Associations between skeletal muscle parameters and metabolic markers in metabolic syndrome: an exploratory cross-sectional study using photon-counting computed tomography

Introduction

Metabolic syndrome (MetS) comprises a cluster of clinical conditions characterized by central obesity, dysglycemia, dyslipidemia, and elevated blood pressure. It represents a critical precursor to the development of type 2 diabetes and cardiovascular diseases, with insulin resistance as its core pathophysiological mechanism (1). Although traditional metabolic risk factors have been extensively studied, accumulating evidence indicates that abnormalities in body composition, particularly sarcopenia and myosteatosis, constitute significant prognostic determinants independent of conventional risk factors (2). However, the conventional body mass index (BMI) fails to differentiate between adipose and skeletal muscle tissues and cannot reflect the degree of intramuscular fat infiltration, thereby presenting a clear limitation in metabolic risk stratification.

Skeletal muscle is not only the largest metabolic organ but also the primary site for glucose uptake and insulin-mediated metabolism. Recent research has progressively shifted focus from “muscle quantity” to “muscle quality”, emphasizing the crucial role of muscle tissue composition, especially the extent of fat infiltration, in metabolic homeostasis (3). Myosteatosis can be quantified by the reduction in skeletal muscle density (SMD) on computed tomography (CT) imaging, which is closely associated with insulin resistance, inflammation activation, and adverse clinical outcomes (4). Notably, existing research has predominantly focused on populations with low muscle mass, whereas a potentially high-risk subgroup—MetS patients with normal or even elevated muscle mass but significantly degraded muscle quality—remains under-investigated. Furthermore, it is currently unclear whether disease status is associated with differences in the intrinsic physiological association patterns between skeletal muscle parameters and circulating metabolic markers.

Photon-counting computed tomography (PCCT), as a new-generation imaging technology, offers a novel platform for the precise and stable quantification of SMD and fat distribution, thanks to its exceptional spatial and contrast resolution (5). Recent studies have attempted to use conventional CT combined with deep learning models to classify metabolic phenotypes based on body composition (6). Building upon this foundation, this study leverages high-precision PCCT imaging with the following aims: (I) to compare the differences in skeletal muscle quantity [skeletal muscle index (SMI)] and quality (SMD) between MetS patients and non-metabolic controls; (II) to systematically evaluate whether the correlation structure between skeletal muscle parameters and metabolic indicators differs between MetS and non-MetS states, thereby revealing the potential impact of metabolic abnormalities on the body’s muscle-metabolism coupling relationship. We present this article in accordance with the STROBE reporting checklist (available at https://qims.amegroups.com/article/view/10.21037/qims-2026-0495/rc).

Methods

Study design and population

This a single-center, retrospective observational study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Institutional Review Board (IRB) of Shaoxing People’s Hospital (approval No. IEC-K-AF-076-1.2). The IRB waived the requirement for obtaining written informed consent due to the retrospective nature of the study, which involved no more than minimal risk to the participants. Adult patients (≥18 years) who underwent abdominal non-contrast PCCT scans at Shaoxing People’s Hospital, Shaoxing, China between January 2025 and October 2025 for health check-ups or non-specific abdominal pain were enrolled. The inclusion criteria were as follows: (I) availability of complete abdominal PCCT non-contrast scans for analysis; and (II) availability of complete fasting metabolic-related blood biochemical indicators within 3 months before or after the scan. The exclusion criteria were as follows: (I) severe hepatic or renal insufficiency; (II) active malignancy; (III) chronic wasting diseases or limb disabilities affecting body composition; and (IV) CT images with significant artifacts or unsuitable for accurate tissue boundary delineation.

