Skip to main content
Download PDF

Prediction of abnormal bone mass with a pericoronary adipose tissue Attenuation model

BMC Geriatrics volume 25, Article number: 261 (2025) Cite this article

Abstract

Background

The aim is to explore the value of pericoronary adipose tissue (PCAT) attenuation in predicting abnormal bone mass by establishing a prediction model.

Materials and methods

361 patients with coronary computed tomography angiography (CCTA) and quantitative computed tomography (QCT) scans were retrospectively recruited. 311 patients from institution 1 from July 2021 to January 2023 were divided into a training cohort (n = 217) and an internal cohort (n = 94). The external cohort comprised 50 patients from institution 2 from January 2023 to August 2023. Clinical variables and PCAT attenuation of the major epicardial vessels were obtained. Univariate and multivariate logistic regression analyses were used to identify factors with statistical significance. Model 1 was constructed based on clinical variables. Model 2 was constructed by combining the clinical variables with the PCAT attenuation. The performances of the models were assessed using receiver operating characteristic curve analysis, calibration curves and decision curve analysis (DCA).

Results

Age, gender, coronary artery disease reporting and data system (CAD-RADS), statins and RCAPCAT were found to be significant predictors of abnormal bone mass. The area under the curve (AUC) of Model 2 was superior to that of Model 1 in the training cohort (AUC: 0.959 vs. 0.920), internal (AUC: 0.943 vs. 0.890) and external validation cohorts (AUC: 0.889 vs. 0.812). The calibration curves and DCA indicated that Model 2 had the higher clinical value.

Conclusion

The model incorporating clinical factors and RCAPCAT has good performance in predicting bone mass abnormalities.

Peer Review reports

Introduction

Osteoporosis is an age-related metabolic bone disease that leads to an increased risk of brittle fractures due to low bone mass and degradation of bone microstructure [1]. These fractures can lead to disability, depression, decreased quality of life, and increased mortality, which can bring huge economic burden to today’s society [2]. Based on the definition of osteoporosis by the World Health Organization, a meta-analysis conducted in 2021 on the epidemiological trends of osteoporosis over the past 20 years showed that the global prevalence of osteoporosis was about 18.3%, with females accounting for about 23.1% and males accounting for about 11.7% [3]. Osteopenia refers to a skeletal state where the bone mineral density (BMD) is abnormal but has not yet reached the definition of osteoporosis. Osteopenia is a precursor to osteoporosis, with a subtle onset [4]. It is also a critical period for timely prevention and treatment of osteoporosis and reducing the risk of fractures. Osteoporosis and osteopenia are collectively referred to as abnormal bone mass [5]. Most researches focus on the population with osteoporosis, and there is insufficient attention paid to osteopenia. Dual energy X-ray absorptiometry (DXA), the gold standard for diagnosing osteoporosis in clinical practice [6], is a two-dimensional imaging acquisition technique that cannot distinguish between cortical and cancellous bone. And factors such as bone hyperplasia, sclerosis, and abdominal aortic calcification can lead to an increased BMD. In the context of the burdens caused by decreased BMD, it is important to develop methods for identify patients at risk of abnormal bone mass so that appropriate actions can be taken.

There is evidence to suggest that the attenuation of pericoronary adipose tissue (PCAT) on coronary computed tomography angiography (CCTA) can serve as a new imaging biomarker, independent of other risk factors, that indicates the presence of active perivascular inflammation [7, 8]. Inflammation plays a key role in the development of coronary atherosclerosis and abnormal bone mass. PCAT can significantly increase due to structural changes caused by inflammation, such as fibrosis and vascular remodeling. IK Jang et al. found through analysis of 578 patients with coronary artery disease that early anti-inflammatory and lipid-lowering treatments may help stabilize plaques [9]. Meanwhile, inflammation also plays an important role in the progression of osteoporosis. Inflammation is an important process of the pathogenesis of bone loss [10]. Inflammatory mediators are activated, generated and released, and then activate osteoclasts, inhibit osteoblasts and cause bone loss finally [11]. On the other hand, adipose tissue is an important mediator of inflammation responses. In the presence of vascular inflammation, proinflammatory cytokines secreted by the inflammatory vascular wall inhibit the accumulation and differentiation of adipocyte lipids in PCAT. In addition, many studies have shown that excessive fat accumulation in the body has a negative impact on BMD [12, 13]. And domestic and foreign studies have shown that different parts of fat accumulation have different effects on bone mass [14]. Subcutaneous fat may play a protective role in osteoporosis [15]. However, visceral fat exhibits a completely different effect. A study found that an increase in visceral fat is closely related to an increased risk of osteoporotic fractures [13]. However, there is currently limited literature on the relationship between pericoronary adipose tissue and bone condition. Given the above background, we assumed that PCAT attenuation may play a potential role in capturing bone mass abnormalities.

Although numerous studies have investigated clinical risk factors or imaging indicators for osteoporosis, few reports have combined clinical data with the PCAT attenuation value for predicting abnormal bone loss. Therefore, the purpose of this study was to explore the predictive value of integrating clinical data and PCAT attenuation captured by routinely applied CCTA on abnormal bone mass and to develop a prediction model accordingly. This will provide a simple and effective method for abnormal bone mass opportunistic screening in patients who have undergone CCTA examination.

