Association between single-slice and whole heart measurements of epicardial and pericardial fat in cardiac MRI

Abstract

Background

Epicardial (ECF) and pericardial fat (PCF) are important prognostic markers for various cardiac diseases. However, volumetry of the fat compartments is time-consuming.

Purpose

To investigate whether total volume of ECF and PCF can be estimated by axial single-slice measurements and in a four-chamber view.

Material and Methods

A total of 113 individuals (79 patients and 34 healthy) were included in this retrospective magnetic resonance imaging (MRI) study. The total volume of ECF and PCF was determined using a 3D-Dixon sequence. Additionally, the area of ECF and PCF was obtained in single axial layers at five anatomical landmarks (left coronary artery, right coronary artery, right pulmonary artery, mitral valve, coronary sinus) of the Dixon sequence and in a four-chamber view of a standard cine sequence. Pearson's correlation coefficient was calculated between the total volume and each single-slice measurement.

Results

Axial single-slice measurements of ECF and PCF correlated strongly with the total fat volumes at all landmarks (ECF: r = 0.85–0.94, P < 0.001; PCF: r = 0.89–0.94, P < 0.001). The best correlation was found at the level of the left coronary artery for ECF and PCF (r = 0.94, P < 0.001). Correlation between single-slice measurement in the four-chamber view and the total ECF and PCF volume was lower (r = 0.75 and r = 0.8, respectively, P < 0.001).

Conclusion

Single-slice measurements allow an estimation of ECF and PCF volume. This time-efficient analysis allows studies of larger patient cohorts and the opportunistic determination of ECF/PCF from routine examinations.

Keywords

Epicardial fat pericardial fat cardiac magnetic resonance imaging DIXON opportunistic imaging

Introduction

Epicardial fat (ECF) is defined as the fat inside the pericardium whereas pericardial fat (PCF) is defined as the complete fat in the mediastinum (1). The amount of ECF has prognostic relevance in various cardiac diseases. Several studies describe an association between ECF and coronary artery disease, with higher amounts of ECF associated with a more frequent occurrence of significant coronary lesions (2 –4). Furthermore, the amount of ECF is associated with the occurrence of fatal as well as non-fatal coronary events (5). Additionally, there is research linking ECF to outcomes after various cardiac interventions. Eberhard et al. (6) showed that the amount of ECF is related to all-cause mortality as well as a standardized early safety endpoint after transcatheter aortic valve implantation (TAVI) in patients with severe aortic valve stenosis. Lu et al. (7) found an association between ECF and the occurrence of major adverse cardiac events (MACE) after percutaneous coronary intervention (PCI). In addition, ECF appears to play a role in remodeling after surgical aortic valve replacement, with greater amounts of ECF associated with more pronounced and severe left ventricular remodeling in patients with severe aortic stenosis (8). Furthermore, the amount of ECF was shown to correlate with the recurrence of atrial fibrillation after ablation therapy (9).

While the clinical and biochemical effects of ECF are well studied, the effects of PCF are much less well known (10,11). Since fat inside and outside of the pericardium differs anatomically and biochemically, the distinction between those two fat depots is important, as discussed in detail in the work of Iacobellis (11).

In clinical practice, the thickness of ECF and PCF can be quickly determined by echocardiography. However, this presents a number of challenges (12). More reliable determination of ECF can be achieved by volumetry in cross-sectional imaging techniques such as cardiac magnetic resonance imaging (MRI) or cardiac computed tomography (CT) (13,14).

Cross-sectional imaging of the heart is performed in patients with heart disease on the basis of a wide variety of clinical indications and questions. Since it has already been demonstrated that the volume of ECF and PCF can be reliably measured in CT and MRI, it is, in principle, possible to gather this data “opportunistically” in clinical routine studies. However, the volumetry of ECF and PCF is very time-consuming (15). In other parts of the body, several studies have already shown that volumes of body compartments, for example the amount of visceral fat, can be reliably estimated from single layer measurements, which drastically reduces the time needed for analysis (16,17).

Hence, the aim of the present study was to investigate whether axial single-layer measurements at different anatomical landmarks allow an estimation of total ECF and PCF in cardiac MRI using the Dixon method.

