RPS 110 - What is new in sarcopenia and body composition

Author Block: M. Thanaj, B. Whitcher, H. Raza, M. Niglas, C. Bell-Bradford, E. L. Thomas, D. Amiras, J. D. Bell; London/UK
Purpose: Glucagon-like peptide-1 receptor agonists (GLP-1RA) are recommended for treatment of type 2 diabetes (T2D) and obesity, but concerns regarding sarcopenia are reported. We developed automated methods to quantify volume and fat infiltration of 20 hip and thigh muscles, evaluating sex- and GLP-1RA–specific differences.
Methods or Background: UK Biobank participants prescribed with GLP-1RA before the baseline imaging visit (N = 67; mean duration 5.9 ± 2.9 years, range 0.1-12) were matched by sex, age (±1.5 years), BMI (±1.5 kg/m²), Townsend index (±1.5), T2D, hypertension, cardiovascular disease, muscle disorders, sedentary lifestyle (>10.6 h/day), and cholesterol medication with non-GLP-1RA controls. Sensitivity analyses included dynapenia (handgrip <16 kg women, <27 kg men) and falls (>1). Twenty hip and thigh muscles were segmented from T1w Dixon MRI scans using a deep learning model with MONAI (Medical Open Network for AI) Swin-UNETR architecture. Median fat fraction per muscle group was calculated. Paired t-tests were performed with false discovery rate (FDR) correction.
Results or Findings: In men (N=60), GLP-1RA use showed a higher pectineus fat fraction (mean diff: 0.013, FDR=0.043) and rectus femoris fat fraction (0.018, FDR=0.038), while adductor longus (-42 mL, FDR=0.038), gluteus medius (-67 mL, FDR=0.038), and rectus femoris (-79 mL, FDR=4×10⁻⁴) volumes were lower, showing no significant differences in women (N=28). Including dynapenia and falls (44M/20F) eliminated differences in muscle volume and fat fraction.
Conclusion: Our scalable deep learning pipeline shows GLP-1RA–related adductor longus, rectus femoris, and gluteus medius volume reduction and pectineus and rectus femoris fat fraction increase in men, with no differences in women or when including dynapenia and falls. These findings suggest possible sex-specific GLP-1RA effects that require validation in larger cohorts.
Limitations: Small sample size, limited longitudinal data, heterogeneity in GLP-1 exposure, and medication type may restrict interpretation
Funding for this study: No funding was recieved
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: Fully anonymised images and participant metadata were obtained through UK Biobank Access Application number 23889. The UK Biobank has approval from the North West Multi-Centre Research Ethics Committee (REC reference: 11/NW/0382), and obtained written informed consent from all participants before the study. All methods were performed in accordance with the relevant guidelines and regulations as presented by the appropriate authorities, including the Declaration of Helsinki.

Author Block: L. Fu, X-G. Peng; Nanjing/CN
Purpose: While the lumbar skeletal muscle index (SMI) predicts outcomes in sepsis, the significance of its dynamic changes remains unclear. This study evaluated the association between longitudinal changes in CT-derived body composition parameters and clinical outcomes in sepsis.
Methods or Background: We retrospectively included sepsis patients from three tertiary centers. Two non-contrast chest CT scans at the L1 level were analyzed: a baseline scan within 48 hours of diagnosis and a follow-up scan ≥5 days later. Parameters included SMI, skeletal muscle density (SMD), subcutaneous and visceral fat area (SFA, VFA) and density (SFD, VFD). X-tile determined the optimal ΔSMI cutoff for mortality. Patients were classified into high and low skeletal muscle wasting (SMW) groups. Survival was analyzed using Kaplan–Meier and log-rank tests. Cox regression identified mortality-associated factors.
