Thoracic ct-derived muscle indices improve prediction of respiratory function beyond standard L3 sarcopenia measures in lung cancer patients

Background. L3 skeletal muscle index (SMI) is the CT standard for whole-body sarcopenia but does not assess thoracic musculature and may miss muscle morphology tied to respiratory mechanics and upper-body function. This study evaluated whether thoracic muscle measures provide independent or complementary information beyond L3 SMI for functional and pulmonary performance in lung cancer patients. Methods. Ninety adults with newly diagnosed lung cancer who completed clinically indicated CT imaging and baseline pre-treatment functional assessments were analyzed. SMI was quantified at L3 and thoracic levels (T4, T6, T8, T10). Functional outcomes included handgrip strength (HGS), gait speed, sit-to-stand, and 6-minute walk distance (6MWT). Pulmonary outcomes included maximal inspiratory and expiratory pressures (MIP, MEP), Forced Vital Capacity (FVC), Forced Expiratory Volume in 1 second (FEV1), and Peak Expiratory Flow (PEF). Outcomes were examined continuously and dichotomized at clinically relevant thresholds: weak HGS (< 28 kg for men; <18 kg for women), impaired 6MWT (<300 m), and respiratory muscle weakness (MIP ≤ 62/83 cmH 2 O and MEP ≤ 81/≤ 109 cmH 2 O in men/women). Correlations, multivariable linear regression (adjusting for age, sex, body mass index (BMI), smoking, and tumor stage), and logistic regression assessed associations; incremental predictive value of thoracic SMI beyond L3 was evaluated using area under the ROC curve (AUC). Results. Thoracic SMI showed moderate-strong correlations with HGS, MIP, MEP, and PEF (r = 0.24-0.63, all p < 0.01). In adjusted models, thoracic SMIs remained associated with HGS (beta = 0.20-0.37, p ≤ 0.007) and several respiratory measures, including MEP at T4/T10 and FVC and PEF at T6/T8. L3 SMI was not independently associated with MEP, MIP, or PEF ( p > 0.10), but was more strongly related to whole-body mobility (TUG and gait speed, both p < 0.02). To connect functional associations with clinical relevance, predictive performance was examined for clinically defined weakness or impairment. For inspiratory muscle weakness, L3 SMI was not predictive (p=0.30; AUC=0.903), and adding thoracic SMI did not improve discrimination (AUCs 0.898 to 0.904). For weak HGS, L3 SMI was not significant (p=0.42, AUC=0.774), whereas adding thoracic SMI increased AUC by 0.06 to 0.12, with T8 showing the largest gain (AUC 0.893). For 6MWT impairment, L3 SMI was predictive (p=0.039, AUC =0.750), while thoracic SMI was not (all p>0.10) and did not meaningfully improve increase AUC (0.75-0.79). Conclusion. Thoracic CT-derived muscle indices provide domain-specific value beyond L3 SMI, improving prediction of upper-body strength but not inspiratory muscle weakness or whole-body endurance. Incorporating thoracic muscle metrics with standard L3 assessment may enhance functional risk stratification and prehabilitation planning in lung cancer.