA

are considered to be HRPFs by virtue of their association with

acute and future coronary events.7e10 After adjusting for

conventional cardiovascular risk factors and BMI, we found

that EFV was related to having at least 1 HRPF, 2 HRPFs, and

also LAP. Our findings are consistent with prior published

work. Schlett et al17 evaluated the relationship of EFV to high-

risk coronary lesions in 358 patients presenting with acute

chest pain. The researchers defined a high-risk coronary

lesion as one having a �50% luminal narrowing and at least 2

high-risk plaque features (from PR, LAP, and spotty calcifica-

tion). In the 13 patients with a high-risk coronary lesion the

researchers showed an independent association to EFV after

adjusting for conventional cardiovascular risk factors and

BMI, but not for Agatston score. In another study, Oka et al16

additionally studied 357 patients undergoing coronary CTA

and examined the relationship of EFV to NCP characteristics.

The researchers showed that EFV was related to PR and LAP

after adjusting for conventional cardiovascular risk factors

and coronary artery Agatston score. In contrast to the present

study, the researchers did not examine PCPs as well as NCPs.

Recently, Ito et al18 have also shown that EFV measured on

cardiac CT is associated with thin-capped fibroatheroma

detected by invasive optical coherence tomography and also

to patients with acute coronary syndrome. The findings from

these studies suggest a possible relationship between EFV and

high-risk plaque structure, which in turn may explain the

observations that EFV is elevated in patients experiencing

Table 4 e Multivariable logistic regression analysis for the prediction of any HRPFs LAP, PR, and concomitant LAP and PR.

Any HRPFs LAP PR LAP þ PR

OR (95% CI) P OR (95% CI) P OR (95% CI) P OR (95% CI) P

Age 1.0 (1.0e1.0) 0.1 1.0 (1.0e1.1) 0.03 1.0 (1.0e1.0) 0.5 1.0 (1.0e1.0) 0.3

BMI 1.0 (0.9e1.1) 0.6 1.0 (0.9e1.0) 0.5 1.0 (0.9e1.1) 0.7 1.0 (0.9e1.1) 0.5

Diabetes mellitus 1.6 (0.9e2.5) 0.04 1.8 (1.1e2.8) 0.02 1.8 (1.1e2.7) 0.01 2.2 (1.3e3.7) 0.002

Hypercholesterolemia 1.5 (0.8e2.6) 0.047 1.5 (0.8e2.8) 0.2 1.8 (1.0e3.1) 0.04 1.7 (0.8e3.6) 0.2

Smoking 1.8 (0.8e4.2) 0.2 2.1 (1.0e4.8) 0.06 1.6 (0.8e3.4) 0.2 2.5 (1.0e6.1) 0.040

Family history 0.9 (0.6e1.6) 0.8 1.0 (0.6e1.8) 0.9 1.1 (0.7e1.8) 0.8 1.3 (0.7e2.6) 0.4

Hypertension 0.6 (0.4e1.1) 0.7 0.7 (0.4e1.3) 0.3 0.8 (0.4e1.3) 0.3 0.5 (0.2e0.9) 0.03

Log EFV 1.7 (0.9e3.4) 0.04 2.4 (1.1e5.1) 0.02 1.8 (1.0e3.4) 0.07 2.6 (1.1e6.2) 0.03

BMI, body mass index (calculated as weight divided by height; kg/m2); EFV, epicardial fat volume; HRPF, high-risk plaque feature; LAP, low

attenuation plaque; PR, positive remodeling.

J o u rn a l o f C a r d i o v a s c u l a r C om p u t e d T omog r a p h y 7 ( 2 0 1 3 ) 1 2 5e1 3 2 131

acute coronary syndrome,4 adverse cardiac events,5 and also

ischemia on single-photon emission CT.3

4.3. Limitations

In the present study we used a tube voltage of 100 or 120 kVp

to minimize the radiation dose. In previous studies that

evaluated HRPFs, higher tube voltages were used, including

135 kVp by Motoyama et al,7 and in the study by Kitigawa

et al,8 all patients were studied with a tube voltage of 120 kVp.

It is possible that the use of lower tube voltage may have

reduced our detection rate of certain HRPFs. Only patients

with good or excellent coronary CTA studies were included.

Limitations in the spatial resolution of coronary CTA may

reduce its capability in accurately describing LAP and PR.

