Evaluation of local control in patients with non-metastatic Ewing sarcoma of the bone: A report from the Children's Oncology Group
10013 Background: Options for local control in patients (pts) with Ewing sarcoma (EWS) include surgery (S), radiation (R), or surgery plus radiation (S+R). The choice of local control depends on factors that also impact outcome in EWS, including tumor site, tumor size, and age. Our objective was to determine if risk of disease progression differs between local control methods, after controlling for potential confounders. Methods: We analyzed the cohort of pts with non-metastatic EWS of the bone treated on Intergroup Study-0091 (Grier et al, NEJM, 2003). Local control decisions were made on an individual basis. We excluded pts with tumors of the skull/face, progression prior to end of local control, or incomplete local control data. Disease progression was recorded from end of local control. Cumulative incidence of disease progression was determined using a competing risks approach. We constructed a Cox proportional hazards model of progression free survival incorporating local control mode and potential confounders into the model. Secondary analyses evaluated overall survival, local failure, and distant failure. Results: 329 pts met eligibility criteria for analysis. 122 pts received S, 142 received R, and 65 received S+R. The cumulative incidence of disease progression at 5 years was 22.1% (95% CI 15.1–29.9%) for S, 36.9% (29.0–44.9%) for R, and 48.1% (35.5- 59.7%) for S+R. Using R as the reference group and controlling for age, tumor size, tumor location, and chemotherapy regimen, the hazard ratio for progression was 0.66 (95% CI 0.38–1.13) for S and 1.55 (0.96–2.49) for S+R. The hazard ratio for death was 0.77 (95% CI 0.44–1.34) for S and 1.46 (0.88–2.42) for S+R. The hazard ratio for local failure was 0.38 (95% CI 0.15–0.98; p=0.04) for S and 0.77 (0.34–1.75) for S+R. The hazard ratio for distant failure was 0.78 (0.42–1.46) for S and 1.60 (0.91–2.82) for S+R. Conclusions: Observed differences in outcome between local control groups are largely due to confounding factors that affect outcome and local control choice. Risk of disease progression, distant failure, or death does not differ between pts who receive S or R. Pts who receive S have a decreased risk of local failure compared to pts who receive R. No significant financial relationships to disclose.