Materials and methods

Study participants and clinical data collection

This study retrospectively collected data obtained from patients with clinically indicated CCTA examinations and chest CT-quantitative computed tomography (QCT) scans performed within 14 days of each other at institution 1 from July 2021 to January 2023 and at institution 2 from January 2023 to August 2023. These patients who seek medical attention for physical examination or chest pain not only underwent CCTA examination, but also underwent routine chest CT screening during hospitalization. This retrospective study was approved by the institutional review board of our institution (ethical approval number: 2023-55), who waived the need for informed consent. The inclusion criteria were as follows: (1) interval between CCTA and chest CT-QCT of less than 14 days; (2) age ≥ 30 years and (3) the CT-QCT scanning range included the T12 or L1 vertebrae. The exclusion criteria were as follows: (1) congenital coronary artery malformation, myocardial infarction, or history of cardiac surgery; (2) other diseases, such as chronic inflammatory diseases, calcium or phosphorus metabolism disorders, or tumors affecting bone metabolism; (3) incomplete clinical data; (4) the CT-QCT scanning range does not include the T12 or L1 vertebrae; and (5) poor CCTA image quality. The study design and patient selection are depicted in Fig. 1.

Fig. 1
figure 1

A flowchart of patient recruitment and study design. CCTA, coronary computed tomography angiography; QCT, quantitative computed tomography

Full size image

All clinical variables were obtained from inpatient medical records by two radiologists. These variables included age, gender, body mass index (BMI), smoking, drinking, hypertension, Coronary Artery Disease Reporting and Data System (CAD-RADS), diabetes, microvascular complications, statins, diuretics, β-blockers, calcium channel blockers (CCB), angiotensin converting enzyme inhibitors (ACEI), insulins, insulin secreting agent, biguanides, α-glucosidase-inhibitors, sodium-dependent glucose transporters 2 (SGLT-2) inhibitors, dyslipidemia, glucose (GLU), blood calcium (Ca), and blood phosphorus (P).

The Coronary Artery Disease Reporting and Data System (CAD-RADS) was divided into six categories [16]: CAD-RADS 0 (0% stenosis: absence of CAD), CAD-RADS 1 (1–24% stenosis: minimal non-obstructive CAD), CAD-RADS 2 (25–49% stenosis: mild non-obstructive CAD), CAD-RADS 3 (50–69% stenosis: moderate stenosis), CAD-RADS 4 (70–99% stenosis or > 50% stenosis in left main coronary artery or ≥ 70% stenosis in 3 vessel: severe stenosis), and CAD-RADS 5 (100% stenosis: total occlusion). In our study, CAD-RADS categories 0–2 were combined and considered as a non-obstructive group. CAD-RADS categories 3–5 were combined and considered as an obstructive group.

QCT scanning and measurement

All patients underwent imaging of the T12-L1 vertebral bodies through chest CT-QCT scans. For the training cohort and internal validation cohort (institution 1), CT examinations were performed with a 256-slice wide detector CT scanner (Revolution HD, GE Healthcare). The scanning parameters were as follows: tube voltage 140 kV, tube current 400 mA, slice thickness and interval 5 mm, thin layer 1.25 mm, and matrix size 512 × 512. For the external validation cohort (institution 2), CT examinations were performed with a different 256-slice wide detector CT scanner (Force, SOMATOM). The scanning parameters were as follows: tube voltage 120 kV, tube current 129 mA, layer thickness and interval 5 mm, thin layer 1.25 mm, and matrix size 512 × 512. The QCT phantom produced by Mindways Company (United States) was used for routine calibration.

The thin-slice CT images were transmitted to the QCT PRO workstation, where the T12-L1 vertebral BMD was measured. Regions of interest (ROIs) were placed in the trabecular bones of the T12 and L1 vertebral bodies while avoiding cortical bone, bone islands and posterior venous plexus; the final sizes of the ROIs were approximately 100 mm². And then built-in software in the workstation was used to automatically calculate the values. Finally, the BMD values of the T12 and L1 vertebral bodies were obtained. The mean values from the T12 and L1 ROIs were used for diagnostic grouping. A junior radiologist completed the QCT measurements under the supervision of a senior radiologist.

According to the thresholds for the diagnosis of osteoporosis using lumbar spine QCT proposed by the International Society for Clinical Densitometry (ISCD) [5] and the American College of Radiology (ACR) [17], the patients were divided into the normal bone mass (> 120 mg/cm3) and abnormal bone mass (including osteopenia and osteoporosis) (≤ 120 mg/cm3) groups.

CCTA acquisition

Patients in the training cohort and internal validation cohort were scanned with a 256-slice wide detector CT scanner (Revolution HD, GE Healthcare). The parameters were as follows: rotation time = 280 ms, tube voltage = 100 kVp, tube current = smart 400–700 mA, reconstructed slice thickness = 0.625 mm, and reconstructed slice interval = 0.5 mm.

Patients in the external validation cohort were scanned with a 256-slice wide detector CT scanner (Force, SOMATOM). The parameters were as follows: rotation time = 250 ms, smart tube voltage, smart tube current, reconstructed slice thickness = 0.75 mm, and reconstructed slice interval = 0.6 mm.

