Myeloid-leukemic cells (AML, MDS, CML) can be differentiated to leukemia-derived dendritic cell [DC (DCleu)] potentially presenting the whole leukemic antigen repertoire without knowledge of distinct leukemia antigens and are regarded as promising candidates for a vaccination strategy. We studied the capability of 6 serum-free DC culture methods, chosen according to different mechanisms, to induce DC differentiation in 137 cases of AML and 52 cases of MDS. DC-stimulating substances were cytokines ("standard-medium", "MCM-Mimic", "cytokine-method"), bacterial lysates ("Picibanil"), double-stranded RNA ["Poly (I:C)"] or a cytokine bypass method ("Ca-ionophore"). The quality/quantity of DC generated was estimated by flow cytometry studying (co) expressions of "DC"antigens, costimulatory, maturation, and blast-antigens. Comparing these methods on average 15% to 32% DC, depending on methods used, could be obtained from blast-containing mononuclear cells (MNC) in AML/MDS cases with a DC viability of more than 60%. In all, 39% to 64% of these DC were mature; 31% to 52% of leukemic blasts could be converted to DCleu and DCleu-proportions in the suspension were 2% to 70% (13%). Average results of all culture methods tested were comparable, however not every given case of AML could be differentiated to DC with 1 selected method. However performing a pre-analysis with 3 DC-generating methods (MCM-Mimic, Picibanil, Ca-ionophore) we could generate DC in any given case. Functional analyses provided proof, that DC primed T cells to antileukemia-directed cytotoxic cells, although an anti-leukemic reaction was not achieved in every case. In summary our data show that a successful, quantitative DC/DCleu generation is possible with the best of 3 previously tested methods in any given case. Reasons for different functional behaviors of DC-primed T cells must be evaluated to design a practicable DC-based vaccination strategy.