BACKGROUND: The local control approach for girls with non-resected vaginal rhabdomyosarcoma (RMS) enrolled onto Intergroup RMS Study Group (IRSG)/Children's Oncology Group (COG) studies has differed from that used at other primary sites by delaying or eliminating radiotherapy (RT) based on response achieved with chemotherapy and delayed primary resection. PROCEDURES: We reviewed locoregional treatment and outcome for patients with localized RMS of the vagina on the two most recent COG low-risk RMS studies. RESULTS: Forty-one patients with localized vaginal RMS were enrolled: 25 onto D9602 and 16 onto Subset 2 of ARST0331. Only four of the 39 with non-resected tumors received RT. The 5-year cumulative incidence of local recurrence was 26% on D9602, and the 2-year cumulative incidence of local recurrence was 43% on ARST0331. Increased local failure rates appeared to correlate with chemotherapy regimens that incorporated lower cumulative doses of cyclophosphamide. Estimated 5-year and 2-year failure free survival rates were 70% (95% CI: 46%, 84%) on D9602 and 42% (95% CI: 11%, 70%) on ARST0331, respectively. CONCLUSIONS: To prevent local recurrence, we recommend a local control approach for patients with non-resected RMS of the vagina that is similar to that used for other primary sites and includes RT. We recognize that potential long-term effects of RT are sometimes unacceptable, especially for children less than 24 months of age. However, when making the decision to eliminate RT, the risk of local recurrence must be considered especially when using a chemotherapy regimen with a total cumulative cyclophosphamide dose of = 4.8 g/m(2).