Prognostic incremental value of perivascular adipose tissue in myocardial infarction with non-obstructive coronary arteries: a multicenter cohort study

Introduction

Perivascular adipose tissue (PVAT) is the adipose tissue that surrounds the blood vessels and secretes a variety of biologically active substances that are essential for the regulation of vascular function (1,2). Analyses based on PVAT characteristics have provided new insights into cardiovascular risk stratification and have demonstrated potential in risk prediction of major adverse cardiovascular events (MACE) (3,4). Myocardial infarction with non-obstructive coronary arteries (MINOCA) is defined as myocardial infarction occurring in the absence of significant obstructive coronary artery disease (stenosis <50%) (5,6), Clinical observations indicate that approximately one-third of MINOCA patients present with metabolic abnormalities, including insulin resistance (7). PVAT is increasingly recognized as a key mediator of vascular inflammation and cardiovascular risk (8). Given its critical role in cardiovascular regulation, the association between PVAT alterations and the high incidence of MACE in MINOCA warrants further investigation. Compared with epicardial adipose tissue (EAT), PVAT more accurately reflects local vascular inflammation and metabolic activity due to its direct anatomical contact with the vascular wall (9,10), and can be assessed noninvasively and quantitatively by cardiac magnetic resonance (CMR) (11). Recent studies have highlighted EAT density as an imaging biomarker in ischemia with non-obstructive coronary arteries, with higher density associated with increased disease risk (12), and have framed atherosclerosis as a bidirectional interplay between cholesterol metabolism and vascular inflammation (13). However, traditional EAT volume measurement remains technically challenging, requiring dedicated post-processing software and expertise. Therefore, this study proposed a simplified CMR-based scoring method for PVAT assessment to enhance its clinical utility in the MINOCA setting.

While PVAT reflects local coronary metabolism, the triglyceride-glucose (TyG) index provides a systemic assessment of metabolic status, enabling sensitive detection of glucose-lipid disorders and insulin resistance at early stages (14,15). However, evidence supporting the association between the TyG index and MACE risk in MINOCA remains limited. This study aimed to evaluate both CMR-based PVAT parameters and the TyG index for risk stratification and to quantify their incremental predictive value for MACE in patients with MINOCA. We hypothesized that the PVAT score would be independently associated with MACE and provide incremental prognostic value beyond established clinical predictors, and that this association would be more pronounced in patients with diabetes mellitus (DM). We present this article in accordance with the STROBE reporting checklist (available at https://qims.amegroups.com/article/view/10.21037/qims-2026-1-0245/rc).

Methods

Study population

The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. This study was approved by the Ethics Committees of Sichuan Provincial People’s Hospital (No. 2024-693), The Affiliated Hospital of Chengdu Medical College (No. 2025CYFYIRB-BA-057) and West China Longquan Hospital of Sichuan University (No. 2025-0119). Informed consent was waived in this retrospective study. All data were anonymized in accordance with Health Insurance Portability and Accountability Act regulations (16). Between January 2015 and December 2024, 272 patients with MINOCA were enrolled from three hospitals (Sichuan Provincial People’s Hospital, The Affiliated Hospital of Chengdu Medical College, West China Longquan Hospital of Sichuan University). Inclusion criteria required that CMR be performed within one week of admission. A total of 204 healthy controls were also recruited, all of whom had no evidence of structural heart disease, myocardial injury, DM, hypertension, or hyperlipidemia as confirmed by CMR. The diagnosis of acute myocardial infarction followed the fourth edition of the “Common Definition of Myocardial Infarction” (17), and MINOCA was defined according to the European Society of Cardiology guidelines and the American Heart Association Scientific Statement (18,19). Eligible patients underwent baseline CMR within 7 days post-admission with preserved left ventricular function. Exclusion criteria were as follows: (I) CMR contraindications or refusal (informed consent not obtained from patients/legal guardians); (II) incomplete CMR data or poor image quality; (III) missing >5% of key clinical variables; (IV) loss to follow-up. To ensure consistency in data collection, CMR image acquisitions were performed according to standardized protocols across all three participating centers. Eligible patients were consecutively enrolled to minimize selection bias (Figure 1).

Figure 1 Flowchart of patients (A) and controls (B) enrollment. CMR, cardiac magnetic resonance.

Metabolic parameters assessed included age, body mass index, blood pressure, blood glucose, triglycerides, and cholesterol. The TyG index is calculated as: ln [fasting triglycerides (mg/dL) × fasting blood glucose (mg/dL)/2] (20). We defined composite endpoints as MACE, including all-cause death, nonfatal reinfarction, ischemic stroke, and readmission for heart failure (21,22). For patients with multiple events, only the earliest event date was recorded.

In the diagnostic phase, CMR was used to exclude alternative non-ischemic diagnoses; patients with subepicardial or mid-wall late gadolinium enhancement (LGE) suggestive of acute myocarditis were excluded from the study. LGE patterns (e.g., subendocardial, transmural, or absence of LGE) were recorded based on routine clinical imaging.