Results or Findings: A total of 561 patients (median age, 69 years; interquartile range, 58–79 years; 378 males) were analyzed. The optimal ΔSMI cutoff for defining high SMW was > 4.04 cm²/m². Patients in the high SMW group had significantly higher 28-day, 3-month, and 1-year mortality Among 561 patients (median age 69; 378 male), the optimal ΔSMI cutoff for high SMW was > 4.04 cm²/m². The high SMW group had significantly higher 28-day (56.5% vs 3.0%), 3-month (82.4% vs 6.8%), and 1-year mortality (88.6% vs 17.4%) (all P<0.001). No other Δparameters correlated with mortality. Multivariate Cox analysis identified ΔSMI, age, Charlson Index, mechanical ventilation, APACHE II, and creatinine as independent predictors. ΔSMI was the strongest predictor (HR: 1.165 per 1 cm²/m² loss; 95% CI: 1.133–1.198; P<0.001).
Conclusion: Longitudinal loss of lumbar muscle mass, quantified by serial CT, independently predicts mortality in sepsis. Monitoring muscle dynamics could enhance risk stratification and guide targeted interventions to improve outcomes.
Limitations: This study requires prospective validation.
Funding for this study: Longitudinal loss of lumbar muscle mass, quantified by serial CT, independently predicts mortality in sepsis. Monitoring muscle dynamics could enhance risk stratification and guide targeted interventions to improve outcomes.
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: 2024ZDSYLL060-p01

Author Block: G. Boscaro, M. Visoná, S. Silvestrin, D. Trevisani, A. Michielin, E. Faccioli, A. Dell'Amore, R. Stramare, C. Giraudo; Padova/IT
Purpose: To assess the role of CT-based body composition in bilateral lung transplant candidates (BLTc).
Methods or Background: We included BLTc referring to our tertiary center who performed at least one HRCT <18 months before the transplant. The paravertebral muscle and subcutaneous tissue were segmented at the level of the 12th thoracic vertebra to extract density (Hu) and area (cm2); the skeletal muscle index (SMI) was then computed. According to Moon et al, the following SMI thresholds, 7.13cm2/m2 for women and 8.67cm2/m2 for men, were used to define sarcopenia. Using the Spearman correlation coefficient, the relationship between muscle strength, hand-grip test, sarcopenia, muscle density, and area was explored. By the ANOVA for repeated measures any change in body composition over three time points (before the transplant, ≤36 months after the transplant, and ≤48 months after the transplant) was investigated. The logistic regression analysis was used to evaluate if demographics, spirometry metrics, and sarcopenia before the transplant predicted the survival.
Results or Findings: Sixty-two BLTc were included (24 female; average age 5313 years; 25 died). Twenty-three (37.1%) BLTc were affected by sarcopenia before the transplant. The ANOVA demonstrated a significant decrease of muscle area at first follow-up after BLT (p<0.001). An inverse, significant but moderate correlation emerged between sarcopenia and muscle strength as well as grip test (r=-.310,p=0.014 and r=-.409,p<0.001, respectively) and a positive moderate correlation occurred between muscle area and muscle strength and grip test (r=445,p<0.001 and r=.342,p=0.007). The multivariate logistic regression showed that sarcopenia predicts mortality (p=0.014, 95%CI for sarcopenia 1.523–45.429).
Conclusion: More than a third of BLTc are affected by sarcopenia which may have an impact on the success of the transplant.
Limitations: Single center study.
Funding for this study: No.
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: local IRB

Author Block: J. Gotta, L. D. Grünewald, V. Koch, S. Mahmoudi, T. Vogl; Frankfurt am Main/DE
Purpose: Advanced CT-derived body composition metrics are physiologically relevant but computationally demanding. This study aimed to evaluate whether attenuation and volume measurements from individual muscle groups, specifically autochthonous, gluteal, and iliopsoas muscles, can serve as surrogates for established metrics of muscle quality and quantity, including skeletal muscle radiodensity (SMRA), myosteatosis, skeletal muscle area (SMA), and muscle volume.
