CNS-Directed Therapy for Childhood Acute Lymphoblastic Leukemia: Childhood ALL Collaborative Group Overview of 43 Randomized Trials

A collaborative meta-analysis was performed to clarify the relative effects on relapse and survival of different types of therapies directed at the CNS in childhood acute lymphoblastic leukemia. Data were sought for each individual patient in all trials started in or before 1993 that included unconfounded randomized comparisons of such treatments. Log-rank survival analyses were performed for each trial, and overall results for groups of trials addressing similar questions were obtained from the totals of the observed minus expected number of events and their variances. Radiotherapy and long-term intrathecal therapy gave similar outcomes, with no significant difference in event-free survival despite random assignment of treatment to 2,848 patients, 1,001 of whom suffered relapse or death. Intravenous methotrexate reduced non-CNS rather than CNS relapses, and hence, the addition of intravenous methotrexate to a treatment regimen including radiotherapy or long-term intrathecal therapy improved event-free survival, with a 17% reduction in the event rate (95% confidence interval, 6% to 27%; P =.003). The event-free survival at 10 years in these trials was 61.9% without intravenous methotrexate and 68.1% with intravenous methotrexate. There was no significant difference in survival (14% death rate reduction; P =.09). There were insufficient randomly assigned patients to adequately address other questions, such as effect of different doses. No evidence was found of differences, between trials or between subgroups of different types of patients, in the relative effects of treatment. Radiotherapy can be replaced by long-term intrathecal therapy. Intravenous methotrexate gives some additional benefit by reducing non-CNS relapses.