The heart rate was controlled within 80 beats/min during the examination. The Coronary Agaston Calcium Score (CACS) was first acquired to assess the total calcification burden. Contrast medium (50 mL to 70 mL; Iomeprol, Iomeron, 400 mg iodine/ml, Bracco) was administered into the antecubital vein in a bolus injection at a flow rate of 4–5 mL/s, followed by an injection of 50 mL of physiological saline at the same flow rate. The scanning system was triggered by a contrast tracking technique for an ROI placed in the ascending aorta.

PCAT Attenuation analysis

The thin-slice CCTA images were imported in Digital Imaging and Communications in Medicine (DICOM) format to the Research Portal V1.1 workstation (United Imaging Intelligence, Co., Ltd.). We selected the proximal three major epicardial coronary arteries: the right coronary artery (RCA), left anterior descending artery (LAD), and left circumflex artery (LCX). The software can automatically define the proximal and distal points of the measurement area (proximal 10–50 mm of the RCA, proximal 40 mm of the LAD and LCX) and measure the respective CT attenuation of adipose tissue within a radial distance from the outer vessel wall equal to the diameter of the vessel based on voxels with an attenuation range from − 190 HU to − 30 HU [18] (Fig. 2). Finally, the attenuation of adipose tissue around RCA, LAD, and LCX was recorded as RCAPCAT, LADPCAT, and LCXPCAT, respectively. Two radiologists with 3 years of experience in cardiovascular imaging who were blinded to the clinical data independently analyzed all PCAT attenuation values.

Fig. 2
figure 2

CCTA image of PCAT attenuation values measurement. (A) RCA, (B) LAD, and (C) LCX. CCTA, coronary computed tomography angiography; PCAT, peri-coronary adipose tissue; LAD, left anterior descending; LCX, left circumflex; RCA, right coronary angiography

Full size image

Statistical analysis

All statistical analyses in the present study were performed with R software (version 3.5.0, http://www.rproject.org). A p value < 0.05 was considered to indicate statistical significance. The Kolmogorov‒Smirnov test was used to assess the normality of the data distribution, and the Levene test was used to assess homogeneity of variance. Quantitative variables are presented as the mean ± standard deviation (SD) or median (interquartile range, IQR). Categorical variables are presented as percentages. The two-sample t test was used to compare normally distributed variables, and the Mann‒Whitney U test was used to compare nonnormally distributed variables between two groups. The chi-square test was used to compare categorical variables.

Clinical data and PCAT attenuation variables with statistical significance were selected by univariate and multivariate logistic regression. Univariate analyses were firstly conducted to determine which of the variables was significantly different between the normal and abnormal bone mass groups for later inclusion in the multivariate logistic regression. Finally, Model 1 was constructed based on patient clinical data, and Model 2 was constructed by combining clinical data and PCAT attenuation. A nomogram was developed based on Model 2. Model performance was evaluated and compared using all the training cohorts and the internal and external validation cohorts. Receiver operating characteristic (ROC) curve analysis was used to evaluate the performance of the models in distinguishing patients with normal bone mass from those with abnormal bone mass, and the corresponding sensitivity, specificity, and accuracy were also calculated. The Delong test was conducted to compare the areas under the curves (AUCs) between the two models. Calibration curves were plotted to assess whether the predictive probability of the models was in accordance with the actual probability. Decision curve analysis (DCA) was used to assess the clinical practicability of the models by quantifying their net benefit.

Results

Patient characteristics

A total of 800 patients were initially reviewed. Among them, 439 patients were excluded for various reasons (see patient inclusion and exclusion flowchart in Fig. 1). Finally, 361 patients (214 males, 147 females; mean age, 59.89 ± 11.08 years), including 168 with abnormal bone mass and 193 with normal bone mass, were included for further analysis. The patients who underwent scans at institution 1 between July 2021 and January 2023 were randomly assigned at a 7:3 ratio to either the training cohort (217 patients; mean age, 60.05 ± 11.42 years; 132 males) or the internal validation cohort (94 patients; mean age, 59.53 ± 11.25 years; 59 males). The patients who underwent scans at institution 2 between January 2023 and August 2023 were included as the external validation cohort (50 patients; mean age, 59.90 ± 9.32 years; 23 males). Compared with the normal bone mass group, the participants in the abnormal bone mass group were older, more female, had higher CAD-RADS scores and higher PCAT values. The patients’ characteristics are summarized in Tables 1 and 2.

Table 1 Patients’ clinical characteristics
Full size table
Table 2 Characteristics of patients in the abnormal bone mass and normal bone mass groups
Full size table

Univariate and multivariate logistic regression analysis

Clinical data and PCAT attenuation variables with statistical significance were selected by univariate and multivariate logistic regression. Among all the variables included in the univariate analysis, age, gender, BMI, CAD-RADS, statins, CCB, insulin secreting agent, α-glucosidase-inhibitors, RCAPCAT, LADPCAT and LCXPCAT were all significant predictors (p < 0.05) of abnormal bone mass in the training cohort and were therefore subjected to multivariate logistic regression analysis. Of these variables, age, gender, CAD-RADS, statins and RCAPCAT were identified as independent risk factors for abnormal bone mass. The details of the logistic regression analysis are presented in Table 3.