Methods or Background: In this retrospective analysis of 25,252 abdominal CT scans, attenuation and volume metrics were extracted for the iliopsoas, autochthonous, and gluteal muscles. Generalized additive models (GAMs) were fitted to assess their association with reference myosteatosis metrics, including L3-SMRA, volumetric muscle radiodensity, intramuscular fat proportion, L3-SMA, and volumetric muscle percentage. All predictors were modeled as continuous spline terms, and model performance was evaluated using pseudo R² (deviance explained). Laterality was excluded. Relationships were visually assessed for linearity.
Results or Findings: Attenuation of the autochthonous muscles was the strongest surrogate for L3-SMRA (R² = 0.88), followed by gluteus medius (R² = 0.73), gluteus maximus (R²=0.70), and iliopsoas (R² = 0.69). For volumetric abdominal myosteatosis, gluteus maximus HU achieved the highest explained variance (R² =0.59), followed by autochthonous (R²=0.57), gluteus medius (R²=0.54), and iliopsoas (R²=0.43). Normalized L3-SMA was best predicted by iliopsoas muscle volume (R²=0.69), followed by autochthonous (R² = 0.64), gluteus medius (R²=0.60), and gluteus maximus (R² = 0.49). Associations with volumetric muscle percentage were consistently weaker across all compartments (R²≤0.35). All results were statistically significant (p < 0.001).
Conclusion: Attenuation metrics of the autochthonous and gluteal muscles are strong linear surrogates for SMRA and myosteatosis, while iliopsoas muscle volume most accurately predicts L3-SMA. These findings suggest that selective muscle-based surrogates can approximate advanced body composition metrics with lower computational cost than full segmentation.
Limitations: Retrospective study
Funding for this study: No funding was received.
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: Under waiver.

Author Block: Y. Zhang, J. Tang, Z. Li; Chengdu/CN
Purpose: This study aimed to use computed tomography (CT) to assess muscle reduction and adipose deposition and examined the predictive values of those indicators on clinical outcomes after kidney transplantation.
Methods or Background: Abnormal body composition has been widely observed in kidney transplant recipients. However, how muscle reduction and adipose deposition affect the prognosis after kidney transplantation is not fully understood, especially in Chinese population. This prospective study examined a consecutive series of patients who underwent their first kidney transplant at our medical center between June 1, 2020 and June 30, 2023. Within 30 days before surgery, skeletal muscle mass index (SMI), skeletal muscle density (SMD), visceral adipose area index (VAI) and subcutaneous adipose area index (SAI) were evaluated using abdominal computed tomography. Multivariable logistic regression and cox regression were used to explore associations between each index and prognosis in terms of infection, delayed graft function and rejection after kidney transplantation.
Results or Findings: The final analysis included 455 patients (147 women, mean age 34.5 ± 10.18 years), who were followed up for a median of 47.67 (IQR, 34.23 – 51.72) months. In the multivariable Cox regression model, each one standard deviation increase in SMD was associated with a 20% lower risk of urinary tract infections (HR, 0.80; 95% CI, 0.69–0.95; p = 0.008), indicating that patients with lower SMD were at significantly higher risk. Elevated VAI independently predicted lung infections (HR, 1.25, 95 CI%, 1.03-1.52, p = 0.022). Patients with high VAI increased delayed graft function risk 1.44-fold (HR 1.44, 95%CI 1.04-1.98, p = 0.028).
Conclusion: Visceral adipose accumulation and skeletal muscle density reduction before kidney transplantation are independently associated with greater risk of poor prognosis on clinical outcomes after kidney transplantation.
Limitations: This was a single-center study.
Funding for this study: None.
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: the Ethics Committee of West China Hospital

Author Block: M. Basile, M. Zanardo, S. Gitto, D. Albano, L. M. Sconfienza; Milano/IT
Purpose: This systematic review and meta-analysis evaluated the impact of sarcopenia on clinical outcomes after percutaneous vertebroplasty (PVP) and kyphoplasty (PKP) in patients with osteoporotic vertebral compression fractures (OVCFs), and examined the heterogeneity in sarcopenia definitions across studies.
