Sensitivities to drugs acting on cells in culture can be measured as dose-response curves, provided a quantitative assay is available for a relevant cell function. We have used two such assays in the study of the blast cells of acute myeloblastic leukemia. Colony formation in culture with methylcellulose detects principally terminal divisions, while growth of clonogenic cells in suspension reflects self-renewal. In a previous study different cytosine arabinoside and 5-azacytidine dose-response curves were obtained with the two assays. For the former the slope of the dose-response curve measured in suspension was steeper than that obtained using the clonogenic assay. For the latter, 5-azacytidine, the relationship between sensitivity in suspension and in methylcellulose was reversed. Further, for cytosine arabinoside, sensitivity in suspension but not in methylcellulose was associated with successful remission-induction. In this article we report an association between 5-azacytidine sensitivity in suspension and successful remission induction, for patients treated only with high-dose cytosine arabinoside. There was no correlation between the 5-azacytidine dose-response curve in methyl-cellulose and clinical outcome. A model is presented that may explain these findings, based on the hypothesis that there are genetic mechanisms responsible for blast cell renewal.