Table 3 Univariate and multivariate logistic regression analysis
Full size table

Establishment and performance evaluation of prediction models

Next, the variables identified as independent predictors of abnormal bone mass were incorporated to create two predictive models. Model 1 (clinical data) included the clinical variables age, gender, CAD-RADS and statins, whereas Model 2 (clinical data and PCAT attenuation) included the above factors plus RCAPCAT. A nomogram was developed based on Model 2 to obtain individual score values based on the patient’s various indicators (Fig. 3).

Fig. 3
figure 3

A nomogram of clinical-PCAT joint model. This nomogram provides a method to calculate the probability of abnormal bone mass, on the basis of a patient’s combination of covariates. CAD-RADS, Coronary Artery Disease - Reporting and Data System; RCA, right coronary angiography; PCAT, peri-coronary adipose tissue

Full size image

The performance of Model 1 and Model 2 can be evaluated through ROC curves. According to the ROC curve analysis (Fig. 4), the AUC of Model 2 in the training cohort was 0.959 (95% CI, 0.923–0.977), which was significantly higher than that of Model 1 (0.920) (95% CI, 0.884–0.956; p = 0.010). Similarly, compared with Model 1, Model 2 had superior performance in the internal and external validation cohorts (AUC = 0.943 (95% CI, 0.906–0.990) vs. 0.890 (95% CI, 0.826–0.954), p = 0.007; 0.889 (95% CI, 0.800-0.979) vs. 0.812 (95% CI, 0.694–0.930), p = 0.042). In terms of the performance metrics, Model 2 also performed well in the internal and external validation cohorts in predicting abnormal bone mass, with the accuracy, sensitivity and specificity were 0.820, 0.855, 0.923 and 0.883, 0.875, 0.808, respectively (Table 4).

Fig. 4
figure 4

The ROC curves of model 1 and model 2 in the training cohort (A), internal validation cohort (B), and external validation cohort (C). AUC, area under the curve

Full size image
Table 4 Diagnostic efficacy of model 1 and model 2 in the three cohorts
Full size table

We further assessed the performance of the combined clinical and PCAT attenuation model (Model 2) with calibration and decision curve analyses. The calibration curves for Model 2 all showed that the predicted probability had good fitness with the actual probability in both the training cohort and the internal and external validation cohorts (Fig. 5). The Hosmer‒Lemeshow test confirmed this finding and further demonstrated that the differences between the predicted and actual probabilities of abnormal body mass were not significantly different for any of the three cohorts (p > 0.05). DCA demonstrated that Model 2 had a higher net benefit over almost the entire threshold probability interval than Model 1 in the three cohorts (Fig. 6).

Fig. 5
figure 5

Calibration curves of model 2 in the training cohort (A), internal validation cohort (B), and external validation cohort (C). A closer fitness to the diagonal line indicated an ideal model

Full size image
Fig. 6
figure 6

Decision curve analysis (DCA) of model 1 and model 2 in the training cohort (A), internal validation cohort (B), and external validation cohort (C). Blue and red lines represent the net benefit of model 1 and model 2, respectively. The gray line represents the assumption that all patients have abnormal bone mass, while the black line represents the assumption that no patients have abnormal bone mass

Full size image

Discussion

Previous studies have shown that osteoporosis is related to atherosclerosis, vascular calcification, inflammation and dyslipidemia [19,20,21,22], but there are few literatures on whether pericoronal fat is closely related to bone condition. In the present study, we constructed and validated machine learning models to predict abnormal bone mass by incorporating clinical data and the CCTA-based PCAT attenuation surrounding the proximal segments of three major epicardial coronary vessels. Discrimination, calibration and performance analyses showed that the combined model had superior performance to the clinical-only model in discriminating abnormal bone mass from normal bone mass. In addition, the similar performance of the combined model in the internal and external validation cohorts indicated its generalizability and reliability. In other words, the results demonstrated that incorporating PCAT attenuation enhanced the predictive efficacy of the clinical model and provided satisfactory predictive power. This model can serve as an evaluation tool for screening opportunistic bone mass abnormalities in individuals undergoing CCTA examination.

In our study, the addition of RCAPCAT may provide additional biological information to the model, which may be directly related to the occurrence of bone mass abnormalities. In addition, there may be interactions between clinical factors and RCAPCAT, and the predictability of model 2 may also be affected by the interaction of these variables. Therefore, the performance of model 2, which combined clinical factors and RCAPCAT, was superior to model 1. This study found that RCAPCAT, as an imaging biomarker, can improve the performance of the model in predicting abnormal bone mass. Inflammation can disrupt bone metabolism homeostasis, exacerbate bone resorption by osteoclasts, inhibit bone formation by osteoblasts, and lead to bone loss [10, 23]. Inflammation also plays a key role in the formation of atherosclerosis via contributions from proinflammatory cytokines [24]. An increase in the degree of PCAT attenuation, which can be captured by CCTA, can reflect the decrease in lipid content of adipocytes caused by vascular inflammation [7, 18]. This indicated that pericoronal fat can reflect bone mass status to some extent. Notably, RCAPCAT, rather than LADPCAT or LCXPCAT, was the only imaging parameter that significantly predicted abnormal bone mass. The reason for this result may be that RCA is the dominant coronary artery in most people, with measurement stability, and is less affected by surrounding nonfatty structures. This is consistent with previous studies, which have reported that the measurement of RCAPCAT is more reproducible and robust than that of LADPCAT and LCXPCAT [18, 25]. Therefore, this study suggested that RCAPCAT may serve as an independent predictor and be the preferred variable for clinical application to reduce workload. Moreover, the segmentation and measurement procedures in PCAT are fully automated, which greatly improves the simplicity and practicality of acquiring the attenuation value. Therefore, RCAPCAT were included when building the model, which proved valuable for enhancing the model’s prediction efficacy of abnormal bone mass in this study.

