Survival after relapse in children and adolescents with rhabdomyosarcoma: A report from the intergroup rhabdomyosarcoma study group

Background: Despite advances in therapy, nearly 30% of children with rhabdo myasarcoma experience progressive or relapsed disease, which is often fatal . Patients and Methods: To facilitate the development of a retrieval therapy protocol, we studied potential risk factors that were predictive of surviva l after first relapse in 605 children who were enrolled onto three consecut ive Intergroup Rhabdomyosarcoma Study Group protocols, Results: The median survival time from first recurrence was 0.8 years; the estimated percentage of patients who survived 5 years from first recurrence was 17% +/- 2% (mean +/- SE). Univariate analysis showed that tumor histol ogy was an important predictor of 5-year survival (P <.001): the 5-year sur vival rate was 64% for patients with botryoid tumors (n = 19), 26% for pati ents with embryonal tumors (n = 313), and 5% for patients with alveolar or undifferentiated sarcoma (n = 273). Further analysis identified prognostic factors within histologic subtypes (P <.001). For patients with embryonal t umors, the estimated 5-year survival rate was 52% for patients who initiall y presented with stage 1 or group I disease, 20% for those with stage 2/3 o r group II/III disease, and 12% for those with group IV disease. For patien ts with stage 1/group I disease, estimated 5-year survival rates were highe r for patients with local (72%) or regional (50%) recurrence than for those with distant (30%) recurrence. Among patients with alveolar or undifferent iated sarcoma, only the disease group predicted outcome: the 5-year surviva l estimate was 40% for group I versus 3% for groups II through IV. We ident ified a "favorable risk" group (approximately 20% of patients) whose 5-year estimated survival rate was near 50%; for all other patients, the estimate d survival was near 10%. Conclusion: This analysis demonstrates that the probability of 5-year survi val after relapse for rhabdomyosarcoma is dependent on several factors at t he time of initial diagnosis, including histologic subtype, disease group, and stage. These findings will form the basis of a multi-institutional risk -adapted relapse protocol for childhood rhabdomyosarcoma. (C) 1999 by Ameri can Society of Clinical Oncology.