However, 2 recent studies have confirmed that the assess-

ment of these adverse plaque characteristics by coronary CTA

correlates well with that measured by intravascular ultra-

sound (IVUS). Marwan et al31 compared 40 predominantly

lipid-rich plaques by using IVUS and dual-source CT and

showed that plaques with a cutoff of 5.5% pixels with an

attenuation of <30 HU had a sensitivity of 95%, a specificity of

80%, and a positive predictive value of 93% for the detection of

lipid-rich plaques by IVUS. Similarly, Gauss et al32 in

a comparative study between IVUS and coronary CTA showed

a sensitivity of 83% and specificity of 78% for the detection of

PR with the use of a remodeling index of 1.1. Because the

present study was retrospective in design, no measurements

of visceral fat or inflammatory markers were available.

Although the mean EFV values in the present study were

lower than in previous studies, this most likely reflects varied

methodologies of EFV quantification and differing patient

cohorts. In addition, the findings of the present study need to

be confirmed in similar cohorts of patients and in other

cohorts with various demographic and risk factor profiles. No

long-term follow-up data were available for major adverse

cardiac events in the present study, and further studies are

indicated to further evaluate the prognostic potential of EFV in

patients undergoing NCT.

5. Conclusions

EFV is elevated in patients with CP, PCP, and NCP and is

associated with severe coronary stenosis. In patients with

NCP and PCP, EFV is independently associated with HRPFs.

Our findings suggest that EFV may be a useful additional

measurement in risk stratifying patients who undergo routine

coronary artery calcium scoring and may indirectly provide

prognostic information above and beyond the assessment of

stenosis severity on coronary CTA.

Acknowledgments

Drs Rajani and Shmilovich contributed equally to this work.

This work was partly supported by grants from the Eisner,

Glazer, and Lincy Foundations (to Dr Berman), from the

National Institute of Biomedical Imaging and Bioengineering

(R21EB006829 to Dr Dey) and from the American Heart Asso-

ciation (12GRNT9450056 to Dr. Dey).

r e f e r e n c e s

1. Gorter PM, de Vos AM, van der Graaf Y, et al. Relation ofepicardial and pericoronary fat to coronary atherosclerosisand coronary artery calcium in patients undergoing coronaryangiography. Am J Cardiol. 2008;102:380e385.

2. Dey D, Wong ND, Tamarappoo B, et al. Computer-aided non-contrast CT-based quantification of pericardial and thoracicfat and their associations with coronary calcium andMetabolic Syndrome. Atherosclerosis. 2010;209:136e141.

3. Tamarappoo B, Dey D, Shmilovich H, et al. Increasedpericardial fat volume measured from noncontrast CTpredicts myocardial ischemia by SPECT. JACC CardiovascImaging. 2010;3:1104e1112.

4. Harada K, Amano T, Uetani T, et al. Cardiac 64-multislicecomputed tomography reveals increased epicardial fatvolume in patients with acute coronary syndrome. Am JCardiol. 2011;108:1119e1123.

5. Cheng VY, Dey D, Tamarappoo B, et al. Pericardial fat burdenon ECG-gated noncontrast CT in asymptomatic patients whosubsequently experience adverse cardiovascular events. JACCCardiovasc Imaging. 2010;3:352e360.

6. Mazurek T, Zhang L, Zalewski A, et al. Human epicardialadipose tissue is a source of inflammatory mediators.Circulation. 2003;108:2460e2466.

7. Motoyama S, Kondo T, Sarai M, et al. Multislice computedtomographic characteristics of coronary lesions in acutecoronary syndromes. J Am Coll Cardiol. 2007;50:319e326.

J o u r n a l o f C a r d i o v a s c u l a r C om p u t e d T omog r a p h y 7 ( 2 0 1 3 ) 1 2 5e1 3 2132

8. Kitagawa T, Yamamoto H, Horiguchi J, et al. Characterization ofnoncalcified coronary plaques and identification of culpritlesions in patients with acute coronary syndrome by 64-slicecomputed tomography. JACCCardiovasc Imaging. 2009;2:153e160.

9. Pflederer T, Marwan M, Schepis T, et al. Characterization ofculprit lesions in acute coronary syndromes using coronarydual-source CT angiography. Atherosclerosis.2010;211:437e444.

10. Motoyama S, Sarai M, Harigaya H, et al. Computedtomographic angiography characteristics of atheroscleroticplaques subsequently resulting in acute coronary syndrome.J Am Coll Cardiol. 2009;54:49e57.

11. Ding J, Kritchevsky SB, Harris TB, et al. The association ofpericardial fat with calcified coronary plaque. Obesity (SilverSpring). 2008;16:1914e1919.

12. Mahabadi AA, Reinsch N, Lehmann N, et al. Association ofpericoronary fat volume with atherosclerotic plaque burdenin the underlying coronary artery: a segment analysis.Atherosclerosis. 2010;211:195e199.