This study incorporated CAD-RADS into the models and obtained good results. Guan et al. found that low BMD is an independent correlate of global coronary atherosclerotic plaque burden [26]. Akin et al. demonstrated that a noncalcified/mixed plaque burden is an independent predictor of osteopenia or osteoporosis [27]. Among the clinical variables included in the model was CAD-RADS category, which significantly differed between the normal and abnormal bone mass groups when treated as a binary, categorical variable. The incidence of abnormal bone mass was significantly different in the obstructive (53%, 110/208) and nonobstructive groups (38%, 58/153) (p = 0.05). Our results are consistent with previous published work, which reported that patients with coronary artery stenosis > 50% have a significantly higher incidence of abnormal bone mass [28].

The statin administration history of the participants was collected in this study and participated in the construction of the model. Hyperlipidemia is not only a pathogenic factor of atherosclerosis, but also can lead to bone loss [29]. Researches have confirmed that hyperlipidemia can increase the expression of various inflammatory factors, stimulate inflammation in bone tissue, and promote osteoclast differentiation. Furthermore, lipid aggregation can induce the generation of oxidized lipids in bone tissue and inhibit osteoblast differentiation [30]. Based on the current research status of the correlation between hyperlipidemia and osteoporosis, statins have also been applied in the research of osteoporosis treatment drugs. Statins is capable of stabilizing plaque and preventing atherosclerosis progression [31]. An integrated epidemiological observation analysis and Mendelian randomization study demonstrated that statins have a protective effect on bone by increasing BMD through reducing low density lipoprotein cholesterol [32]. In our study, statins were identified as having a significantly protective effect against abnormal bone mass and thus was included in the model.

In addition to the factors described above, aging and female were both identified in our study as predictive risk factors for abnormal bone mass. It is widely known that the incidences of abnormal bone mass and coronary atherosclerosis increase as the value of aging increases [33]. In this study, the risk of abnormal bone mass increased by 0.16 times for one year of age increase. Additionally, the presence of female is more likely to be associated with the development of abnormal bone mass. This finding is consistent with a previous study showing that the incidence of abnormal bone mass increases in patients with postmenopausal female [34]. Estrogen participates in regulating bone metabolism processes, reducing bone turnover and loss, and promoting bone formation by inhibiting osteoclasts and promoting osteoblasts. Postmenopausal women experience a decrease in estrogen levels, with bone resorption exceeding bone formation, leading to bone loss and ultimately developing osteoporosis [35]. Estrogen deficiency can also increase the production of inflammatory cytokines, which not only inhibit the differentiation and activity of osteoblasts, but also stimulate the activity of osteoclasts and reduce their apoptosis [36].

When we turned our attention to other common factors, such as smoking, drinking, hypertension, diabetes and dyslipidemia [37, 38], none of these variables were involved in model development, as they were not statistically significant in the univariate logistic regression analysis. At present, the impact of smoking on BMD is still controversial. A Norwegian cohort study followed 34,856 adults for 3 years and found that smokers had an increased risk of hip fractures compared to non-smokers. Interestingly, the impact of smoking on fracture risk appeared to be largely unrelated to BMD, as the risk ratios (RR) associated with hip fractures after adjusting for age were similar to those associated with various risk factors, regardless of whether BMD was adjusted or not [39]. Therefore, although smoking is the main lifestyle risk factor for osteoporosis, the potential mechanisms of smoking related bone loss are still unclear [40]. In this study, more work is needed in the future to confirm the reversibility of smoking related bone conditions. A previous Korean study that included 2657 men and 2080 women found that the group who drank low alcohol had the highest whole-body BMD [41]. Compared to moderate to high drinkers, low drinkers have higher BMD, indicating a positive effect of low alcohol consumption on BMD [42]. However, many scholars believe that long-term excessive alcohol consumption is a high-risk factor for osteoporosis, leading to the occurrence of alcoholic osteoporosis [43]. Excessive alcohol consumption leads to metabolic imbalance between bone growth and resorption, disruption of bone microstructure, and ultimately increases the risk of fractures in patients. In this study, the reason for this result may be due to the lack of classification of alcohol consumption.

Similarly, patients with dyslipidemia commonly present with abnormal bone mass and atherosclerosis. In osteoporosis, as in atherosclerosis, lipids are deposited under the intima of the bone vascular system. The accumulation of lipids in the bone marrow may place external pressure on microcirculation, leading to cortical hypoxia and ischemia [29]. However, dyslipidemia was not incorporated in the model because there are too many factors affecting blood lipids. But we acknowledge the importance of blood lipids in bone health and will explore the role of dyslipidemia in this context in future research.