Methods or Background: A comprehensive literature search was conducted in PubMed, Embase, and Scopus through June 2025. Eligible studies assessed sarcopenia in OVCF patients treated with PVP or PKP. Extracted data included diagnostic modalities (CT, MRI, DEXA, BIA), anatomical landmarks, muscle indices (skeletal muscle index [SMI], psoas muscle index, psoas-to-lumbar vertebral index, appendicular muscle index), functional parameters (handgrip strength, gait speed), clinical outcomes (refracture, residual low back pain [RBP]), and statistical measures.
Results or Findings: Eighteen studies (3,531 patients; 1,357 sarcopenic) were included. CT was the most frequently employed diagnostic tool (72%), followed by MRI (27%), DEXA (11%), and BIA (0.2%). The most common reference levels were L3 (28%) and T12 (28%). SMI was reported in 7 studies (38%), with cut-off values ranging from 29–30.6 cm²/m² in males and 36–42.6 cm²/m² in females. Six studies (33%) assessed functional parameters, while only three (17%) fully adhered to EWGSOP2 criteria. Pooled analysis demonstrated that sarcopenic patients had a markedly higher risk of refracture (HR 3.61; 95% CI 2.49–5.23; p<0.05) and an increased likelihood of RBP (OR 2.78; 95% CI 1.50-5.00).
Conclusion: Sarcopenia significantly worsens outcomes after PVP and PKP in OVCF patients. Adoption of a standardized diagnostic framework—integrating CT-based muscle mass quantification (particularly SMI at T12), muscle strength (handgrip), and physical performance (gait speed)—is recommended to enhance risk stratification. Tailored management strategies for sarcopenic patients may improve postoperative outcomes.
Limitations: Heterogeneity in sarcopenia definitions, the predominance of retrospective study designs, and limited generalizability underscore the need for further prospective, standardized, multicenter studies.
Funding for this study: Not applicable
Has your study been approved by an ethics committee? Not applicable
Ethics committee - additional information:

Author Block: D. Vogele, S. Waibel, K. Mueller, T. Ettrich, M. J. Beer; Ulm/DE
Purpose: Sarcopenia is associated with poor oncologic outcomes, increased risk of falls, hospitalization, and reduced quality of life. However, diagnostic approaches remain heterogeneous. This study evaluated a simple radiological method for sarcopenia detection in colorectal cancer patients using routine staging CT scans, and assessed its prognostic relevance. The role of radiomic features combined with machine learning for diagnosing sarcopenia and predicting tumor progression was also analyzed.
Methods or Background: CT scans of 57 rectal and 46 sigmoid carcinoma patients undergoing chemotherapy were retrospectively reviewed. At the third lumbar vertebra, muscle areas of the psoas major, quadratus lumborum, erector spinae, and rectus abdominis were measured. The Psoas Muscle Index (PMI) was calculated as psoas muscle area/height². Sarcopenia was defined using Bahat et al. cut-offs. Tumor response was assessed by RECIST at diagnosis (T1) and after 9 months (T3). Sex and tumor entity differences were evaluated. Radiomic features of skeletal muscles were extracted, and feature selection performed by lasso regression to train a neural network.
Results or Findings: Psoas muscle assessment yielded high imaging quality in 97.7% of cases; unilateral and bilateral measures were equivalent (p=0.064). Low PMI showed a non-significant trend toward predicting progressive disease (PD) in univariate regression (OR=0.79; p=0.07), but not in multivariate analysis (OR=1; p>0.5). No association was observed between PMI and PD-free survival. Machine learning achieved 0.71±0.08 accuracy (AUC=0.79±0.07) for sarcopenia detection, but only 0.63±0.04 (AUC=0.66±0.06) for PD prediction.
Conclusion: The psoas muscle is a simple, reproducible CT marker for sarcopenia in colorectal cancer. Although PMI was not an independent predictor of PD, a protective trend was observed. Machine learning showed promise for sarcopenia detection but limited value for predicting progression. Larger prospective studies are warranted.
