Difference of pericoronary adipose tissue attenuation between culprit and non-culprit lesions in acute coronary syndrome: a systematic review and meta-analysis

Abstract

Background

Pericoronary adipose tissue (PCAT) attenuation, a non-invasive marker of coronary inflammation, predicts adverse cardiovascular events. However, the extent of localized inflammation at the culprit lesion site in acute coronary syndrome (ACS) has not been systematically quantified. We aimed to meta-analyze the difference in PCAT attenuation between culprit and non-culprit lesions in patients with ACS.

Methods

We systematically searched PubMed and Embase for studies comparing PCAT attenuation between culprit and non-culprit lesions/vessels in ACS patients. A random-effects model was used to calculate the pooled mean difference (MD). Pre-specified subgroup and meta-regression analyses were performed to explore heterogeneity.

Results

Twelve studies with 826 patients were included. Meta-analysis was performed in 11 studies comprising 12 cohorts (461 culprit/731 non-culprit lesions and 388 culprit/487 non-culprit vessels). PCAT attenuation was significantly higher around culprit sites compared to non-culprit sites (MD 4.95 HU, 95% CI: 2.62–7.27, p < 0.001), with substantial heterogeneity (I²=89%). Subgroup analysis revealed a significantly greater effect size when PCAT was measured at the perilesional level (MD 7.50 HU) versus the vessel level (MD 2.43 HU). Meta-regression identified the analysis lesion unit (p = 0.002) and ACS subtype (p = 0.003) as major sources of heterogeneity.

Conclusion

This meta-analysis provides consistent evidence that culprit lesions in ACS are characterized by a measurable and highly localized increase in pericoronary inflammation. A perilesional measurement approach may better capture this focal inflammatory signal, highlighting the potential of PCAT attenuation for identifying culprit lesion in ACS.

Graphical Abstract

Supplementary Information

The online version contains supplementary material available at 10.1186/s12872-026-05785-x.

Introduction

Acute coronary syndrome (ACS) remains a leading cause of morbidity and mortality worldwide, fundamentally driven by the rupture or erosion of vulnerable atherosclerotic plaques [1]. There is a growing consensus that inflammation is a pivotal mechanism in all stages of atherosclerosis, from plaque initiation to its destabilization and thrombotic complications [2]. Consequently, the non-invasive quantification of coronary artery inflammation has become a key objective in cardiovascular imaging to improve risk stratification and guide personalized therapies [3]. Pericoronary adipose tissue (PCAT) is a metabolically active fat depot encasing the coronary arteries [4]. As a special part of epicardial adipose tissue, it has emerged as a powerful imaging biomarker [5, 6]. Through paracrine signaling, inflamed plaques alter the composition of adjacent PCAT, a change that can be quantified on standard coronary computed tomography angiography (CCTA) as an increase in tissue attenuation [7].

Previous cohort studies have established that PCAT attenuation is a novel important predictor of future adverse cardiovascular events, providing prognostic value incremental to traditional risk factors and luminal stenosis [8, 9]. Subsequent multiple studies have further linked high PCAT attenuation to the presence of high-risk plaque features in patients with stable coronary disease [10–12]. These findings have solidified the role of PCAT attenuation as a biomarker of vessel inflammation. However, ACS is often caused by a single, highly inflamed culprit lesion within a coronary tree that may harbor multiple other plaques [13]. While several studies have observed higher PCAT attenuation around culprit lesions compared to non-culprit lesions in ACS patients, the overall magnitude of this effect and the factors contributing to it remain poorly defined [14–17]. A systematic synthesis of this evidence is currently lacking, leaving a critical knowledge gap regarding the local inflammatory heterogeneity within the coronary arteries at the time of an acute cardiovascular event.

Therefore, we conducted this systematic review and meta-analysis to quantitatively synthesize the existing evidence comparing PCAT attenuation between culprit and non-culprit lesions or lesion vessels in patients with ACS. We hypothesized that PCAT attenuation would be significantly higher at the site of the culprit lesion, reflecting a focal inflammatory hotspot. We further aimed to explore potential sources of heterogeneity through pre-specified subgroup and meta-regression analyses, specifically investigating the impact of the anatomical unit of measurement (perilesional level vs. lesion vessel level), ACS subtype, and PCAT quantification methodology. This study has been registered with PROSPERO (CRD 42024540021) and no amendments have been made.

