Stakeholders ask for flexibility, more examples in FDA guidance on overall survival in cancer trials
Stakeholders asked the US Food and Drug Administration for more examples in comments on a recently issued guidance on the use of overall survival as an endpoint in oncology drug clinical trials. They also asked for more flexibility when choosing what endpoint to use and alignment with international guidelines.The guidance noted that overall survival should be prioritized as the primary endpoint when feasible but also acknowledged that there are situations when that may not be possible. (RELATED: FDA proposes guidance on using overall survival endpoints in cancer drug trials, Regulatory Focus 18 August 2025)
PhRMA
Several stakeholders representing researchers, manufacturers, and advocacy groups wrote to FDA expressing appreciation for the agency’s acknowledgment that using overall survival as a primary endpoint isn't always feasible or practical. The Pharmaceutical Research and Manufacturers of America (PhRMA) said that while the guidance includes examples where overall survival as an endpoint is infeasible, it wants the agency to provide more examples and additional methodologies for assessing overall survival.
PhRMA also said it appreciates FDA's reference to the International Council for Harmonisation (ICH) guidelines but asked the agency to ensure further alignment.
"As medical product development is an increasingly global endeavor, we note that continued FDA discussion and alignment with other health authorities regarding recommendations for assessing overall survival in oncology clinical trials, as appropriate, will be critical to ensuring the efficiency of current and future medical product development programs," said the group. "To this end, we recommend revising the draft guidance to align more closely with terminology in ICH E9(R1), as appropriate, e.g., use of the term 'risk' instead of harm, clarification of the phrase 'supplementary analyses.'"
PhRMA also proposed several changes to the guidance regarding trial design considerations. For instance, the group asked FDA to include language recognizing that using immature overall survival data may hamper researchers' ability to evaluate harm, especially regarding long-term safety and mortality outcomes. It also proposed changes to how the agency approaches interim analyses of overall survival, such as clarifying how soon or how often a futility and/or harm analysis should be performed.
"As overall survival is both an efficacy and a safety endpoint, the draft guidance suggests specifying futility and harm analyses; however, some passages of the guidance note that 'futility or harm should be included,' while others recommend including “efficacy, futility, and harm'" said PhRMA. "We recommend clarifying FDA’s preferred timing for conducting these analyses as well as defining 'futility analysis' and 'harm analysis,' including when and why each is warranted and how they are related.
"To this end, PhRMA cautions that additional looks at survival data, especially early in a trial when data are immature, can increase the likelihood of observing a spurious harm signal that is not representative of the true survival effect," the group added.
PhRMA also has asked for clarification on is on the use of independent data monitoring committees (DMC). Echoing past comments, the group recommended using a DMC charter to clarify the purpose and role of the specific DMC, including when it can access trial data, in order to ensure sensitive information is handled carefully.
"As DMC charters may be posted publicly, and the schedule and basis of planned interim analyses may constitute highly sensitive information, we recommend noting that DMC charters may reference the protocol and statistical analysis plan (SAP) as containing the schedule and basis of planned interim analyses, rather than including this information directly in the DMC charter," said PhRMA. "To preserve the integrity of blinding and to protect highly sensitive and proprietary information, however, we ask the Agency to clarify that the specific decision criteria threshold that the DMC will use to guide decisions about stopping early for futility or efficacy should not be included in documents that are made publicly available (e.g., DMC charter)."
Another important area that PhRMA asked for updates on is statistical analysis considerations. More specifically, the group asked FDA to more closely align its guidance with ICH E9(R1) on terms such as 'supplementary analysis.' It also asked for clarification on the methodologies the agency will use to assess harm and flexibility regarding the timing of harm assessment procedures.
"If overall survival is not the primary endpoint, for example, and overall survival event rates are low, the assessment of harm might ideally be performed at interim analyses planned for other endpoints in the study," said PhRMA."
It also asked for changes to the subgroup considerations section of the draft guidance to allow for more flexibility to address situations where subgroup outcomes are highly unpredictable.
"A sponsor cannot always predict the degree of varying survival benefit or detriment for different subgroups in a study, and there can be considerable uncertainty regarding the 'expected contribution' of the subgroups to the 'overall treatment effect," said PhRMA. "Including an early interim analysis to assess potential harm in a subgroup may not be realistic due to poor operating characteristics caused by high variability.
"We recommend revising this section of the draft guidance to account for scenarios in which information regarding survival for different subgroups is highly limited," it added.
FoCR
In its comments on the guidance, the Friends of Cancer Research (FoCR) asked for more flexibility when considering overall survival as a clinical endpoint in trials and said the agency should further clarify that it can be interpreted as a component of a broader benefit-risk framework. The group also asked for more examples where overall survival can be evaluated for potential harm and the key considerations that go into making such an evaluation.
“Because [overall survival] reflects the ultimate outcome but may lag behind other indicators of harm, relying on survival status alone can provide an incomplete or delayed understanding of safety,” said FoCR. “Integrating additional data collected within the trial (such as serious adverse events, dose modifications, treatment discontinuations, or quality of life trends) allows earlier and more interpretable insights while avoiding undue emphasis on immature [overall survival] data.”
“We encourage FDA to emphasize a multidimensional approach to interpreting [overall survival] and harm: one that considers statistical, clinical, and patient-reported information together to inform a balanced and clinically relevant assessment of benefit and risk,” the group added.
Cancer Research UK
While most comments supported FDA's attempt to balance the use of overall survival as a primary endpoint when feasible with other potential endpoints, the Cancer Research UK & UCL Cancer Trials Centre said it may not go far enough to ensure researchers feel that they have more flexibility.
"The Background section states 'Overall survival [OS] should be prioritized as the primary endpoint when feasible,'" said the UK researchers. "Although this does not mandate OS as the primary outcome measure for all trials, it could be viewed as such by some investigators.
"There are certainly situations where OS should be used as the major outcome but there are also examples where showing benefits on OS has been impossible, e.g. high numbers of crossovers or multiple lines of subsequent therapies," they added. "‘Feasible’ is interpreted as logistically possible (e.g. study size), but allowing crossover is a clinical feature of a trial, not logistical one. We suggest clarifying the statement, e.g.: 'Overall survival [OS] should be prioritized as the primary endpoint when feasible and clinically interpretable given the study design.'"
The group also noted that FDA's guidances have a wider impact beyond commercial clinical trials since non-commercial trial researchers use them to ensure their studies don't encounter major issues during evaluation.
Novartis
Novartis wrote to the FDA emphasizing that it's important to note that crossover in clinical trials can occur when there is a crossover to an approved therapy or when the crossover is for an investigational drug.
"Crossover to an approved therapy mirrors real-world clinical practice and, as such, does not confound the interpretation of overall survival (OS); the OS results for control arm patients who cross over would simply reflect outcomes expected in routine care," said the drugmaker. "Conversely, crossover to an investigational drug does not align with clinical practice since patients outside the trial would not have access to the investigational treatment. In this context, estimating the treatment effect as if protocol-specified crossover had not occurred provides a more meaningful perspective for prescribers and patients than a treatment policy strategy."
EMD Serono
EMD Serono noted that while the guidance focuses on situations when overall survival is not the primary endpoint, it also addresses situations when it should be considered the primary endpoint. With that in mind, the drugmaker said it would be useful to integrate the FDA's estimand framework into the guidance. It said that including the framework could help researchers address intercurrent events after a trial has started that potentially affect how they interpret the results.
EMS Serono also asked for changes and clarifications to how FDA assess trial designs, statistical analysis, and subgroup considerations.
Public docket
