Expansion Cohort

An expansion cohort is an additional group of participants enrolled at a dose already selected during the dose-escalation portion of a Phase 1 trial — typically the maximum tolerated dose (MTD) or a recommended Phase 2 dose (RP2D) — in order to further characterize that dose’s safety profile and generate preliminary evidence of anti-tumor or other biological activity before the drug moves into later-phase testing. Expansion cohorts are most common in first-in-human (FIH) oncology trials, where a single protocol runs dose escalation followed by one or more expansion cohorts rather than opening a separate Phase 2 study.

How an expansion cohort fits into trial design

A conventional Phase 1 oncology trial has two conceptual stages under one protocol:

• Dose escalation — small cohorts (commonly using a 3+3 or rolling-six design, or a model-based method such as the continual reassessment method) are dosed in sequence to identify the MTD and/or RP2D. See Phase 1 Trial for how that stage works.
• Expansion — once a dose is selected, additional participants are enrolled specifically at that dose. Because the escalation cohorts were small and often heterogeneous in tumor type, the expansion stage exists to answer questions escalation cohorts weren’t powered to answer: is the safety profile consistent in a larger, more homogeneous group, and is there a detectable signal of clinical activity?

A trial with a single expansion cohort simply enrolls more patients at the selected dose, usually within one tumor type or indication. A trial with a multiple expansion cohort (MEC) design opens several expansion cohorts in parallel — for example, separate cohorts for different tumor types, or cohorts defined by a biomarker or mutation status — each evaluated against the same dose but analyzed as its own group, often with its own stopping rule (a Simon two-stage design is common, though FDA’s guidance notes other designs are also acceptable).

FDA framework: the 2022 expansion cohort guidance

FDA’s expectations for this design are set out in Expansion Cohorts: Use in First-in-Human Clinical Trials to Expedite Development of Oncology Drugs and Biologics — Guidance for Industry, finalized and announced in the Federal Register on March 2, 2022 (finalizing a draft version issued in August 2018). The guidance covers:

• which drug characteristics make a candidate suitable for a combined dose-escalation-plus-expansion design rather than separate studies;
• what to include in the IND submission to justify opening each individual expansion cohort;
• when sponsors should interact with FDA as the protocol evolves (expansion cohorts are frequently added, and their scope changed, via protocol amendment rather than a new IND); and
• safeguards for subjects enrolled in expansion cohorts — FDA is explicit that the added complexity and risk of running several expansion cohorts in one FIH protocol means this design is appropriate only for serious or life-threatening diseases where the risk-benefit calculus supports it, and that oversight (adequate stopping rules, independent safety review, informed consent tailored to each cohort) has to scale with the number and complexity of cohorts running at once.

Because expansion cohorts and dose escalation share one IND and one protocol, sponsors typically amend the protocol to add or modify cohorts rather than filing a new study — this is what makes the design faster than running dose escalation and a separate Phase 2 as sequential, independently-initiated trials.

Expansion cohorts vs. related multi-arm designs

Expansion cohort designs are frequently discussed alongside other “master protocol” designs, but they are not the same thing:

• An expansion cohort trial is a single-drug, single-protocol design where the defining feature is the sequence — dose escalation first, then expansion at a fixed dose. It may include multiple expansion cohorts by tumor type or biomarker, but all cohorts receive the same investigational product at (or near) the same dose.
• An umbrella trial instead tests multiple different drugs against a single disease, with patients assigned to a biomarker-matched arm — the multi-arm structure is organized around matching drug to biomarker, not around confirming one dose. Umbrella and basket trials are two of the master-protocol design types described in FDA’s separate 2022 master protocols guidance for oncology.

In practice, a multiple expansion cohort trial and a basket trial can look superficially similar (both enroll patients across several tumor types in parallel cohorts), but a basket trial is organized around a shared biomarker across diseases as the entry criterion for a single treatment, while a multiple expansion cohort trial is organized around confirming a dose selected earlier in the same protocol, with tumor type or biomarker status used to define distinct expansion cohorts rather than as the trial’s core structuring principle.

Illustrative example

A representative pattern in oncology FIH trials: a novel targeted therapy completes 3+3 dose escalation across five dose levels, identifying an RP2D. The protocol then opens four parallel expansion cohorts at that RP2D — one each for non-small cell lung cancer, melanoma, and two biomarker-defined subgroups — with each cohort using its own sample size and activity threshold to decide whether that indication warrants further development. This is a composite, illustrative description of how the design is commonly used in practice, not a specific named trial.

Counter-example

A dose-escalation cohort itself — the 3+3 cohorts used to climb from one dose level to the next while still searching for the MTD — is not an expansion cohort, even though both occur under the same protocol. The distinguishing feature of an expansion cohort is that the dose is already fixed; participants are added to characterize that dose, not to help find it. Likewise, a basket trial enrolling patients across tumor types by shared biomarker, with no preceding within-protocol dose-escalation stage of its own, is a master-protocol design rather than an expansion cohort trial.

Related terms

• Phase 1 Trial — covers dose-escalation methodology (3+3, rolling six, MTD, RP2D) in depth.
• Umbrella Trial — a related but distinct master-protocol design matching multiple drugs to one disease by biomarker.