Limitations: Small cohort (n=103), retrospective design, single measurement timepoint
Funding for this study: None
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: The study was approved by the local ethics committee (no. 219/20).

Author Block: Y. Stohldreier¹, M. Dieckmeyer², O. Dietrich¹, S. Schläger¹, J. Seißler¹, U. Ferrari¹, N. Hesse¹, A. Gersing¹; ¹Munich/DE, ²Bern/CH
Purpose: To evaluate associations between metabolic status and proton density fat fraction (PDFF) and second-order texture features of vertebral bone marrow (VBM) and autochthonous (AM) and psoas muscles on magnetic resonance imaging (MRI) in premenopausal women with and without recent gestational diabetes mellitus (GDM).
Methods or Background: Thirty-six premenopausal women (GDM, n=19; healthy controls (HC), n=17; mean age, 36.3±3.9 years) underwent MRI at 11.0±2.4 months postpartum. Multivariable logistic regression models adjusting for age and BMI were used to assess associations between GDM and metabolic-syndrome parameters (triglycerides, HDL, waist circumference, fasting plasma glucose and blood pressure) and differences in muscle and bone marrow quality, measured using texture analysis of PDFF maps.
Results or Findings: PDFF values of the thoracic (PDFF T9-12) and lumbar (PDFF L1-4) VBM were significantly higher in the GDM group compared to HC (PDFF T9-12 41.6±12.3% vs. 33.9±7.6%, p=0.03; PDFF L1-4 46.7±12.4% vs. 39.5±8.1%, p<0.05), whereas AM and psoas showed no significant PDFF difference (p>0.05). After adjustment for age and BMI, PDFF of VBM was significantly associated with GDM status (PDFF T9-12 p=0.02; PDFF L1-4 p=0.03). In addition, several muscle second-order texture features also remained associated with GDM status (dissimilarity AM (OR 4.30 [95% CI 1.61, 16.78], p=0.01), homogeneity AM (OR 0.26 [95% CI 0.07, 0.63], p=0.01), dissimilarity psoas (OR 3.98 [95% CI 1.57, 14.25], p=0.01) and homogeneity psoas (OR 0.29 [95% CI 0.09, 0.71], p=0.01)), indicating a reduced muscle quality in GDM patients.
Conclusion: Women with recent GDM exhibit more heterogenous paraspinal skeletal muscle tissue after adjusting for age and BMI, suggesting a reduced muscle quality in patients with GDM compared to healthy controls, which indicates that muscle PDFF may be a useful biomarker for muscle health in patients with metabolic impairment.
Limitations: Retrospective monocentric study.
Funding for this study: The author(s) declare financial support was received for the research, authorship, and/or publication of this article. This work was funded by the Munich Clinician Scientist Program (MCSP) of the University of Munich (LMU; grant number ACS-10), LMU Klinikum, the German Center for Diabetes Research (DZD), and the Helmholtz Zentrum München.
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: The study was approved by the local institutional review board (Ethics Commission of the Medical Faculty, Ludwig-Maximilians Universität München) and all study participants provided written informed consent prior to their participation in the study, which was conducted in accordance with the declaration of Helsinki.

Author Block: M. Ahmad, D. Amiras, J. Shalhoub, A. Davies, A. G. Rockall; London/UK
Purpose: Vision transformer (ViT) is a recent development in the world of deep learning models (AI) and is an alternative to existing convolutional neural networks. ViTs aim to classify, detect and segment images. Diabetic foot disease (DFD) is a complex disease and is associated with lower limb amputation. Magnetic resonance imaging (MRI) is commonly used in patients with DFD. Our aim was to explore the use of deep learning models to identify predictors on MRI for limb loss in patients with DFD by using sarcoepenia as a potential surrogate marker by stratifying the extent of sarcopenia on MRI images of the foot.
