Outgrowth of minimal residual disease (MRD) in acute myeloid leukaemia (AML) is responsible for the occurrence of relapses. MRD can be quantified by immunophenotyping on a flow cytometer using the expression of leukaemia-associated phenotypes. MRD was monitored in follow-up samples taken from bone marrow (BM) of 72 patients after three different cycles of chemotherapy and from autologous peripheral blood stem cell (PBSC) products. The MRD% in BM after the first cycle (n=51), second cycle (n=52) and third cycle (n=30), as well as in PBSC products (n=39) strongly correlated with relapse-free survival. At a cutoff level of 1% after the first cycle and median cutoff levels of 0.14% after the second, 0.11% after the third cycle and 0.13% for PBSC products, the relative risk of relapse was a factor 6.1, 3.4, 7.2 and 5.7, respectively, higher for patients in the high MRD group. Also, absolute MRD cell number/ml was highly predictive of the clinical outcome. After the treatment has ended, an increase of MRD% predicted forthcoming relapses, with MRD assessment intervals of < or =3 months. In conclusion, MRD parameter assessment at different stages of disease is highly reliable in predicting survival and forthcoming relapses in AML.