Pericoronary fat attenuation index predicts vulnerable plaque and adverse outcomes in coronary heart disease

Abstract

Background

Noncalcified coronary plaque is prone to rupture and may lead to major adverse cardiovascular events (MACEs). The pericoronary fat attenuation index (FAI), derived from coronary computed tomographic angiography (CCTA), is an emerging marker of coronary inflammation. This study aimed to assess the predictive value of the FAI for vulnerable plaque and prognosis in patients with coronary heart disease (CHD).

Methods

We retrospectively enrolled 453 patients with CHD who underwent CCTA and were followed for 3 years. Patients were divided into a MACE group (n=103) and a control group (n=350) based on the occurrence of MACEs. The FAI was measured using artificial intelligence software at the site of the most severe coronary stenosis. Clinical characteristics, coronary plaque burden, and FAI were compared between groups. Multivariate logistic regression identified independent predictors of MACEs, and a nomogram prediction model was developed and validated.

Results

Patients with MACEs had significantly higher age, greater total and noncalcified plaque burden, lower left ventricular ejection fraction (LVEF), higher FAI, and a greater prevalence of the multivessel disease. Independent predictors of MACEs included age ≥80 years [relative risk (RR) 12.39, 95% confident interval (CI): 5.75–26.69], LVEF <50% (RR 8.73, 95% CI: 4.10–18.58), total coronary plaque burden >33.3% (RR 4.27, 95% CI: 2.23–8.18), increased FAI (RR 1.08, 95% CI: 1.05–1.11), and multivessel disease (RR 3.14, 95% CI: 1.67–5.90) with all P<0.001. The nomogram model demonstrated strong predictive performance, with area under the curve (AUC) values of 0.920 and 0.862 in the training and validation sets, respectively. FAI was significantly correlated with noncalcified plaque burden (r=0.234, P<0.001).

Conclusions

FAI is associated with coronary noncalcified plaque burden and is an independent predictor of MACEs in patients with CHD. A prediction model incorporating the FAI demonstrated promising efficacy in identifying high-risk patients, supporting its potential role in personalized risk stratification.

Highlight box.

Key findings

• Higher fat attenuation index (FAI) was independently associated with coronary noncalcified plaque burden and adverse cardiovascular outcomes in patients with coronary heart disease (CHD).

• A nomogram prediction model incorporating the FAI demonstrated promising efficacy in identifying patients at increased risk of major adverse cardiovascular events (MACEs).

What is known and what is new?

• Noncalcified coronary plaque is more vulnerable to rupture and is a key contributor to MACEs. Coronary inflammation plays a central role in plaque instability.

• The FAI, a coronary computed tomographic angiography (CCTA)-derived marker of pericoronary inflammation, is significantly associated with noncalcified plaque burden and independently predicts MACEs. A novel nomogram model using the FAI improves individualized risk stratification.

What is the implication, and what should change now?

• FAI may be a valuable imaging biomarker for identifying patients at high risk of MACEs.

• Intervention measures should be strengthened for patients at high risk of MACEs.

Introduction

In recent years, the incidence of hypertension, diabetes, and hyperlipidemia has risen steadily due to changes in diet and lifestyle. Consequently, the prevalence of coronary heart disease (CHD) has also increased, making it one of the leading causes of mortality among middle-aged and older adults worldwide (1,2). Coronary artery disease (CAD) plays a central role in the development and progression of CHD. Coronary plaque can be broadly classified into calcified and noncalcified plaque. While calcified plaques contribute to luminal narrowing, myocardial ischemia, hypoxic remodeling, and eventually heart failure (3,4), noncalcified plaques are particularly prone to rupture, potentially leading to acute myocardial infarction (5,6). An increased coronary plaque burden has been identified as a major risk factor for major adverse cardiovascular events (MACEs) (7).

Identifying factors associated with coronary plaque burden is crucial for improving risk stratification and guiding prevention strategies in patients with CHD. In this context, recent studies have explored the role of the pericoronary fat attenuation index (FAI), an emerging imaging biomarker derived from coronary computed tomography angiography (CCTA), in cardiovascular diseases. FAI has been linked to abnormalities in glucose metabolism, vascular inflammation, and obesity (8-10), all of which are risk factors for plaque progression and vulnerability.

