In therapy study ALL-BFM 81 633 previously untreated patients with acute lymphoblastic leukemia (ALL) less than 18 years of age have been recruited from April 1, 1981 to September 30, 1983 and treated in 37 institutions throughout West-Germany and Austria. Here only therapy results of 611 patients with non-B-ALL are presented. Patients with ALL of B-type are described elsewhere. In this fourth consecutive trial of the BFM study group three major questions have been asked: 1. Is it possible to assess the individual risk for relapse more accurately by the use of a risk factor rather than by the risk score which was the discriminator in studies ALL-BFM 76 and ALL-BFM 79? Does this risk factor discriminate more precisely patients at the highest risk for relapse? Offers more intensive risk-adapted therapy to this patient group a better chance for disease-free survival? 2. In patients at a standard risk for relapse with a risk factor below 1.2--approximately 60% of patients with non-B-ALL--can radiotherapy for prevention of CNS disease effectively be replaced by chemotherapy (intermediate dose Methotrexate)? 3. It is possible to reduce duration of maintenance therapy by 6 months to a total duration of 18 months with no unfavorable effect? To assess the radiation problem in standard risk patients and to evaluate the importance of duration of maintenance therapy two randomisations have been utilized. After a median duration of study ALL-BFM 81 of 4 1/2 years and 3 1/4 years after the study had been closed (date of evaluation January 1, 87) the answers are as follows: 1. For the majority of patients risk-adapted therapy had a curative effect. The probability for event-free survival (EFS) in standard risk patients in slightly above 70%, in medium risk patients 67%. In high-risk patients risk-adapted therapy did not improve prognosis, the EFS being still in the order of 50%. A good assessment of the individual risk for relapse is possible by the newly introduced risk factor. This principle is superior to the risk score used in former studies ALL-BFM 76 and ALL-BFM 79 because a low risk group (risk factor below 0.8) could be identified including approximately 25% of all patients with non-B ALL. Selection, quality, and timing of therapy elements remain the decisive prognostic factors, however. 2. Standard risk patients with a risk factor below 0.8 can effectively be protected for CNS relapse by treatment with intermediate dose Methotrexate.(ABSTRACT TRUNCATED AT 400 WORDS)