The Dixon method offers a robust technique to separate fat from water and is relatively insensitive to inhomogeneities of the static magnetic field (18,19). Since a 3D-Dixon sequence is not part of every standard cardiac MRI examination, we additionally studied the relationship between fat area measurements in a four-chamber view of a standard cine sequence and the total fat volume.

Material and Methods

Study population

This retrospective study was approved by the institutional review board with a waiver of written informed consent. The study population (n = 113) consisted of volunteers, composed of healthy control groups from previous studies, as well as consecutive patients who received a cardiac MRI to rule out various cardiac diseases such as cardiomyopathy or ischemia and where a 3D-Dixon sequence was acquired. Medical records were reviewed and baseline characteristics like patient and volunteer age at the time of examination as well as body weight and body height were obtained where available.

Imaging protocol

All patients/volunteers were examined on a 1.5-T whole-body MRI-system (Ingenia; Philips Healthcare, Best, The Netherlands) utilizing a 32-channel torso coil for signal reception. All patients received standard functional cine imaging with a balanced steady-state free precession (bSSFP) sequence (field of view [FOV] = 350 × 350 mm², slice thickness = 8 mm, repetition time [TR] = 3.1 ms, echo time [TE] = 1.6 ms, flip angle = 60°) and an axial ECG-triggered and respiratory navigator gated magnetization prepared mDixon-sequence (FOV = 350 × 302 × 180 mm³; TR = 5.4 ms, TE1/TE2 = 1.8/4.0 ms, flip angle = 20°) (14).

Epi- and pericardial fat measurements

Image datasets were retrieved from the institutional picture archiving and communication system and imported into an image analysis tool on a standard workstation. Areas of ECF, defined as the adipose tissue within the visceral layer of the pericardium, as well as PCF, defined as ECF plus the pericardial fat outside of the pericardium, were measured with an in-house software for volumetry and planimetry written in MATLAB (MathWorks, Natick, MA, USA) (16).

The two fat compartments were delineated manually in the 3D-Dixon dataset from the bifurcation of the pulmonal artery down to the most inferior slice of the myocardium by a board-certified cardiologist with >5 years of experience in cardiac MRI. The different mDIXON images (fat-only, water-only, in-phase, opposed-phase) could be switched during the delineation process. A fat fraction map was calculated and an appropriate threshold for voxels containing predominantly fat was obtained based on the signal intensity of fat only voxels compared to the signal intensity of the myocardium. Details have been described before (14). Finally, fat volumes were computed based on the number of fat voxels and the voxel size.

We chose the right pulmonary artery (RPA), the origin of the left main coronary artery (LMCA), the origin of the right coronary artery (RCA), the coronary sinus (CS) as well as the middle of the mitral valve (MV) as anatomical landmarks for measuring the ECF and PCF areas to investigate which single-slice area measurement shows the highest correlation with the total ECF and PCF volume. The locations of the anatomical landmarks were determined manually while the heart was displayed in three orthogonal planes. Part of the landmarks have been examined before in cardiac CT (15).

The relationship between the ECF and PCF volume and a single slice measurement was additionally investigated in a mid-ventricular slice of a standard four-chamber view of a balanced steady-state free precision (bSSFP) cine sequence that is part of every cardiac MRI examination. To create a realistic clinical scenario, the ECF and PCF areas were delineated manually in end-diastolic images on a standard clinical workstation (IMPAX; Agfa Healthcare, Mortsel, Belgium). Small non-fatty components inside the fat area like coronary arteries were not excluded from the analysis to facilitate an easy and fast analysis.

Intra- and inter-reader agreement

A second reader with two years of experience in cardiac MRI additionally performed ECF/PCF measurements in a subset of 25 randomly selected individuals twice at the LMCA level and in the four-chamber view, with at least two weeks between repeated measurements. Both intraclass correlation coefficient (ICC) and Bland–Altman analysis were calculated.

Statistical analysis

Statistical analysis was performed with the Software Package R and a specific plotting library (ggplot2 package; Hadley Wickham, New York, NY, USA and R Foundation for Statistical Computing, Vienna, Austria). The data were tested for normal distribution by the Shapiro–Wilk test. Continuous variables are presented as mean ± SD. Categorical data are given as frequencies. Pearson's correlation coefficient was obtained for the fat area at every landmark and the respective total fat volume. Continuous data with a non-parametric distribution were compared using the Mann–Whitney U test. Standard linear regression was applied between the single-slice measurements at the LMCA level and the total ECF/PCF volume using the “stats” package in R. Correlation coefficients were compared using Williams t-Test and Fisher's z-Transformation using the “cocor” package in R (20). A P value <0.05 was considered statistically significant.

