Modeling of in vitro drug activity and prediction of clinical outcome in acute myeloid leukemia.

Abstract

The objectives of this study were to develop a population pharmacodynamic model describing the in vitro drug sensitivity of tumor cells and to relate in vitro parameters to clinical outcome. Cell samples from 179 patients with acute myelocytic leukemia were exposed to cytosine arabinoside and daunorubicin, and cytotoxicity was analyzed using the fluorometric microculture cytotoxicity assay. A sigmoid E(max)-model for daunorubicin and an E(max)-model for cytosine arabinoside described the data. The model predicted drug potency (EC(50)) adequately from 1 concentration measurement. A logistic regression on individual in vitro parameters of 46 patients treated with the daunorubicin plus cytosine arabinoside regimen showed that the probability of complete response was significantly (P < .05) related to the product of the E(max)/EC(50) ratio of the two drugs. The findings demonstrate the value of population pharmacodynamic modeling of in vitro drug sensitivity data and a significant relationship between the in vitro parameters and clinical outcome.