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 final blog in the series.

Despite targeting certain types of cancer in an oncology trial, the overall enrolled population remains highly heterogeneous based on demographics, clinical history, and molecular characteristics of the disease.  Analyses conducted using the overall population can mask the extent of the benefit-risk in subgroups of patients.  Thus, subgroup analyses are critical to fully understand the benefit-risk in pivotal oncology trials and guide regulatory decisions for approval and labelling.  However, subgroup analyses also present several challenges especially for overall survival (OS) where larger sample sizes and longer follow-up are needed, and where early analyses of immature data may be conducted.

Challenges with subgroup analyses include small sample sizes resulting in increased variability and reduced power for formal testing, multiplicity increasing likelihood of varying results across subgroups and/or endpoints, and lack of biological rationale for subgroup results. Such factors can undermine the credibility of OS subgroup results. How can one appropriately use and interpret OS subgroup analyses and increase the credibility of the results?

Below are 6 keys to managing the benefits and risks of subgroup analyses of OS:

• Pre-specification of subgroup analyses and adequate sample size for formal hypothesis testing in key subgroups leads to increased credibility and interpretability.
• Post-hoc subgroup analyses should be interpreted with caution. These analyses can still be valuable for hypothesis generation and assessing robustness especially when considered with results across studies or when multiple subgroups within a study exhibit a consistent observed treatment benefit/risk.
• Post-hoc subgroup analyses of a ‘failed’ study should be considered for hypothesis generation only but not for a claim for efficacy.
• Strong biological rationale for a subgroup increases credibility of results.
• OS assesses benefit and risk so one must understand that although subgroup analyses can identify groups exhibiting greater benefit, groups that may be subject to increased harm may also be seen, and
• Interpretation of subgroup results should consider various statistical, biological, and practical factors such as study design, disease setting, biologic rationale of subgroup, maturity of data, and evidence of benefit/harm from other studies.

Several recent oncologic drug approvals have restricted an indication to a subgroup of patients or are cases where identification of substantial benefit in a subgroup has been a gateway to expansion of the indication to a wider population or other subgroups. When conducted appropriately, subgroup analyses are a powerful tool in identifying patients for which agents will be both safe and effective.

Patricia Stephenson, Sc.D., Associate Director, Biostatistics is a Harvard graduate with over 10 years of experience working in oncology, including working with researchers at the Dana-Farber Cancer Institute. Dr. Stephenson has served as the lead statistician for multiple oncology studies, including a Phase 1 study pivotal for an NDA submission for accelerated approval. As a result, she was involved in several submission activities including supporting the statistical preparation and review for the Summary of Clinical Efficacy (SCE) and Summary of Clinical Safety (SCS) as well as preparation for a FDA Advisory Committee Meeting.  Previously, Dr. Stephenson also served as the lead statistician for multiple Phase 1 and 2 studies in ovarian cancer, renal cell carcinoma, gastrointestinal stromal tumors, and non-small cell lung cancer.