Results

In total, cardiac MRI scans of 79 patients and 34 healthy volunteers were analyzed. Of the 113 individuals, 44 were female (39%) with an age range of 21–82 years (mean age = 55.4 ± 15.0 years, mean age of men = 55.4 ± 17.7 years, mean age of women = 52.7 ± 14.0 years; P = 0.61). In 47 individuals of this cohort, the body mass index (BMI) could be obtained retrospectively, and the BMI was in the range of 14–42 kg/m² (mean BMI = 26.5 ± 5.2 kg/m², mean BMI of men = 27 ± 5.0 kg/m², mean BMI of women = 25.4 ± 5.8 kg/m²; P = 0.36). Of the individuals with available BMI, 62% were overweight (BMI >25 kg/m²; male: 66%, female: 53%). All anatomical landmarks (RPA, LMCA, RCA, CS, MV) were easily identified in all cases. Examples of axial MRI images at the different anatomical landmarks with segmented ECF and PCF can be found in Fig. 1.

Fig. 1.

Overview of axial slices at the different landmarks and a coronal reconstruction of the Dixon fat-only image with the fat segmentation as an overlay (red, epicardial fat; blue, pericardial fat) and a Dixon water-only image for anatomical orientation. CS, coronary sinus; LMCA, left main coronary artery; MV, mitral valve; RCA, right coronary artery; RPA, right pulmonary artery.

Correlation between epicardial and pericardial fat area and total fat volumes

The ECF and PCF areas measured at each anatomical landmark in the axial Dixon dataset correlated strongly with the total ECF volume (r = 0.85–0.94; P < 0.001) as well as the total PCF volume (r = 0.89–0.94; P < 0.001), respectively. The highest correlation coefficients of the association between ECF and PCF area and total ECF and PCF volume were observed for the LMCA level (r = 0.94 in both cases; P < 0.001). There was a statistically significant difference between the correlation at the LMCA level and every other level except the RCA level for ECF. For PCF, there was no statistically significant difference between LMCA and the other landmarks, except for LMCA versus RPA (see Table 1). A scatterplot with a linear fit at LMCA grouped by sex is presented in Fig. 2. Determined parameters of the ungrouped linear model for the association of the fat area at the LMCA level and the total fat volume are as follows:

Total epicardial fat volume [{mm}^{3}] = 32801.5 {mm}^{3} + 56.2 mm * LMCA epicardial fat area [{mm}^{2}]

Total pericardial fat volume [{mm}^{3}] = 66029.9 {mm}^{3} + 94.7 mm * LMCA pericardial fat area [{mm}^{2}]

Fig. 2.

Correlation between (a) epicardial and (b) pericardial fat volume and the respective fat area at the origin of the LMCA grouped by sex (Pearson's correlation coefficients for epicardial fat: r_female = 0.91, r_male = 0.95; Pearson's correlation coefficients for pericardial fat: r_female = 0.91, r_male = 0.93). LMCA, left main coronary artery.

Table 1.

Epicardial and pericardial fat area measurements at five different anatomical landmarks and their correlation with total epicardial and pericardial fat volume (n = 113).*

Anatomical landmark	r for epicardial fat^†	P value vs. LMCA	r for pericardial fat^†	P value vs. LMCA
RPA	0.86	<0.001	0.89	<0.001
LMCA	0.94	-	0.94	-
RCA	0.93	0.31	0.94	0.98
CS	0.85	<0.001	0.92	0.06
MV	0.89	<0.001	0.91	0.19

*The highest correlation coefficient was found at the LMCA level for epicardial fat as well as pericardial fat (r = 0.94). For epicardial fat, there was a statistically significant difference between the correlation coefficient of every landmark and the LMCA level except for the RCA level. For pericardial fat, there was no statistically significant difference between the correlation coefficients for the LMCA level and all other landmarks, except for the RPA level.

^†

r, Pearson's correlation coefficient; highest correlation coefficients in bold.

CS, coronary sinus; LMCA, left main coronary artery; MV, mitral valve; RCA, right coronary artery; RPA, right pulmonary artery.