Methods

This systematic review and meta-analysis was performed according to the Preferred Reporting Items for Systematic reviews and Meta-Analyses 2020 (PRISMA 2020) guidelines (Fig. 1) [18]. The PRISMA 2020 checklist is provided in the Supplementary Material PRISMA checklist.

Fig. 1.

Search strategy and source of included studies according to PRISMA. ACS, acute coronary syndrome; PCAT, pericoronary adipose tissue

Search strategy

From database inception to October 1, 2025, a systematic search was conducted on the PubMed and Embase databases to retrieve all human studies. Papers not published in English were also reviewed. The literature search strategy was designed by X.H and H.M. The search, carried out by X.H. and H.H., included the search strings (“peri-coronary adipose tissue” OR “pericoronary adipose tissue” OR “peri-coronary fat” OR “pericoronary fat” OR “perivascular fat” OR “peri-vascular adipose tissue” OR “perivascular adipose tissue” OR “fat attenuation index”).

Eligibility criteria and study selection

We have developed eligibility criteria for our meta-analysis based on the PICOS principle (Table 1). Studies were assessed as eligible if the population was patients with ACS, including STEMI, NSTEMI, or UA. Intervention and control are respectively culprit lesions or vessels and non-culprit lesions or vessels. Outcomes include mean PCAT attenuation values.

Table 1.

PICOS principles and corresponding eligibility criteria

	Differences in PCAT attenuation values between culprit and non-culprit lesions
P (population)	Patients with acute coronary syndrome
I (Intervention)	Culprit lesions or vessels
C (Comparison)	Non-culprit lesions or vessels
O (Outcomes)	Mean PCAT attenuation values
S (Study design)	Case-control studies or cross-sectional studies

PCAT pericoronary adipose tissue

The studies retrieved from the literature search were preliminarily screened by X.H and H.H independently based on titles and abstracts. In cases where there was disagreement between the authors regarding the eligibility assessment, a third author, H.M, was involved in determining the inclusion or exclusion of relevant studies. Only articles meeting the eligibility criteria underwent full-text screening and data extraction. Studies lacking sufficient data required for meta-analysis were excluded. Unpublished or duplicate studies were also excluded. Study designs included eligible cross-sectional studies and case-control studies.

Data extraction and endpoints

Two experienced assessors independently extracted relevant data from eligible studies, and any inconsistent data were assessed by a third assessor to reach a final consensus. Firstly, we extracted study characteristics, including author(s), year of publication, population demographics, sample size (number of patients and lesions or vessels), and study design. Secondly, we extracted detailed information regarding CT and PCAT, including the detailed CT parameters, PCAT attenuation measurement methods, traced PCAT coronary segments and PCAT attenuation analysis unit (lesions or lesion vessels).

It’s crucial to note that our meta-analysis solely focuses on PCAT attenuation. Data on PCAT thickness, ratio, volume and radiomic features are excluded. PCAT mean attenuation (PCATma) refers to the mean CT attenuation of adipose tissue within a specified coronary region measured in Hounsfield units (HU). The fat attenuation index (FAI) is a specialized metric derived from the attenuation distribution that accounts for technical parameters to improve the sensitivity of detecting vascular inflammation [8, 19].

Sample sizes, mean values and standard deviations (SD) of PCAT attenuation from different groups were extracted to compare effect sizes. For two studies (Lin A [20, 21] and Hoshino M [22, 23]) that only provided the median and interquartile range (IQR) for different groups, we utilized specific methods to convert them into estimated mean values and SD (Cochrane Handbook Chap. 6.5.2.9). Specifically, we applied the method described by Luo et al. to estimate the mean and the method by Wan et al. to estimate the SD [24, 25]. If a study analyzed multiple cohorts separately, each was treated as a separate study only if the patient populations were confirmed to be mutually exclusive with no overlap. If primary articles lacked necessary results, supplementary materials were searched for additional information extraction.

Statistical analysis

All statistical analyses were performed using the statistical programming environment Review Manager (RevMan) 5.4.1 and R 4.4.3. Statistical analyses were conducted separately for each endpoint using inverse variance weights in a random-effects model. We used forest plots to visually display the effect sizes and combined estimates with their 95% CI for each study.