CMR protocol

All patients underwent CMR examinations on a Siemens Healthineers 3.0T clinical scanner. The standardized protocol included balanced steady-state free precession dynamic imaging techniques to assess cardiac function and structure, covering the left ventricular short-axis view and standard long-axis views. LGE images were acquired using a phase-sensitive inversion recovery sequence 10–15 minutes after intravenous gadolinium contrast administration to detect myocardial fibrosis or necrosis. High-resolution images suitable for PVAT assessment were obtained via standard dynamic sequences in the four-chamber and basal short-axis views. Key imaging parameters ranged as follows: repetition time =3.33 ms, echo time =1.92 ms, flip angle =20°, slice thickness =8 mm, field of view =122 mm ×
242 mm, matrix size =152 × 256.

CMR imaging and PVAT score

CMR cine short-axis images assessed global left ventricular function and structure in all patients. PVAT thickness measurements followed established methods (23,24). On four-chamber views, the left atrioventricular groove (LAVG), right atrioventricular groove (RAVG), and anterior interventricular groove (AIVG) were identified according to standardized anatomical criteria (25,26) (Figure S1A). The basal short-axis view was used to measure PVAT thickness at the superior interventricular groove (SIVG) and inferior interventricular groove (IIVG) (Figure S1B). All measurements normalized to body surface area (BSA). PVAT measurements were performed by two independent cardiologists blinded to clinical outcomes. Intra- and interobserver correlation coefficients for PVAT thickness are presented in Table S1, demonstrating good reproducibility. To integrate multi-regional PVAT information into a clinically applicable scoring tool, the PVAT score (range 0–5) was developed as follows: The five anatomical sites (LAVG, RAVG, AIVG, SIVG, IIVG) were selected for their clear identifiability on standard CMR views, directly overlie the major coronary arteries, and capture regional heterogeneity in perivascular fat distribution.

The 95th percentile of each PVAT index (LAVGi, RAVGi, AIVGi, SIVGi, IIVGi) derived from the 204 healthy controls was set as the upper normal limit. This threshold was chosen to conservatively define values exceeding those observed in 95% of the healthy population as abnormal, ensuring robustness to data distribution. For each MINOCA patient, an index exceeding its corresponding limit was assigned 1 point (otherwise 0), based on the premise that PVAT expansion beyond the healthy range reflects abnormal perivascular fat deposition or inflammation linked to elevated cardiovascular risk. Points from the five regions were then summed to yield the total PVAT score, with higher scores indicating a greater extent of abnormal PVAT accumulation. Finally, for straightforward clinical interpretation, patients were stratified into low- [0–1], medium- [2–3], and high-risk [4–5] categories. CMR analysis was performed by experienced radiologists blinded to clinical outcomes and patient group assignment.

Statistical analysis

Continuous variables are presented as mean ± standard deviation or median (interquartile range), with between-group comparisons using Student’s t-test or Mann-Whitney U test. Categorical variables were compared using χ² tests or Fisher’s exact tests. Multivariable Cox proportional hazards regression analyzed MACE risk factors. Predictor variables for the multivariable model were selected based on clinical relevance and results from univariable analyses (P<0.10), with age and LGE positivity forced into the model as established prognostic factors. Multicollinearity among predictor variables was assessed using the variance inflation factor (VIF), with a threshold of >5 indicating significant collinearity. Variables exhibiting high collinearity (e.g., DM and lipid parameters with the TyG index) were not simultaneously included in the multivariable model to ensure stability. The proportional hazards assumption of the Cox proportional hazards regression model was validated using the Schoenfeld residual test. Test results are presented via Chi-squared statistics and corresponding P values, with visualized assessments provided through scaled Schoenfeld residual plots for each covariate. Incremental predictive value was assessed through: likelihood ratio tests comparing nested models; receiver operating characteristic (ROC) curves with C-statistics; with differences in area under the curve (AUC) compared using DeLong’s test; net reclassification index (NRI) and integrated discrimination improvement (IDI); and decision curve analysis (DCA) quantifying net clinical benefit. Risk stratification was evaluated using Kaplan-Meier survival curves with log-rank tests. Interaction tests were employed to assess the statistical significance of the association between biomarkers and ischemic heart disease with the risk of MACE, and to quantify differences in risk gradients across different subgroups. Data analysts were blinded to group allocation during statistical analysis. Missing data were handled using complete case analysis, as the proportion of missing values was low (<1%) after excluding patients with >5% missing key variables. All analyses used R software (version 4.4.2; R Foundation for Statistical Computing), with two-sided P<0.05 considered significant.

Results

Baseline characteristics

Table 1 presents the baseline characteristics of the study population, comparing the MINOCA group with the control group. A significantly higher proportion of MINOCA patients carried traditional cardiovascular risk factors, including dyslipidemia and glycemia, compared with controls. All region-specific PVAT indicators were significantly higher in the MINOCA group than in healthy controls (P<0.05), with particularly significant between-group differences in LAVGi, RAVGi, and SIVGi (P<0.001). In the overall MINOCA cohort, the median PVAT score was 1.5 [interquartile range (IQR) 0.9–2.5], with 112 patients (41.2%) classified as low‑risk (score 0–1), 124 (45.6%) as medium‑risk (score 2–3), and 36 (13.2%) as high‑risk (score 4–5).

