Sui Huang Institute for Systems Biology The single major reason for therapy failure of cancer treatment is recurrence of a drug-resistant tumor which is currently explained by somatic Darwinian evolution of tumor cells. But why is relapse inevitably associated with qualitatively novel, sophisticated traits of increased aggressiveness for which there is no selection pressure? In this talk I will present a systems biology approach and explain, in a permissively simplified manner, the mathematical and molecular basis of the "attractor landscape" established by the gene regulatory network- a modern version of Waddington's epigenetic landscape. The landscape formalisms is more than just a metaphor. In this formal framework cancer arises because cells are "stuck" in side-valleys (cancer attractors), failing to descend to the bottom attractors of the mature states. This concept also predicts how phenotype plasticity can allow tumor cells to quickly acquire coherent, more malignant phenotypes because the latter also represent latent normally not realised attractors on the epigenetic landscape. Applied to tumor progression in response to therapy, this model offers new insights on how tumor cell populations as a whole adapt to therapy stress: Because of inherent non-genetic heterogeneity in cancer cell populations, not all cells die when we treat a tumor. The surviving cells, however, are not just innocent bystanders carrying pre-existing mutations that make them resistant, as postulated by the paradigm of Darwinian selection. Instead, as theory predicts, and our single-cell resolution measurements show, the surviving cancer cells become even more aggressive by mounting a stress-response which corresponds to a transition into preexisting nearby but unoccupied attractor states that encode a stem-like, “ancient” cell phenotype. Thus, rather than following Darwin's "Survival of the fittest", tumor cells follow Nietzsche's "What does not kill me makes me stronger". Specific molecular pathways underlying this fateful behavior and their relevance for new therapeutic strategies that minimize this active cellular adaptation to drugs will be discussed.