When a new cancer drug is reported to "improve progression-free survival by 7.2 months," what does that actually mean for a patient sitting in a clinic? The language of clinical trial endpoints — PFS, OS, ORR, DFS, DOR — is the vocabulary through which oncology communicates treatment benefit. But these terms are not interchangeable, and the choice of endpoint shapes how we should interpret every drug listed on PipelineEvidence.
Overall Survival: The Gold Standard
Overall survival (OS) measures the time from randomization until death from any cause. It is the most objective endpoint in oncology because the event — death — is unambiguous and requires no interpretation. An OS benefit is the clearest possible evidence that a drug extends life.
The drawback of OS is time. In diseases with long natural histories (like HR-positive breast cancer or indolent lymphomas), waiting for enough deaths to demonstrate a statistically significant OS difference can take many years. Additionally, OS can be diluted by crossover — when patients in the control arm switch to the experimental drug after progression, the OS benefit of the experimental drug in the intent-to-treat analysis may be diminished even if the drug truly extends life.
Progression-Free Survival: The Practical Workhorse
Progression-free survival (PFS) measures the time from randomization until the cancer progresses (grows or spreads) or the patient dies from any cause. Progression is typically assessed by regular imaging scans, evaluated using Response Evaluation Criteria in Solid Tumors (RECIST) — a standardized system that defines progression as a 20% increase in the sum of target lesion diameters or the appearance of new lesions.
PFS is the most commonly used primary endpoint in modern oncology trials because it matures faster than OS (progression events outnumber deaths), requires smaller sample sizes, is not confounded by crossover or subsequent therapies, and directly measures the drug's effect on tumor biology.
The limitations are important. PFS depends on how often patients are scanned (more frequent scanning catches progression earlier, potentially making PFS appear shorter), can be subject to assessment bias (which is why many trials use blinded independent central review, or BICR), and a PFS benefit does not guarantee an OS benefit. A drug might delay progression by months without ultimately changing how long patients live, particularly if effective subsequent therapies are available.
The Supporting Cast: DFS, ORR, DOR, and EFS
Disease-free survival (DFS) is the equivalent of PFS in the adjuvant setting — after a patient's tumor has been surgically removed, DFS measures the time until cancer recurrence or death. Drugs like adjuvant osimertinib (ADAURA trial) and adjuvant abemaciclib (monarchE trial) were approved based on DFS benefits.
Overall response rate (ORR) measures the percentage of patients whose tumors shrink by a defined amount (complete or partial response by RECIST). ORR is a snapshot, not a time-to-event measure. It is most useful in single-arm trials where there is no comparator — many accelerated approvals are based on ORR plus duration of response (DOR), which measures how long the response is maintained.
Event-free survival (EFS) has gained prominence in the neoadjuvant setting, where treatment is given before surgery. EFS typically includes disease progression that prevents surgery, failure to achieve a defined pathological response, recurrence after surgery, or death. The recent rise of neoadjuvant immunotherapy trials (such as pembrolizumab in KEYNOTE-522 for triple-negative breast cancer) has made EFS an increasingly important endpoint.
How the FDA Uses Endpoints for Drug Approval
The FDA recognizes OS as the definitive evidence of clinical benefit, but it accepts PFS and other endpoints through two pathways. Regular approval requires evidence of clinical benefit — usually OS or an endpoint reasonably likely to predict clinical benefit (like PFS in many settings). Accelerated approval allows drugs to be approved based on surrogate endpoints (like ORR) that are reasonably likely to predict clinical benefit, with a requirement for confirmatory trials.
This system enables faster patient access to promising drugs while maintaining accountability. However, it also means that some drugs on the market have not yet demonstrated an OS benefit — something to keep in mind when reviewing the clinical evidence for therapies on PipelineEvidence.
A Practical Framework for Evaluating Endpoints
When reading about any approved cancer therapy, consider: Was the primary endpoint OS or PFS? If PFS, is there supportive OS data or a trend? What was the absolute improvement in months, not just the hazard ratio? Was the response assessment blinded (BICR) or investigator-assessed? Was crossover allowed, and could it have diluted the OS signal?
These questions transform a headline result into actionable understanding. For detailed endpoint data on each approved oncology drug, explore the individual drug pages in our Drug Library, where we link directly to ClinicalTrials.gov registrations and primary publications.