The aim of the present study was to analyze the incidence of AML cases displaying more than one blast cell subpopulation by immunophenotype at diagnosis, since, any of them, although minimal, can be responsible for the relapse. For this purpose we have prospectively investigated the immunophenotype of blast cells from 40 de novo AML patients at diagnosis with a large panel of monoclonal antibodies in double and triple staining combinations analyzed at flow cytometry. The discrimination between the different cell populations was based on: (1) the existence of aberrant phenotypes; (2) differences in light-scatter characteristics; and (3) the expression of differentiation-associated antigens (CD34, CD117, HLADR, CD33, CD15, CD14, CD11b and CD4). More than one blast cell subpopulation was identified in 34 patients (85%), two subpopulations in 12 patients (30%), three in three cases (7.7%), four in 13 patients (32.5%) and five populations in six cases (15%). The most common criteria for discrimination of blast cell subpopulations was based on the expression of maturation-associated antigens and, interestingly, the blast subpopulations defined by higher reactivity for myeloid differentiation-associated markers had a more mature FSC/SSC pattern. In 53% of the patients at least one of the subpopulations identified was minimal (< 10% of the total leukemic cells). Regarding the existence of aberrant phenotypes three situations were observed: (1) none of the subpopulations had antigenic aberrations (10 cases); (2) coexistence of normal and aberrant subpopulations (five cases); and (3) all the subpopulations displayed aberrant phenotypes (19 cases). In 17 of the 23 patients (74%) who had two or more blast cell subpopulations with phenotypic aberrations, at least one aberrant criteria was common to all the subpopulations; this criteria by itself would permit the simultaneous identification of all subpopulations in minimal residual disease (MRD) studies. In the remaining cases the investigation of MRD should be based on the phenotypic characteristics of each subpopulation.