Heterogeneity was assessed using the I² statistic, with values of 25%, 50%, and 75% considered to represent low, moderate, and high heterogeneity, respectively [26]. A p-value < 0.05 was considered significant. Given that most included studies used a within-patient design but did not report the correlation coefficients or the standard deviation of the paired differences, we analyzed the data using group means and standard deviations. This approach assumes independence and ignores potential positive within-patient correlations, resulting in conservative estimates with wider confidence intervals than if the paired structure were explicitly modeled.

To test the stability of our primary results, two sensitivity analyses were conducted by: (1) restricting the analysis to studies with a strict case-control design, (2) excluding studies where mean PCAT attenuation values were derived from data conversion, and (3) excluding studies with a prolonged time interval (≥ 6 months) between CTA and invasive coronary angiography (ICA).

To further explore sources of heterogeneity and assess the consistency of our findings, we conducted several additional analyses. A pre-specified subgroup analysis was performed based on the anatomical unit of PCAT measurement, stratifying studies into two groups: those measuring PCAT at the perilesional level (lesions) and those measuring it at the vessel level (lesion vessels).

Furthermore, a random-effects meta-regression analysis was conducted to evaluate the influence of key study-level characteristics on the overall effect size. The following categorical covariates were examined as potential moderators: the composition of the ACS population (AMI, NSTE-ACS or unspecified ACS), the PCAT measurement modality (FAI vs. mean attenuation), CT scanner technology (≤ 128 slices vs. >128 slices), and tube voltage (100 kVp vs. 120 kVp).

Quality and risk of bias assessment

All studies were assessed using the Newcastle-Ottawa Quality Assessment Scale (NOS) for case-control studies. We calculated quality scores for the three main components of included studies: selection of study groups (0–4 points), comparability of study groups (0–2 points), and determination of the outcome of interest (0–3 points). Higher scores indicate higher methodological quality. Quality assessment for all included studies was conducted by X.H and H.H, with any discrepancies resolved through consensus. Detailed information on the criteria for calculating points and how points are allocated to studies can be found in Supplementary Table S2. Additionally, funnel plots and Egger’s regression test were used to assess publication bias.

Results

Study selection and characteristics

A total of 2757 publications were screened and 132 studies were identified for full-text assessment. Following a meticulous screening process detailed in the PRISMA flow diagram (Fig. 1), 12 studies with 826 patients met the predefined inclusion criteria for the systematic review. Notably, while our search strategy included non-English publications, no non-English articles met the specific inclusion criteria; thus, all included studies were published in English. One study by Antonopoulos et al. was excluded from the quantitative synthesis as it did not report specific PCAT attenuation values [19]. Consequently, 11 studies, comprising 12 distinct cohorts, were included in the final meta-analysis.

The key characteristics of the included populations, study designs, and PCAT assessment methodologies are summarized in Table 2. Most studies focus on patients with ACS, NSTE-ACS, or AMI, without further analysis of ACS subtypes. Only the study by Dong X et al. conducted further subgroup analysis of AMI and UA in ACS patient subtypes [27]. Methodologically, culprit lesions were generally identified by combining angiographic findings with clinical evidence. CTA was consistently performed prior to ICA. The temporal interval between imaging modalities varied with the majority of studies acquiring data in the acute phase within hours to days of presentation. In contrast, two cohorts utilized baseline imaging acquired months prior to the event [14, 16]. For the anatomical analysis unit of PCAT measurement, six cohorts focused on the immediate pericoronary space around the lesion, encompassing 461 culprit and 731 non-culprit lesions. The remaining six cohorts assessed PCAT attenuation along the lesion vessel segment, including 388 culprit and 487 non-culprit lesion vessels. The specific CT parameters and PCAT-related definitions of studies included in the analysis are summarized in Supplementary Table S2.

Table 2.