Methods or Background: Two-dimensional images of the foot at the base of the 1st metatarsal were classified as having mild, moderate or severe sarcopenia. A subset of 50 images were also graded by a musculoskeletal radiologist to establish the inter-rater reliability. 824 images were annotated. Following data pre-processing and data augmentation, 1740 images were available for the deep learning models which were split into a 70:20:10 ratio for training:validation:testing. A ViT model was applied to classify the images as mild, moderate or severe.
Results or Findings: The inter-rater reliability was 0.827 [95% CI 0.726-0.928; p-value <0.001]. ViT had an accuracy of 78.7% with an F1 score of 79.9% in classifying sarcopenia severity on two-dimensional MRI pictures. The model had a high precision (81.5%) and recall (78.7%). The confidence threshold could be set to 51% without any deterioration in the model's performance.
Conclusion: ViT is a useful deep learning model in classifying the severity of sarcopenia in patients with diabetic foot disease.
Limitations: Further external validation is required to test the robustness of the model on different datasets.
Funding for this study: Imperial College Healthcare NHS Trust radiology pump priming fund.
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: Ethics approval received from Health Research Authority UK

Author Block: H. Abbasguliyev¹, A. Yalçın², M. S. ATEŞ ÇETİN², F. Alper²; ¹Muğla/TR, ²Erzurum/TR
Purpose: This study aimed to quantify intra-abdominal visceral adipose tissue (VAT) using CT and assess its relationship with BMI. Additionally, we evaluated the association between meniscal damage on knee MRI and both BMI and intra-abdominal fat, while also testing the reproducibility of automated fat quantification.
Methods or Background: From January 2010 to December 2023, patients who underwent abdominal CT and also had knee MRI within one year were retrospectively identified. BMI values were classified into four groups (underweight, normal, overweight, obese). Meniscal signal changes on T2WI were graded in five stages (1, 2a, 2b, 2c, 3). Subcutaneous adipose tissue (SCAT) and VAT volumes were calculated using 3D Slicer software. Correlations between BMI, SCAT, VAT, and meniscal damage were analyzed (p < 0.05).
Results or Findings: Forty-one patients were included (21 men, 20 women; mean age 30 ± 9 years; mean BMI 25.9 ± 4.6). Mean SCAT volume was 260.31 ± 145.74 and VAT volume 315.74 ± 226.96. VAT showed a negative correlation with BMI (r = –0.3, p = 0.01), while SCAT did not (r = 0.2, p = 0.8). Medial meniscus damage correlated strongly with BMI (r = 0.7, p = 0.01), but not with SCAT. Lateral meniscus damage showed no significant association with BMI, SCAT, or VAT (p = 0.6, 0.8, 0.8, respectively).
Conclusion: This study demonstrates a novel link between VAT, BMI, and medial meniscus damage. Automated VAT quantification shows promise as a reproducible index, warranting larger studies to clarify its potential as a complement to BMI.
Limitations: The retrospective design required exclusion of patients with motion artifacts on knee MRI or abdominal CT, as well as those failing to meet at least one inclusion criterion.
Funding for this study: N/A
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: Atatuk University Non-Interventional Clinical Research Ethics Committee Decision No: B.30.2.ATA.0.01.00/306

Author Block: M. PAVOŠEV¹, D. Wieman², J. Menzinga², C. Chimbunde³, F. Ernst⁴, L. Sardinha⁵, G. D. J. Colombo⁶, C. D' Helft⁷; ¹Ljubljana/SI, ²Groningen/NL, ³Bloemfontein/ZA, ⁴Cormondrèche/CH, ⁵Lisbon/PT, ⁶Milan/IT, ⁷Dublin/IE
Purpose: Ultrasound (US) assessment of muscle thickness might be a safe and accessible method for the evaluation of sarcopenia. US can reliably measure muscle thickness during relaxation, but its functional meaning is limited. Therefore, this study investigates whether US-measured upper arm muscle thickness during contraction, incorporating neuromuscular activation, can also be assessed reliably, and how it is correlated to strength across physical activity levels and arm dominance.
