OS, PFS, ORR and EFS explained: what each cancer trial endpoint measures, when FDA accepts it for approval, and why survival now carries more weight in 2026.
Last updated: November 18, 2026
Overall survival (OS) is the gold standard cancer endpoint because death is unambiguous. Progression free survival (PFS) and event free survival (EFS) measure time to tumor events, and objective response rate (ORR) measures shrinkage. FDA accepts each in different settings, with OS weighed more heavily since 2025.
Scope: the main efficacy endpoints in oncology trials, how FDA’s 2018 endpoints guidance treats each one, what the August 2025 draft guidance on overall survival changes, and how recent 2026 decisions map onto the rules.
What is the difference between OS, PFS and ORR?
Overall survival counts how long patients live. Progression free survival counts how long they live without their cancer growing. Objective response rate counts how many patients see their tumors shrink by a set amount. Each answers a different question, and FDA weighs them differently.
FDA’s final guidance, Clinical Trial Endpoints for the Approval of Cancer Drugs and Biologics (December 2018, issued by the Oncology Center of Excellence, CDER and CBER), gives the working definitions:
- Overall survival: time from randomization until death from any cause, measured in the intent to treat population.
- Progression free survival: time from randomization until objective tumor progression or death, whichever comes first.
- Time to progression: the same clock, but deaths are not counted as events. FDA prefers PFS because it captures deaths.
- Objective response rate: the proportion of patients with tumor shrinkage of a predefined amount for a minimum period. FDA generally defines it as partial responses plus complete responses. Stable disease does not count.
- Disease free survival (DFS): time from randomization to recurrence or death, mainly used after surgery or radiotherapy (the adjuvant setting).
- Event free survival (EFS): time from randomization, before definitive surgery or radiotherapy, to progression that rules out surgery, recurrence, or death.
Why is overall survival called the gold standard?
FDA’s guidance calls OS the most reliable cancer endpoint. It is precise and easy to measure, it is documented by the date of death, and, in the guidance’s words, bias is not a factor in measuring it.
Its weaknesses are practical. OS trials need large populations and long follow up. Later treatments muddy the comparison, because patients who progress go on to other drugs. When control patients cross over to the investigational drug, the survival gap can shrink even if the drug works. OS also counts deaths that have nothing to do with cancer.
FDA’s guidance says OS evidence should come from randomized controlled trials, and that external control data are seldom reliable for time to event endpoints. That is why a single arm study can never show an OS benefit.
When does FDA accept progression free survival for approval?
It depends on the disease and the size of the effect. FDA’s guidance says PFS can serve as a surrogate for accelerated approval, as a surrogate for traditional approval, or as a direct clinical benefit, and that a large PFS improvement has supported traditional approval in select cancers.
The guidance names the factors that decide which: effect magnitude, the toxicity of the drug, the location of metastases, and whether delaying progression in critical sites such as the brain or spine, or delaying more toxic therapy, matters to patients.
PFS has real measurement problems, which is why the guidance spends pages on them:
- Progression is judged on scans, so open label trials risk assessment bias. FDA generally wants tumor based primary endpoints verified by central reviewers blinded to treatment, or a prespecified blinded audit.
- Missed or late scans change the date of progression. The guidance tells sponsors to define visit windows and to run sensitivity analyses.
- PFS gains do not always translate into longer survival. FDA lists that lack of correlation as a core disadvantage.
The November 2026 ivonescimab decision shows the limit. Summit Therapeutics reported a PFS hazard ratio of 0.52 in its pivotal lung cancer trial, yet FDA issued a complete response letter, saying substantial evidence was not established without a statistically significant OS result and that PFS alone did not establish a positive benefit risk. Our full coverage is in FDA rejects ivonescimab over missing overall survival win.
Can a drug be approved on response rate alone?
Yes, usually through accelerated approval. FDA’s guidance says ORR has been the most commonly used surrogate endpoint for accelerated approval, and because it measures drug activity directly, it can be evaluated in a single arm trial.
Single arm ORR trials are accepted for accelerated approval when there is no available therapy and tumor shrinkage can be attributed to the drug. FDA assesses ORR by its magnitude, its duration and the share of complete responses. Durable responses have occasionally supported traditional approval.
The latest example landed this week. GSK’s Nelkarmi (neladalkib) won accelerated approval on November 17, 2026 for pretreated ALK positive lung cancer, based on a confirmed response rate of 49% after alectinib and 32% after two or more ALK inhibitors, with the label stating the approval rests on tumor response rate and duration of response. See our Nelkarmi approval analysis.
