Automated Pretreatment Thoracic CT-Based Body Composition Analysis Predicts Progression-Free Survival in Head and...

Background/Objectives: To evaluate the prognostic significance of automated, volumetric body composition analysis (BCA) derived from pretreatment thoracic computed tomography (CT) scans in patients with head and neck cancer (HNC). Methods: We retrospectively assessed 160 patients (median age: 63 years; 26.9% women) undergoing primary treatment. BCA quantified various tissue volumes, including bone (B), skeletal muscle (SM), and subcutaneous adipose tissue (SAT). Optimal sex-specific cutoffs for BCA metrics were established via maximally selected log-rank tests. Internal validation of BCA cutoffs was conducted via bootstrap resampling. Kaplan-Meier survival analysis and Cox proportional hazards modeling were used to investigate progression-free survival (PFS). Results: The median PFS for all patients was 51.7 months (95% confidence interval (CI): 31.4-68.8). Among the continuous BCA parameters, only SM/B was significant across the total cohort (hazard ratio (HR): 0.23; 95%CI: 0.12-0.46; p < 0.0001, males (p = 0.0009), females (p = 0.004)). Internal validation of gender-specific cutoffs demonstrated strong-to-intermediate stability for SM/B across both sexes and for SAT/B in males. In contrast, SAT/B exhibited only weak stability among female participants. In univariate PFS analysis, dichotomized SM/B, SAT/B, Union for International Cancer Control (UICC) stage, Eastern Cooperative Oncology Group (ECOG) status, higher body mass index (BMI), normal albumin, and Charlson Comorbidity Index were identified as significant predictors of PFS. Multivariable analysis identified high SM/B (HR: 0.53; 95% CI: 0.3-0.93; p = 0.026) and high SAT/B (HR: 0.58; 95% CI: 0.35-0.95; p = 0.029) as independent prognostic factors, alongside lower UICC stage (p = 0.045) and lower Charlson Comorbidity Index (p = 0.038). Patients with high SM/B and SAT/B ratios had the longest median PFS (65.9 months, 95%CI: 51.7-.), compared to 36.4 months (95%CI: 19.4-.) for high SM/B or SAT/B and 12.6 months (95%CI: 4.2-25.1) for low SM/B and SAT/B (p < 0.0001). Conclusions: Although the BCA parameters SM/B and, to a lesser extent, SAT/B appear to be promising biomarkers, external validation and investigation within well-defined patient subgroups are warranted to establish their generalizability in clinical practice.