Methods or Background: Twenty-four subjects completed a digital survey capturing demographics, sporting habits and arm dominance. US and maximal isometric strength measurements were taken respectively by the SonoScape Doppler E2 and MicroFET2 dynamometer, in Fowler’s position with the arm stabilized in ~135° elbow flexion. Reliability was assessed using the intraclass correlation coefficient (ICC). Correlations were computed for the relaxed and contracted state between muscle thickness and strength for the entire study population, and subgroups based on physical activity level and arm dominance.
Results or Findings: ICC values indicated good to excellent intra-observer (>0.837) and inter-observer (>0.781) reliability in both relaxed and contracted states. Significant correlations (p<0.001) between muscle thickness and strength were observed: rs=0.737 (relaxed) and rs =0.748 (contracted). In the low-to-moderate activity group (n=15), correlations were respectively rs =0.829 and rs =0.850. No significant correlations were found in the high activity group (n=9). For arm dominance (n=21), significant correlations (p<0.001) were found for both dominant and non-dominant arms: rs =0.748 and rs =0.700 (relaxed), rs =0.670 and rs =0.778 (contracted).
Conclusion: Upper arm muscle thickness can be reliably measured using US in both relaxed and contracted states, and shows strong correlation with strength.
Limitations: Population size is limited and age is unevenly distributed, consisting mainly of young adults.
Funding for this study: No funding was received for this study.
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: Joint approval of the University of Ljubljana and UCD

Author Block: P. Platzek, R. Graf, J. S. Kirschke, J. H. W. Bodden; Munich/DE
Purpose: Paraspinal intramuscular fat content is associated with chronic low back pain (CLBP). This study explores the relationship between paraspinal muscle fat infiltration, subcutaneous (SAT), and visceral adipose tissue (VAT) with CLBP using MRI data from the NAKO database.
Methods or Background: Fully automated whole-body segmentations were obtained in all NAKO MRI datasets using an in-house developed segmentation framework. Paraspinal muscle (erector spinae medialis (ESM) and lateralis (ESL), psoas major (PS), and quadratus lumborum (QL)) proton density fat fraction (PDFF, ‰) values were extracted and SAT and VAT volumes (cm3) were obtained. PDFF, SAT and VAT were assessed as standard deviation from the cohort mean. Multivariate logistic regression models were used to investigate associations between SAT, VAT, muscle PDFF, and self-reported CLBP (no CLBP; low, medium, or high CLBP), adjusted for age, weight, and height.
Results or Findings: VAT and SAT were measured in 24,634, and PDFF in ≥22,875 participants. Higher intramuscular fat was associated with greater odds for any degree of CLBP, and chances increased with CLBP severity (low CLBP: OR=1.09-1.22; medium CLBP: OR=1.08-1.28; high CLBP: OR=1.14-1.36; p < .001). PDFF of ESM, ESL, and QL further distinguished participants with non-CLBP (duration <3 months) (OR ≥1.07, p ≤ .001). In contrast, VAT and SAT (p=0.34 and 0.72) were not associated with significantly increased odds for CLBP and VAT was negatively associated with CLBP duration <3 months (OR=0.89, p=0.001).
Conclusion: Paraspinal muscle PDFF predicted CLBP presence and severity, while VAT and SAT did not. Results withstood adjustments for age, height, and weight.
Limitations: Due to fat-water swaps, PDFF measurements were not available in a minority of participants. It remains unclear whether CLBP promotes intramuscular fat accumulation or if intramuscular fat itself contributes to pain and instability.
Funding for this study: Funded by European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (101045128 — iBack-epic — ERC-2021-COG)
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: This study was approved by the TUM institutional review board (593/21 S-NP).

Author Block: C. J. M. L. Edeline, P. Margain, P. Omoumi; Lausanne/CH
Purpose: Assessing muscle volume is crucial for understanding how training, aging, immobilization, and disease affect muscle health. Whole-body Dixon MRI enables robust tissue characterization, but manual segmentation for volume estimation is time-consuming. This study developed and evaluated a deep learning model for automated volume estimation of nine thigh and leg muscle compartments, assessing accuracy versus manual labels and test–retest repeatability.