What are DFS and EFS used for?
They measure whether a treatment given around surgery keeps cancer from coming back. DFS is the standard in the adjuvant setting, after surgery. EFS is used when treatment starts before surgery, in the neoadjuvant or perioperative setting.
FDA’s guidance says DFS has supported traditional approval in breast, colorectal, gastrointestinal stromal tumor, melanoma and renal cell carcinoma, and that it can act as a surrogate or a direct benefit depending on effect size and toxicity. Both endpoints carry bias risks if follow up schedules differ between arms or patients drop out unevenly.
Bladder cancer gives a clean 2026 example. FDA approved durvalumab plus enfortumab vedotin for cisplatin ineligible muscle invasive bladder cancer on November 13 based on the VOLGA trial, where EFS hazard ratio was 0.51 and OS hazard ratio was 0.58. Coverage: FDA approves Imfinzi plus Padcev for cisplatin ineligible MIBC.
What other endpoints does FDA accept in oncology?
Several, each tied to specific diseases. The 2018 guidance names them:
- Complete response in leukemia has been a clinical endpoint for traditional approval, because it is linked to fewer infections, less bleeding and fewer transfusions.
- Pathologic complete response (pCR) has been used as a surrogate for accelerated approval in early high risk breast cancer.
- Minimal residual disease has served as a surrogate for accelerated approval in acute lymphoblastic leukemia.
- Metastasis free survival has been a clinical endpoint for traditional approval in non metastatic castration resistant prostate cancer.
- Symptom endpoints and patient reported outcomes count as direct clinical benefit, but FDA flags bias in open label trials and the difficulty of separating cancer symptoms from drug side effects.
- Time to treatment failure is generally not recommended as a regulatory endpoint for new molecular entities.
How do the endpoints compare side by side?
| Endpoint | What it measures | Trial design | Typical regulatory role | Main weakness |
|---|---|---|---|---|
| Overall survival | Death from any cause | Randomized only | Clinical benefit, traditional approval | Large, slow, confounded by later therapy and crossover |
| Progression free survival | Progression or death | Randomized | Surrogate (accelerated or traditional) or direct benefit, case by case | Scan timing and reader bias; may not predict OS |
| Objective response rate | Tumor shrinkage (partial plus complete) | Single arm possible | Most common accelerated approval surrogate | Varying definitions; may not predict OS |
| DFS / EFS | Recurrence, progression or death around surgery | Randomized | Traditional approval in several adjuvant settings; surrogate in others | Follow up schedule and dropout bias |
| Pathologic complete response | No residual cancer found at surgery | Randomized | Accelerated approval surrogate in early breast cancer | Link to long term outcome varies |
| Patient reported symptoms | How patients feel and function | Ideally blinded | Direct clinical benefit | Open label bias; toxicity overlap |
Our summary of FDA’s December 2018 endpoints guidance. Roles depend on disease, effect size and available therapy.
What did FDA’s 2025 overall survival guidance change?
It raised the weight of OS even in trials where OS is not the primary endpoint. In August 2025 FDA’s Oncology Center of Excellence issued the draft guidance Approaches to Assessment of Overall Survival in Oncology Clinical Trials (docket FDA-2024-D-5850), which emphasizes analyzing OS as a prespecified safety endpoint.
The Federal Register notice, published August 19, 2025, says OS should be prioritized as a primary endpoint when feasible and calls it a gold standard endpoint in oncology. It acknowledges when OS is impractical: indolent diseases, highly effective therapies where long follow up would be needed, and single arm trials, where OS cannot be adequately interpreted. It also names crossover, later therapy and other intercurrent events as confounders. The draft covers trial design, statistical analysis, subgroup evaluation and regulatory submission. The guidance remains a draft and is not for implementation.
The practical message for sponsors: even if PFS is your primary endpoint, plan to show OS is not moving the wrong way, and plan your data maturity and analysis timing with that in mind. Our complete response letter explainer covers what happens when FDA decides the evidence falls short.
How do endpoints connect to accelerated versus traditional approval?
Traditional approval rests on demonstrated clinical benefit, or on a surrogate known to predict it. FDA’s guidance says survival improvement is considered an appropriate measure of clinical benefit in oncology. Accelerated approval rests on a surrogate reasonably likely to predict benefit, and it comes with postmarketing studies to verify that benefit.
