Progression-free survival (PFS) offers sponsors the opportunity to assess treatment benefit earlier than overall survival (OS), potentially accelerating development timelines and bringing promising therapies to patients sooner. However, while PFS is frequently used as a primary endpoint in oncology studies, its interpretation is often more complex than many stakeholders realize. This blog explores the opportunities and challenges of using PFS in oncology clinical trials, highlighting key regulatory guidance and best practices for demonstrating robust treatment benefit.
For regulators, demonstrating a statistically significant improvement in PFS is only one part of the story. Sponsors must also show that the observed treatment effect is robust, clinically meaningful, and not driven by assumptions made during the analysis process. This is where sensitivity analyses play a critical role.
Why PFS Is Used in Oncology Clinical Trials
Overall survival remains the gold standard endpoint in oncology because it directly measures the outcome that matters most: whether patients live longer. Regulatory authorities, including the FDA, EMA, and PMDA, consistently identify OS as the most objective and clinically meaningful measure of treatment benefit.¹⁻⁴
Despite its strengths, OS presents practical challenges. Many oncology trials require lengthy follow-up periods before a sufficient number of survival events occur, potentially delaying regulatory and clinical decision-making. In addition, subsequent lines of therapy administered after disease progression can complicate interpretation of overall survival results.
PFS helps address these challenges by measuring the time from randomization or treatment initiation until either:
- Disease progression
- Death from any cause
whichever occurs first.² Because progression events typically occur earlier than death events, PFS can provide an earlier indication of treatment effect and support more efficient drug development.
What Do Regulators Expect?
Although regulatory agencies accept PFS as an endpoint in many oncology settings, they continue to emphasize that its value depends on the wider clinical context.
According to FDA guidance, whether PFS is considered a direct measure of clinical benefit, a surrogate endpoint supporting accelerated approval, or evidence supporting traditional approval depends on multiple factors, including:²
- Magnitude of treatment effect
- Disease setting
- Available treatments
- Risk-benefit profile
- Clinical consequences of delaying progression
The EMA similarly recognizes the importance of PFS while stressing the need for rigorous methodology and robust analyses. The agency notes that sponsors should carefully evaluate assumptions related to progression assessment, censoring, protocol deviations, and missing data when designing and analyzing studies.¹
Meanwhile, PMDA guidance highlights a particular challenge in immuno-oncology. Certain immunotherapies may exhibit delayed treatment effects or transient tumor progression before a therapeutic response becomes apparent. In these situations, PFS may underestimate the true clinical benefit of treatment.⁴
Taken together, these positions demonstrate a consistent message from regulators: PFS can be a valuable endpoint, but sponsors must clearly demonstrate the reliability of their findings.
Why PFS Is Challenging to Analyze
Unlike death, which occurs at a clearly defined point in time, disease progression is usually identified through scheduled radiological assessments. As a result, progression events are rarely observed exactly when they occur. Instead, progression is typically known to have happened between two assessments. This creates several important analytical challenges:
Each of these situations can influence estimates of treatment effect and potentially affect study conclusions.
Standard Statistical Approaches to PFS
PFS is a time-to-event endpoint and is commonly analyzed using established survival analysis techniques.
Kaplan-Meier Estimates
Kaplan-Meier curves provide estimates of the probability of remaining progression-free over time and allow visual comparison between treatment groups.
Log-Rank Tests
The log-rank test is commonly used to compare progression-free survival distributions between treatment arms.
Cox Proportional Hazards Models
Cox regression models estimate the hazard ratio (HR), which quantifies the relative risk of progression or death between study groups.
Typically:
HR < 1 indicates a reduced risk in the treatment arm
HR = 1 indicates no difference between groups
HR > 1 indicates a higher risk in the treatment arm
While these methods are well understood, their results depend heavily on how progression events and censored observations are handled.
The Critical Role of Sensitivity Analyses
Given the assumptions required for PFS analyses, regulators expect sponsors to evaluate whether their conclusions remain consistent under alternative scenarios. FDA guidance recommends that statistical analysis plans specify both a primary analysis and one or more sensitivity analyses designed to evaluate the robustness of results.² Similarly, the EMA states that sensitivity analyses should be sufficiently comprehensive to demonstrate that trial conclusions are robust and that any differences observed across analyses should be clearly explained.¹ At its core, a sensitivity analysis asks a simple question:
If reasonable alternative assumptions were applied to the data, would the trial reach the same conclusion?
