Cancer trial terms, in plain language

Cancer trial terms in our research (single-arm and randomized trials, response rate, duration of response, survival, RECIST) explained simply.

Cancer drugs are judged with their own set of measurements, and FDA letters about them use those terms precisely. This page explains them in plain language, in the same three parts as the FDA approval terms: our explanation, the FDA's words in a shaded box with a link, and an example from our research.

Most examples come from Replimune's RP1, a treatment for advanced melanoma (a skin cancer) made from an engineered virus that is injected into tumours and given together with nivolumab, an established cancer drug. Its two FDA letters use nearly every term on this page. Open Replimune's question web →

How cancer trials are built#

Phase 1, 2 and 3

Drugs are tested in steps. Phase 1 looks at safety and dose in a small group. Phase 2 looks for signs that the drug works. Phase 3 is the large trial meant to prove it. In cancer, even Phase 1 is done in patients, not healthy volunteers.

“Clinical trials follow a typical series from early, small-scale, Phase 1 studies to late-stage, large scale, Phase 3 studies.” · “However, if a new drug is intended for use in cancer patients, researchers conduct Phase 1 studies in patients with that type of cancer.”FDA · Step 3: Clinical Research

RP1's approval rested on a single-arm trial, RPL-001-16; its Phase 3 trial, RP1-104, was still running. Read it →

Randomized controlled trial

Patients are assigned by chance to get either the new treatment or a comparison treatment (the control). Because chance decides, the two groups are alike, so differences in outcome can be credited to the treatment. It is the FDA's preferred design.

“Given the limitations of single-arm trials, a randomized controlled trial is the preferred approach to support an application for accelerated approval.”FDA draft guidance · Clinical Trial Considerations to Support Accelerated Approval of Oncology Therapeutics (March 2023)

Single-arm trial

Every patient gets the new treatment; there is no comparison group in the trial. It is faster and cheaper, and tumour shrinkage can still be measured, because tumours rarely shrink on their own. But with nothing to compare against, it is harder to know how much of the result is due to the drug.

“Response rate is a marker of drug activity because malignant tumors do not typically regress on their own, and because this endpoint can be interpreted in single-arm trials for monotherapy oncology drug regimens.”FDA draft guidance · Clinical Trial Considerations to Support Accelerated Approval of Oncology Therapeutics (March 2023)

“… we ‘would not recommend that the Sponsor submit a BLA based on the results of a single-arm study’.”FDA letter to Replimune, April 10, 2026, page 3, quoting a 2021 meeting

The FDA's concern about RP1's single-arm trial dates back to 2021. Read it →

Control arm

The comparison group in a trial. It may get a standard treatment, a placebo, or the doctor's choice of treatment (physician's choice or investigator's choice). What the control arm gets matters: if it gets something weak, the new drug looks better than it is.

“The study design may not isolate the contribution of vusolimogene oderparepvec due to the inclusion of single-agent chemotherapy options in the investigator’s choice control arm.”FDA letter to Replimune, July 21, 2025, page 2

The control arm of RP1's Phase 3 trial was under discussion from the first letter onwards. Read it →

External control and historical control

When a trial has no control arm, results are sometimes compared with patients outside the trial, such as patients treated in the past (a historical control) or records from a disease registry (natural history). The risk is that those patients differed in ways that affect the comparison.

“In an externally controlled trial, outcomes in participants receiving the test treatment according to a protocol are compared to outcomes in a group of people external to the trial who had not received the same treatment.” · “The external control arm can be a group of people, treated or untreated, from an earlier time (historical control), or it can be a group of people, treated or untreated, during the same time period (concurrent control) but in another setting.”FDA draft guidance · Considerations for the Design and Conduct of Externally Controlled Trials (February 2023)

REGENXBIO's letter questioned whether its natural-history external control was comparable. See the concept preview →

Heterogeneity

How different the patients in a group are from each other: how many earlier treatments they had, how far the disease had spread. A very mixed group makes comparisons with other groups unreliable.

“However, comparing response rates cross-trial from RPL-001-16 to historical controls is challenging due to the heterogeneity of the patient populations in RPL-001-16 and the historical controls (e.g., the type of prior therapy received, such as single agent anti-PD-1 or combination checkpoint inhibitor, duration of prior therapy …)”FDA letter to Replimune, April 10, 2026, page 5

Combination therapy and “contribution of each component”

When two drugs are given together and patients improve, which drug did it? If a trial gives everyone both, it cannot tell. The FDA often asks for a design that separates each drug's effect.

“The RPL-001-16 trial was also not designed to isolate the contribution of vusolimogene oderparepvec to the observed ORR when administered in combination with nivolumab.”FDA letter to Replimune, July 21, 2025, page 1

Intent-to-treat population, and other populations

Results depend on which patients are counted. The intent-to-treat population is everyone who entered the trial, whatever happened next. Narrower groups (for example only patients who could be evaluated, or who met an extra condition) can give quite different numbers from the same trial.

