Tumor Size Modifies the Survival Benefit of Chemotherapy in Localized Soft Tissue Sarcomas: A Propensity-Matched...

Background/Introduction: Soft tissue sarcomas (STS) represent a diverse group of rare cancers that have variable responses to chemotherapy. Although tumor size is an established prognostic factor, its influence on the benefit of chemotherapy within specific histologies is not well understood. Methods: We conducted a retrospective analysis of 3890 patients with five STS subtypes using SEER data from 2000 to 2021. Patients were stratified by tumor size (<5 cm, 5-10 cm, >10 cm) and propensity score matched within each subtype-size cohort to control for confounders. Cox regression assessed the impact of chemotherapy on overall survival, with results presented as hazard ratios (HR) and 95% confidence intervals (95%-CI). Inverse probability of treatment weighting (IPTW) was used to improve selection bias. Results: Chemotherapy use in UPS demonstrated worse survival in smaller tumors <5 cm (HR = 2.65, 95%-CI = 1.19-5.92, p = 0.018) and 5-10 cm tumors (HR = 1.45, 95%-CI = 1.03-2.04, p = 0.031). In larger UPS tumors (>10 cm), a directionally protective association observed in matched analysis attenuated after inverse probability of treatment weighting (IPTW) (HR = 0.82, 95%-CI = 0.60-1.12, p = 0.211). Fibromyxosarcoma 5-10 cm tumors demonstrated worse survival with chemotherapy (matched HR = 3.74, 95%-CI = 2.30-6.10, p < 0.001), which remained consistent after IPTW (HR = 4.47, 95%-CI = 2.63-7.60, p < 0.001), along with >10 cm tumors (IPTW HR = 2.16, 95%-CI = 1.07-4.34, p = 0.031). DDLPS >10 cm tumors demonstrated a directionally harmful association (HR = 1.49, 95%-CI = 0.96-2.29, p = 0.073). Synovial sarcoma 5-10 cm tumors demonstrated a directionally protective trend that remained statistically non-significant across analyses. Conclusions: The effect of chemotherapy on survival in localized STS depends on both histologic subtype and tumor size. However, subgroup estimates with confidence intervals approaching 1.0 should be interpreted cautiously.