The extent of clearance of leukemic cells from the blood or bone marrow during the early phase of therapy is an independent prognostic factor in acute lymphoblastic leukemia (ALL). Several methods are available to measure the minimal residual disease (MRD) remaining after initial intensive chemotherapy. The most promising are flow cytometric detection of aberrant immunophenotypes and polymerase chain reaction analysis of clonal antigen-receptor gene rearrangements. When applied together, these techniques enable one to monitor MRD in virtually all cases of ALL. Patients who achieve an 'immunologic' or 'molecular' remission (ie leukemic involvement of <0.01% of nucleated bone marrow cells at the end of remission induction therapy) are predicted to have a better clinical outcome than patients whose remission is defined solely by morphologic criteria. In studies to date, patients with MRD at a level of 10(-2) or more at the end of induction have fared almost as poorly as those with > or =5% blast cells in the bone marrow (ie induction failures). Sequential monitoring of MRD can improve the clinical utility of risk assessment still further. Additional studies are needed to determine the critical levels of MRD at various times of treatment and whether therapeutic intervention based on MRD findings can improve clinical outcome.