Rho experts recently attended the FDA-AACR-ASA Workshop on the use of overall survival as an endpoint in oncology trials.  In a series of five blogposts, they will discuss key topics to consider when designing such a study as presented at the workshop. These topics will include study design, endpoint construction, subgroup considerations, and analysis interpretation. This is the second blog in the series.

Overall survival (OS) is considered the gold standard endpoint for oncology trials and is both an important efficacy & safety endpoint. However, reviewing medical literature one will frequently see the phrase “unmet medical need” in reference to oncology treatments which requires sponsors to consider study designs with the primary analysis utilizing intermediate or surrogate endpoints, like progression free survival, so that safe treatments can be developed quicker. In these settings, overall survival is often a post-hoc analysis as it is descriptive in nature with no Type I error control.

Post-hoc analysis of overall survival brings about a variety of concerns for regulatory bodies. Most notably, given the inherent lag in data between a surrogate endpoint and OS, the OS endpoint will frequently be less mature meaning low statistical power and a small sample size at the time of analysis. This poses a challenge for regulators that would like to make a robust benefit-risk assessment for the proposed treatment.

To remedy these concerns, a variety of options should be considered including:

1) Efficacy Perspectives

Exploratory analyses which test the sensitivity of results to a variety of assumptions (e.g., effect of intercurrent events) can be helpful. When a strong relationship between the study’s primary endpoint & OS can be established, there will be more confidence in any signal that is observed.

2) Safety Perspectives

Descriptive analyses should always be presented with corresponding confidence intervals. Additional efforts can be placed on not only ruling out specific benchmarks (e.g., a hazard ratio of 2.0), but also the probability of achieving desirable benchmarks (e.g., an 80% chance of having a hazard ratio less than 1). Qualitative information can also be valuable here if we consider subject narratives surrounding study deaths and/or adverse events of special interest.

3) Contextual Placement of Analyses

Regardless of the analyses performed, placing the results and the treatment itself into the proper clinical context is essential. What is clinical benefit of treatment? If the main benefit is overall survival, strong links to any surrogate endpoint will need to be established. However, if the clinical benefit is something different (e.g., a better quality of life), analyses focused on harm reduction will be of value. Similarly, treatments that are first-in-class or a new line of therapy will change the landscape and offer a different risk/benefit profile compared to a product in a more crowded indication/therapy space.

As illustrated, when analyzing overall survival in a post-hoc manner sponsors need to carefully consider the types of analyses to present and have a thorough understanding of both the regulatory precedent and market landscape which their product will enter. Partnering with a CRO that interacts often with regulatory authorities and understands these challenges can help minimize pain points and ensure a smoother path to market.

Scott Mollan, Associate Director, Biostatistics, has over 17 years of experience in clinical and non-clinical statistics across the CRO & pharmaceutical industry. Having led studies of all phases (pilot, pivotal, post-market, phases I-IV) and assisting clients during both the pre-submission phase and FDA approval via the NDA/PMA process (2 NDAs/10 PMAs), he has a wealth of experience to draw upon to support clients. Armed with graduate degrees in business and statistics, Mr. Mollan has been able to leverage his understanding of the clinical trial process via a diverse range of indications to publish on the medical device trial process, lung cancer diagnostics, and women’s health while similarly offering industry presentations on missing data analysis strategies and the use of adaptive trial designs within medical devices studies.