A national childhood acute lymphoblastic leukemia (ALL) study was initiated in Israel in 1984 with the aim of improving results in difficult aspects of treatment including: high-risk groups, the problems of late relapses, and the effect of cranial irradiation for CNS prophylaxis in leading to late neuropsychiatric sequelae and secondary tumors. Induction of chemotherapy with a combination of 6 drugs (vincristine, cyclophosphamide, cytosine arabinoside, adriamycin, prednisone and L-asparaginase), followed by intensification with methotrexate and L-asparaginase, was introduced in both the usual and the high-risk groups. In a selected group with better prognostic factors, therapy was reduced. In an attempt to minimize the sequelae of CNS prophylactic therapy, cranial irradiation was omitted in half the patients and intrathecal (IT) triple therapy was given instead. Following 2 years of unsatisfactory preliminary results in a very high-risk group (VHR; non-T- and T-cell leukemia with WBC counts of greater than 100,000 and greater than 20,000, respectively), treatment was modified and intensified. Between Nov. 1984 and Feb. 1989, 143 patients were enrolled from 10 hospitals. During follow-up of a median of 2.5 years, there were 32 failures (2 failed remissions, 27 relapsed and 3 died of bleeding and sepsis). 107 patients are alive in first remission and an additional 8 in second and third remissions. By Kaplan-Meier life table analysis, the rates of leukemia-free interval (LFI) and event-free interval (EFI) for 4 years were 60% and 57%, respectively. Improved LFI results of 71% for 4 years were achieved in a group with non-T-cell ALL with WBC less than 100,000 (the largest group, 65% of the patients). In the small "good risk" group (10% of patients), and the T-cell group with WBC less than 100,000, LFI for 4 years were 56% and 54%, respectively. In the VHR group, modification seemed to have improved results: LFI of 41% for 3 years. CNS prophylaxis with IT triple therapy was as effective as cranial irradiation in the standard risk group. In 1 out of 33 children on this protocol a single CNS relapse occurred, as compared to 2 out of 35 matched controls with cranial irradiation. These results warrant extension of IT triple therapy to higher risk groups of childhood ALL. As for systemic treatment, increased up-front high-dose intensive therapy is recommended for all groups with ALL, but with reduction of cumulative dose to minimize late effects.