Follow-up and outcomes

During follow-up, 57 patients (20.9%) met the composite endpoints: all-cause death in 12 (4.4%); reinfarction in 10 (3.7%); heart failure in 29 (10.6%); and ischemic stroke in 6 (2.2%). Patients who developed MACE demonstrated a significantly higher cardiovascular risk burden: elevated fasting blood glucose, triglycerides, TyG index and lower high-density lipoprotein cholesterol levels (P<0.001). Patients who developed MACE exhibited significantly higher PVAT measurements, including PVAT score and LAVGi (P<0.001) (Table 2).

Univariable and multivariable analyses

In univariable Cox regression analysis, age, LGE, TyG index, PVAT score, LAVGi, RAVGi, AIVGi, SIVGi, IIVGi, fasting blood glucose and triglycerides were associated with an increased risk of MACE. After adjustment for multivariable Cox regression analysis, age [hazard ratio (HR): 1.09, 95% confidence interval (CI): 1.05–1.13, P<0.001], LGE positivity (HR: 1.98, 95% CI: 1.16–3.39, P=0.012), TyG index (HR: 4.28, 95% CI: 1.92–9.51, P<0.001), a higher PVAT score (HR: 3.85, 95% CI: 1.52–9.72, P=0.004) were independent predictors of composite MACE in patients after MINOCA (Table 3). Multicollinearity was assessed using VIF. In the full model (age, LGE, TyG index, and PVAT score), all VIF values were below 2 (age: 1.12 years, LGE: 1.08, TyG index: 1.35, PVAT score: 1.28), indicating no significant collinearity. The proportional hazards assumption was tested using the Schoenfeld residual test. All covariates in the final Cox model satisfied the assumption: age (χ²=0.032, P=0.86), positive LGE (χ²=0.842, P=0.36), TyG index (χ²=0.717, P=0.40), and PVAT score (χ²=0.973, P=0.32). A global test also confirmed the proportional hazards assumption for the entire model (χ²=1.84, P=0.77). Scalar Schoenfeld residual plots further visually confirmed that the HRs for each covariate remained constant throughout the follow-up period (Figure S2).

Incremental predictive value

Since the predictive value of age and LGE for clinical endpoint events in MINOCA has been confirmed by previous literature (27), the TyG index and PVAT score were relatively new indicators. Therefore, four prediction models with nested relationships were developed, namely, Model 1 (age, LGE), Model 2 (age, LGE, TyG index), Model 3 (age, LGE, PVAT score) and Model 4 (age, LGE, PVAT score, TyG index). Model 4 had the highest AUC value (Figure 2A); compared with model 1, the C-indexes of models 2, 3, and 4 increased by 0.092, 0.082 and 0.118, respectively (all P<0.05) (Figure 2B), indicating that it outperformed the other models in terms of discrimination ability. Pairwise comparisons by DeLong’s test confirmed significant improvements in discrimination when adding TyG index or PVAT score to the base model (all P<0.001), and between the full model and both Base + TyG (P=0.028) and Base + PVAT score (P=0.013) models (Table S2). As shown in Table 4, NRI and IDI analysis demonstrated that adding either TyG index or PVAT score alone to the base model significantly improved the predictive performance of the model, as evidenced by the significant improvement in categorical NRI, continuous NRI and IDI (all P<0.05). The full model constructed by incorporating both TyG and PVAT scores demonstrated the largest performance increment compared to the base model (categorical NRI: 0.51; continuous NRI: 1.05; IDI: 0.26; P<0.001). In addition, model 4 (χ²=97.8) had an optimal increment of goodness-of-fit compared with model 1 (χ²=46.1), model 2 (χ²=89.5) and model 3 (χ²=83.3) (Figure S3). The clinical decision curve also showed the most significant incremental gain for model 4 compared to other models (Figure 2C,2D), indicating TyG index and PVAT score have incremental value for model discrimination, goodness of fit and clinical applicability.

Figure 2 Incremental predictive value analysis for discrimination and clinical applicability. Four prediction models with nested relationships were developed: Model 1 (age, LGE) (Base), Model 2 (Base + TyG index), Model 3 (Base + PVAT score) and Model 4 (Base + TyG index and PVAT score). Model 4 had the highest AUC value (A). Compared with model 1, the C-indexes of models 2, 3, and 4 increased by 0.092, 0.082, and 0.118, respectively (B). *, P<0.05. The clinical decision curve also showed the most significant incremental gain for Model 4 compared to other models (C,D). AUC, area under the curve; CI, confidence interval; LGE, late gadolinium enhancement; PVAT, perivascular adipose tissue; ROC, receiver operating characteristic; TyG, triglyceride-glucose.

Risk stratification and survival analysis

Patients were categorized into low-risk, medium-risk and high-risk groups based on the tertiles of each region-specific PVAT index. Kaplan-Meier curves demonstrated statistically significant differences in the probability of event-free survival for LAVGi (P=0.023), AIVGi (P=0.050) and SIVGi (P=0.018) (Figure 3A-3C); while RAVGi and IIVGi were not significantly different among the risk groups (Figure 3D,3E). In particular, the probability of event-free survival based on PVAT score was significantly lower in the high-risk group than in the medium-risk group versus the low-risk group (P<0.001), demonstrating superior risk stratification compared with individual regional PVAT index (Figure 3F).

