Optimizing Prostate Cancer Treatment: High-Frequency Adaptive Therapy

We utilize game-theoretic multi-population models of prostate cancer to show how a high-frequency treatment via adaptive therapy can significantly increase life expectancy while drastically reduce the necessary dosage of chemotherapeutic agents. Rather than pursuing an extended treatment regimen aimed at maximal tumor reduction, our approach focuses on stabilizing cancer size and effectively arresting cancer progression through extreme brief, concentrated treatment episodes interspersed with short resting intervals. Here we consider two distinct evolutionary scenarios: one in which drug-resistant cells are present from the beginning, and another in which the drug-resistant cells can spontaneously emerge through stress-induced mutagenesis. We present analytical estimations and justifications for high-frequency adaptive therapy, and provide proposals for how it can be implemented through open-loop observations of patients in clinical practice. In addition to our analytical investigations, we introduce a Bayesian optimization search to efficiently navigate the intricate landscape of cancer treatment strategies, i.e. exploring the parameter space associated with potential multi-drugs treatment policies to maximize life expectancy.