Purpose We evaluated 8-year survival and late neuropsychologic toxicity in children with acute lymphoblastic leukemia treated in a randomized clinical trial to test whether hyperfractionated (twice daily) cranial radiation therapy (CRT) can reduce incidence and severity of late toxicities associated with 18 Gy of CRT. Patients and Methods Between 1987 and 1995, 369 children treated on two consecutive Dana-Farber Cancer Institute Consortium protocols for high-risk acute lymphoblastic leukemia were randomly assigned to conventionally fractionated CRT (CFX) or hyperfractionated CRT (HFX) to a total dose of 18 Gy. Neuropsychologic testing was completed for 125 of 287 children in continuous complete remission. Event-free and overall survival, as well as neuropsychologic function, were compared for the two arms of the protocol. Results Eight-year event-free survival ( SE) was 80% 3% for children randomly assigned to CFX and 72% 3% for HFX (P .06). Overall survival was 85% 3% for CFX and 78% 3% for HFX (P .06). CNS relapses occurred in 2.8% of patients receiving CFX and 2.7% receiving HFX (P .99). Cognitive function for both groups was solidly in the average range, with no group differences in intelligence, academic achievement, visuospatial reasoning, or verbal learning. Children on the HFX arm exhibited a modest advantage for visual memory (P .05). Conclusion HFX provides no benefit in terms of cognitive late effects and may compromise antileukemic efficacy. HFX should not be substituted for conventionally dosed CRT in children who require radiation therapy for treatment of acute lymphoblastic leukemia.