Figure 3 Risk stratification and survival analysis. Kaplan-Meier curves demonstrated statistically significant differences in the probability of event-free survival for LAVGi, AIVGi, and SIVGi (A-C); while RAVGi and IIVGi were not significantly different among the risk groups (D,E). The probability based on PVAT score was significantly lower in the high-risk group than in the medium-risk and low-risk group, with demonstrated superior risk stratification compared to individual regional PVAT index (F). *, P≤0.05; ***, P<0.001. AIVGi, anterior interventricular groove index; IIVGi, inferior interventricular groove index; LAVGi, left atrioventricular grooves index; PVAT, perivascular adipose tissue; RAVGi, right atrioventricular groove index; SIVGi, superior interventricular groove index.

DM interaction effects

Subgroup analyses stratified by DM status revealed consistent patterns for both biomarkers: MINOCA patients with DM exhibited significantly higher MACE risks than their non-diabetic counterparts across all TyG index levels and PVAT score strata (all P<0.001, Figure 4A,4B). Both biomarkers demonstrated dose-dependent relationships with MACE risk in diabetic and non-diabetic subgroups, as evidenced by increasing predicted probabilities with elevated TyG indices and higher PVAT scores. Notably, the presence of DM amplified the risk gradient, with diabetic patients showing steeper risk escalation per unit increase in either biomarker compared to non-diabetic patients.

Figure 4 Subgroup analyses based on DM status. MINOCA patients with DM exhibited significantly higher MACE risks than their non-DM counterparts across all TyG index levels (P<0.001) (A) and PVAT score strata (P<0.001) (B). DM, diabetes mellitus; MACE, major adverse cardiovascular events; MINOCA, myocardial infarction with non-obstructive coronary arteries; PVAT, perivascular adipose tissue; TyG, triglyceride-glucose.

Discussion

To our knowledge, this multicenter study is the first to establish the prognostic value of CMR-based PVAT assessment in MINOCA through developing a novel scoring tool. Notably, PVAT score tertiles effectively stratified patients into distinct risk categories in patients with MINOCA. Subgroup analyses revealed that DM amplified the dose-dependent associations of TyG index and PVAT score with MACE risk in MINOCA patients, with steeper risk escalation per unit increase in either biomarker observed in diabetic compared with non-diabetic patients.

The value of PVAT assessment extends beyond simple thickness measurements; regional specificity plays a key role in predicting prognosis in MINOCA. By integrating information from key regions—such as the LAVG and the AIVG—through CMR multiregional quantification, we found that the PVAT score emerged as the strongest independent predictor of MACE, offering a novel tool for individualized risk stratification. Other studies have shown that PVAT functional status in specific coronary artery regions (28,29) is more sensitive to localized microvascular injury and more closely correlates with the risk of adverse cardiovascular events, making it a more effective prognostic indicator than overall thickness. PVAT abnormalities were strongly associated with metabolic disorders including obesity and insulin resistance. These comorbidities might exacerbate PVAT inflammation and adipokine imbalance, forming a vicious cycle that collectively impairs microvascular function and worsens MINOCA prognosis (30,31). Although regional localized fat attenuation indices derived from coronary computed tomography (CT) angiography have been shown to be a reliable marker of coronary inflammation and plaque vulnerability (32,33) and might improve the prediction of the risk of cardiovascular events in MINOCA (33,34), the associated radiation exposure limits their use for serial monitoring. In contrast, radiation-free CMR offers significant advantages for long-term risk assessment (6,35,36), aligning with the emerging paradigm of contrast-free CMR, where artificial intelligence (AI) has recently been shown to further enhance image quality and diagnostic accuracy (37). Previous studies have demonstrated the significant prognostic value of the RAVG fat thickness index quantified by CMR in patients with metabolic syndrome, and this imaging biomarker independently predicts the risk of major cardiovascular events and is expected to provide an early warning of the risk of reinfarction and hospitalization for heart failure in patients with MINOCA (38). Accordingly, the present study not only assessed multiple region-specific, radiation-free PVAT metrics based on CMR, but also established a simple and validated PVAT scoring tool that integrates these multidimensional regional CMR-PVAT data to improve prognostic accuracy in MINOCA, particularly in patients with concomitant DM.

While PVAT characteristics reflect localized coronary metabolic status, the TyG index serves as a systemic metabolic marker strongly associated with coronary atherosclerosis progression and acute myocardial infarction risk (39,40). In clinical practice, relying on fasting blood glucose alone might miss the diagnosis of metabolic abnormalities in early stage (40), whereas the TyG index could effectively make up for the shortcomings of fasting glucose. In MINOCA patients, elevated TyG index not only indicates metabolic disorders but also predicts adverse cardiovascular events (41), which may involve insulin resistance-mediated endothelial dysfunction, microvascular inflammation and increased plaque vulnerability (42). Notably, a significant proportion of MINOCA patients have under-recognized abnormalities of glucose metabolism, and these hidden metabolic disturbances constitute an important source of residual risk (43). The results of the present study suggest that TyG index also has an early warning value for MINOCA patients with concomitant diabetes. When elevated TyG index coexists with DM, the two have a synergistic amplification effect: the diabetic state may further impair coronary microcirculatory function and promote the spread of perivascular fatty inflammation, which not only exacerbates myocardial damage, but may also multiply the risk of recurrent myocardial infarction, hospitalization for heart failure and cardiovascular death in patients with MINOCA (44).

