Abstract
Clinical trials are costly and time-intensive endeavors, with a high rate of drug candidate failures. Moreover, the standard approaches often evaluate drugs under a limited number of protocols. In oncology, where multiple treatment protocols can yield divergent outcomes, addressing this issue is crucial. Here, we present a computational framework that simulates clinical trials through a combination of mathematical and statistical models. This approach offers a means to explore diverse treatment protocols efficiently and identify optimal strategies for oncological drug administration. We developed a computational framework with a stochastic mathematical model as its core, capable of simulating virtual clinical trials closely recapitulating the clinical scenarios. Testing our framework on the landmark SOLO-1 clinical trial investigating Poly-ADP-Ribose Polymerase maintenance treatment in high-grade serous ovarian cancer, we demonstrate that managing toxicity through treatment interruptions or dose reductions does not compromise treatment's clinical benefits. Additionally, we provide evidence suggesting that further reduction of hematological toxicity could significantly improve the clinical outcomes. The value of this computational framework lies in its ability to expedite the exploration of new treatment protocols, delivering critical insights pivotal to shaping the landscape of upcoming clinical trials.