Progression-Free Survival (PFS) refers to the time interval in clinical trials from a defined starting point (usually randomization or treatment initiation) to objective evidence of disease progression or death—whichever occurs first. PFS is a central efficacy endpoint in many oncology trials, but is increasingly used in other indications with clearly measurable deterioration of disease status. As a near-surrogate endpoint, PFS is frequently employed to capture the direct impact of a therapy on tumor growth or disease dynamics.
Importantly, PFS is not simply “time until deterioration,” but a precisely operationalized measure. Study documentation must clearly define which assessments count, which criteria trigger progression, how missing assessments are handled, and which events are considered death for the purposes of the endpoint. Without this clarity, deviations between sites can arise in practice, leading to inconsistencies in the dataset and to discussions during audits or inspections.
Definition, Measurement Timepoint, and Typical Censoring
For analysis, PFS is evaluated as a time-to-event variable. The “event” is typically radiologically or clinically confirmed progression according to predefined criteria (e.g., RECIST in oncology) or death. Participants without an event at the time of analysis are censored, e.g., at the last tumor-free assessment. For robust results, it is critical that assessment timepoints, the progression definition, and handling of missing measurements are precisely specified in the clinical trial protocol.
Why PFS Is Frequently Used as a Primary Endpoint
PFS can be observed earlier than overall survival, making trials shorter and feasible with less follow-up. This is particularly relevant when effective subsequent therapies are available after progression, which may dilute a survival benefit. In such situations, PFS can be a more sensitive marker of direct treatment effect. At the same time, PFS is often less influenced by subsequent therapies than overall survival.
Methodology: Analysis, Bias Risks, and Data Quality
Analysis is typically performed using Kaplan-Meier curves and hazard ratios from a Cox model, supplemented by sensitivity analyses. PFS is susceptible to assessment bias because progression determination depends on imaging and its interpretation. Therefore, standardized imaging intervals, investigator training, and—in many programs—blinded independent central review are important. In practice, it must also be ensured that all assessments are documented promptly and that underlying imaging data are completely archived, so that audits and inspections can verify endpoint derivation.
A common operational challenge is so-called assessment timing: when imaging occurs outside planned windows or documents are submitted late, the apparent progression timepoint can shift. This directly affects PFS estimates. Sponsors and CROs should therefore focus monitoring on completeness of imaging, documentation of schedule deviations, and consistency between radiological findings, eCRF entries, and source documents. For robust analysis, traceability in the Trial Master File is also relevant (e.g., versions of assessment guidelines, training records, central review processes).
Distinction from Related Endpoints
PFS differs from “Time to Progression” (TTP), where death without documented progression often does not count as an event. PFS is distinguished from “Disease-Free Survival” (DFS) because DFS is typically used after curative therapy in adjuvant settings and includes recurrence or death. “Event-Free Survival” (EFS) is usually defined more broadly and may include additional events (e.g., treatment change, lack of response). Clear terminology prevents misunderstandings between sponsor, CRO, site, and statistical analyst.
In reports and publications, it is helpful to transparently present not only the endpoint but also the clinical interpretation: Which types of progression were dominant? Was there evidence of pseudoprogression (e.g., under immunotherapies) that could complicate assessment? Were new lesions, clinical deterioration, and death consistently captured? These detailed questions are critical to ensure that PFS is not only statistically significant but also clinically plausible.
Regulatory Context (EU/DE) and Marketing Authorization Relevance
From a regulatory perspective, PFS is generally accepted as an endpoint when it is clinically meaningful, validly collected, and the interpretation of treatment effect remains comprehensible. For EU-based programs, requirements for study design, data integrity, and endpoint definition are anchored in ICH E6 (Good Clinical Practice); additionally, for trials in the EU, the provisions of Clinical Trials Regulation (EU) No. 536/2014 must be observed. For registration-enabling studies, it is also important that the statistical analysis plan, handling of intercurrent events, and endpoint derivation are consistently documented.
In practice, this means: the endpoint definition must be formulated in the protocol so that site, monitor, and analysis team interpret it identically. Furthermore, the statistical analysis plan should contain clear rules for censoring, handling of missing assessments, and sensitivity analyses. For dossiers and Clinical Study Reports, endpoint derivation must be described comprehensibly (e.g., in endpoint algorithms, data listings, and deviation analyses) so that authorities can assess the robustness of the results.
FAQ
Which Starting Points Are Typically Used for PFS?
Randomization is most common because it ensures comparability of groups. In single-arm studies, the first day of treatment may also serve as the starting point, which must then be clearly justified in the report.
Why Can PFS Be Problematic Despite Objective Criteria?
Even small differences in imaging intervals, missed assessments, or inconsistent interpretation can shift the timepoint of progression. Therefore, standardized processes and often central review are particularly important.
Is a Benefit in PFS Always Clinically Relevant?
Not necessarily. Clinical relevance depends on the absolute prolongation, the adverse event profile, quality of life, and the treatment context. Regulatory assessment and benefit-risk evaluation therefore consider multiple lines of evidence.
Regulatory References (Selection): ICH E6(R3) Good Clinical Practice; Regulation (EU) No. 536/2014 (Clinical Trials Regulation).