Methods or Background: Nine compartments were analyzed: posterior, anterior, and medial thigh; sartorius; gluteal muscles; and superficial posterior, deep posterior, anterior tibial, and lateral tibial leg compartments. Whole-body Dixon MRI (1.95×1.95×4.00 mm; ~5 min) from 21 asymptomatic participants (12 males, 9 females; 44.2 ± 16.9 y; BMI 22.3 ± 2.5 kg/m²) from the Lausanne Knee Study were used. Manual segmentations were created on four sequences (fat-sat, water-sat, in-phase, out-of-phase). A 3D U-Net trained on in-phase images with five-fold cross-validation was evaluated for accuracy (MAE, Pearson r) and scan–rescan repeatability; test–retest metrics (ICC and coefficient of variation [CV]) were computed on n = 9 scans.
Results or Findings: High accuracy was observed, with mean absolute errors (MAE) of 4.4–30.4 cm³ (1.1–6.3% of mean volume). Pearson correlations exceeded 0.95 for most compartments (p < 0.001). Test–retest repeatability (n = 9 scans) was excellent: ICC = 0.98–0.99 and CV = 0.33–1.75% across the nine compartments.
Conclusion: A 3D U-Net can accurately and repeatably quantify muscle compartment volumes on whole-body Dixon MRI. The combination of high accuracy and excellent reliability supports application in large-scale cohorts and longitudinal or clinical studies, pending external validation.
Limitations: Limitations include data derived from a single center and a single protocol, as well as the use of an asymptomatic cohort. Therefore, the model should be validated in participants with relevant conditions, such as sarcopenia.
Funding for this study: This work was funded by the Swiss National Science Foundation, Switzerland (SNSF Grant #CRSII5_177155)
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: The study was approved by the Cantonal Ethics Committee for Research on Human Beings (CER-VD).

Author Block: F. Abubacker Sulaiman, A. A. Basheer Ahmed; Chennai/IN
Purpose: To evaluate the feasibility of assessing sarcopenia and body composition using low-dose CT (LDCT) data obtained from routine ^18F-FDG PET/CT scans and to correlate these measurements with clinical parameters and treatment outcomes in oncologic patients.
Methods or Background: A prospective study was conducted on 70 patients who underwent whole-body ^18F-FDG PET/CT for staging or response assessment. Cross-sectional muscle and fat areas were analyzed at the third lumbar vertebral (L3) level using LDCT images reconstructed from PET/CT datasets. Skeletal muscle index (SMI), visceral fat area (VFA), and subcutaneous fat area (SFA) were quantified using semi-automated segmentation software. Sarcopenia was defined using established sex-specific SMI cutoffs. Associations between SMI, BMI, serum albumin, and treatment response were statistically analyzed using Pearson’s correlation and regression models.
Results or Findings: Sarcopenia was identified in 32 patients (45.7%), predominantly among older and low-BMI individuals. Mean SMI values showed strong correlation with serum albumin levels (r = 0.72, p < 0.001) and moderate correlation with treatment response on follow-up PET/CT (r = 0.64, p < 0.01). Patients with sarcopenia exhibited higher incidence of adverse events and poorer metabolic response. LDCT-based measurements were reproducible, with excellent interobserver agreement (ICC = 0.92). Radiation dose from LDCT did not exceed 2.5 mSv per scan.
Conclusion: Low-dose CT from routine PET/CT provides a reliable, opportunistic tool for assessing sarcopenia and body composition without additional imaging or radiation exposure, adding valuable prognostic information in oncology.
Limitations: Single-center design, limited sample size, and absence of longitudinal survival analysis. Larger multicentric studies are warranted for validation.
Funding for this study: No external funding was received for this study.
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: Institutional ethical committee approval was obtained.