Differences in epicardial and pericardial fat area between sexes

We observed a significant difference between the ECF and PCF areas measured at LMCA between male and female individuals (ECF: mean = 1933 ± 1267 mm² and 1111 ± 726 mm², respectively; PCF: mean = 2856 ± 1905 mm² and 1502 ± 913 mm², respectively; P < 0.001 for both (Fig. 3). While the LMCA level showed the best correlation in male as well as in female individuals for ECF (r = 0.91 for both), in men the best correlation for PCF was found at the RCA level (r = 0.94), while in women the highest correlation was again found at the LMCA level (r = 0.95) (Table 2). There was no statistically significant difference between men and women at this landmark (P = 0.18).

Fig. 3.

Violin plots of the (a) epicardial and (b) pericardial fat area at the origin of the LMCA grouped by sex. A significant difference between the amount of epicardial and pericardial fat between men and women was observed (P < 0.001 for both, Mann–Whitney U). LMCA, left main coronary artery.

Table 2.

Pearson's correlation coefficients for epicardial and pericardial fat area measurements at different anatomical landmarks with the total epicardial and pericardial fat volume grouped by sex.*

	Epicardial fat		Pericardial fat
Anatomical landmark	r for male patients	r for female patients	r for male patients	r for female patients
RPA	0.68	0.77	0.89	0.87
LMCA	0.91	0.91	0.93	0.95
RCA	0.88	0.90	0.94	0.94
MV	0.88	0.89	0.90	0.83
CS	0.83	0.86	0.93	0.89

*r, Pearson's correlation coefficient; highest correlation coefficients in bold.

CS, coronary sinus; LMCA, left main coronary artery; MV, mitral valve; RCA, right coronary artery; RPA, right pulmonary artery.

BMI analysis

Only a moderate correlation was present between the PCF and ECF areas at the different landmarks and the BMI (r = 0.52–0.61 and r = 0.49–0.66, respectively) (Fig. 4). There was a significant difference in both ECF and PCF areas at the LMCA level between overweight (BMI >25 kg/m²) and non-overweight individuals. The mean LMCA ECF area for non-overweight individuals was 961 ± 780 mm² and 2092 ± 1404 mm² for overweight (P = 0.003, Mann–Whitney U test). The mean LMCA PCF area was 1265 ± 975 mm² and 3095 ± 2170 mm², respectively (P = 0.002, Mann–Whitney U test). For overweight and non-overweight individuals, the strongest associations between single-slice and fat volume measurements were found at different anatomical landmarks (Table 3). For PCF, the LMCA level showed the best correlation for non-overweight individuals (r = 0.96), while the RCA and CS levels showed the highest correlation in overweight (r = 0.94). For ECF, the RCA and CS levels correlated best with the total fat volume in non-overweight individuals (r = 0.95), while in overweight individuals the highest correlation was found for the CS level (r = 0.95).

Fig. 4.

Correlation between the BMI and the (a) epicardial and (b) pericardial fat area at the origin of the LMCA grouped by sex (Pearson's correlation coefficients for epicardial fat: r_female = 0.59, r_male = 0.55; Pearson's correlation coefficients for pericardial fat: r_female = 0.51, r_male = 0.51). BMI, body mass index; LMCA, left main coronary artery.

Table 3.

Pearson's correlation coefficients in patients with available BMI (n = 47) for epicardial and pericardial fat area measurements at different anatomical landmarks and the total epicardial and pericardial fat volume grouped by overweight (BMI > 25 kg/m²).*

	Epicardial fat		Pericardial fat
Anatomical landmark	r for non-overweight patients	r for overweight patients	r for non-overweight patients	r for overweight patients
RPA	0.88	0.78	0.95	0.85
LMCA	0.91	0.94	0.96	0.92
RCA	0.95	0.94	0.94	0.94
MV	0.87	0.83	0.90	0.92
CS	0.95	0.95	0.94	0.94

*r, Pearson's correlation coefficient; highest correlation coefficients in bold.

CS, coronary sinus; LMCA, left main coronary artery; MV, mitral valve; RCA, right coronary artery; RPA, right pulmonary artery.

Four-chamber view

The ECF and PCF areas measured in the four-chamber view of standard cine image data showed a strong but lower correlation with the total ECF and PCF volumes, compared to the axial slices. The correlation coefficient was r = 0.75 for ECF and r = 0.80 for PCF (Fig. 5). An example of the fat segmentation in the four-chamber view can be found in Fig. 6.