Main characteristics of the studies included in the systematic review and meta-analysis

Study	Country	Study design	Population (n for patients)	Culprit lesion definition	Time interval (CTA to ICA)	PCAT attenuation type	PCAT attenuation analysis unit	Main findings
Antonopoulos AS et al. 2017[19]*	United Kingdom	Case- control study	AMI (n = 10)	NA	NA (CTA within 72 h of admission)	FAI	lesion	The FAI of culprit lesions were higher than that of non-culprit lesions
Balcer B et al. 2018[23]	Germany	Case- control study	AMI (n = 46)	ICA	median 7 days	PCAT mean attenuation	lesion vessel	No significant difference in PCAT attention between culprit and non-culprit lesion vessels (p = 0.77)
Goeller M et al. 2018[17]	Germany	Cross-sectional study	NSTE-ACS (n = 19)	ICA + ECG or Echocardiography	within 14 days	PCAT mean attenuation	lesion	The PCAT attention of culprit lesions were higher than that of non-culprit lesions (p = 0.01)
Lin A et al. 2020[20]**	Australia	case-control study	AMI (n = 60) (STEMI 5; NSTEMI 55)	ICA + ECG or Echocardiography	NA (CTA within 48 h of admission before CAG)	PCAT mean attenuation	lesion	The PCAT attention of culprit lesions were higher than that of non-culprit lesions (p = 0.01)
Nakajima A et al. 2022[28]	Japan	Case- control study	NSTE-ACS (n = 198; 107 for plaque rupture and 91 for plaque erosion) (NSTEMI 153; UA 45)	NA	median 4.0 h (IQR: 2.0–27.0 h)	PCAT mean attenuation	lesion vessel	The PCAT attention of culprit and non-culprit lesions were not directly compared
Kuneman JH et al. 2023[16]	The Netherlands	Case- control study	ACS (n = 66) (STEMI 6; NSTEMI 19; UA 41)	ICA	within 48 h	PCAT mean attenuation	lesion	The PCAT attention of culprit lesions were higher than that of non-culprit lesions (p < 0.001)
Huang M et al. 2023[14]	China	Case- control study	ACS (n = 30) (STEMI 6; NSTEMI 9; UA 15)	NA	NA (CTA within 24 h of admission before CAG)	FAI	lesion	The PCAT attention of culprit lesions were higher than that of non-culprit lesions (p < 0.001)
Dong X et al. 2023[27]	China	Cross-sectional study	ACS (n = 173) (AMI 23; UA 150)	ICA signs of plaque rupture	within 6 months	FAI	lesion	The PCAT attention of culprit and non-culprit lesions were not directly compared
Hoshino M et al. 2023[22]**	Japan	Cross-sectional study	NSTE-ACS (n = 69)	ICA + ECG or Echocardiography	within 2 years	PCAT mean attenuation	lesion vessel	The PCAT attention of culprit and non-culprit lesions were not directly compared
Qi L et al. 2023[29]***	China	Case- control study	AMI (n = 60)	ICA + ECG	NA (CTA within 1 h of presentation before CAG)	FAI	lesion vessel	The PCAT attention of culprit and non-culprit lesions were not directly compared
Li L et al. 2024[15]	China	Case- control study	ACS (n = 60)	ICA + ECG, and stent	within 1 month	FAI	lesion	The PCAT attention of culprit lesions were higher than that of non-culprit lesions (p < 0.001)
Ocana G et al. 2025[30]	Switzerland	Case- control study	ACS (n = 35)	Histopathology	whin 1 day (autopsies)	FAI	lesion vessel	The PCAT attention of culprit lesions were higher than that of non-culprit lesions (p = 0.003)

ACS acute coronary syndrome, AMI acute myocardial infarction, ICA invasive coronary angiography, CT computed tomography angiography, FAI fat attenuation index, LAD left anterior descending artery, LCX left circumflex artery, NA not mentioned, NSTEMI non-ST-segment elevation myocardial infarction, PCAT pericoronary adipose tissue, RCA right coronary artery, STEMI ST-segment elevation myocardial infarction, UA unstable angina

* This study was only included in the systematic review and not in the meta-analysis

** These studies reported the median and quartiles, and we converted them

*** This study respectively reported the values of the three main coronary arteries where the lesions were located, and we summarized and converted them

Differences in PCAT attenuation between culprit and non-culprit lesion or lesion vessels

In the original studies, seven reported significantly higher PCAT attenuation at culprit lesions or in culprit vessels compared to the non-culprit group, whereas one study found no significant difference. Four studies did not directly perform this comparison.

In a pooled meta-analysis using a random-effects model, PCAT attenuation was significantly higher around culprit lesions or vessels compared to non-culprit groups in patients with ACS. The mean MD was 4.95 HU (95% CI: 2.62–7.27, p < 0.001), although high heterogeneity was observed among studies (I² = 89%) (Fig. 2).

