Advances in the molecular and immunologic characterization of leukemic cells have greatly aided the diagnosis and risk assignment of ALL, as well as the monitoring of bone marrow samples for minimal residual disease. Currently, 75% of childhood cases have biologically and therapeutically relevant genetic abnormalities. Although gene discoveries in ALL have not been directly translated into effective therapy, there is every reason to believe that this disease will eventually yield to molecular intervention. In the meantime, efforts are being made to enhance the efficacy of existing regimens while reducing their toxic side effects. We have learned, for example, the following: high-dose methotrexate is more effective than lower-dose methotrexate, especially for T-cell ALL; patients who need drastic adjustment of mercaptopurine dosage due to thiopurine S-methyltransferase deficiency can be prospectively identified; dexrazoxane (ICRF-187) could reduce anthracycline cardiotoxicity; granulocyte colony-stimulating factor can shorten hospital stays for febrile neutropenia after intensive remission induction therapy; and prolonged low-dose epipodophyllotoxin treatment may reduce the risk of therapy-induced acute myeloid leukemia without compromising treatment efficacy. The challenge now is to identify specific treatments for genetically defined subtypes of ALL.