There are still several limitations of this study. First, because of the retrospective nature of the study, some clinical variables were not sufficiently comprehensive. For example, the intake of diet and dietary supplements is an important factor in bone health. These covariates were not included in this study. We will expand our dataset to include more comprehensive dietary information. Future research can ensure that all relevant clinical variables are thoroughly captured through prospective design studies, and further explore the relationship between PCAT and bone health. Furthermore, the sample size was relatively small. A larger sample size would allow for a more robust analysis. Future research should aim to replicate our study with a larger and more diverse sample size.

Conclusions

After incorporation of clinical factors and RCAPCAT, the model proposed in this study has good performance in predicting bone mass abnormalities. This study not only provides a new guidance basis for understanding the relationship between pericoronal adipose tissue and bone health, but also helps to construct a more accurate bone health evaluation system for patients in clinical practice.

Data availability

The data-set generated and analyzed during the current study is available from the corresponding author on a reasonable request.

Abbreviations

BMD:

bone mineral density

PCAT:

peri-coronary adipose tissue

CCTA:

coronary computed tomography angiography

QCT:

quantitative computed tomography

BMI:

body mass index

CAD-RADS:

Coronary Artery Disease Reporting and Data System

ROI:

Regions of interest

RCA:

right coronary artery

LAD:

left anterior descending artery

LCX:

left circumflex artery

ROC:

receiver operating characteristic

AUC:

area under the curve

DCA:

decision curve analysis

References

  1. Aibar-Almazán A, Voltes-Martínez A, Castellote-Caballero Y, Afanador-Restrepo DF, Carcelén-Fraile MDC, López-Ruiz E. Current status of the diagnosis and management of osteoporosis. Int J Mol Sci. 2022;23(16).

  2. Johnston CB, Dagar M. Osteoporosis in older adults. Med Clin North Am. 2020;104(5):873–84.

    Article  PubMed  Google Scholar 

  3. Salari N, Darvishi N, Bartina Y, et al. Global prevalence of osteoporosis among the world older adults: a comprehensive systematic review and meta-analysis. J Orthop Surg Res. 2021;16(1):669.

    Article  PubMed  PubMed Central  Google Scholar 

  4. Karaguzel G, Holick MF. Diagnosis and treatment of osteopenia. Rev Endocr Metab Disord. 2010;11(4):237–51.

    Article  CAS  PubMed  Google Scholar 

  5. Engelke K, Adams JE, Armbrecht G, et al. Clinical use of quantitative computed tomography and peripheral quantitative computed tomography in the management of osteoporosis in adults: the 2007 ISCD official positions. J Clin Densitom. 2008;11(1):123–62.

    Article  PubMed  Google Scholar 

  6. Dimai HP. Use of dual-energy X-ray absorptiometry (DXA) for diagnosis and fracture risk assessment; WHO-criteria, T- and Z-score, and reference databases. Bone. 2017;104:39–43.

    Article  PubMed  Google Scholar 

  7. Lin A, Dey D, Wong DTL, Nerlekar N. Perivascular adipose tissue and coronary atherosclerosis: from biology to imaging phenotyping. Curr Atheroscler Rep. 2019;21(12):47.

    Article  PubMed  PubMed Central  Google Scholar 

  8. Goeller M, Achenbach S, Duncker H, Dey D, Marwan M. Imaging of the pericoronary adipose tissue (PCAT) using cardiac computed tomography: modern clinical implications. J Thorac Imaging. 2021;36(3):149–61.

    Article  PubMed  Google Scholar 

  9. Fujimoto D, Kinoshita D, Suzuki K, et al. Relationship between calcified plaque burden, vascular inflammation, and plaque vulnerability in patients with coronary atherosclerosis. JACC Cardiovasc Imaging. 2024;17(10):1214–24.

    Article  PubMed  Google Scholar 

  10. Agca R, Heslinga SC, Rollefstad S, et al. EULAR recommendations for cardiovascular disease risk management in patients with rheumatoid arthritis and other forms of inflammatory joint disorders: 2015/2016 update. Ann Rheum Dis. 2017;76(1):17–28.

    Article  CAS  PubMed  Google Scholar 

  11. Ko Y-J, Wu J-B, Ho H-Y, Lin W-C. Antiosteoporotic activity of Davallia formosana. J Ethnopharmacol. 2012;139(2):558–65.

    Article  CAS  PubMed  Google Scholar 

  12. Lin Y, Wang X, Wu R, Zhou J, Feng F. Association between segmental body composition and bone mineral density in US adults: results from the NHANES (2011–2018). BMC Endocr Disord. 2023;23(1):246.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  13. Li L, Zhong H, Shao Y, Zhou X, Hua Y, Chen M. Association between lean body mass to visceral fat mass ratio and bone mineral density in united States population: a cross-sectional study. Arch Public Health. 2023;81(1):180.

    Article  PubMed  PubMed Central  Google Scholar 

  14. Rao VN, Bush CG, Mongraw-Chaffin M, et al. Regional adiposity and risk of heart failure and mortality: the Jackson heart study. J Am Heart Assoc. 2021;10(14):e020920.

    Article  PubMed  PubMed Central  Google Scholar 

  15. Malkov S, Cawthon PM, Peters KW, et al. Hip fractures risk in older men and women associated with DXA-Derived measures of thigh subcutaneous fat thickness, Cross-Sectional muscle area, and muscle density. J Bone Min Res. 2015;30(8):1414–21.