Fig. 2.

Forest plot of the difference in PCAT attenuation values between culprit and non-culprit lesion groups. IV, inverse variance; SD, standard deviation; CI, confidence interval; * Entries Nakajima A 2022a and Nakajima A 2022b represent different cohorts from the same studies conducted by Nakajima A. They were treated as independent studies for analysis purposes

Sensitivity analyses confirmed the consistency of this finding. The primary result remained consistent after restricting the analysis to studies with a strict case-control design, excluding studies where mean PCAT values were derived from data conversion or excluding studies with a prolonged time interval between CTA and angiography (Supplementary Figures S1 to S3).

We further performed a pre-specified subgroup analysis based on the lesion analysis unit of PCAT measurement. A significantly greater difference in PCAT attenuation was observed when measurements were confined to the perilesional area (MD 7.50 HU, 95% CI: 4.80–10.21; I² = 83%) compared to when they were averaged across the lesion vessel (MD 2.43 HU, 95% CI: 0.47–4.38; I² = 67%) (Fig. 3).

Fig. 3.

Forest plot for subgroup analysis based on lesion analysis unit (lesion or lesion vessels). IV, inverse variance; SD, standard deviation; CI, confidence interval; * Entries Nakajima A 2022a and Nakajima A 2022b represent different cohorts from the same studies conducted by Nakajima A. They were treated as independent studies for analysis purposes

Meta-regression analysis

To further explore the origins of the observed heterogeneity, we conducted an exploratory meta-regression analysis and 3 key moderators were identified. The results suggested potential moderators influencing the effect size. Firstly, the composition of the ACS population appeared to influence the effect size (QM(df = 2) = 11.73, p = 0.003), statistically accounting for explaining approximately 54.9% of the inter-study variance (R² = 54.9%). Studies enrolling unspecified ACS population reported a greater PCAT attenuation difference than those limited to either AMI or NSTE-ACS (Fig. 4A). Additionally, consistent with our subgroup findings, the unit of PCAT analysis (lesions vs. lesion vessels) was a significant moderator (QM(df = 1) = 9.47, p = 0.002), accounting for 50.9% of the heterogeneity (R² = 50.9%). A perilesional measurement method was associated with a larger effect size (Fig. 4B). Finally, the type of PCAT attenuation used showed a borderline significant effect (QM(df = 1) = 4.15, p = 0.042). Studies utilizing the FAI tended to report a larger magnitude of difference compared to those using PCAT mean attenuation (Fig. 4C).

Fig. 4.

Meta-regression of (a) Lesion analysis unit, (b) PCAT attenuation type, and (c) ACS subtype in the difference of PCAT attenuation values between culprit and non-culprit lesion groups. ACS, acute coronary syndrome; AMI, acute myocardial infarction; FAI, fat attenuation index; NSTE-ACS, non-ST-segment acute coronary syndrome; PCATa, pericoronary adipose tissue attenuation; PCATma, pericoronary adipose tissue mean attenuation.Meta-regression p value = 0.002 (a), p = 0.042 (b), p = 0.003 (c)

In contrast, no significant association was found between the PCAT attenuation difference and number of CT scan slices (≤ 128 slices vs. >128 slices; p = 0.629) or tube voltage (100 kVp vs. 120 kVp; p = 0.432).

Quality assessment and risk of bias

The methodological quality of the included studies assessed using the NOS scale was shown in Supplementary Table S1. All of the 12 studies demonstrated high quality, with scores ranging from 7 to 8. A point was consistently deducted for the selection criterion, as the control groups (non-culprit lesions/vessels) were derived from the same patients rather than a separate community-based population, representing an inherent design feature of such case-control studies.

Visual inspection of the funnel plot did not display obvious asymmetry (Supplementary Figure S4). While Egger’s regression test was not statistically significant (z = -0.75, p = 0.46), these results should be interpreted with caution. Given the limited number of included studies (n = 12), these tests are underpowered to detect subtle bias. Therefore, the possibility of selective reporting in this rapidly evolving imaging field cannot be definitively excluded.

