Local Failure in Non-Metastatic Ewing Sarcoma Patients Treated with Definitive Radiation Therapy on AEWS1031: A Report from the Children's Oncology Group

Abstract

<h3>Purpose/Objective(s)</h3> To evaluate clinical and treatment variables associated with local failure for Ewing sarcoma (ES) patients treated with definitive radiation therapy (RT) on AEWS1031. <h3>Materials/Methods</h3> AEWS1031 was a randomized phase 3 trial comparing two interval compressed chemotherapy regimens. Local control modality (surgery alone; RT; surgery + radiation) was at the discretion of the treating investigator. Local failure (LF) was defined as LF only or combined LF and distant failure. Standardized uptake values from PET obtained prior to (SUV1) and after neoadjuvant chemotherapy (SUV2) were used to define chemotherapy response. Good PET response was defined as tumor SUV2:1 ≤ 0.5. Event-free survival (EFS) and cumulative incidence of LF from local control in the cohort treated with RT only was estimated using the method of Fine and Gray. <h3>Results</h3> Patients in the cohort (n=159) had 60 pelvis (38%), 24 extremity (15%), 46 spine/chest wall/skull (29%), and 29 (12%) extraosseous tumors. 79% of patients were <18 years. Tumor volume was available in 148 patients with 105 tumors (71%) <200 mL at diagnosis. SUV2:1 ratio was available in 65 patients and 51 (78%) had a good PET response. Central RT plan review was conducted in 155 cases with 6 cases (3.8%) logged as a major protocol deviation. With a median follow up of 67 months after RT, five-year post-local control EFS was 81.8% (95% CI, 74.8–87.1%) and cumulative incidence of LF was 7.8% (95% CI, 4.5-13%). The cumulative incidence of combined local and distant relapse was 0.65% (95% CI, 0.09-4.5%). LF incidence by tumor site was 8.6% for pelvis, 12.5% for extremities, 6.7% for spine/chest wall/skull, and 3.6% for extraosseous tumors (p=0.61). LF incidence was higher for tumors ≥200 mL (14.3%) compared to tumors <200 mL (5.9%, p=0.04). There was no significant difference in LF for good PET responses vs. not (6% vs. 0%, p=0.30) and by patient age (12.9%, ≥18 years vs. 6.5%, <18 years; p=0.32). A major RT protocol deviation was associated with a higher LF incidence of 33.3% vs. 7% for none and/or minor deviations (p=0.03). <h3>Conclusion</h3> Local tumor control for patients treated with definitive RT on AEWS1031 (7.8% LF) was superior compared to prior COG ES studies (15.3% LF from combined INT0091, INT0154, and AEWS0031 analysis). In patients with unresectable primary tumors, PET response, tumor site, and patient age were not associated with LF. Tumor volume ≥200 mL at diagnosis and major RT protocol deviations were associated with higher LF. These data highlight the need for quality assurance when RT is the primary modality of local control.