Beyond its statistical incremental value, the PVAT score has potential clinical implications. Derived from routine CMR images, this simple score may aid MINOCA management in several ways. First, in terms of risk stratification and follow-up strategies, the PVAT score enables refinement of risk categories: patients with high scores [4–5], particularly those with concomitant diabetes, may benefit from more intensive surveillance—such as closer outpatient follow-up and more frequent cardiac function assessment—while those with low scores [0–1] might be spared unnecessary medical burden. Second, regarding metabolic therapy guidance, the strong interaction between PVAT score and diabetes suggests that this imaging biomarker may help identify candidates for targeted metabolic interventions. For diabetic MINOCA patients with elevated PVAT scores, clinicians might consider prioritizing cardioprotective glucose-lowering agents [e.g., sodium-glucose cotransporter 2 (SGLT2) inhibitors or glucagon-like peptide-1 (GLP-1) receptor agonists] that not only improve glycemic control but also exert anti-inflammatory effects potentially targeting the perivascular pathology reflected by PVAT. Third, the PVAT score holds promise as an enrichment tool for future clinical trials, enabling selection of high-risk patients most likely to benefit from novel anti-inflammatory or metabolic therapies, thereby increasing trial efficiency and power.

MINOCA encompasses a spectrum of underlying etiologies, including atherothrombotic, vasospastic, embolic, and spontaneous coronary artery dissection. In the present study, we used CMR to help rule out alternative non-ischemic diagnoses (e.g., acute myocarditis) during patient enrollment. LGE patterns were recorded descriptively, but no formal etiological subclassification was performed. Given the retrospective design and the absence of routine intracoronary imaging (optical coherence tomography or intravascular ultrasound) to confirm specific etiologies, we did not perform independent prognostic analyses for different etiological subgroups. Instead, this study focused on the prognostic value of the PVAT score in the overall MINOCA cohort. Future studies incorporating advanced intracoronary imaging and comprehensive CMR protocols are warranted to further elucidate the relationship between PVAT and specific MINOCA etiologies. Recent multimodal imaging approaches, including CMR, have demonstrated improved diagnostic accuracy and risk stratification in MINOCA patients (36). Moreover, machine learning approaches offer promising tools for integrating multi-modal imaging and clinical data to achieve even more precise risk stratification in MINOCA (45).

Nevertheless, several limitations of this study warrant consideration. First, the retrospective observational design precludes causal inferences; our findings should be interpreted as hypothesis-generating, and residual confounding from unmeasured variables—such as medication use (e.g., statins, antiplatelet agents), systemic inflammatory markers (e.g., high-sensitivity C-reactive protein, interleukin-6), and coronary microvascular dysfunction—cannot be entirely excluded despite multivariable adjustment. Second, we did not exclude potential heart failure with preserved ejection fraction, which may confound MACE outcomes. Third, the relatively short follow-up duration necessitates longer-term validation in future studies. Fourth, the PVAT scoring tool was derived using healthy control thresholds from the same cohort; it requires prospective validation in larger, independent populations, as well as external validation in community-based settings, diverse ethnic groups, and non-tertiary centers to confirm its generalizability. Future prospective studies with comprehensive data collection are warranted to address these limitations and validate our findings.

Conclusions

In this multicenter cohort study, region-specific PVAT indicators and PVAT score were increased in patients with MINOCA and were associated with a poorer prognosis. Beyond established clinical predictors including age and LGE, the PVAT score independently predicted MACE with significant incremental prognostic utility, further enhanced by the addition of the TyG index. This incremental value was particularly pronounced in diabetic MINOCA patients. By enabling refined risk stratification and identifying a high-risk phenotype potentially amenable to targeted metabolic interventions, this simple, radiation-free imaging biomarker holds promise for guiding personalized surveillance and therapeutic strategies in MINOCA. Prospective validation in independent cohorts is warranted to facilitate its translation into clinical practice.

Acknowledgments

We thank all the researchers who contributed to this article.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://qims.amegroups.com/article/view/10.21037/qims-2026-1-0245/rc

Data Sharing Statement: Available at https://qims.amegroups.com/article/view/10.21037/qims-2026-1-0245/dss

Funding: This study was supported by the Clinical Research Funding of Chinese Society of Cardiovascular Disease (CSC) of 2019 (No. HF-CSC2019B01), the Project of Chengdu Science and Technology (No. 2022-GH03-00011-HZ), the National Natural Science Foundation of China (Nos. 82572196 and 82471970), the Key Project of Natural Science Foundation of Sichuan Province (No. 2026NSFSCZY0011), and the Chengdu Science and Technology Program (Nos. 2026-YF08-00166-GX and 2026-XT00-00013-GX).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://qims.amegroups.com/article/view/10.21037/qims-2026-1-0245/coif). The authors have no conflicts of interest to declare.