Based on this evidence, we hypothesized that the attenuation index of FAI is associated with coronary plaque burden in patients with CHD and may serve as a predictor of MACEs. However, studies specifically examining this association remain limited. Therefore, the present study aimed to investigate the predictive value of the FAI for identifying vulnerable coronary plaques and forecasting prognosis in patients with CHD. We present this article in accordance with the TRIPOD reporting checklist (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-775/rc).

Methods

Study population

This was a retrospective cohort study that included 453 patients with CHD [acute coronary syndrome (ACS)] who were admitted to the Wuhan Fourth Hospital between January 2020 and December 2021. Upon admission, all patients underwent CCTA to assess coronary artery anatomy and the FAI. Patients were followed for 3 years (the main risk of recurrence after PCI is within 3 years) and categorized, based on the occurrence of MACEs, into a MACE group and a control group. The inclusion criteria were: (I) confirmed diagnosis of CAD; (II) age ≥18 years; and (III) availability of complete clinical and CCTA data. The exclusion criteria were: (I) presence of malignancy; (II) concurrent structural heart disease (e.g., congenital heart disease, severe valvular heart disease, nonischemic cardiomyopathy, infiltrative heart disease); (III) severe dysfunction of vital organs (e.g., liver, kidney); and (IV) loss to follow-up. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. This study was approved by the Ethics Committee of Wuhan Fourth Hospital (No. 202501242). Due to the retrospective design, the requirement for individual informed consent was waived.

Data collection

The following clinical and imaging data were collected: (I) demographic and clinical variables including age, gender, history of hypertension, diabetes mellitus, and hyperlipidemia; (II) inflammation markers including high-sensitivity C-reactive protein (hs-CRP); (III) cardiac biomarkers including N-terminal pro-B-type natriuretic peptide (NT-proBNP), creatine kinase and creatine kinase isoenzymes; (IV) CCTA findings including left ventricular ejection fraction (LVEF), number of coronary arteries involved, stent type, degree of coronary artery stenosis, total coronary plaque burden, calcified plaque burden, noncalcified plaque burden (The components of plaques included lipid plaques, fibrous plaques and calcified plaques. The volume of noncalcified plaque was equal to the volume of lipid plaques plus the volume of fibrous plaques. Calcified plaque burden = calcified plaques/vascular volume; noncalcified plaque burden = noncalcified plaque/vascular volume), and FAI; and (V) medications including antihypertensive, lipid-lowering, and hypoglycemic agents.

Clinical outcome, MACEs, were defined as the occurrence of any of the following events during the 3-year follow-up: myocardial infarction, unstable angina, heart failure, cardiac arrest, or life-threatening arrhythmia including sustained monomorphic ventricular tachycardia, ventricular fibrillation, and appropriate device therapy.

Treatment

All patients underwent percutaneous coronary intervention (PCI) with coronary stent implantation. Post-procedural management included 12 months of dual antiplatelet therapy, followed by maintenance of either aspirin or P2Y12 receptor inhibitor therapy. Guideline-directed medical therapy was prescribed as clinically indicated.

Imaging protocol and FAI measurement

CCTA was performed using a dual-source computed tomography (CT) scanner (Siemens Healthineers, Erlangen, Germany). Scanning parameters: 120 kV, 300 mAs; reconstruction layer thickness: 0.625 mm; layer spacing of 0.5 mm. A nonionic iodinated contrast agent, Iohexol 350 mg/mL (GE HealthCare, Chicago, IL, USA), was administered intravenously. After image acquisition, diastolic and systolic phases were reconstructed. Perivascular fat analysis software (Beijing Shukun Technology Co., Ltd., Beijing, China) was used to quantify the average FAI at the site of left anterior descending branch, left circumflex branch and right coronary artery (every scan was performed with 120 kV). Fat attenuation was measured in Hounsfield units (HU), using a standard density range of –190 to –30 HU and within a radial distance from the vessel wall equal to the vessel diameter to define FAI (Figure 1) (11).

Figure 1.

Schematic diagram for attenuation index of pericoronary adipose tissue measurement (the FAI of the region of interest in this male patient was –63 HU). FAI, fat attenuation index; HU, Hounsfield units.

Statistical analysis

Statistical analyses were performed using SPSS 26.0 (IBM Corp., Armonk, NY, USA). A P value of <0.05 indicates a statistically significant difference. Continuous variables are presented as mean ± standard deviation (SD) and compared using an independent t-test. Categorical variables are expressed as counts (n) and percentages (%) and compared using the chi-squared test. Pearson correlation analysis was used to evaluate the association between the FAI and coronary plaque burden. Multivariate logistic regression analysis was conducted to identify independent predictors of MACEs.

