Density dependence and evolvability limit adaptive therapy in non-small cell lung cancer mouse model

Abstract

Abstract Using NSG mice grafted with H3122 non-small cell lung cancer cells, we compared outcomes from (1) continuous therapy with alectinib, (2) adaptive therapy where alectinib was cycled on and off based on a mouse’s tumor burden, and (3) no treatment. Adaptive therapy has proven successful in mouse models and clinical trials. The success of adaptive therapy improves when either there is density-dependent feedbacks favoring competition by sensitive over resistant cancer cells when therapy is off, or when the rate at which the population of cancer cells becomes more sensitive to therapy during periods of no treatment is faster than the rate at which resistance increases when on therapy. We found tumor growth rates were lowest under continuous therapy and highest under no therapy with adaptive therapy in between. We fitted the data to three separate game-theory models of resistance evolution where we let resistance be a quantitative trait consistent with resistance mechanisms to alectinib. All models successfully identified why adaptive therapy proved less successful than continuous: there was no evidence for favorable density-dependent feedbacks; and there was no significant difference in the rate of evolution towards or away from resistance when therapy was on or off.