An experimental scheme is suggested for determining optimal combinations of two drugs. A first experiment uses three doses, chosen to cover the effective range (assumed known), of each of the pure drugs, and three corresponding doses of each of three mixtures. A second experiment is conducted in the region defined by the observed and estimated optimal treatments from the first experiment, and a quadratic surface fitted to selected results of the second experiment, or of both experiments, is used to obtain a revised estimate of the optimal treatment; if this estimate lies outside the range of treatments used in the second experiment, a third should be performed. An example is given, using data derived from life-spans of leukemic mice treated with Amethopterin and Terephthalanilide
Sequences of Combination Chemotherapy Experiments
Addelman, S., Gaylor, DW., & Bohrer, RE. (1966). Sequences of Combination Chemotherapy Experiments. Biometrics, 22(4), 730-746.