The same endpoint can sit on either side of that line. A large PFS gain in one disease can be clinical benefit, while a modest one in another is only a surrogate. That is why sponsors should agree on endpoints with FDA before the pivotal trial starts, which the guidance explicitly recommends for emerging endpoints. Confirmatory obligations are covered in our accelerated approval rules explainer.
Our framework: an endpoint checklist for sponsors and investors
- Match the endpoint to the setting. Metastatic: OS or PFS. Perioperative: EFS or DFS. Refractory with no options: ORR in a single arm may work.
- Ask whether OS is mature enough to rule out harm. After 2025, a PFS win with an OS hazard ratio drifting above 1 is a red flag.
- Check who read the scans. Blinded independent central review carries more weight than investigator assessment in open label trials.
- Look at durability, not just response. The share of responders past 12 months often decides an ORR based filing.
- Model crossover. If control patients can switch to your drug, your OS result may understate benefit. Address it in the statistical plan.
- Get FDA alignment in writing. End of Phase 2 meeting minutes that accept the endpoint are worth more than any consultant’s opinion.
- Plan the confirmatory trial early. For accelerated approval, FDA can require it to be underway before approval, and it should test the drug where the market is.
What to watch next
- Whether FDA finalizes the 2025 overall survival draft guidance, and any changes from the comments.
- How Summit responds to the ivonescimab CRL on overall survival.
- The lidERA adjuvant giredestrant decision due November 30, 2026, a test of how FDA reads an adjuvant endpoint in early breast cancer.
- Confirmatory trial progress for 2026 accelerated approvals such as Nelkarmi.
FAQ
What does PFS mean in cancer trials?
Progression free survival is the time from randomization until the tumor grows by a defined amount or the patient dies, whichever happens first.
What does OS mean in cancer trials?
Overall survival is the time from randomization until death from any cause. FDA calls it the most reliable cancer endpoint.
Is PFS a surrogate endpoint?
Sometimes. FDA’s guidance says PFS can be a surrogate for accelerated or traditional approval, or a direct clinical benefit, depending on the disease and effect size.
What is a good objective response rate?
There is no fixed threshold. FDA judges ORR by its size, how long responses last, the share of complete responses, and what therapy already exists.
Why can’t single arm trials show overall survival benefit?
Without a randomized control group, there is nothing reliable to compare survival against. FDA says external controls are seldom reliable for time to event endpoints.
What is the difference between PFS and TTP?
Time to progression ignores deaths. PFS counts deaths as events, which is why FDA prefers it.
What is the difference between DFS and EFS?
DFS starts after surgery or radiotherapy and tracks recurrence or death. EFS starts before definitive treatment and also counts progression that prevents surgery.
What is pathologic complete response?
pCR means no residual cancer is found in tissue removed at surgery after preoperative treatment. FDA has used it as an accelerated approval surrogate in early high risk breast cancer.
Does FDA require overall survival for cancer approval?
Not always. But its August 2025 draft guidance says OS should be prioritized as a primary endpoint when feasible and analyzed as a prespecified safety endpoint when it is not.
Why did FDA reject ivonescimab despite a PFS win?
Summit said FDA concluded substantial evidence was not established without a statistically significant OS result, and that PFS alone did not establish a positive benefit risk.
What is blinded independent central review?
Radiologists who do not know which treatment patients received read the scans. FDA generally wants it for tumor based primary endpoints to limit bias.
Is the 2025 overall survival guidance final?
No. It was issued as a draft in August 2025 and is labeled not for implementation.
Sources
- Primary: FDA, Clinical Trial Endpoints for the Approval of Cancer Drugs and Biologics, final guidance, December 2018 (PDF).
- Primary: FDA, Approaches to Assessment of Overall Survival in Oncology Clinical Trials, draft guidance, August 2025.
- Primary: Federal Register notice 2025-15796, published August 19, 2025.
- Primary: GSK, Nelkarmi approval release, November 17, 2026, and the US prescribing information.
- Primary, via prior BioMed Nexus coverage: Summit Therapeutics CRL release (November 15, 2026) and FDA’s VOLGA approval notice (November 13, 2026).
Related coverage
- FDA rejects ivonescimab over missing overall survival win
- FDA approves Imfinzi plus Padcev for cisplatin ineligible MIBC
- FDA accelerated approval: confirmatory trial rules explained
- What is an FDA complete response letter?
- What is an oral SERD? ESR1 breast cancer drugs compared
- SELLAS Phase 3 REGAL: 41% lower death risk in AML maintenance
Get the daily biotech, pharma and medtech briefing: subscribe to BioMed Nexus.