The MetS group was defined using a modified International Diabetes Federation (IDF) criteria combined with quantitative abdominal CT indices. Given that this is an imaging-based body composition analysis study, CT-measured visceral fat area (VFA) was used as an imaging surrogate for waist circumference to provide a more direct and objective reflection of visceral fat burden. Cases meeting at least three of the following five criteria were included in the MetS group: (I) VFA ≥100 cm² as measured by CT; (II) fasting plasma glucose (FPG) ≥6.1 mmol/L or previous diagnosis of diabetes; (III) triglycerides (TG) ≥1.70 mmol/L or receiving relevant treatment; (IV) high-density lipoprotein cholesterol (HDL-C) <1.0 mmol/L (men) or <1.3 mmol/L (women); (V) systolic blood pressure ≥130 mmHg or diastolic blood pressure ≥85 mmHg or receiving antihypertensive treatment. The threshold of VFA ≥100 cm² has been validated as a surrogate for central obesity in Asian populations, showing good agreement with conventional waist circumference criteria (7). Those not meeting these criteria were included in the healthy control (HC) group.

Among the 38 participants in the MetS group, 12 (31.6%) had a pre-existing clinical diagnosis of MetS or at least one of its components before the CT scan; the remaining 26 (68.4%) were newly identified based on the study assessment.

PCCT image acquisition and body composition analysis

All scans were performed using a NAEOTOM Alpha photon-counting CT scanner (Siemens Healthineers, Erlangen, Germany). Scanning parameters were set as follows: tube voltage 120 kV, automatic tube current modulation, detector collimation 144×0.4 mm, and reconstruction slice thickness 0.625 mm. Images were transferred under standard body window settings to a body composition analysis workstation seamlessly connected to in-house developed quantitative software.

The third lumbar vertebra (L3) level was chosen for measurement, as its skeletal muscle parameters have been widely validated as reliable proxies for whole-body muscle mass and quality (8). Body composition measurements were performed by Radiologist A, with ≥3 years of experience in abdominal imaging diagnosis. Under blinded conditions, manual segmentation was performed on the axial image at the mid-level between the superior and inferior endplates of the L3 vertebral body for each case, including:

• Skeletal muscle area (SMA): the total cross-sectional area (cm²) delineated along the outer fascial border of the abdominal wall and paraspinal core muscle groups.
• SMD: the mean CT attenuation value [Hounsfield units (HU)] calculated within the segmented muscle region, reflecting the degree of intramuscular fat infiltration.
• SMI: calculated using the formula SMI = SMA/height² (m²).
• VFA: the area (cm²) of intra-abdominal fat delineated along the inner peritoneal border at the same level. A representative example of this manual segmentation is illustrated in Figure 1.

Figure 1 Schematic illustration of body composition segmentation on a photon-counting computed tomography axial image at the L3 vertebral level. Yellow shading indicates the VAT area, delineated along the inner peritoneal border. Blue shading indicates the SMA, which includes the abdominal wall muscles (rectus abdominis, transversus abdominis, internal and external obliques) and paraspinal muscles (psoas, erector spinae, quadratus lumborum). The SMD is derived as the mean CT attenuation (in HU) within the blue-shaded region. CT, computed tomography; HU, Hounsfield units; SMA, skeletal muscle area; SMD, skeletal muscle density; VAT, visceral adipose tissue.

To evaluate measurement reliability, a two-stage reliability test was conducted: (I) radiologist A re-measured 20 randomly selected images after a 2-week interval to assess intra-observer consistency; (II) another radiologist (B) of comparable experience independently repeated measurements on 30 randomly selected images to assess inter-observer consistency. Intraclass correlation coefficients (ICCs) were calculated using a two-way random-effects model, with ICC >0.75 indicating good agreement.

Clinical and biochemical data

Demographic and anthropometric data were collected, including age, sex, height, and weight (used to calculate BMI). Laboratory indicators included FPG, glycated hemoglobin (HbA1c), TG, HDL-C, low-density lipoprotein cholesterol (LDL-C), total cholesterol (TC), and uric acid (UA). In the MetS group, 11 patients (28.9%) were receiving glucose-lowering medications; in the control group, 6 patients (7.6%) were on such medications.