R 4.0.3 Statistical software (The R Foundation for Statistical Computing) was used to complete the construction and validation of the prediction model. A nomogram prediction model was constructed based on the significant variables identified in the regression analysis. According to the principle of a completely random number table, the patients were assigned into a training set (n=226) and a validation set (n=227). The model’s performance was evaluated using receiver operating characteristic (ROC) curves and area under the curve (AUC) metrics in both a training set and a validation set. Calibration of the model was assessed using calibration curves and the Hosmer-Lemeshow goodness-of-fit test.

Results

Study population and patient characteristics

A total of 453 patients with CHD were included in the study, of whom 103 (22.7%) experienced MACEs during the 3-year follow-up period. The remaining 350 patients (77.3%) did not develop MACEs and were assigned to the control group. The enrollment process is outlined in Figure 2.

Figure 2.

Flowchart of the inclusion of patients with coronary heart disease. MACEs, major adverse cardiovascular events.

The baseline clinical characteristics of both groups are summarized in Table 1. Compared with the control group, patients in the MACE group were significantly older (76.9±7.7 vs. 64.0±10.2 years; P<0.001) and had lower LVEF (50.6%±4.4% vs. 56.6%±4.9%; P<0.001). The MACE group also had a higher total coronary plaque burden (40.1%±5.6% vs. 32.4%±3.1%; P<0.001), greater calcified plaque burden (22.4%±5.2% vs. 20.7%±2.5%; P<0.001), greater noncalcified plaque burden (17.8%±2.0% vs. 11.8%±1.8%; P<0.001), and a higher FAI (–61.0±11.1 vs. –68.8±10.6 HU; P<0.001). Additionally, the prevalence of multivessel CAD was significantly higher in the MACE group (44.7% vs. 20.0%; P<0.001).

Table 1. Comparison of the main clinical characteristics of the two groups.

Variable	MACE group (n=103)	Control group (n=350)	t/χ2 value	P value
Age (years)	76.9±7.7	64.0±10.2	11.890	<0.001
Gender			0.250	0.62
Male	62 (60.2)	201 (57.4)
Female	41 (39.8)	149 (42.6)
Body mass index (kg/m2)
Hypertension	62 (60.2)	221 (63.1)	0.295	0.59
Diabetes	48 (46.6)	189 (54.0)	1.746	0.19
Hyperlipidemia	82 (79.6)	281 (80.3)	0.023	0.88
LVEF (%)	50.6±4.4	56.6±4.9	11.146	<0.001
NT-proBNP (pg/mL)	593.7±204.8	606.8±198.5	0.585	0.56
Creatine kinase (U/L)	472.93±109.48	458.93±110.82	1.130	0.26
Creatine kinase isoenzymes (U/L)	404.92±98.83	397.72±92.72	0.682	0.50
Number of coronary arteries			25.404	<0.001
Single	57 (55.3)	280 (80.0)
Multivessel	46 (44.7)	70 (20.0)
Stent type			0.430	0.51
Drug-eluting stent	92 (89.32)	320 (91.43)
Bare mental stent	11 (10.68)	30 (8.57)
Total coronary plaque burden (%)	40.1±5.6	32.4±3.1	18.029	<0.001
Calcified plaque burden (%)	22.4±5.2	20.7±2.5	4.507	<0.001
Noncalcified plaque burden (%)	17.8±2.0	11.8±1.8	28.805	<0.001
FAI (HU)	–61.0±11.1	–68.8±10.6	6.471	<0.001
hs-CRP (mg/L)	4.7±1.9	4.9±1.9	0.990	0.32
Antihypertensive agents	62 (60.2)	221 (63.1)	0.295	0.59
Hypoglycemic agents	48 (46.6)	189 (54.0)	1.746	0.19
Lipid-lowering agents	82 (79.6)	281 (80.3)	0.023	0.88

Data are presented as mean ± standard deviation or n (%). FAI, fat attenuation index; hs-CRP, high-sensitivity C-reactive protein; HU, Hounsfield units; LVEF, left ventricular ejection fraction; MACE, major adverse cardiovascular event; NT-proBNP, N-terminal pro-B-type natriuretic peptide.

No significant differences were observed between the groups in terms of gender distribution, hypertension, diabetes mellitus, hyperlipidemia, NT-proBNP levels, or hs-CRP levels (P>0.05 for all).