Fig. 5.

Correlation between the single slice measurements in a four-chamber view for (a) epicardial and (b) pericardial fat and the total epicardial and pericardial fat volume (Pearson's correlation coefficient for epicardial fat: r = 0.75; for pericardial fat: r = 0.8; P < 0.001 for both).

Fig. 6.

Example of (a) pericardial and (b) epicardial fat segmentation in a single slice of a four-chamber view of a standard balanced steady-state free precision (bSSFP) cine sequence. The segmentation was performed on a standard workstation (IMPAX; Agfa Healthcare, Mortsel, Belgium).

Intra- and inter-reader agreement

ECF measurements at the LMCA level showed a high intra- as well as inter-reader agreement with ICCs of 0.998 and 0.969, respectively, and mean differences of −83 mm² (95% confidence interval [CI] −192 to 25 mm²) and −5.3 mm² (95% CI −164 to 154 mm²), respectively. Both intra- as well as inter-reader agreement for PCF were also good with ICCs of 0.997 and 0.993 and mean differences of −43 mm² (95% CI = −126 to 40 mm²) and −115 mm² (95% CI = −232 to 2 mm²), respectively.

ICCs for ECF measurements in the four-chamber view were 0.815 (intra-reader) and 0.823 (inter-reader). The mean differences were 7.6 mm² (95% CI = −97 to 113 mm²) and −77 mm² (95% CI = −187 to 35 mm²), respectively. For PCF, ICCs were 0.904 and 0.846 with mean differences of 22 mm² (95% CI = −202 to 247 mm²) and 48 mm² (95% CI = −220 to 317 mm²).

Discussion

The aim of the present study was to investigate whether the ECF and PCF areas measured in axial cardiac MRI slices at easy-to-recognize anatomical landmarks as well as in a single slice of a four-chamber view of a standard cine sequence allow the estimation of the total ECF and PCF volumes. We found a strong correlation between the total fat volumes and the single-slice measurements at every landmark measured. Overall, the highest linear correlation was found at the LMCA level for ECF as well as PCF. The correlation between the fat areas in a single slice of a four-chamber view and the total fat volumes was lower, but still strong.

While the LMCA level shows the best correlation for the patient/volunteer collective as a whole, male and female as well as overweight and non-overweight individuals show different anatomical landmarks for the highest correlation between axial fat area and total fat volume. However, since the correlation coefficients at the LMCA level show a strong correlation between the single-slice measurements and the total fat volumes in all subgroups, we recommend performing single-slice measurements for estimation of ECF/PCF volumes at the LMCA level, independent of specific patient characteristics.

The correlation of ECF volume with ECF area measurements at different anatomical landmarks have already been studied before by Oyama et al. (15). In this CT-based study, which included only non-obese Japanese patients, the strongest association between axial single-slice and volume measurements was also found at the LMCA level (r = 0.92), which is in line with our findings. Our MRI study enrolled a larger number of patients to investigate the relationship between volume and area measurements in a more general population consisting of patients and healthy participants with a wide range of BMIs and different ethnic backgrounds. The large standard deviation observed for ECF and PCF reflects the diverse subject collective. Another study by Tran et al. (4) reported a similar correlation coefficient of r = 0.89 between the ECF area at the LMCA level and the total ECF volume in cardiac CT.

To our knowledge, this is the first work studying the possibility of estimating both total ECF and PCF volume from area measurements in a single axial slice in cardiac MRI. A previous study has already proven the feasibility of measuring ECF/PCF volume in cardiac MRI using the Dixon method (14). Although all examinations were performed on a field strength of 1.5 T, the findings should extend to 3 T as well.

Although ECF and PCF measurements in a four-chamber view have been applied before by other researchers (21), its association with total fat volumes has not been investigated yet. Therefore, our study is the first to show that fat area measurements in a four-chamber view may also be used as a surrogate parameter for the total fat volume, although the correlations with total volumes are lower compared to measurements performed on axial slices.

Our results also indicate that the BMI correlates only moderately with the total ECF and PCF volume. This observation is in line with previous studies (9,22,23) and underlines that estimation of ECF and PCF volumes by mere measurement of the BMI is not reliable. Of note, due to its local expression of inflammatory mediators, ECF plays a special role in cardiac diseases and is not to be equated with visceral fat in general (24).

