Impact of chemotherapy on survival in surgically resected retroperitoneal sarcoma.

BACKGROUND: The role of systemic chemotherapy (CT) in the multimodality treatment strategy for retroperitoneal sarcomas (RPS) remains controversial. We hypothesized that chemotherapy does not improve overall survival for patients with surgically resected RPS. METHODS: The National Cancer Database was used to identify all patients with RPS that underwent surgical resection from 1998 to 2011. Univariate and multivariable Cox proportional hazards modeling were used to assess overall survival (OS) and logistic regression was used for associations. Propensity score (PS) modeling was performed to create balanced cohorts for analysis. RESULTS: A total of 8653 patients with surgically resected RPS were identified; 1525 (17.6%) received CT; 10.6% of patients (n = 163) in the neoadjuvant setting. Factors associated with receipt of CT included moderate (OR 2.3) to poorly differentiated (OR 4.3) tumors, leiomyosarcoma (OR 1.8) or undifferentiated pleomorphic sarcoma (OR 2.3) histology, and R2 resection status (OR 2.2) (all p < 0.05). Unadjusted median OS for patients receiving CT compared to surgery alone was 40 vs 68.2 months respectively (p < 0.01). Following propensity score matching, worse median OS persisted among the CT cohort (40 vs 52 months, p = 0.002). Receipt of chemotherapy was not associated with improved long term survival in adjusted models for the raw and propensity matched cohorts (HR 1.17, 95% CI: 1.04-1.31; p = 0.009). CONCLUSION: Current available chemotherapy regimens for RPS do not confer a survival benefit. Routine use of chemotherapy for RPS should be discouraged until new effective systemic agents become available.