13. Nakazato R, Dey D, Cheng VY, et al. Epicardial fat volume andconcurrent presence of both myocardial ischemia andobstructive coronary artery disease. Atherosclerosis.2012;221:422e426.

14. Iwasaki K, Matsumoto T, Aono H, Furukawa H, Samukawa M.Relationship between epicardial fat measured by 64-multidetector computed tomography and coronary arterydisease. Clin Cardiol. 2011;34:166e171.

15. Wang TD, Lee WJ, Shih FY, et al. Association of epicardialadipose tissue with coronary atherosclerosis is region-specific and independent of conventional risk factors andintra-abdominal adiposity. Atherosclerosis. 2010;213:279e287.

16. Oka T, Yamamoto H, Ohashi N, et al. Association betweenepicardial adipose tissue volume and characteristics of non-calcified plaques assessed by coronary computedtomographic angiography. Int J Cardiol. 2012;161:45e49.

17. Schlett CL, Ferencik M, Kriegel MF, et al. Association ofpericardial fat and coronary high-risk lesions as determinedby cardiac CT. Atherosclerosis. 2012;222:129e134.

18. Ito T, Nasu K, Terashima M, et al. The impact of epicardial fatvolume on coronary plaque vulnerability: insight from opticalcoherence tomography analysis. Eur Heart J CardiovascImaging. 2012;13:408e415.

19. Agatston AS, Janowitz WR, Hildner FJ, Zusmer NR,Viamonte M Jr, Detrano R. Quantification of coronary arterycalcium using ultrafast computed tomography. J Am CollCardiol. 1990;15:827e832.

20. Dey D, Suzuki Y, Suzuki S, et al. Automated quantitation ofpericardiac fat from noncontrast CT. Invest Radiol.2008;43:145e153.

21. Yoshizumi T, Nakamura T, Yamane M, et al. Abdominal fat:standardized technique for measurement at CT. Radiology.1999;211:283e286.

22. Kvist H, Chowdhury B, Grangard U, Tylen U, Sjostrom L. Totaland visceral adipose-tissue volumes derived frommeasurements with computed tomography in adult men andwomen: predictive equations. Am J Clin Nutr.1988;48:1351e1361.

23. Sjostrom L, Kvist H, Cederblad A, Tylen U. Determination oftotal adipose tissue and body fat in women by computedtomography, 40K, and tritium. Am J Physiol.1986;250:E736eE745.

24. Nakazato R, Shmilovich H, Tamarappoo BK, et al. Interscanreproducibility of computer-aided epicardial and thoracic fatmeasurement from noncontrast cardiac CT. J CardiovascComput Tomogr. 2011;5:172e179.

25. Dey D, Lee CJ, Ohba M, et al. Image quality and artifacts incoronary CT angiography with dual-source CT: initialclinical experience. J Cardiovasc Comput Tomogr.2008;2:105e114.

26. Gutstein A, Dey D, Cheng V, et al. Algorithm for radiation dosereduction with helical dual source coronary computedtomography angiography in clinical practice. J CardiovascComput Tomogr. 2008;2:311e322.

27. Ferencik M, Ropers D, Abbara S, et al. Diagnostic accuracy ofimage postprocessing methods for the detection of coronaryartery stenoses by using multidetector CT. Radiology.2007;243:696e702.

28. Cheng V, Gutstein A, Wolak A, et al. Moving beyond binarygrading of coronary arterial stenoses on coronarycomputed tomographic angiography: insights for theimager and referring clinician. JACC Cardiovasc Imaging.2008;1:460e471.

29. Ueno K, Anzai T, Jinzaki M, et al. Increased epicardial fatvolume quantified by 64-multidetector computedtomography is associated with coronary atherosclerosis andtotally occlusive lesions. Circ J. 2009;73:1927e1933.

30. Alexopoulos N, McLean DS, Janik M, Arepalli CD, Stillman AE,Raggi P. Epicardial adipose tissue and coronary artery plaquecharacteristics. Atherosclerosis. 2010;210:150e154.

31. Marwan M, Taher MA, El Meniawy K, et al. In vivo CTdetection of lipid-rich coronary artery atheroscleroticplaques using quantitative histogram analysis: a head tohead comparison with IVUS. Atherosclerosis.2011;215:110e115.

32. Gauss S, Achenbach S, Pflederer T, Schuhback A, Daniel WG,Marwan M. Assessment of coronary artery remodelling bydual-source CT: a head-to-head comparison withintravascular ultrasound. Heart. 2011;97:991e997.