Selecting an appropriate primary endpoint for a Phase 2 trial in recurrent or metastatic head and neck squamous cell carcinoma (R/M HNSCC) is a consequential decision with direct implications for regulatory acceptability, trial feasibility, and the ability to advance a program into pivotal development. FDA and EMA have signaled distinct preferences depending on trial design, whether accelerated or traditional approval is sought, and the evidentiary context established by prior approvals in the indication. Misalignment between the chosen endpoint and agency expectations can result in a trial that generates data insufficient to support a registration pathway or that requires costly amendments.
This analysis examines the primary endpoints used and accepted in Phase 2 R/M HNSCC trials by reviewing ClinicalTrials.gov registrations, FDA approval packages and review documents for key agents in the indication, relevant FDA oncology guidance on endpoint acceptability, and the EMA European Public Assessment Report for pembrolizumab in HNSCC. The goal is to characterize which endpoints dominate in practice, what regulatory rationale underpins each, and how design considerations such as single-arm versus randomized architecture influence endpoint selection.
Want Rhizome's help on your own question? Try it for free.
Primary endpoints for Phase 2 trials in recurrent/metastatic head and neck squamous cell carcinoma
Bottom line
For most Phase 2 R/M HNSCC trials, especially single-arm studies in the platinum-refractory setting, objective response rate (ORR) by RECIST 1.1 is the primary endpoint of choice, with duration of response (DoR) as an essential companion measure. This is what the ClinicalTrials.gov landscape shows in practice (ORR is the most frequently named Phase 2 primary endpoint) 899594221, and it matches the regulatory precedent: pembrolizumab's first R/M HNSCC indication was an accelerated approval based on ORR from the KEYNOTE-012 single-arm cohort, assessed by independent central review under RECIST 1.1 3032. Randomized Phase 2 studies more often use PFS as the primary endpoint 81113, while OS is generally reserved for randomized registrational (Phase 3) trials because of feasibility and follow-up demands.
Endpoint-by-endpoint assessment
Objective response rate (ORR) — the default Phase 2 primary endpoint
ORR (confirmed complete plus partial responses under a predefined criterion such as RECIST 1.1) is the workhorse Phase 2 primary endpoint in this disease. FDA guidance treats ORR as a direct measure of antitumor activity that can be assessed in a single-arm trial, because in a refractory tumor with no available therapy the response is directly attributable to the drug 112136. It is the most frequently used endpoint to support accelerated approval in oncology single-arm trials 120. In R/M HNSCC specifically, ORR supported pembrolizumab's accelerated approval (KEYNOTE-012) 3032, and it is the primary endpoint most commonly named across current Phase 2 R/M HNSCC records on ClinicalTrials.gov, including anti-EGFR and first-line combination studies 899594221. EMA relied on ORR for the single-arm supportive HNSCC studies (KEYNOTE-012 and KEYNOTE-055) as well 146.
Design points that regulators expect:
- Use standardized, pre-specified response criteria (RECIST 1.1 is the cited example) fixed in the protocol before the study starts 136.
- Report confirmed responses, and pair ORR with the percentage of complete responses and with DoR, because the clinical meaning of ORR depends on the magnitude and durability of response 136113.
- Low-magnitude response rates are generally not considered reasonably likely to predict clinical benefit, so power the study to detect a response rate that is clinically compelling for the line of therapy 120.
Duration of response (DoR) — the required companion to ORR
FDA repeatedly ties the interpretability of ORR to durability: DoR is measured from initial response to documented progression, and for accelerated approval the responses should be durable 136113. Durable ORR has itself been used to support approval in some settings 131. In practice DoR is best specified as a co-primary or key secondary endpoint alongside ORR rather than left as an afterthought.
Progression-free survival (PFS) — primary in randomized Phase 2
When a Phase 2 study is randomized (for example, an experimental regimen versus a control or versus standard chemotherapy), PFS is a common primary endpoint and appears as a Phase 2 R/M HNSCC primary measure in the trial landscape 81. FDA accepts time-to-event tumor endpoints such as PFS to support approval when the treatment effect is substantial, statistically robust, and clinically meaningful in context; the agency declines to fix a universal effect-size threshold because context matters 113. PFS avoids the confounding of post-progression therapy that complicates OS in a heavily pretreated, multi-line disease, which makes it attractive for a randomized signal-seeking Phase 2.
