BACKGROUND During the fourth Intergroup Rhabdomyosarcoma (RMS) Study (IRS IV, 1991-97), a preoperative staging system was evaluated prospectively for the first time. The authors evaluated this staging system and the role of surgery in extremity RMS in contemporary multimodal therapy. METHODS A total of 139 patients (71 girls; median age, 6 years) were entered in IRS IV with extremity-site RMS. Stage was assigned by the IRSG Preoperative Staging System. Postsurgical group was determined by tumor status after initial surgical intervention. Multivariate analysis was performed using all pretreatment factors that were significant by univariate analysis, including clinical Group (i.e., I through IV), tumor invasiveness (T1,T2), nodal status (N0,N1), and tumor size (< or > or =5 cm). Failure-free survival rates (FFS) and survival rates were estimated using the Kaplan and Meier method. RESULTS Preoperative staging and clinical group distribution were as follows: Stage 2, n = 34; Stage 3, n = 73; Stage 4, n = 32; Group I, n = 31; Group II, n = 21; Group III, n = 54; Group IV, n = 33. Three-year FFS was 55%, and the overall survival rate was 70%. Eighty-seven patients had either unresectable, gross residual disease (Group III) or metastases (Group IV). FFS was significantly worse for these patients with advanced disease, compared with that for patients with complete resection or with only microscopic residual tumor (i.e., Group I or II; Group I, 3-year FFS, 91%; Group II, 72%; Group III, 50%; Group IV, 23%; P<.001). Lymph nodes were evaluated surgically in 76 patients with positive results in 38. Clinically, 13 additional patients had nodal disease. Both stage and group were highly predictive of outcome and were highly correlated. By multivariate analysis, none of the other variables were predictors of FFS. CONCLUSIONS This review confirms the utility of pretreatment staging for stratification of patients with extremity RMS with widely different risks of relapse, thereby paving the way for development of risk-based therapy. Group (operative staging) remains the most important predictor of FFS, emphasizing the importance of complete gross resection at initial surgical intervention, when feasible without loss of limb function. The high incidence of nodal disease in the patients who had lymph node biopsy confirms the need for surgical evaluation of lymph nodes to ensure accurate staging in children with extremity rhabdomyosarcoma.