Risk factors for MACEs

Multivariate logistic regression analysis identified the following independent risk factors for MACEs (Table 2): age ≥80 years [relative risks (RR) 12.39, 95% confidence interval (CI): 5.75–26.69, P<0.001], LVEF <50% (RR 8.73, 95% CI: 4.10–18.58, P<0.001), total coronary plaque burden >33.3% (RR 4.27, 95% CI: 2.23–8.18, P<0.001), increased FAI (RR 1.08, 95% CI: 1.05–1.11, P<0.001), and multivessel CAD (RR 3.14, 95% CI: 1.67–5.90).

Table 2. Analysis of risk factors for MACEs in patients with coronary heart disease.

Variable	B value	Standard error	Wald	P value	RR (95% CI)
Age ≥80 years	2.517	0.392	41.295	<0.001	12.39 (5.75–26.69)
LVEF <50%	2.167	0.385	31.589	<0.001	8.73 (4.10–18.58)
Coronary artery plaque burden >33.3%	1.452	0.331	19.188	<0.001	4.27 (2.23–8.18)
FAI (HU)	0.077	0.015	26.238	<0.001	1.08 (1.05–1.11)
Multivessel CAD	1.143	0.322	12.566	<0.001	3.14 (1.67–5.90)

33.3% was the median of the coronary artery plaque burden. CAD, coronary artery disease; CI, confident interval; FAI, fat attenuation index; HU, Hounsfield units; LVEF, left ventricular ejection fraction; MACE, major adverse cardiovascular event; RR, relative risk.

Development and validation of the MACE prediction model

A nomogram prediction model was developed based on the significant risk factors identified in the logistic regression analysis (Figure 3).

Figure 3.

Nomogram model for predicting MACEs in patients with coronary heart disease. HU, Hounsfield units; MACEs, major adverse cardiovascular events.

For model validation, the dataset was randomly split into a training set (n=226) and a validation set (n=227). The model demonstrated high predictive performance with a training set AUC of 0.920 (95% CI: 0.875–0.964) and a validation set AUC of 0.862 (95% CI: 0.803–0.920) (Figure 4). The model’s calibration of the validation set was assessed using the Hosmer-Lemeshow goodness-of-fit test, which yielded a chi-squared value of 10.351 and a P value of 0.24, indicating good agreement between predicted and observed outcomes (Figure 5).

Figure 4.

The value of the prediction model for predicting MACEs in patients with coronary heart disease (A: training set; B: validation set). AUC, area under the curve; MACEs, major adverse cardiovascular events.

Figure 5.

Calibration curves of the model in predicting MACEs in patients with coronary heart disease (A: training set; B: validation set). MACEs, major adverse cardiovascular events.

Correlation between FAI and coronary plaque burden

Pearson correlation analysis revealed a significant correlation between the FAI and noncalcified plaque burden (r=0.234, P<0.001). However, the FAI showed a weaker correlation with calcified plaque burden (r=0.097, P=0.04). These results suggest that the FAI may be more closely associated with noncalcified, vulnerable plaque rather than overall coronary plaque burden (Figure 6).

Figure 6.

Heatmap for the correlation between FAI and coronary plaque burden, left ventricular ejection fraction, and age in patients with coronary heart disease. The P values of the association between FAI and age, LVEF, calcified plaque burden, non-calcified plaque burden were 0.003, 0.008, 0.04 and <0.001. FAI, fat attenuation index; HU, Hounsfield units; LVEF, left ventricular ejection fraction.

Discussion

Coronary atherosclerosis is a progressive inflammatory disease that leads to plaque formation, coronary stenosis, and subsequent restriction of the myocardial blood supply. Additionally, rupture of coronary atheromatous plaque can abruptly reduce myocardial perfusion leading to acute myocardial infarction and MACEs (12). Emerging evidence suggests that pericoronary adipose tissue attenuation, as measured by CT, may serve as a noninvasive biomarker of coronary inflammation and plaque vulnerability (13-15). In this study, we demonstrated that an elevated FAI was significantly associated with noncalcified coronary plaque burden and independently predicted MACEs in patients with CHD with RR of 1.08 (95% CI: 1.05–1.11). Furthermore, we developed and validated a nomogram prediction model incorporating the FAI, which exhibited high accuracy in predicting future cardiovascular events.

