Project Summary Prostate cancer (PCa) is the most prevalent cancer in men in the US. A major obstacle in PCa therapy is that continuous treatment at maximum tolerable doses often renders the tumor resistant. PCa is comprised of androgen-independent cancer stem cells (PCaSC) and more differentiated, androgen-dependent PCa cells (PCaC) that make up the bulk of the tumor. Treatment-induced enrichment in PCaSC appears to confer therapy resistance. Continuous treatment neglects the evolutionary dynamics where competition, adaptation and selection between treatment-sensitive and -resistant cells contribute to therapy failure. Intermittent androgen deprivation therapy (IADT) with on-and off-treatment cycles may counteract competitive release of androgen-independent cancer cells and delay time to progression (TTP). Successful clinical implementation of IADT requires identification of resistance mechanisms, prediction of responses, and determination of clinically actionable triggers for pausing and resuming IADT cycles. We propose to integrate our mathematical, biological, clinical, and statistical expertise to test the hypothesis that PCaSC dynamics underlie response to therapy and evolution of resistance in IADT. By fitting different mechanistic mathematical models to retrospective longitudinal data of individual patients in a training data set we can determine clinically plausible model parameter distributions. From treatment response dynamics in early treatment cycles, we aim to simulate and reliably forecast an individual patient's response to subsequent treatment cycles in a validation data set. Then, we will use the validated model to simulate IADT protocols with different cycle intervals for each patient. Nominal and relative cutoffs for PSA levels to pause and resume IADT will be simulated and TTP will be determined. PSA cutoffs that maximize TTP will be correlated with model-derived PCaSC dynamics and used to identify optimal patient-specific IADT protocols. Compared to androgen deprivation alone, co-treatment with docetaxel (DOC) improves patient survival with the survival benefit dependent on treatment timing. We will simulate DOC therapy initialized at different IADT cycles to determine DOC timing- dependent TTP in correlation to patient-specific PCaSC dynamics parameters to further improve PCa treatment outcomes. This exploratory high-risk and high-reward project may provide a significant conceptual advance in PCa treatment, away from continuous androgen deprivation at maximum tolerable dose until the tumor becomes resistant towards an IADT protocol, using triggers based on individual patients' response dynamics. If successful, the findings of this proposal will inform the optimal protocol and required sample size of a subsequent first-in-kind clinical trial of personalized adaptive IADT that delays TTP.

Narrative Intermittent androgen deprivation therapy (IADT) is a promising strategy to counteract evolution of resistance in prostate cancer patients. However, successful implementation of IADT requires identification of resistance mechanisms, prediction of responses, and determination of clinically actionable triggers of when to pause and when to resume IADT cycles. Here we propose to develop a mathematical framework to explore the contribution of prostate cancer stem cell dynamics to evolving resistance, and to use these dynamics in computer simulations to reliably forecast the response in subsequent treatment cycles on a per patient basis.