The most important prognostic factor in current clinical management of acute myeloid leukemia (AML) is the leukemic karyotype, identifying patients with relatively favorable, intermediate and adverse prognosis. Unfortunately, the majority of patients are assigned to the intermediate risk group although they represent a broad spectrum regarding prognosis. While molecular mapping may diminish this problem in the future, an approach to prognostic stratification of patients within the intermediate risk group is urgently needed in order to develop risk-adapted treatment strategies and evaluate the efficacy of new treatments. The aim of this retrospective study was to evaluate the prognostic impact of diagnostic multiparametric flow cytometry (MFC) in the cytogenetic intermediate risk group of AML patients. Diagnostic samples were obtained from 162 AML patients (median age 66 years, range 0–88; cytogenetic risk group favorable/intermediate/adverse/unknown 11/91/17/43; bone marrrow/blood 141/21). Four-color flow cytometry and a panel of 35 monoclonal antibodies was used to characterize the immunophenotype of AML blast cells, defined by side scatter and CD45 expression. For each antigen in the panel, mean fluorescence intensity (MFI) of the blast cell population was recorded. For each antigen, the expression pattern was characterized as homogeneous or heterogeneous. Heterogeneous antigen expression was defined as the presence of two or more subpopulations, within the blast cell gate, expressing the antigen in question with distinctly different intensities. Flow cytometry data, biological and clinical variables for the cytogenetic intermediate risk group and for patients with unknown cytogenetic risk were analysed by univariate Cox regression analysis and selected variables were analysed by multivariate Cox regression analysis (backward likelihood-ratio statistics, p<0.05 considered significant) with overall survival as outcome variable. Four conventional variables and four blast cell phenotype variables significantly predicting inferior survival were identified: FAB type other than 0/1 (p<0.03), presence of extramedullary disease (p<0.007), high white blood count (p<0.0005) and high age (p<0.0005); a heterogeneous CD117 expression (p<0.009), heterogeneous CD90 expression (p<0.03) as well as high MFI of CD56 expression (p<0.0005) and low MFI of CD58 expression (p<0.001) for CD45/SSC gated blast cells. The prognostic model identified two subsets of cytogenetic intermediate risk patients with a median overall survival of 0.1 and 1.5 years respectively. The subset of cytogenetic intermediate risk patients with high risk according to the prognostic model had an outcome inferior to that of the cytogenetically defined adverse risk group. This prognostic model was also valid for patients with unknown cytogenetic risk and for patients assigned to the favorable and adverse risk groups by cytogenetics. Informative MFC was applicable in 96% of the cases analysed. The prognostic model should be evaluated in a prospective clinical multicentre trial.