The accuracy of drug-sensitivity assays for neoplastic cells depends on the ability to sustain cell survival in vitro, but cells in most cases of acute lymphoblastic leukemia (ALL) rapidly die by apoptosis in culture. We recently reported that allogeneic bone marrow stromal layers support long-term culture of ALL cells. We now describe the standardization of a novel drug-sensitivity assay for ALL lymphoblasts maintained on stromal layers, in which the viability of treated and untreated cells is compared using flow cytometry. Cultures were performed in U-bottomed 96-well plates: 2 x 10 (4) stromal cells per well produced rapidly confluent layers which supported the survival of ALL blasts. In four ALL cases studied, 68.8-106%(median 94.8%) of the lymphoblasts originally seeded were recovered after 7 days of culture, in contrast to< 5% recovered in the absence of stroma. Sensitivity to five antileukemic drugs was tested at the indicated concentrations: vincristine (0.001-1 microgram/ml); dexamethasone (0.001-100 microM); 6-thioguanine (0.078-20 micrograms/ml); teniposide (0.001-10 microM); cytosine arabinoside (0.039-10 microM). After 4 days of culture, cells were labeled with CD19 monoclonal antibody and analyzed by flow cytometry. The concentration of each antileukemic agent producing 50% cytotoxicity (LC50) was determined for each patient by fitting a sigmoid model to the drug concentrations versus percentage viability data. Clinically relevant drug concentrations produced cytotoxic effects, and ALL blast sensitivity varied considerably among patients. We conclude that this is an informative method of systematically evaluating drug sensitivity in ALL patients.