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 Committees of Sichuan Provincial People’s Hospital (No. 2024-693), The Affiliated Hospital of Chengdu Medical College (No. 2025CYFYIRB-BA-057) and West China Longquan Hospital of Sichuan University (No. 2025-0119). Informed consent was waived in this retrospective study.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Queiroz M, Sena CM. Perivascular adipose tissue in age-related vascular disease. Ageing Res Rev 2020;59:101040. [Crossref] [PubMed]
2. Tai GJ, Ma YJ, Feng JL, Li JP, Qiu S, Yu QQ, Liu RH, Wankumbu SC, Wang X, Li XX, Xu M. NLRP3 inflammasome-mediated premature immunosenescence drives diabetic vascular aging dependent on the induction of perivascular adipose tissue dysfunction. Cardiovasc Res 2025;121:77-96. [Crossref] [PubMed]
3. Chatterjee D, Shou BL, Matheson MB, Ostovaneh MR, Rochitte C, Chen MY, Dewey M, Ortman J, Cox C, Lima JAC, Arbab-Zadeh A. Perivascular fat attenuation for predicting adverse cardiac events in stable patients undergoing invasive coronary angiography. J Cardiovasc Comput Tomogr 2022;16:483-90. [Crossref] [PubMed]
4. Sun X, Zhu Y, Zhang N, Yuan K, Ling J, Ye J. Prognostic value of serial coronary computed tomography angiography-derived perivascular fat-attenuation index and plaque volume in patients with suspected coronary artery disease. Clin Radiol 2024;79:599-607. [Crossref] [PubMed]
5. Thygesen K, Alpert JS, Jaffe AS, Chaitman BR, Bax JJ, Morrow DA, White HD. Executive Group on behalf of the Joint European Society of Cardiology (ESC)/American College of Cardiology (ACC)/American Heart Association (AHA)/World Heart Federation (WHF) Task Force for the Universal Definition of Myocardial Infarction. Fourth Universal Definition of Myocardial Infarction (2018). Circulation 2018;138:e618-51. [Crossref] [PubMed]
6. Reynolds HR, Maehara A, Kwong RY, Sedlak T, Saw J, Smilowitz NR, et al. Coronary Optical Coherence Tomography and Cardiac Magnetic Resonance Imaging to Determine Underlying Causes of Myocardial Infarction With Nonobstructive Coronary Arteries in Women. Circulation 2021;143:624-40. [Crossref] [PubMed]
7. Hu P, Liu XB, Liang J, Zhu QF, Pu CX, Tang MY, Wang JA. A hospital-based survey of patients with severe valvular heart disease in China. Int J Cardiol 2017;231:244-7. [Crossref] [PubMed]
8. Trimarchi G, Carerj ML, Zito C, Bella GD, Taverna G, Cusmà Piccione M, Crea P, Lo Giudice S, Buonpane A, Bonanni M, Restelli D, Paradossi U, Monteleone A, Micari A, Carerj S. Epicardial Adipose Tissue: A Multimodal Imaging Diagnostic Perspective. Medicina (Kaunas) 2025;61:961. [Crossref] [PubMed]
9. Kim HW, Shi H, Winkler MA, Lee R, Weintraub NL. Perivascular Adipose Tissue and Vascular Perturbation/Atherosclerosis. Arterioscler Thromb Vasc Biol 2020;40:2569-76. [Crossref] [PubMed]
10. Agabiti-Rosei C, Saxton SN, De Ciuceis C, Lorenza Muiesan M, Rizzoni D, Agabiti Rosei E, Heagerty AM. Influence of Perivascular Adipose Tissue on Microcirculation: A Link Between Hypertension and Obesity. Hypertension 2024;81:24-33. [Crossref] [PubMed]
11. Lind L, Strand R, Michaëlsson K, Ahlström H, Kullberg J. Voxel-wise Study of Cohort Associations in Whole-Body MRI: Application in Metabolic Syndrome and Its Components. Radiology 2020;294:559-67. [Crossref] [PubMed]
12. Sun F, Tian Y, Yu W, Song X, Zhang F, Wang J, Shao X, Liu B, Yang X, Wan P, Chen Y, Li S, Wang Y. A modifiable imaging biomarker: epicardial adipose tissue density in ischemia with non-obstructive coronary arteries. Front Cardiovasc Med 2026;13:1686602. [Crossref] [PubMed]
13. Zeng RX, Li S, Li JJ. Cholesterol-Inflammation Fusion Hypothesis in Atherosclerosis: An Evolving Paradigm in Pathogenesis and Therapy. JACC Asia 2026;6:284-96. [Crossref] [PubMed]
14. Cho YK, Kim HS, Park JY, Lee WJ, Kim YJ, Jung CH. Triglyceride-Glucose Index Predicts Cardiovascular Outcome in Metabolically Unhealthy Obese Population: A Nationwide Population-Based Cohort Study. J Obes Metab Syndr 2022;31:178-86. [Crossref] [PubMed]
15. Brahimaj A, Rivadeneira F, Muka T, Sijbrands EJG, Franco OH, Dehghan A, Kavousi M. Novel metabolic indices and incident type 2 diabetes among women and men: the Rotterdam Study. Diabetologia 2019;62:1581-90. [Crossref] [PubMed]