FAI is part of epicardial adipose tissue, but the two have different morphological and functional characteristics, with FAI being closer to the coronary arteries and more sensitive to paracrine inflammatory signals than to autocrine signals released from fat deposits. The FAI in patients with CHD is an imaging biomarker for the noninvasive quantification of coronary artery inflammation. The release of inflammatory signals from the blood vessel wall directly diffuses into the pericoronary fat, which prevents lipid accumulation by affecting biological processes such as adipocyte differentiation, proliferation, and decomposition. It can also lead to increased microvascular permeability and promote perivascular edema, transforming the lipid phase to the aqueous phase, increased attenuation in CT images, and a higher FAI (16). Meanwhile, vascular inflammation is the principal cause of endothelial dysfunction, which can promote the progression and instability of atherosclerotic plaques; in turn, this results in an increase in coronary plaque burden, especially noncalcified plaque burden (13,14), and ultimately MACEs.

One study demonstrated that the FAI is a valuable parameter for diagnosing patients with suspected CHD, as patients diagnosed with CHD exhibited significantly higher attenuation indices (15). Li et al. further confirmed that an elevated FAI could predict the occurrence of ACS and was associated with lumen stenosis (17). Another study reported that the FAI was correlated with the development of unstable angina (18). Additionally, Zuo et al. showed that a persistent increase in the FAI following coronary artery stent implantation was associated with an increased coronary plaque burden (19). Consistent with our findings, several other studies have also demonstrated that the FAI is associated with the presence of coronary artery lesions and MACEs (20-24). Our study also showed a weak correlation between the FAI and noncalcified plaque burden (r=0.234).

Interestingly, while the FAI correlated significantly with noncalcified plaque burden, its association with calcified plaque burden was weaker. Previous studies also confirmed that the FAI was correlated with the noncalcified plaque burden (24,25). This observation is consistent with the hypothesis that pericoronary adipose tissue reflects active inflammatory changes rather than chronic atherosclerosis and supports the finding in the previous studies (26,27). Noncalcified plaques are more prone to rupture, leading to acute thrombotic events, whereas calcified plaques are generally more stable. These findings support the potential clinical utility of the FAI as a marker for identifying patients at risk of plaque rupture and MACEs. However, van Rosendael et al. found that FAI had no significant predictive value for long-term MACEs events, and more research is needed to confirm the role of FAI in MACEs events (28).

The integration of the FAI into cardiovascular risk assessment may offer additional prognostic information beyond traditional markers such as plaque burden and LVEF. Our predictive nomogram model, which included the FAI, demonstrated excellent discriminative ability (AUC =0.920 in the training set, 0.862 in the validation set), highlighting its potential for use in individualized risk stratification. These findings suggest that patients with elevated FAI may benefit from more intensive secondary prevention strategies, including early initiation of anti-inflammatory therapies (e.g., colchicine, IL-1 inhibitors), tighter lipid control. Future prospective studies are warranted to evaluate whether targeting patients with high FAI for aggressive intervention improves clinical outcomes (29,30).

This study has several strengths. First, it employs quantitative CCTA analysis with AI-based software, ensuring objective and reproducible FAI measurements. Second, the 3-year follow-up period provides robust longitudinal data. Third, the use of a predictive nomogram model allows for individualized risk assessment, which is clinically valuable.

This study has several limitations that should be acknowledged. First, this study was a single-center, retrospective study, which may introduce selection bias and limit the generalizability. Second, only patients with ACS were included; thus, the findings may not apply to patients with chronic coronary syndrome. Third, direct inflammatory biomarkers (e.g., interleukin-6, tumor necrosis factor-alpha) were not assessed, and histopathologic validation of plaque characteristics was not performed, which limits the biological correlation of imaging findings. Finally, external validation in broader populations is needed to confirm the utility of the proposed prediction model. Fourth, the control group was larger than the MACE group, this could introduce bias.

Conclusions

This study provides evidence that the FAI is an independent predictor of MACEs and correlates with noncalcified coronary plaque burden. A novel nomogram prediction model incorporating the FAI demonstrated high predictive accuracy, suggesting its potential clinical utility in risk stratification and personalized treatment strategies. Further validation studies are warranted to establish the role of the FAI in guiding cardiovascular risk management.

Supplementary

The article’s supplementary files as

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. This study was approved by the Ethics Committee of Wuhan Fourth Hospital (No. 202501242). Due to the retrospective design, the requirement for individual informed consent was waived.

Data Sharing Statement

Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-775/dss

jtd-18-02-153-dss.pdf ^{(99KB, pdf)}

DOI: 10.21037/jtd-2025-775