Discussion

This systematic review and meta-analysis was the first to quantitatively synthesize evidence on intracoronary inflammatory heterogeneity in acute coronary syndrome. This study suggested that pericoronary adipose tissue attenuation is significantly elevated around culprit lesions relative to non-culprit lesions. This finding identifies a focal inflammatory hotspot at the site of plaque rupture. Notably, our subgroup and meta-regression analyses indicate that this inflammatory signal is highly localized. The magnitude of the difference appeared greater when PCAT was measured at the perilesional level compared with the vessel level. Furthermore, our exploratory analyses suggest that ACS subtype, anatomical unit of analysis, and PCAT measurement method may be factors influencing the observed effect size, offering refined insights into this important imaging biomarker.

PCAT attenuation as a specific marker of culprit lesion

The critical role of inflammation in the pathogenesis and clinical outcomes of cardiovascular disease is increasingly recognized [2]. PCAT attenuation has emerged as an important imaging biomarker reflecting coronary artery inflammation [4]. CRISP-CT study as a landmark study, has established that elevated RCA PCAT attenuation predicts future adverse cardiovascular events [8]. A recent meta-analysis further linked PCAT attenuation across RCA and LCX to MACE in patients with stable coronary artery disease [9]. Despite high heterogeneity in our analysis, the direction of the association remained consistent, with 9 of the 12 included cohorts showing higher PCAT attenuation in culprit lesions. Our study extends these findings by demonstrating that the culprit lesion itself is associated with localized pericoronary inflammation, with a pooled mean difference of 4.95 HU compared to non-culprit lesions.

However, the clinical interpretation of this magnitude requires careful consideration regarding measurement physics. While a difference of ~ 5 HU is relatively small compared to the full Hounsfield scale, it is highly consistent with differences reported in landmark prospective studies [17, 19]. Crucially, the validation study on inter-observer and intra-observer variability suggested that the measurement error for PCAT attenuation is typically within the range of ± 1.5 to 2.0 HU [31]. Therefore, the observed 4.95 HU difference in our meta-analysis is more than double the threshold of technical noise, supporting it as a genuine biological signal of localized inflammation [19]. Nevertheless, given the potential for measurement overlap and the influence of technical factors, we emphasize that PCAT attenuation should be interpreted as a quantitative marker of plaque instability risk rather than a standalone diagnostic tool for individual patients.

Biological mechanism of coronary artery perilesional inflammation

Culprit lesions are often associated with higher degrees of high-risk plaque and inflammation. The pronounced elevation of PCAT attenuation surrounding the culprit lesion is mechanistically plausible, reflecting bidirectional intense paracrine signaling between the high-risk plaque and its adjacent adipose tissue [12]. This process is likely driven by an inside-out signaling cascade, where unstable plaques rich in lipids and inflammatory cells release a plethora of pro-inflammatory cytokines, specifically tumor necrosis factor-α, interleukin-6, and interferon-γ [32]. These cytokines traverse the vessel wall, altering the phenotype of the adjacent PCAT by inhibiting adipocyte differentiation and stimulating local lipolysis [33]. This process, described as a localized cachexia-type response, results within the adipocytes shifting from a lipid-rich to a lipid-poor phenotype with smaller cell size [34]. From a radiomic perspective, this histological change alters the tissue lipid-to-water ratio. Since lipid has a significantly lower density (approximately − 100 HU) than the aqueous phase (approximately 0 HU), the depletion of intracellular lipid combined with interstitial edema translates to a higher tissue density and a less negative Hounsfield unit value on computed tomography [7]. Furthermore, this localized browning of fat creates a pro-inflammatory microenvironment that may, through an outside-in feedback loop, further destabilize the plaque and culminate in its rupture [35].