16. Cohen IG, Mello MM. HIPAA and Protecting Health Information in the 21st Century. Jama 2018;320:231-2. [Crossref] [PubMed]
17. DeFilippis AP, Chapman AR, Mills NL, de Lemos JA, Arbab-Zadeh A, Newby LK, Morrow DA. Assessment and Treatment of Patients With Type 2 Myocardial Infarction and Acute Nonischemic Myocardial Injury. Circulation 2019;140:1661-78. [Crossref] [PubMed]
18. Byrne RA, Rossello X, Coughlan JJ, Barbato E, Berry C, Chieffo A, et al. 2023 ESC Guidelines for the management of acute coronary syndromes: Developed by the task force on the management of acute coronary syndromes of the European Society of Cardiology (ESC). European Heart Journal 2023;44:3720-826. Erratum in: Eur Heart J 2024;45:1145.
19. Rao SV, O'Donoghue ML, Ruel M, Rab T, Tamis-Holland JE, Alexander JH, et al. 2025 ACC/AHA/ACEP/NAEMSP/SCAI Guideline for the Management of Patients With Acute Coronary Syndromes: A Report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines. Circulation 2025;151:e771-862. [Crossref] [PubMed]
20. Xia W, Cai Y, Zhang S, Wu S. Association between different insulin resistance surrogates and infertility in reproductive-aged females. BMC Public Health 2023;23:1985. [Crossref] [PubMed]
21. Kelshiker MA, Seligman H, Howard JP, Rahman H, Foley M, Nowbar AN, Rajkumar CA, Shun-Shin MJ, Ahmad Y, Sen S, Al-Lamee R, Petraco RCoronary Flow Outcomes Reviewing Committee. Coronary flow reserve and cardiovascular outcomes: a systematic review and meta-analysis. Eur Heart J 2022;43:1582-93. [Crossref] [PubMed]
22. Eichhorn C, Koeckerling D, Reddy RK, Ardissino M, Rogowski M, Coles B, Hunziker L, Greulich S, Shiri I, Frey N, Eckstein J, Windecker S, Kwong RY, Siontis GCM, Gräni C. Risk Stratification in Nonischemic Dilated Cardiomyopathy Using CMR Imaging: A Systematic Review and Meta-Analysis. JAMA 2024;332:1535-50. [Crossref]
23. Rodriguez-Granillo GA, Raggio IM, Deviggiano A, Bellia-Munzon G, Capunay C, Nazar M, Martinez JL, Carrascosa P, Martinez-Ferro M. Impact of pectus excavatum on cardiac morphology and function according to the site of maximum compression: effect of physical exertion and respiratory cycle. Eur Heart J Cardiovasc Imaging 2020;21:77-84. [Crossref] [PubMed]
24. Sepehri Shamloo A, Schoene K, Stauber A, Darma A, Dagres N, Dinov B, Bertagnolli L, Hilbert S, Müssigbrodt A, Husser D, Bollmann A, Hindricks G, Arya A. Epicardial adipose tissue thickness as an independent predictor of ventricular tachycardia recurrence following ablation. Heart Rhythm 2019;16:1492-8. [Crossref] [PubMed]
25. Sung KT, Kuo JY, Yun CH, Lin YH, Tsai JP, Lo CI, Hsiao CC, Lai YH, Tsai CT, Hou CJ, Su CH, Yeh HI, Chien CY, Hung TC, Hung CL. Association of Region-Specific Cardiac Adiposity With Dysglycemia and New-Onset Diabetes. J Am Heart Assoc 2021;10:e021921. [Crossref] [PubMed]
26. Tornvall P, Beltrame JF, Nickander J, Sörensson P, Reynolds HR, Agewall S. How to Use Cardiac Magnetic Resonance Imaging in Myocardial Infarction With Nonobstructive Coronary Arteries. Circ Cardiovasc Imaging 2024;17:e016463. [Crossref] [PubMed]
27. Ricci F, Khanji MY, Bisaccia G, Cipriani A, Di Cesare A, Ceriello L, Mantini C, Zimarino M, Fedorowski A, Gallina S, Petersen SE, Bucciarelli-Ducci C. Diagnostic and Prognostic Value of Stress Cardiovascular Magnetic Resonance Imaging in Patients With Known or Suspected Coronary Artery Disease: A Systematic Review and Meta-analysis. JAMA Cardiol 2023;8:662-73. [Crossref] [PubMed]
28. Toya T, Corban MT, Imamura K, Bois JP, Gulati R, Oh JK, Lerman LO, Lerman A. Coronary perivascular epicardial adipose tissue and major adverse cardiovascular events after ST segment-elevation myocardial infarction. Atherosclerosis 2020;302:27-35. [Crossref] [PubMed]
29. Guglielmo M, Penso M, Carerj ML, Giacari CM, Volpe A, Fusini L, et al. DEep LearnIng-based QuaNtification of epicardial adipose tissue predicts MACE in patients undergoing stress CMR. Atherosclerosis 2024;397:117549. [Crossref] [PubMed]
30. Koenen M, Hill MA, Cohen P, Sowers JR. Obesity, Adipose Tissue and Vascular Dysfunction. Circ Res 2021;128:951-68. [Crossref] [PubMed]
