By using rapid flow cytometric techniques capable of detecting one leukemic cell in 10(4) normal cells, we prospectively studied minimal residual disease (MRD) in 195 children with newly diagnosed acute lymphoblastic leukemia (ALL) in clinical remission. Bone marrow aspirates (n = 629) were collected at the end of remission induction therapy and at 3 intervals thereafter. Detectable MRD (ie, > or = 0.01% leukemic mononuclear cells) at each time point was associated with a higher relapse rate (P < .001); patients with high levels of MRD at the end of the induction phase (> or = 1%) or at week 14 of continuation therapy (> or = 0.1%) had a particularly poor outcome. The predictive strength of MRD remained significant even after adjusting for adverse presenting features, excluding patients at very high or very low risk of relapse from the analysis, and considering levels of peripheral blood lymphoblasts at day 7 and day 10 of induction therapy. The incidence of relapse among patients with MRD at the end of the induction phase was 68% +/- 16% (SE) if they remained with MRD through week 14 of continuation therapy, compared with 7% +/- 7% if MRD became undetectable (P = .035). The persistence of MRD until week 32 was highly predictive of relapse (all 4 MRD(+) patients relapsed vs 2 of the 8 who converted to undetectable MRD status; P = .021). Sequential monitoring of MRD by the method described here provides highly significant, independent prognostic information in children with ALL. Recent improvements in this flow cytometric assay have made it applicable to more than 90% of all new patients. (Blood. 2000;96:2691-2696)