If the answer is yes, confidence in the treatment effect increases.
Common Sensitivity Analyses for PFS
Several sensitivity analyses are routinely conducted in oncology studies.
Alternative Progression Dating Rules
The FDA recommends assigning progression to the earliest date at which progression is observed and censoring at the last radiological assessment demonstrating no progression.² Sensitivity analyses may explore different approaches to progression dating and assess their impact on study results.
Missing Data Analyses
Alternative assumptions regarding missing assessments can be examined to determine whether study conclusions change materially.
Unscheduled Assessment Analyses
Progression events detected during unscheduled visits may be reassigned according to scheduled assessment windows to understand their influence on treatment effect estimates.¹
Independent Central Review
Sponsors frequently compare investigator-determined progression outcomes with results from blinded independent central review to confirm consistency.
Protocol Deviation Analyses
Sensitivity analyses may exclude patients with major protocol deviations to assess whether those deviations influenced the primary findings.
Collectively, these analyses help build confidence that observed benefits are genuine rather than artifacts of analytical decisions.
Is PFS Always a Reliable Surrogate for Overall Survival?
The widespread use of PFS often assumes a strong relationship between PFS and OS. However, evidence suggests that this relationship is not always consistent. A more in depth discussion of this question and related research will be included in a separate blog post. However key to note here is that these findings do not invalidate PFS as an endpoint. Rather, they highlight the importance of interpreting PFS within the broader clinical context and alongside additional efficacy, safety, and patient-centered outcomes.
Looking Beyond Survival
Another important consideration is that longer survival does not automatically equate to better patient outcomes. From a patient perspective, quality of life can be just as important as quantity of life. A treatment that extends survival while maintaining functional status and minimizing toxicity may provide greater overall value than one that prolongs survival at substantial cost to quality of life. This broader perspective is reflected in Health Technology Assessment (HTA) processes, which often incorporate Quality-Adjusted Life Years (QALYs) when evaluating the value of new therapies. By combining survival and quality of life into a single measure, QALYs can support more holistic decision-making regarding treatment adoption and reimbursement.
Conclusion
PFS remains a cornerstone endpoint in oncology drug development, offering sponsors an opportunity to evaluate treatment efficacy earlier than would be possible using overall survival alone. Regulatory agencies recognize its value across many cancer settings, but they also expect sponsors to demonstrate that PFS findings are reliable and clinically meaningful.¹⁻⁴
Sensitivity analyses have therefore become an essential component of modern oncology trials. By testing key assumptions related to censoring, missing data, progression timing, and protocol deviations, sponsors can provide regulators with greater confidence that observed treatment effects are robust and reproducible.
As oncology trials continue to evolve and novel treatment modalities reshape the development landscape, rigorous sensitivity analyses will remain critical for ensuring that PFS-based conclusions stand up to both regulatory scrutiny and scientific scrutiny.
From trial design and endpoint strategy to statistical analysis and regulatory submissions, we help sponsors generate oncology evidence that stands up to regulatory scrutiny.
References
1.European Medicines Agency (EMA). Appendix 1 to the Guideline on the Evaluation of Anticancer Medicinal Products in Man: Methodological Consideration of Using Progression-Free Survival or Disease-Free Survival in Confirmatory Trials.
U.S. Food and Drug Administration (FDA). Clinical Trial Endpoints for the Approval of Cancer Drugs and Biologics: Guidance for Industry.
U.S. Food and Drug Administration (FDA). Clinical Trial Endpoints for the Approval of Non-Small Cell Lung Cancer Drugs and Biologics: Guidance for Industry.
Pharmaceuticals and Medical Devices Agency (PMDA). Guidelines on Cancer Immunotherapy Development.
Hess LM, Brnabic A, Mason O, Lee P, Barker S. Relationship Between Progression Free Survival and Overall Survival in Randomized Clinical Trials of Targeted and Biologic Agents in Oncology. Journal of Cancer. 2019;10(16):3717-3727.
Milrod B, et al. Correlation Between PFS and OS in FL: Implications for Clinical Trial Endpoint Selection. Blood. 2023;142(Supplement 1):4395.
Ciani O, Davis S, Tappenden P, et al. Validation of Surrogate Endpoints in Advanced Solid Tumors: Systematic Review of Statistical Methods, Results, and Implications for Policy Makers. International Journal of Technology Assessment in Health Care. 2014;30(3):312-324.