Statistical analysis

Continuous variables are presented as median (interquartile range); categorical variables are presented as frequency (percentage). Comparisons of continuous variables between the MetS and control groups were performed using the Mann-Whitney U test, whereas categorical variables were compared using the Chi-squared test. Spearman’s rank correlation analysis (Spearman’s rho) was employed to assess the within-group relationships between muscle parameters and metabolic indicators. To control the false positive risk from multiple comparisons, P values from correlation tests were adjusted using the false discovery rate (FDR) (Benjamini-Hochberg) method. To formally compare the strength of correlation coefficients between the two groups, Fisher’s r-to-z transformation was applied. A two-tailed P value <0.05 was considered statistically significant for this comparison. All statistical analyses were performed using R language (version 4.2.0; R Foundation for Statistical Computing, Vienna, Austria). A two-sided FDR-adjusted P value <0.05 was considered statistically significant.

Results

Baseline characteristics of the study population

A total of 117 patients were included in the final analysis, comprising 38 in the MetS group and 79 in the HC group. There were no statistically significant differences between the two groups in terms of age, sex, or height (all P>0.05).

As shown in Table 1, the MetS group exhibited typical features of metabolic dysregulation: median levels of weight, BMI, FPG, HbA1c, TG, and UA were all significantly higher than those in the HC group, whereas the median level of HDL-C was significantly lower (all P<0.05). There were no significant differences in median serum TC or LDL-C levels between the groups.

Regarding body composition, the median SMI was significantly higher in the MetS group than it was in the HC group (42.73 vs. 39.16 cm²/m², P<0.001). In contrast, the SMD, reflecting muscle quality, showed a decreasing trend in the MetS group (30.54 vs. 35.22 HU), although the between-group difference did not reach the conventional threshold for statistical significance (P=0.074). This distinct body composition phenotype is visually summarized in Figure 2.

Figure 2 Comparison of skeletal muscle parameters between the metabolic syndrome and control groups. (A) Comparison of SMD between the MetS group and the HC group. (B) Comparison of SMI between the MetS group and the healthy control group. All skeletal muscle parameters were measured at the L3 level using PCCT. Data are presented as median (interquartile range), and between-group comparisons were performed using the Mann-Whitney U test. HC, healthy control; MetS, metabolic syndrome; PCCT, photon-counting computed tomography; SMD, skeletal muscle density; SMI, skeletal muscle index.

Measurement reliability analysis

The reliability analysis demonstrated excellent reproducibility of body composition quantification based on PCCT.

For intra-observer consistency, Radiologist A achieved ICC [95% confidence interval (CI)] as follows: SMA 0.98 (0.95–0.99), SMD 0.96 (0.91–0.99), and VFA 0.97 (0.93–0.99).

Inter-observer consistency was similarly excellent, with ICCs for SMA, SMD, and VFA being 0.95 (0.90–0.98), 0.93 (0.86–0.97), and 0.96 (0.92–0.98), respectively.

Subgroup correlation analysis: association patterns between muscle parameters and metabolic markers by metabolic state

To control for false positives arising from multiple comparisons, all P values from correlation analyses were adjusted using the Benjamini-Hochberg FDR method.

To provide an overall reference framework, correlation analysis was first performed in the entire cohort (N=117). The results are summarized in Table S1. Overall, skeletal muscle quantity parameters (SMA and SMI) showed significant positive correlations with BMI (SMA: ρ=0.549; SMI: ρ=0.464; both FDR-adjusted P<0.001) and UA (SMA: ρ=0.301, FDR-adjusted P=0.005; SMI: ρ=0.372, FDR-adjusted P<0.001). SMA and SMI were also positively correlated with TG (SMA: ρ=0.293, FDR-adjusted P=0.006; SMI: ρ=0.250, FDR-adjusted P=0.020), while SMI was negatively correlated with HDL‑C (ρ=−0.238, FDR-adjusted P=0.026). For muscle quality, SMD demonstrated a negative correlation with HbA1c (ρ=−0.249, FDR-adjusted P=0.020) and a positive correlation with UA (ρ=0.225, FDR-adjusted P=0.034). Notably, no significant association was observed between SMA or SMI and FPG in the overall cohort (both FDR-adjusted P>0.05).