Localized inflammation and clinical implications for multivessel disease

An important finding from our subgroup analysis and meta-regression is that perilesional PCAT measurements yielded a nearly threefold greater effect size than lesion vessel assessments. This observation likely reflects a combination of biological focal inflammation and the physical principles of spatial sampling. This indicates that pericoronary inflammation in ACS is not only a diffuse, pan-coronary phenomenon but also a highly localized process, analogous to a “volcanic eruption” at the plaque site [35]. Methodologically, measuring PCAT at the vessel level inevitably includes healthy adipose tissue, which dilutes the focal inflammatory signal and introduces heterogeneity through spatial averaging. Conversely, perilesional sampling maximizes the signal-to-noise ratio by focusing specifically on the inflammatory hotspot. These findings align with the evolving concept of pan-coronary vulnerability which suggests that acute coronary syndrome involves widespread coronary inflammation [36]. While patients with acute coronary syndrome often exhibit a diffuse inflammatory background, our meta-analysis identifies a significant incremental increase in pericoronary adipose tissue attenuation specifically at the culprit site. This indicates that a focal hotspot of intense inflammation superimposed on a generalized inflammatory milieu may contribute to plaque instability [37]. Overall, our findings may reinforce the vulnerable plaque hypothesis, which posits that cardiovascular events are driven by a few high-risk lesions rather than generalized atherosclerosis [38]. Previous studies have also shown that PCAT attenuation around lesions was associated with functional coronary artery stenosis and had a higher predictive value for future cardiovascular events [39, 40]. Consequently, these results indicate that for ACS risk assessment, lesion-specific PCAT attenuation quantification rather than vascular mean values may be more valuable.

The clinical implications of these findings may be substantial for the management of multivessel coronary artery disease in the setting of acute coronary syndrome. Current European Society of Cardiology guidelines recommend complete revascularization for patients with multivessel disease [1]. However, identifying which non-culprit lesions pose a risk of future events remains challenging. Recent large-scale evidence underscores the benefits of complete revascularization but highlights the need for refined criteria to select lesions for intervention beyond physiological ischemia [41, 42]. Our findings suggest that localized PCAT attenuation could serve as a novel marker to discriminate high-risk vulnerable plaques from stable lesions. This distinction is particularly relevant when differentiating the acute inflammatory instability characterizing ACS from the stable atherosclerosis observed in chronic coronary syndromes [36]. Consequently, integrating PCAT assessment into the diagnostic workflow may refine risk stratification and guide revascularization strategies for non-culprit lesions. Furthermore, future developments in diagnosis and treatment may leverage this biomarker for precision medicine. A recent comprehensive overview by Imbesi et al. highlights the critical link between inflammation and clinical outcomes [36]. Beyond guiding revascularization, PCAT attenuation could serve as a non-invasive tool to select high-risk patients who would benefit most from emerging anti-inflammatory pharmacotherapies [43]. This approach aligns with the paradigm shift toward treating biological plaque vulnerability rather than solely anatomic stenosis.

Interpreting heterogeneity of ACS subtypes and PCAT attenuation measurements

The substantial heterogeneity observed in our analysis warrants careful interpretation. This high level of statistical variability was frequently observed in imaging meta-analyses and likely reflects a combination of methodological, clinical, and technical factors. A primary driver of this heterogeneity was likely the clinical diversity of the patient populations, which included individuals with varying degrees of plaque instability and distinct risk profiles. Furthermore, the relatively small sample sizes of the included cohorts inherently contributed to the observed statistical variance. Additionally, variations in baseline pharmacotherapy across studies likely contributed to the observed heterogeneity. Statin therapy is known to reduce pericoronary adipose tissue attenuation [21]. Although the within-patient design of the included studies controls for systemic medication effects, differences in medication protocols across cohorts may still influence the absolute magnitude of the recorded attenuation values. Our meta-regression analysis identified ACS subtype as a potential moderator. However, given the limited number of studies (n = 12) and the high statistical heterogeneity, these findings should be interpreted as hypothesis-generating rather than definitive. The classification used in our analysis (AMI, NSTE-ACS, unspecified ACS) was necessitated by the lack of granular reporting in primary studies, which precluded a mutually exclusive stratification into STEMI, NSTEMI, and UA. The effect size was most pronounced in cohorts with unspecified ACS, while it did not reach statistical significance in the limited number of studies focused solely on AMI. This could be attributed to the small number of AMI-specific studies and potential methodological outliers, such as the study by Balcer B et al. using a non-standard HU range for PCAT definition [23]. It is important to recognize that ACS encompasses a spectrum of pathologies, including STEMI, NSTEMI, and UA, each with distinct pathophysiological mechanisms [13]. For instance, plaque rupture, more common in STEMI, has been associated with higher PCAT attenuation compared to plaque erosion [28]. Therefore, future large-scale studies are warranted to meticulously investigate the differences in PCAT attenuation across various ACS subtypes and their respective diagnostic and prognostic utility.