31. Cheng CK, Ding H, Jiang M, Yin H, Gollasch M, Huang Y. Perivascular adipose tissue: Fine-tuner of vascular redox status and inflammation. Redox Biol 2023;62:102683. [Crossref] [PubMed]
32. Nomura CH, Assuncao-Jr AN, Guimarães PO, Liberato G, Morais TC, Fahel MG, Giorgi MCP, Meneghetti JC, Parga JR, Dantas-Jr RN, Cerri GG. Association between perivascular inflammation and downstream myocardial perfusion in patients with suspected coronary artery disease. Eur Heart J Cardiovasc Imaging 2020;21:599-605. [Crossref] [PubMed]
33. Wang C, Wu X, Liu Y, Qiu X, Su H, Wang Z, Gao J. Association between coronary plaque vulnerability features and multiparametric pericoronary fat indices on coronary computed tomography angiography: a cross-sectional study. Quant Imaging Med Surg 2025;15:6897-909. [Crossref] [PubMed]
34. Gaibazzi N, Martini C, Botti A, Pinazzi A, Bottazzi B, Palumbo AA. Coronary Inflammation by Computed Tomography Pericoronary Fat Attenuation in MINOCA and Tako-Tsubo Syndrome. J Am Heart Assoc 2019;8:e013235. [Crossref] [PubMed]
35. Bergamaschi L, Foà A, Paolisso P, Renzulli M, Angeli F, Fabrizio M, et al. Prognostic Role of Early Cardiac Magnetic Resonance in Myocardial Infarction With Nonobstructive Coronary Arteries. JACC Cardiovasc Imaging 2024;17:149-61. [Crossref] [PubMed]
36. Taverna G, Canton L, Zilio L, Calabrese V, Cecere A, Savo MT, et al. A novel multimodal imaging approach for working diagnosis of acute myocardial infarction with non-obstructive coronary arteries: a promising diagnostic strategy. Front Cardiovasc Med 2025;12:1646418. [Crossref] [PubMed]
37. Cicek V, Bagci U. AI-powered contrast-free cardiovascular magnetic resonance imaging for myocardial infarction. Front Cardiovasc Med 2024;11:1457498. [Crossref] [PubMed]
38. Liang KW, Tsai IC, Lee WJ, Lee IT, Lee WL, Lin SY, Wan CJ, Fu CP, Ting CT, Sheu WH. MRI measured epicardial adipose tissue thickness at the right AV groove differentiates inflammatory status in obese men with metabolic syndrome. Obesity (Silver Spring) 2012;20:525-32. [Crossref] [PubMed]
39. Chen N, Xu Y, Xu C, Duan J, Zhou Y, Jin M, Xia H, Yuan W, Chen R. Effects of triglyceride glucose (TyG) and TyG-body mass index on sex-based differences in the early-onset heart failure of ST-elevation myocardial infarction. Nutr Metab Cardiovasc Dis 2024;34:590-7. [Crossref] [PubMed]
40. Aimo A, Chiappino S, Clemente A, Della Latta D, Martini N, Georgiopoulos G, Panichella G, Piagneri V, Storti S, Monteleone A, Passino C, Chiappino D, Emdin M, Gimelli A, Neglia D. The triglyceride/HDL cholesterol ratio and TyG index predict coronary atherosclerosis and outcome in the general population. Eur J Prev Cardiol 2022;29:e203-4. [Crossref] [PubMed]
41. Gao S, Ma W, Huang S, Lin X, Yu M. Impact of triglyceride-glucose index on long-term cardiovascular outcomes in patients with myocardial infarction with nonobstructive coronary arteries. Nutr Metab Cardiovasc Dis 2021;31:3184-92. [Crossref] [PubMed]
42. Yang R, Hu Y, Lee CH, Liu Y, Diaz-Canestro C, Fong CHY, Lin H, Cheng KKY, Pravelil AP, Song E, Lam KSL, Xu A. PM20D1 is a circulating biomarker closely associated with obesity, insulin resistance and metabolic syndrome. Eur J Endocrinol 2021;186:151-61. [Crossref] [PubMed]
43. Karakasis P, Stalikas N, Patoulias D, Pamporis K, Karagiannidis E, Sagris M, Stachteas P, Bougioukas KI, Anastasiou V, Daios S, Apostolidou-Kiouti F, Giannakoulas G, Vassilikos V, Fragakis N, Giannopoulos G. Prognostic value of stress hyperglycemia ratio in patients with acute myocardial infarction: A systematic review with Bayesian and frequentist meta-analysis. Trends Cardiovasc Med 2024;34:453-65. [Crossref] [PubMed]
44. Gao S, Huang S, Lin X, Xu L, Yu M. Prognostic implications of stress hyperglycemia ratio in patients with myocardial infarction with nonobstructive coronary arteries. Ann Med 2023;55:990-9. [Crossref] [PubMed]
45. Cicek V, Cikirikci EHK, Babaoğlu M, Erdem A, Tur Y, Mohamed MI, Cinar T, Savas H, Bagci U. Machine learning for prognostic prediction in coronary artery disease with SPECT data: a systematic review and meta-analysis. EJNMMI Res 2024;14:117. [Crossref] [PubMed]