Formal comparison of correlation coefficients using Fisher’s r-to-z transformation revealed a statistically significant difference in the SMA-FPG correlation between the control group (ρ=0.001) and the MetS group (ρ=0.411; z=−2.133, P=0.033), suggesting a difference in association patterns between groups.

However, in multivariable linear regression analysis including an interaction term (SMA × metabolic status), the interaction did not reach statistical significance (β=0.051, P=0.215). Similarly, in the MetS group, the association between SMA and FPG was not statistically significant after adjustment for age, sex, and BMI (β=0.056, P=0.451).

Group-specific association patterns between skeletal muscle parameters and metabolic markers are visually presented in the comparative heatmaps (Figure 3).

Figure 3 Comparative correlation heatmaps of skeletal muscle-metabolic association patterns between the healthy control and metabolic syndrome groups. Panel A represents the healthy control group, and panel B represents the MetS group. The heatmaps display the Spearman correlation coefficients (ρ) between skeletal muscle parameters (SMA, SMI, and SMD), anthropometric measures (height, body weight, and BMI), and circulating metabolic markers (FPG, HbA1c, TG, UA, etc.) in each group. The color gradient indicates the direction and strength of the correlations, with red representing positive correlations and blue representing negative correlations. Statistically significant correlations after FDR correction are indicated by asterisks (*). The heatmaps are intended for visual illustration; statistical significance was formally assessed using correlation analysis with FDR correction and Fisher’s r‑to‑z comparison. Visual differences between panels do not imply statistically significant differences between groups. BMI, body mass index; FDR, false discovery rate; FPG, fasting plasma glucose; HbA1c, glycated hemoglobin; MetS, metabolic syndrome; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; SMA, skeletal muscle area; SMD, skeletal muscle density; SMI, skeletal muscle index; TG, triglycerides; UA, uric acid.

Observed association patterns between muscle quality and glycemic control

In the HC group (Table 2), SMD showed a negative correlation trend with HbA1c (Spearman’s ρ=–0.265, unadjusted P=0.018), which did not remain statistically significant after FDR correction (adjusted P=0.072). In the MetS group (Table 3), this correlation was no longer present (ρ=–0.087, adjusted P=0.997).

Differences in the association between muscle quantity and glycemia across groups

The correlation between SMA and FPG differed markedly between groups. No significant association was observed in the HC group (ρ≈0.001, adjusted P=0.990). Conversely, in the MetS group, SMA showed a significant positive correlation with FPG (ρ=0.411, unadjusted P=0.010, adjusted P=0.035).

Other specific changes in associations

The positive correlation between SMI and TG was significant only in the HC group (ρ=0.289, adjusted P=0.024) and was not significant in the MetS group.

SMD showed a positive correlation trend with UA in the control group (ρ=0.258, adjusted P=0.072), which also did not reach statistical significance in the MetS group.

Stable structural associations

Both SMA and SMI maintained significant positive correlations with height, weight, and BMI in both groups, and these remained statistically significant after FDR correction (all adjusted P<0.05).

Discussion

This study is an exploratory cross-sectional analysis and does not imply causality. Thus, the current findings do not establish independent or causal biological mechanisms. This study, leveraging the high-precision capabilities of PCCT, suggests differences in the association patterns between quantitative skeletal muscle parameters and circulating metabolic markers in individuals with MetS (5). The principal findings can be summarized as follows: (I) patients with MetS exhibit a distinct body composition phenotype characterized by increased skeletal muscle quantity accompanied by a trend toward lower SMD; (II) muscle-metabolism associations are strongly context-dependent, as evidenced by a negative correlation trend between muscle quality (SMD) and long-term glycemic control (HbA1c) in healthy individuals that is not observed in MetS patients; and (III) most importantly, the association between muscle quantity (SMA) and FPG shows a group‑specific difference, with no detectable relationship in controls and a positive correlation observed in the MetS group.

