Beginning in 1984, we started a prospective study to evaluate the role of postoperative, pre-radiation chemotherapy in the treatment of infants and poor prognosis children with medulloblastoma. The study was designed to evaluate the role of pre-radiation chemotherapy in two specific patient populations: (a) children under the age of 2 years in which there was an attempt to delay definitive radiation and thus reduce the risk of toxicity to the developing nervous system; and (b) children over age 2 years with Stage T3 and T4 disease who were known to have a relatively poor prognosis with surgery and radiation. The five patients under age 2 years received cisplatinum (100 mg/m2) every 3 weeks and weekly vincristine (1.5 mg/m2) for a total of 9 weeks. Nitrogen mustard (6 mg/m2), procarbazine (100 mg/m2), and vincristine (1.5 mg/m2) (MOP) were given in 28 day cycles as long as there was no disease progression or until the child's second birthday, at which time the children were referred for radiation therapy. The 13 patients over 2 years of age received the 9 week course of cisplatinum and vincristine and then began radiation. Responses measured by computed tomography were obtained in 10 of 12 children with measurable disease at the start of chemotherapy. With a median follow-up of 22 months, 15 of 18 children were alive and free of disease. Except for mild ototoxicity in one child, the acute side effects have been well tolerated. In conclusion, it appears that some infants can have their radiation delayed until the age of 2 years. Although the follow-up time was short, all but three patients were free of disease, time exceeding the median time to failure with radiation alone. Pre-radiation chemotherapy might improve local control and survival in children with advanced stage medulloblastoma.