    Article  CAS  Google Scholar 

  16. Cury RC, Leipsic J, Abbara S, et al. Tomography (SCCT), the American College of Cardiology (ACC), the American College of Radiology (ACR), and the North America Society of Cardiovascular Imaging (NASCI). J Cardiovasc Comput Tomogr. 2022;16(6):536–57. CAD-RADS™ 2.0–2022 Coronary Artery Disease-Reporting and Data System: An Expert Consensus Document of the Society of Cardiovascular Computed.

  17. Yu JS, Krishna NG, Fox MG, et al. ACR appropriateness Criteria® osteoporosis and bone mineral density: 2022 update. J Am Coll Radiol. 2022;19(11S):S417–32.

    Article  PubMed  Google Scholar 

  18. Oikonomou EK, Marwan M, Desai MY, et al. Non-invasive detection of coronary inflammation using computed tomography and prediction of residual cardiovascular risk (the CRISP CT study): a post-hoc analysis of prospective outcome data. Lancet. 2018;392(10151):929–39.

    Article  PubMed  PubMed Central  Google Scholar 

  19. Shaya GE, Leucker TM, Jones SR, Martin SS, Toth PP. Coronary heart disease risk: Low-density lipoprotein and beyond. Trends Cardiovasc Med. 2022;32(4):181–94.

    Article  CAS  PubMed  Google Scholar 

  20. Warburton DER, Nicol CW, Gatto SN, Bredin SSD. Cardiovascular disease and osteoporosis: balancing risk management. Vasc Health Risk Manag. 2007;3(5):673–89.

    PubMed  PubMed Central  Google Scholar 

  21. Khandkar C, Vaidya K, Karimi Galougahi K, Patel S. Low bone mineral density and coronary artery disease: A systematic review and meta-analysis. Int J Cardiol Heart Vasc. 2021;37:100891.

    PubMed  PubMed Central  Google Scholar 

  22. Lee SN, Cho JY, Eun YM, Song SW, Moon KW. Associations between osteoporosis and coronary artery disease in postmenopausal women. Climacteric. 2016;19(5):458–62.

    Article  CAS  PubMed  Google Scholar 

  23. Wang S-Y, Jiang J-H, Liu S-Y, et al. Interleukin 6 promotes BMP9-induced osteoblastic differentiation through Stat3/mTORC1 in mouse embryonic fibroblasts. Aging. 2023;15(3):718–33.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  24. Soehnlein O, Libby P. Targeting inflammation in atherosclerosis - from experimental insights to the clinic. Nat Rev Drug Discov. 2021;20(8):589–610.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  25. Syu D-K, Hsu S-H, Yeh P-C, et al. The association between coronary artery disease and osteoporosis: a population-based longitudinal study in Taiwan. Arch Osteoporos. 2022;17(1):91.

    Article  PubMed  PubMed Central  Google Scholar 

  26. Guan X-Q, Xue Y-J, Wang J, et al. Low bone mineral density is associated with global coronary atherosclerotic plaque burden in stable angina patients. Clin Interv Aging. 2018;13:1475–83.

    Article  PubMed  PubMed Central  Google Scholar 

  27. Akin MN, Altun I. Associations of coronary plaque characteristics and coronary calcification with bone mineral density in postmenopausal women. Eur Rev Med Pharmacol Sci. 2022;26(20):7616–22.

    CAS  PubMed  Google Scholar 

  28. Xu R, Cheng X-C, Zhang Y, Lai H-M, Yang H-N. Association of severity of coronary lesions with bone mineral density in postmenopausal women. Arq Bras Cardiol. 2018;110(3):211–6.

    CAS  PubMed  PubMed Central  Google Scholar 

  29. Anagnostis P, Florentin M, Livadas S, Lambrinoudaki I, Goulis DG. Bone health in patients with dyslipidemias: an underestimated aspect. Int J Mol Sci. 2022;23(3).

  30. Mandal CC. High cholesterol deteriorates bone health: new insights into molecular mechanisms. Front Endocrinol (Lausanne). 2015;6:165.

    Article  PubMed  Google Scholar 

  31. Blumenthal RS, Kapur NK. Can a potent Statin actually regress coronary atherosclerosis? JAMA. 2006;295(13):1583–4.

    Article  CAS  PubMed  Google Scholar 

  32. Li GH-Y, Cheung C-L, Au PC-M, Tan KC-B, Wong IC-K, Sham P-C. Positive effects of low LDL-C and Statins on bone mineral density: an integrated epidemiological observation analysis and Mendelian randomization study. Int J Epidemiol. 2020;49(4):1221–35.

    Article  PubMed  Google Scholar 

  33. Azeez TA. Osteoporosis and cardiovascular disease: a review. Mol Biol Rep. 2023;50(2):1753–63.

    Article  CAS  PubMed  Google Scholar 

  34. Black DM, Rosen CJ. Clinical practice. Postmenopausal osteoporosis. N Engl J Med. 2016;374(3):254–62.

    Article  CAS  PubMed  Google Scholar 

  35. Frenkel B, Hong A, Baniwal SK, et al. Regulation of adult bone turnover by sex steroids. J Cell Physiol. 2010;224(2):305–10.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  36. Adami G. Regulation of bone mass in inflammatory diseases. Best Pract Res Clin Endocrinol Metab. 2022;36(2):101611.