The analysis also suggests potential methodological influences. The borderline significance favoring the FAI over mean attenuation (p = 0.042) should be interpreted with caution given the multiple comparisons performed. While this result aligns with the hypothesis that FAI may represent a more consistent index by adjusting for technical confounders [19], it remains preliminary and requires validation in larger studies. While we did not observe a significant impact of CT slice number or tube voltage on the difference, this does not negate their known influence on absolute attenuation values [44]. The lack of association in our analysis likely reflects that these technical factors affected both culprit and non-culprit measurements similarly within the same patient, thus minimizing their impact on the calculated difference. These findings collectively underscore the critical importance of standardizing PCAT measurement protocols to ensure reliable and reproducible results in future studies.

Consequently, to reduce inter-study variability and facilitate clinical translation, we urgently call for a standardized consensus on PCAT measurement protocols. Specifically, future investigations must harmonize: (1) HU thresholds for defining adipose tissue; (2) sampling strategies, prioritizing lesion-centered over vessel-based measurements given our findings; (3) CT acquisition parameters (tube voltage, reconstruction kernels) and metric harmonization (e.g., FAI vs. raw attenuation); and (4) reporting standards, including precise definitions of culprit lesions, CTA-to-event timing, and concurrent anti-inflammatory therapies.

Limitations

This study has several limitations. First, our study was restricted to major international databases and the exclusion of specific local-language databases may introduce selection bias. Second, as the primary outcome is a mean difference in PCAT attenuation, our findings demonstrate an association but cannot establish diagnostic accuracy or a definitive cut-off value for identifying culprit lesions. Third, we observed substantial statistical heterogeneity, likely reflecting methodological disparities across studies. Given this variability, calculating a precise prediction interval for future studies would be unreliable. Although we explored sources of heterogeneity through subgroup and meta-regression analyses, unmeasured technical and clinical variables, specifically variations in background chronic medical therapy, likely contributed to this residual heterogeneity. Fourth, the predominantly retrospective design and inclusion of multiple vessel segments in some control arms introduce potential bias and preclude causal inference, although the latter likely yields conservative statistical estimates. Furthermore, the estimation of mean values from medians in two studies may introduce bias given the potential skewness of biological data, although excluding these studies did not alter our primary findings. Additionally, while our sensitivity analysis confirmed the robustness of results after excluding studies with long CT-to-event intervals, a formal meta-regression to quantify the impact of the CTA-to-ICA time lag was precluded by the inconsistent reporting of time intervals across the included studies. Nevertheless, the dynamic nature of PCAT attenuation means that temporal misalignment in some cohorts remains a potential confounder. Finally, subgroup analyses for specific ACS pathologies were limited by the inconsistent reporting of ACS subtypes across studies, which prevented a rigorous stratification into mutually exclusive categories, underscoring the need for more detailed reporting in future studies.

Conclusion

In conclusion, this meta-analysis provides further evidence that culprit lesions in ACS exhibit a measurable and highly localized increase in pericoronary inflammation, quantified by PCAT attenuation. The magnitude of this inflammatory signature is influenced by both the anatomical site of PCAT attenuation measurement and the clinical subtypes of ACS. These findings support the pathophysiological role of focal inflammation and suggest the potential clinical application value of PCAT attenuation as a key biomarker for ACS personalized risk assessment and management.

Authors' contributions

The conceptualization of the study was led by X.H. and G.Y. The statistical methodology and analysis were conducted by X.H. and H.M., while H.H. and G.S. validated the statistical results. Formal analysis was conducted by G.Y., Z.B., and Z.Y. The investigation was carried out by X.H., H.M. and H.H. Data curation was handled by H.X., and H.H. The original draft preparation was undertaken by H.X., and the writing—review and editing was completed by Z.B., Z.Y. and G.Y. Visualization efforts were led by H.X. Supervision of the project was conducted by Z.Y. and G.Y., while project administration was managed by Z.B., Z.Y., and G.Y. Finally, G.Y. was responsible for funding acquisition.

Funding

This study is supported by the National Key Research and Development Program of China 2021YFA1000200 and 2021YFA1000204.

Data availability

All data and materials used in this study are freely available in electronic databases including PubMed and Embase databases, and specific data are sourced from references and Supplementary materials in this article.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.