Overall survival (OS) — the preferred endpoint, but usually a Phase 3 choice
OS is FDA's preferred and most reliable oncology endpoint when feasible, and it should be evaluated in randomized controlled studies 117. In R/M HNSCC it was the primary endpoint of the pivotal randomized registrational trials: OS for nivolumab in CheckMate-141 (with PFS and ORR as secondary) 14, OS in the pembrolizumab KEYNOTE-040 pivotal trial per the EMA assessment 146, and OS as one of the co-primary endpoints (with PFS) in the first-line KEYNOTE-048 study, tested sequentially across PD-L1 CPS subgroups 232235238243. Cetuximab's R/M HNSCC approval likewise rested on OS together with PFS and ORR 5658. For a Phase 2 trial, OS is rarely the primary endpoint because the sample size and follow-up needed to show a survival effect are generally beyond Phase 2 scope; OS is more appropriately collected as a secondary/exploratory endpoint to inform later development.
Disease control rate (DCR) — use with caution
DCR (adding stable disease to responders) is sometimes proposed for Phase 2, but FDA guidance is explicit that stable disease should not be included in ORR and that "clinical benefit rate" measures which fold in stable disease should not be used as the efficacy endpoint 135136. DCR can be reported descriptively, but it is not a regulator-preferred primary endpoint.
How the choice maps to trial design
| Phase 2 design | Typical primary endpoint | Regulatory rationale |
|---|---|---|
| Single-arm, platinum-refractory / later-line | ORR (RECIST 1.1, confirmed, ICR/BICR) with DoR | Direct measure of activity attributable to drug in refractory disease; basis for accelerated approval when response is large and durable 112113120136; precedent = KEYNOTE-012 3032 |
| Randomized Phase 2 (vs control/standard) | PFS | Time-to-event tumor endpoint acceptable when effect is substantial and clinically meaningful 113; avoids post-progression confounding 81 |
| Signal-seeking with survival intent | OS as co-primary or key secondary | OS is preferred/most reliable but usually needs Phase 3 scale 117; collect to de-risk registrational design |
| Any design proposing DCR | Avoid DCR as primary | Stable disease should not anchor the efficacy endpoint 135136 |
Regulatory context you should build into the protocol
- Accelerated vs regular approval. ORR (with durable responses) can support accelerated approval, particularly from single-arm trials in refractory disease with no available therapy; a confirmatory randomized trial evaluating PFS or OS is then expected to verify benefit 112113120135142. Design the Phase 2 with the confirmatory pathway already in mind.
- PD-L1 (CPS) stratification. The first-line pembrolizumab program used PD-L1 combined positive score to define efficacy-testing subgroups and the hypothesis-testing sequence (CPS >=20, then CPS >=1, then ITT) 232235238243. Even in Phase 2, pre-specifying CPS-defined analyses aligns the study with how the disease's endpoints are now interpreted.
- Independent review. Registrational ORR in this disease was assessed by independent central / blinded independent central review under RECIST 1.1 30232; building ICR/BICR into a Phase 2 ORR endpoint strengthens its regulatory weight.
- EMA alignment. The European assessment used OS for the pivotal randomized trial and ORR for single-arm supportive studies 146, so an ORR-primary Phase 2 supported by DoR travels well across FDA and EMA.
Limitations and gaps
The ClinicalTrials.gov extraction reflects registry text that is inconsistently completed: many Phase 2 R/M HNSCC records describe an objective ("evaluate efficacy and safety") without naming a discrete primary endpoint, so the ORR-dominant pattern is directional rather than an exhaustive census 7677787983879799. The EU Clinical Trials Register returned relevant Phase 2 titles but did not expose primary-endpoint fields in the retrieved records, so European trial-level endpoint frequencies could not be tallied here 167171172. Pathological complete response (pCR) and event-free survival (EFS) appear in HNSCC development but belong to the resectable/neoadjuvant setting, not R/M disease 150. A natural follow-up is to profile a specific comparator set (for example, all single-arm anti-PD-1 combination Phase 2 studies) and pull their exact ORR thresholds, DoR definitions, and statistical assumptions.