For cases where an actual value of the ECF/PCF volume is required, we provide a formula for calculating the ECF/PCF volume from the ECF/PCF area at the LMCA level based on a simple linear model. Similar to other studies (6,25,26), we found a statistically significant difference of ECF/PCF between sexes with a higher amount of ECF/PCF in men.

PCF and ECF are two fundamentally different fat depots. ECF has a similar origin as mesenterial and omental fat (27) and shares the blood supply with the myocardium. While the biochemical and clinical effects of ECF are well studied (10,28), the effects of PCF are less known yet, but possibly equally important (11,29). For example, a recent study links a CT-derived biomarker called “pericardial fat enhancement” to coronary artery disease (29). Therefore, the distinction between the different fat depots around the heart is necessary. Previous studies only investigated the relationship between single-slice measurements of ECF and the complete ECF volume but did not investigate whether the same relationship holds true for PCF as well. Our study is the first that shows a strong correlation between the PCF area in a single-slice and total PCF volume.

Commandeur et al. (30) recently proposed a fully automated approach for assessing ECF in cardiac CT with a 3D deep learning model. However, the trained model is not freely available and a fully automated end-to-end approach for measuring ECF/PCF volumetrically in a 3D dataset still presents a significant challenge. In contrast, measuring the fat area in a single slice allows the estimation of fat burden “opportunistically” based on routine examinations, analogous to the opportunistic imaging of sarcopenia explained by Lenchik et al. (31,32).

The present study has some limitations. Due to its retrospective nature, not all patient variables could be controlled. However, the participants included in this retrospective study represent a very heterogeneous cohort strengthening the generalizability of the observations. Additionally, in our collective of 113 participants, BMI values were available retrospectively in a subgroup of 47 individuals only, which, however, may only limit the generalizability of the results of the BMI subgroup analyses.

In conclusion, the total ECF/PCF burden of a patient can be estimated by single-slice measurements. Such measurements should ideally be performed in axial Dixon images at the level of the left main coronary artery. For clinical routine, estimation of total fat volume based on a standard four-chamber cine image is also feasible. Single-slice analysis facilitates further investigation of ECF and PCF in large patient cohorts.

Footnotes

Data availability

The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Alois M Sprinkart

References

Bertaso

Bertol

Duncan

, et al. Epicardial fat: definition, measurements and systematic review of main outcomes. Arq Bras Cardiol 2013;101:18–28.

Schlett

Ferencik

Kriegel

, et al. Association of pericardial fat and coronary high-risk lesions as determined by cardiac CT. Atherosclerosis 2012;222:129–134.

Bachar

Dicker

Kornowski

, et al. Epicardial adipose tissue as a predictor of coronary artery disease in asymptomatic subjects. Am J Cardiol 2012;110:534–538.

Tran

Small

Cocker

, et al. A single slice measure of epicardial adipose tissue can serve as an indirect measure of total epicardial adipose tissue burden and is associated with obstructive coronary artery disease. Eur Heart J Cardiovasc Imaging 2014;15:423–430.

Mahabadi

Berg

Lehmann

, et al. Association of epicardial fat with cardiovascular risk factors and incident myocardial infarction in the general population: the Heinz Nixdorf recall study. J Am Coll Cardiol 2013;61:1388–1395.

Eberhard

Stocker

Meyer

, et al. Epicardial adipose tissue volume is associated with adverse outcomes after transcatheter aortic valve replacement. Int J Cardiol 2019;286:29–35.

Jia

Wang

. Association between epicardial adipose tissue and adverse outcomes in coronary heart disease patients with percutaneous coronary intervention. Biosci Rep 2019;39:BSR20182278 doi.org/10.1042/BSR20182278.

Coisne

Ninni

Ortmans

, et al. Epicardial fat amount is associated with the magnitude of left ventricular remodeling in aortic stenosis. Int J Cardiovasc Imaging 2019;35:267–273.

Tabl

Torky

Farid

. The relation between total epicardial fat volume assessed by cardiac CT and the presence of atrial fibrillation. Egypt Heart J 2016;68:97–102.

10.

Sacks

Fain

. Human epicardial adipose tissue: a review. Am Heart J 2007;153:907–917.

11.