The coexistence of a higher SMI with a declining trend in SMD observed in the MetS group delineates a high-risk body composition phenotype that is inadequately captured by the traditional “sarcopenia” framework. This phenotype has been variably described as myosteatotic obesity or normo-muscular sarcopenia, emphasizing preserved or increased muscle quantity with compromised muscle quality. The identification of this phenotype is consistent with accumulating evidence highlighting the importance of CT-derived quantitative metrics—such as VFA—for the evaluation of metabolic dysregulation (9). Moreover, international consensus statements, including the Writing Group for the European Working Group on Sarcopenia in Older People 2 (EWGSOP2) guidelines, explicitly recognize muscle quality as being equally important as muscle quantity in the assessment of muscular health and metabolic risk (10).

Importantly, the accurate delineation of this phenotype in the present study was facilitated by the high reproducibility of PCCT-based measurements, as reflected by the excellent ICC (>0.95). This allowed for reliable detection of subtle differences in muscle composition. Previous mechanistic studies have established that myosteatosis—particularly intramyocellular lipid accumulation—is a central pathological process impairing insulin signal transduction and promoting peripheral insulin resistance (11,12). From this perspective, the declining SMD observed in the MetS cohort is not merely an imaging-derived feature but represents a visually quantifiable biomarker of skeletal muscle-level metabolic dysfunction. It is noteworthy that, although the between-group difference in SMD did not reach conventional statistical significance, the SMD level in the MetS group in this study was markedly lower than the reference ranges (e.g., approximately 40–41 HU) commonly used in prior studies to suggest a risk of myosteatosis, indicating a potential widespread tendency toward lower SMD in this population.

The central contribution of this study lies in describing the differential muscle–metabolism association patterns across metabolic states. Under healthy physiological conditions, higher muscle quality—reflecting lower fat infiltration—is consistently associated with improved glycemic control, reinforcing the role of skeletal muscle as a primary insulin-sensitive organ (13). In contrast, this physiological coupling may be disrupted in MetS. We hypothesize that when systemic metabolic insults, such as chronic low-grade inflammation, lipotoxicity, and oxidative stress, accumulate beyond a critical threshold, skeletal muscle adaptations may transition from reversible functional alterations to more persistent structural differences in associations and functional inactivation (14). In such a state, myosteatosis reflected by SMD may represent a relatively stable structural lesion, whereas its dynamic linkage to short-term glucose homeostasis becomes attenuated or lost. Importantly, in the overall cohort, no significant association was observed between muscle quantity parameters and FPG, whereas a significant positive association emerged in the MetS subgroup. This pattern supports a metabolic status-dependent difference in association patterns.

The most clinically instructive observation is the emergence of a significant positive correlation between SMA and FPG in the MetS group. This counterintuitive finding challenges the conventional assumption that increased muscle mass is uniformly metabolically protective. Instead, it suggests the presence of “BMI-adjusted association” or metabolically ineffective muscle differences in associations under conditions of insulin resistance. This interpretation is supported by recent large-scale imaging-based epidemiological studies demonstrating that increasing blood glucose levels are associated with greater trunk muscle area but lower muscle density (15). Collectively, these findings carry important clinical implications. In individuals with MetS, therapeutic strategies aimed at improving metabolic health—particularly exercise-based interventions—should shift from a singular focus on increasing muscle mass toward an integrated goal of enhancing muscle quality, restoring metabolic function, and improving insulin sensitivity (16).

A key strength of this study is the high reproducibility of PCCT-based muscle density measurements, as reflected by the excellent ICC (>0.95). This reproducibility facilitated the detection of subtle association patterns that might otherwise have been obscured by measurement noise. Several limitations should be acknowledged. First, the cross-sectional design precludes causal inference and limits interpretation to associative relationships. Second, the relatively small size of the MetS cohort (n=38) may have reduced statistical power, potentially masking biologically relevant differences and limiting subgroup analyses; therefore, non-significant findings should be interpreted with caution. Post‑hoc power analysis indicated that, with the current MetS sample size (n=38), the study had approximately 80% power to detect a moderate correlation (ρ≥0.45). For smaller effect sizes, such as the observed SMD-HbA1c correlation (ρ=−0.087), the lack of statistical significance may be attributable to limited power rather than a true absence of association. In addition, the heterogeneity within the MetS group (e.g., varying combinations of diagnostic components) was not specifically analyzed, which may have influenced association patterns.