    Article  PubMed  Google Scholar 

  37. Murray CE, Coleman CM. Impact of diabetes mellitus on bone health. Int J Mol Sci. 2019;20(19).

  38. Tarantino U, Cariati I, Greggi C et al. Skeletal system biology and smoke damage: from basic science to medical clinic. Int J Mol Sci. 2021;22(12).

  39. Forsén L, Bjørndal A, Bjartveit K, et al. Interaction between current smoking, leanness, and physical inactivity in the prediction of hip fracture. J Bone Min Res. 1994;9(11):1671–8.

    Article  Google Scholar 

  40. Wong PKK, Christie JJ, Wark JD. The effects of smoking on bone health. Clin Sci (Lond). 2007;113(5):233–41.

    Article  CAS  PubMed  Google Scholar 

  41. Cho Y, Choi S, Kim K, Lee G, Park SM. Association between alcohol consumption and bone mineral density in elderly Korean men and women. Arch Osteoporos. 2018;13(1):46.

    Article  PubMed  Google Scholar 

  42. Jang H-D, Hong J-Y, Han K, et al. Relationship between bone mineral density and alcohol intake: A nationwide health survey analysis of postmenopausal women. PLoS ONE. 2017;12(6):e0180132.

    Article  PubMed  PubMed Central  Google Scholar 

  43. Luo Z, Liu Y, Liu Y, Chen H, Shi S, Liu Y. Cellular and molecular mechanisms of alcohol-induced osteopenia. Cell Mol Life Sci. 2017;74(24):4443–53.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

Download references

Acknowledgements

We thank all the volunteers for the participation and personnel for their contribution in the study.

Funding

The present study was supported by grants from Academic Promotion Programme of Shandong First Medical University (No.2019QL017), the Science and Technology Development Plan Project of Tai ‘an (2021NS251), the Science and Technology Development Plan Project of Tai ‘an (2020NS133) and 2022 Jining Key Research and Development Plan Project (2022YXNS033).

Author information

Author notes

  1. Yanbo Liang and Xiaoqing Yuan contributed equally to this work.

Authors and Affiliations

  1. Department of Radiology, The Second Affiliated Hospital of Shandong First Medical University, No.366 Taishan Street, Tai’an City, Shandong Province, 271000, China

    Yanbo Liang, Qiang Shi, Hui Yang, Minghao Che, Yue Chen, Changqin Li & Jian Qin

  2. Chinese Institute for Medical Research, Capital Medical University, Beijing, China

    Xiaoqing Yuan

  3. Beijing Friendship Hospital, Capital Medical University, Beijing, China

    Xiaoqing Yuan

  4. Department of Radiology, Affiliated Hospital of Jining Medical University, Jining, 272000, China

    Luping Zhao

  5. Department of Radiology, Beijing Chaoyang Hospital, Capital Medical University, Beijing, 100020, China

    Qi Yang

Authors
  1. Yanbo Liang
    View author publications

    You can also search for this author inPubMed Google Scholar

  2. Xiaoqing Yuan
    View author publications

    You can also search for this author inPubMed Google Scholar

  3. Qiang Shi
    View author publications

    You can also search for this author inPubMed Google Scholar

  4. Hui Yang
    View author publications

    You can also search for this author inPubMed Google Scholar

  5. Luping Zhao
    View author publications

    You can also search for this author inPubMed Google Scholar

  6. Minghao Che
    View author publications

    You can also search for this author inPubMed Google Scholar

  7. Yue Chen
    View author publications

    You can also search for this author inPubMed Google Scholar

  8. Changqin Li
    View author publications

    You can also search for this author inPubMed Google Scholar

  9. Qi Yang
    View author publications

    You can also search for this author inPubMed Google Scholar

  10. Jian Qin
    View author publications

    You can also search for this author inPubMed Google Scholar

Contributions

YBL and XQY contributed to conceptualization, data collection, data curation, formal analysis, writing the original draft, and review and editing of the paper. QS and HY contributed to review and editing of the paper. LPZ contributed to data collection, data curation. MHC and YC contributed to data collection, data curation. CQL contributed to funding acquisition and supervision. QY and JQ contributed to methodology, project administration, supervision, and review and editing of the paper. All authors have read and approved the manuscript.

Corresponding authors

Correspondence to Qi Yang or Jian Qin.

Ethics declarations

Ethics approval and consent to participate

Ethics committee/IRB of The Second Affiliated Hospital of Shandong First Medical University approved this study (ethical approval number: 2023-55) and informed consent waiver. Due to the research was a retrospective study, there was no informed consent form in our study.

Consent for publication

Not Applicable.

Competing interests

The authors declare no competing interests.

Clinical trial number

Not applicable.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Liang, Y., Yuan, X., Shi, Q. et al. Prediction of abnormal bone mass with a pericoronary adipose tissue Attenuation model. BMC Geriatr 25, 261 (2025). https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12877-025-05928-3

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12877-025-05928-3

Keywords

BMC Geriatrics

ISSN: 1471-2318

Contact us