Iacobellis

. Epicardial and pericardial fat: close, but very different. Obesity 2009;17:625–625.

12.

Gulgun

Genç

. Measurement of epicardial fat thickness by echocardiography presents challenges. Arq Bras Cardiol 2016;107:497–498.

13.

Marwan

Achenbach

. Quantification of epicardial fat by computed tomography: why, when and how? J Cardiovasc Comput Tomogr 2013;7:3–10.

14.

Homsi

Meier-Schroers

Gieseke

, et al. 3D-Dixon MRI based volumetry of peri- and epicardial fat. Int J Cardiovasc Imaging 2016;32:291–299.

15.

Oyama

Goto

Ito

, et al.

Single-slice epicardial fat area measurement: do we need to measure the total epicardial fat volume?

Jpn J Radiol 2011;29:104–109.

16.

Faron

Luetkens

Schmeel

, et al. Quantification of fat and skeletal muscle tissue at abdominal computed tomography: associations between single-slice measurements and total compartment volumes. Abdom Radiol 2019;44:1907–1916.

17.

Faron

Pieper

Schmeel

, et al. Fat-free muscle area measured by magnetic resonance imaging predicts overall survival of patients undergoing radioembolization of colorectal cancer liver metastases. Eur Radiol 2019;29:4709–4717.

18.

Börnert

Koken

Nehrke

, et al. Water/fat-resolved whole-heart Dixon coronary MRA: an initial comparison. Magn Reson Med 2014;71:156–163.

19.

Dixon

. Simple proton spectroscopic imaging. Radiology 1984;153:189–194.

20.

Diedenhofen

Musch

. Cocor: a comprehensive solution for the statistical comparison of correlations. PloS One 2015;10:e0121945.

21.

Rado

Lorbeer

Gatidis

, et al. MRI-based assessment and characterization of epicardial and paracardial fat depots in the context of impaired glucose metabolism and subclinical left-ventricular alterations. Br J Radiol 2019;92:20180562.

22.

Dey

Wong

Tamarappoo

, et al. Computer-aided non-contrast CT-based quantification of pericardial and thoracic fat and their associations with coronary calcium and metabolic syndrome. Atherosclerosis 2010;209:136–141.

23.

Aitken-Buck

Moharram

Babakr

, et al. Relationship between epicardial adipose tissue thickness and epicardial adipocyte size with increasing body mass index. Adipocyte 2019;8:412–420.

24.

Mazurek

Zhang

Zalewski

, et al. Human epicardial adipose tissue is a source of inflammatory mediators. Circulation 2003;108:2460–2466.

25.

Dagvasumberel

Shimabukuro

Nishiuchi

, et al. Gender disparities in the association between epicardial adipose tissue volume and coronary atherosclerosis: a 3-dimensional cardiac computed tomography imaging study in Japanese subjects. Cardiovasc Diabetol 2012;11:106.

26.

Mancio

Pinheiro

Ferreira

, et al. Gender differences in the association of epicardial adipose tissue and coronary artery calcification: ePICHEART study. Int J Cardiol 2017;249:419–425.

27.

Marchington

Mattacks

Pond

. Adipose tissue in the mammalian heart and pericardium: structure, foetal development and biochemical properties. Comp Biochem Physiol Part B Comp Biochem 1989;94:225–232.

28.

Iacobellis

Corradi

Sharma

. Epicardial adipose tissue: anatomic, biomolecular and clinical relationships with the heart. Nat Clin Pract Cardiovasc Med 2005;2:536–543.

29.

Kallianos

Lorca

Moraes

, et al. Pericardial fat enhancement: new computed tomography biomarker influences the relationship between pericardial fat and coronary artery disease. J Thorac Imaging 2020;35:270–275.

30.

Commandeur

Goeller

Razipour

, et al. Fully automated CT quantification of epicardial adipose tissue by deep learning: a multicenter study. Radiol Artif Intell 2019;1:e190045.

31.

Lenchik

Boutin

. Sarcopenia: beyond muscle atrophy and into the new frontiers of opportunistic imaging, precision medicine, and machine learning. Semin Musculoskelet Radiol 2018;22:307–322.

32.

Luetkens

Faron

Geissler

, et al. Opportunistic computed tomography imaging for the assessment of fatty muscle fraction predicts outcome in patients undergoing transcatheter aortic valve replacement. Circulation 2020;141:234–236.