The clonogenic assay as a reproducible in vitro system to study predictive parameters of treatment outcome in acute nonlymphoblastic leukemia.

Seven institutions studied the cloning pattern of leukemic cells from pretherapy bone marrows of 273 patients with newly presenting and relapsed acute nonlymphoblastic leukemia. The cloning assay was done in all centers using an identical double-layer agar method with a common source of colony-stimulating factor. Cells were incubated for seven days, and clones were identified visually with an inverted microscope. All centers were able to obtain clonal growth in a substantial proportion of patients. Differences in growth pattern were observed between the major contributing center and the pooled results for all other centers. However, an analysis of clinical results suggested that in vitro differences were more likely related to differences in the patient populations than to variability in laboratory technique. The proportion of marrows in which leukemic cells formed colonies (greater than 40 cells) and large clusters (20-40 cells) was greater in relapsed patients than in newly presenting patients (P less than .06). A progressive improvement in induction treatment outcome was seen with decreasing clonal growth. Patients whose marrows did not produce clones had a complete remission (CR) rate of 83% versus 50% for those patients whose marrow leukemic cells formed colonies and/or large clusters (P = .05). In vitro drug sensitivity studies with cytosine arabinoside and adriamycin were performed on bone marrow cells of patients treated with this combination clinically. The percent killing of clonogenic cells in this assay correlated with remission induction outcome. Complete remission was obtained in 73% of 15 patients whose bone marrow leukemic cells showed greater than 30% killing by in vitro drug exposure, and CR was obtained in only 30% of 23 patients whose marrows showed less than 30% killing (P less than .01). The results indicate that the clonogenic assay correlates with treatment outcome and can be used for drug sensitivity testing in patients with acute nonlymphoblastic leukemia. The limitations of the assay are discussed.