Autologous stem cell transplantation using marrow or peripheral blood is routinely used to consolidate patients with acute myelocytic leukemia in complete remission. The situation is less clear for adult acute lymphocytic leukemia in which results achieved with all strategies are disappointing. In acute myelocytic leukemia, autografts should be done in patients with good and standard risk factors. Patients with high-risk acute myelocytic leukemia defined by poor cytogenetics or failure to achieve remission with the first induction course, should proceed to allogeneic stem cell transplantation with the best available human leukocyte antigen-identical donor (family or unrelated), and the nature of the conditioning regimen (myelo-ablative or non-myelo-ablative) should be decided in relation to age, and the patient's clinical condition. Results of autografting in acute myelocytic leukemia rely strongly on the quality of the graft. Higher doses of infused stem cells translate into lower relapse rates. Marrow purging with cyclophosphamide derivatives also diminishes the relapse incidence. Autologous stem cell transplantations using peripheral blood are presently preferred to marrow as the source of stem cells, but an aggressive prior in vivo purge (high-dose consolidation course(s) before cytaphereses) is then mandatory. In good-risk acute myelocytic leukemia, autografting is superior to high-dose ARA-C; in standard-risk acute myelocytic leukemia, both are supposedly equivalent. There is no prospective randomized study testing the two approaches in the good-standard-risk population. We presently test the combination of marrow and blood both purged by mafosfamide. In adult acute lymphocytic leukemia, good-risk patients get the best benefit from autografting over conventional chemotherapy. Maintenance chemotherapy after transplant is likely to bring benefit. Research in progress aims at facilitating access of the largest number of patients to autografting and at introducing posttransplant immunomodulation maneuvers such as tumor vaccination. (C) 2002 Lippincott Williams Wilkins, Inc.