To evaluate the toxicity and tumor efficacy of the halopyrimidine IUdR (NSC #39661, IND 22475) as a chemical modifier of radiation response when used in a high dose short time infusion versus the acceptable 4 day infusion. In August 1987 we initiated a prospective study in patients with newly diagnosed anaplastic astrocytoma and glioblastoma. The study was designed to have a fixed dose of radiation (60.16 Gy = 1.88 Gy in 32 fractions in 6.5 weeks) but varying the dose schedule of IUdR, keeping the total dose between 21 and 24 g/m2. IUdR was delivered in a 96, 48, or 24 hr continuous intravenous infusion per week for 6.5 weeks during radiation treatment. The study was closed for patient accrual on October 1, 1991. Twenty-two patients were treated on the 96 hrs, 32 on the 48 hr and 25 on the 24 hr schedules. The incidence of glioblastoma ranged between 68 and 75% in the three arms. Seventy percent of the patients had a Karnofsky of 80-90% at the onset of treatment. Over 50% of the patient population were under age 55. Drug tolerance was related to the duration of the IUdR infusion. Toxicities were most pronounced in the 96 hr IUdR infusion schedule where 27.4% of the patients reported a grade 3 drug toxicity. No fatal or grade 4 toxicities were observed. More patients on the 24 and 48 hr schedule received at least 80% of the IUdR dose specified per protocol. We did not observe a trend in acute normal tissue radiation reactions in any of the three arms. The median survivals calculated from the Kaplan-Meier plot are 13.4, 10.5, and 11 months, respectively, for the 96, 48, and 24 hr infusions. The Cox Proportional Hazards model showed that any difference in survival can be attributed to histological grade, type of previous surgery and, to some extent, age of the patient. Dose schedule was not a significant predictor of survival, although statistically nonsignificant trend toward longer survival is seen in those patients with glioblastoma treated in the "long" 4 day schedule. Overall, our treatment combination, particularly for patients with glioblastoma, has not shown convincing evidence of an improvement in survival. Of interest, however, it is the 2 year survival rate of 68% for patients with anaplastic astrocytoma. In our experience, the administration of IUdR is laborious, time consuming and with bothersome acute gastrointestinal and hematological toxicities.