OBJECTIVES: The use of primary endpoints other than overall survival (OS) in cancer trials supports timely approval of new drugs, but creates challenges for market access as evidence for improvements in OS is often sought for approval. The challenge can be exacerbated if the OS signal is diluted by the effects of subsequent therapies. We demonstrate how clinical trial simulation (CTS) can support understanding the potential effects of subsequent therapy on OS outcomes in a hypothetical trial in nonmetastatic castration resistant prostate cancer (nmCRPC).

METHODS: A hypothetical trial was simulated in which patients with nmCRPC were randomly assigned to an experimental treatment or placebo. The hypothetical trial was powered for a primary endpoint of bone metastasis-free survival and a hazard ratio (HR) for the experimental treatment of 0.70. A published model in metastatic castration resistant prostate cancer (mCRPC) was used to simulate treatment and OS following metastasis. Following metastasis, patients could continue onto prednisone or a subsequent treatment with a OS HR of 0.74. Scenarios of subsequent treatment use were simulated to determine the relationship between subsequent treatment and OS signal. OS results underwent time to event analyses to produced KM data and estimate HR and statistical significance.

RESULTS: In a scenario in which 75% of patients experiencing metastasis receive subsequent treatments, OS for the experimental treatment started to become significant at four years of follow-up. As the level of subsequent treatment use lowered, time to show benefit in OS shortened.

CONCLUSIONS: CTS can help understand how trial outcomes relevant for market access may be influenced by subsequent treatment, even where the primary outcome is not impacted, and provide insights to the timing of when outcomes become clinically meaningful.