Although Fisher’s r-to-z transformation suggested a significant difference in correlation coefficients between groups, this finding was not supported by interaction term analysis in multivariable regression. This discrepancy may reflect limited statistical power, particularly in the MetS subgroup (n=38), and indicates that the observed differences in correlation patterns should be interpreted with caution. Therefore, the evidence for betweengroup differences in associations should be interpreted as exploratory rather than confirmatory. Third, the absence of direct measurements of insulin sensitivity (e.g., hyperinsulinemic-euglycemic clamp) and objective assessments of muscle function (e.g., handgrip strength or performance-based tests) prevents direct functional validation of the proposed muscle metabolic dysfunction. Fourth, glucose-lowering medications were used by 11 participants (28.9%) in the MetS group and 6 participants (7.6%) in the HC group. Although these proportions were modest, the presence of medication users may have influenced fasting glucose and HbA1c levels, potentially biasing correlation estimates. Due to the limited sample size, we were unable to perform sensitivity analyses excluding these participants. Therefore, the reported associations should be interpreted with caution. Fifth, the temporal window of up to three months between imaging and laboratory measurements may introduce variability in metabolic markers, potentially affecting correlation estimates. Sixth, this study did not exploit the spectral or material decomposition capabilities of PCCT; muscle density measurements were based on conventional HU, which are also obtainable with standard CT. The added value of PCCT here primarily lies in improved measurement stability and reproducibility. Finally, body composition assessment was confined to a single axial level at L3. Although this approach is widely accepted and validated (8), it does not capture regional heterogeneity in muscle distribution or quality, particularly in the appendicular musculature.

Taken together, these findings suggest that although differences in muscle–glucose association patterns between metabolic states may exist, the current evidence remains exploratory and requires confirmation in larger, well-powered studies.

Hou et al. recently demonstrated that CT-derived body composition parameters are valuable for predicting MetS in adults, supporting the broader role of quantitative body composition analysis in metabolic research (17).

Conclusions

In this exploratory cross-sectional study, MetS was associated with higher SMI and a trend toward lower muscle density. Differences in correlations between skeletal muscle parameters and metabolic markers were observed descriptively but were not consistently supported by adjusted analyses. These findings should be considered hypothesis‑generating and warrant confirmation in larger prospective studies. Future research should focus on longitudinal validation of PCCT-derived muscle quality metrics as predictors of cardiometabolic outcomes, as well as on targeted interventions aimed at improving muscle quality.

Acknowledgments

We are deeply grateful to the Key Laboratory of Functional Molecular Imaging in Oncology at Shaoxing People’s Hospital, Zhejiang, China, for their generous financial assistance and technical expertise, which significantly contributed to this study.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://qims.amegroups.com/article/view/10.21037/qims-2026-0495/rc

Data Sharing Statement: Available at https://qims.amegroups.com/article/view/10.21037/qims-2026-0495/dss

Funding: This work was supported by grants from the Shaoxing Health Commission (Nos. 2023SKY035 and 2024SKY013), the Zhejiang Provincial Health Commission (No. 2025HY1294), and Shaoxing University (No. Y20250293).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://qims.amegroups.com/article/view/10.21037/qims-2026-0495/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. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Institutional Review Board (IRB) of Shaoxing People’s Hospital (approval No. IEC-K-AF-076-1.2). The IRB waived the requirement for obtaining written informed consent due to the retrospective nature of the study, which involved no more than minimal risk to the subjects.