“Overall survival is defined as the time from randomization until death from any cause and is measured in the intent-to-treat population.”FDA guidance · Clinical Trial Endpoints for the Approval of Cancer Drugs and Biologics (December 2018)

RP1's response rate is reported as 34%, a confirmed 32.9% and 24.2%: three figures for three populations. Read it →

How results are measured#

Endpoint

The outcome a trial is designed to measure, decided before the trial starts. The primary endpoint is the main one; a trial “meets” or “misses” it.

Outlook's NORSE EIGHT trial missed its week-8 endpoint. Read it →

Overall survival OS

How long patients live. The clearest measure of benefit, but it takes years to collect.

Progression-free survival PFS

How long patients live without the cancer growing or spreading.

“PFS is defined as the time from randomization until objective tumor progression or death, whichever occurs first.”Same guidance

Objective response rate ORR

The share of patients whose tumours shrank by at least a set amount. “A 34% response rate” means about one patient in three responded. It is a surrogate endpoint: shrinking tumours is a good sign, but is not the same as living longer.

“ORR is defined as the proportion of patients with tumor size reduction of a predefined amount and for a minimum time period.”Same guidance

“Because ORR is directly attributable to drug effect, single-arm trials conducted in patients with refractory tumors where no available therapy exists provide an accurate assessment of ORR.”Same guidance, PDF

RP1 received accelerated approval on response rate and duration of response. See the answer and its reasoning →

Duration of response DoR

For patients who responded, how long the response lasted before the cancer grew again. A high response rate matters less if responses are brief.

“Response duration usually is measured from the time of initial response until documented tumor progression.”Same guidance

Replimune's April 2026 8-K reported a median duration of response of 24.8 months.

RECIST and response criteria

The rulebook for deciding whether a tumour “responded”: how to measure it on scans and how much it must shrink. If a trial uses its own rules instead of the standard ones, its response rate may not be comparable with other trials.

“A variety of response criteria have been considered appropriate (e.g., revised RECIST guideline (version 1.1))” · “The response criteria should be predefined in the protocol before the start of the study.”FDA guidance · Clinical Trial Endpoints for the Approval of Cancer Drugs and Biologics (December 2018)

“In the Type A meeting dated September 16, 2025, FDA also communicated that the response criteria used in RPL-001-16 were not consistent with RECIST v1.1 and may not be comparable to response rates reported in the historical literature.”FDA letter to Replimune, April 10, 2026, page 4

Confirmed response

A response seen on one scan and seen again on a later scan, so it is not a one-off measurement. Confirmed response rates are usually a little lower than unconfirmed ones. (Our explanation.)

Replimune's 10-K: “The confirmed ORR was 32.9%”, against 34% in its April 8-K. Read it →

Injected and non-injected lesions

Some treatments, like RP1, are injected straight into tumours (lesions). If only the injected tumours shrink, the effect may be local. Shrinkage in tumours that were not injected shows the treatment working across the body (systemic activity), which is what the FDA wanted to see.

“Thus, over half (53%) of patients with objective response did not have noninjected target lesions to assess systemic anti-tumor activity.”FDA letter to Replimune, April 10, 2026, page 4

The concern traces back to a 2021 meeting. At approval, Replimune's 8-K reported “91 patients with at least one non-injected lesion included in the efficacy-evaluable population”. Read it →

Anti-PD-1 and checkpoint inhibitors; refractory

Checkpoint inhibitors are cancer drugs that release a brake on the immune system so it can attack tumours. Anti-PD-1 drugs, such as nivolumab, are one kind. A patient whose cancer kept growing despite such a drug is said to have progressed on it, and the cancer may be called refractory: these patients have few options, which is why the FDA may accept less conventional evidence for them. (Our explanation.)

“… the numerically higher response rate cannot be adequately interpreted due to heterogeneity of the RPL-001-16 trial patient population, including the type of prior therapy received, such as single agent anti-PD-1 or combination checkpoint inhibitor …”FDA letter to Replimune, July 21, 2025, page 1

Who reviews cancer drugs#

Oncology Center of Excellence OCE

The FDA centre that brings together its cancer experts across drugs, biologics and devices. Its experts often take part in reviews of cancer applications.

“The Oncology Center of Excellence (OCE) was authorized by the 21st Century Cures Act of 2016 and established on January 19, 2017.”FDA · Oncology Center of Excellence

The FDA's second letter to Replimune says its review team, supervisory leadership and “subject matter experts from Oncology Center of Excellence unanimously determined” the data insufficient. Read it →

Sources#

FDA definitions were checked against the FDA pages and guidance documents linked beside them on October 7, 2026. Quotations from FDA letters are matched to the letters on download.open.fda.gov by the same checks as every deep research answer (see How deep research is verified). Entries marked “our explanation” have no single FDA definition; read them as a starting point.