Prestudy patient characteristics and specific therapy of all eligible patients with rhabdomyosarcoma entered into Intergroup Rhabdomyosarcoma (RMS) Studies I (IRS-I) (1972 to 1978, n = 686) or II (IRS-II) (1978 to 1984, n = 1,002) were examined for their relationship to survival within each of the four clinical groups using univariate and multivariate analyses. The estimated survival at 5 years from the start of treatment was 56% in IRS-I and 62% in IRS-II (P = .006). The largest survival difference between studies was in patients with group III tumors (52% v 65%). The clinical group was the most important patient characteristic related to survival in both studies. Survival progressively decreased for patients from clinical group I (localized disease, completely resected) to group IV (metastatic disease at the onset). In clinical group I, the only patient characteristic consistently related to survival was histology. Patients with alveolar tumors had the poorest survival, while those with botryoid/embryonal lesions had the best survival. In clinical group II, no characteristic was consistently related to survival. In clinical group III, an orbital primary site was associated with a favorable survival. In clinical group IV, patients with genitourinary tumors had a significant survival advantage. Use of disease-free survival as an end point gave very similar results. This information, from the largest available data base on prognostic indicators in childhood RMS in the context of aggressive multimodal therapies, is being used to plan therapy in the forthcoming study (IRS-IV).