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. Cornier MA, Dabelea D, Hernandez TL, Lindstrom RC, Steig AJ, Stob NR, Van Pelt RE, Wang H, Eckel RH. The metabolic syndrome. Endocr Rev 2008;29:777-822. [Crossref] [PubMed]
2. Roh E, Choi KM. Health Consequences of Sarcopenic Obesity: A Narrative Review. Front Endocrinol (Lausanne) 2020;11:332. [Crossref] [PubMed]
3. Yuan S, Larsson SC. Epidemiology of sarcopenia: Prevalence, risk factors, and consequences. Metabolism 2023;144:155533. [Crossref] [PubMed]
4. Aubrey J, Esfandiari N, Baracos VE, Buteau FA, Frenette J, Putman CT, Mazurak VC. Measurement of skeletal muscle radiation attenuation and basis of its biological variation. Acta Physiol (Oxf) 2014;210:489-97. [Crossref] [PubMed]
5. Willemink MJ, Persson M, Pourmorteza A, Pelc NJ, Fleischmann D. Photon-counting CT: Technical Principles and Clinical Prospects. Radiology 2018;289:293-312. [Crossref] [PubMed]
6. Pooler BD, Garrett JW, Lee MH, Rush BE, Kuchnia AJ, Summers RM, Pickhardt PJ. CT-Based Body Composition Measures and Systemic Disease: A Population-Level Analysis Using Artificial Intelligence Tools in Over 100,000 Patients. AJR Am J Roentgenol 2025;224:e2432216. [Crossref] [PubMed]
7. Li X, Katashima M, Yasumasu T, Li KJ. Visceral fat area, waist circumference and metabolic risk factors in abdominally obese Chinese adults. Biomed Environ Sci 2012;25:141-8. [PubMed]
8. Kong M, Geng N, Zhou Y, Lin N, Song W, Xu M, et al. Defining reference values for low skeletal muscle index at the L3 vertebra level based on computed tomography in healthy adults: A multicentre study. Clin Nutr 2022;41:396-404. [Crossref] [PubMed]
9. Sun Y, Lin X, Zou Z, Zhou Y, Liu A, Li X, Du Y, Ji X, Li Z, Wu X, Wang Y, Lv X, Li T, Zhang J, Guo Z, Li H, Li Y. Association between visceral fat area and metabolic syndrome in individuals with normal body weight: insights from a Chinese health screening dataset. Lipids Health Dis 2025;24:57. [Crossref] [PubMed]
10. Cruz-Jentoft AJ, Bahat G, Bauer J, Boirie Y, Bruyère O, Cederholm T, Cooper C, Landi F, Rolland Y, Sayer AA, Schneider SM, Sieber CC, Topinkova E, Vandewoude M, Visser M, Zamboni M. Sarcopenia: revised European consensus on definition and diagnosis. Age Ageing 2019;48:16-31. [Crossref] [PubMed]
11. Oh E, Cho NJ, Kang H, Kim SH, Park HK, Kwon SH. Computed tomography evaluation of skeletal muscle quality and quantity in people with morbid obesity with and without metabolic abnormality. PLoS One 2023;18:e0296073. [Crossref] [PubMed]
12. Shulman GI. Ectopic fat in insulin resistance, dyslipidemia, and cardiometabolic disease. N Engl J Med 2014;371:1131-41. [Crossref] [PubMed]
13. Petersen MC, Shulman GI. Mechanisms of Insulin Action and Insulin Resistance. Physiol Rev 2018;98:2133-223. [Crossref] [PubMed]
14. Fan J, Zuo L, Li F, Wang B, An Y, Yu D. Patients With Type 2 Diabetes Mellitus and Early Diabetic Kidney Disease Exhibit Lower Computed Tomography-measured Skeletal Muscle Attenuation Values: A Propensity Score-matched Study. J Ren Nutr 2024;34:509-18. [Crossref] [PubMed]
15. Warner JD, Blake GM, Garrett JW, Lee MH, Nelson LW, Summers RM, Pickhardt PJ. Correlation of HbA1c levels with CT-based body composition biomarkers in diabetes mellitus and metabolic syndrome. Sci Rep 2024;14:21875. [Crossref] [PubMed]
16. Goodpaster BH, Sparks LM. Metabolic Flexibility in Health and Disease. Cell Metab 2017;25:1027-36. [Crossref] [PubMed]
17. Hou B, Li Y, Liu C, Zhang Y, Wen D, Li X. Assessment of body compositions in the prediction of metabolic syndrome in adults. Quant Imaging Med Surg 2024;14:5891-901. [Crossref] [PubMed]