Response rates: ORR, CBR and DCR
In one sentence
Response rates are proportions of a defined study population meeting specified tumor-response or disease-control criteria within the protocol's assessment rules.
The intuition
A headline percentage is a fraction with its label partly hidden. Before comparing two percentages, recover three things: what counted, who was counted and when. A larger fraction can result from a broader definition or a smaller denominator rather than a better treatment.
How it works
Response Evaluation Criteria in Solid Tumours (RECIST) supplies rules for assessing overall disease. Complete response (CR) and partial response (PR) are defined response categories. Stable disease (SD) means the overall assessment does not meet response or progression criteria under the applicable rules; it is not necessarily unchanged size.
| Label | Usual components | Definition to recover |
|---|---|---|
| Objective response rate (ORR), sometimes called overall response rate | CR plus PR | Criteria, confirmation and analysis population |
| Disease control rate (DCR) | Often CR, PR and SD | Assessment time and any minimum duration required for SD |
| Clinical benefit rate (CBR) | Often CR, PR and qualifying durable SD | Exact duration rule and which categories qualify |
CBR and DCR are not universal definitions. Their names also do not prove symptom, quality-of-life or survival benefit. Stable disease can be valuable, but in a single-arm study some stability may reflect the disease's untreated course. ORR excludes SD. FDA oncology endpoint guidance, 2018.
“Confirmed” response means a later assessment meets the required confirmation rules. An initial qualifying scan can contribute to an unconfirmed rate while failing to become confirmed. RECIST 1.1 requires confirmation in nonrandomized trials whose primary endpoint is response; protocols and other designs have their own specified rules. Preserve the reported distinction. Eisenhauer et al., 2009.
Why it matters in cancer
Two androgen-receptor research studies illustrate endpoint variation. Gucalp's 2013 advanced, hormone-receptor-negative breast-cancer trial defined CBR using CR, PR or SD lasting more than six months. Traina's 2018 advanced, androgen-receptor-expressing triple-negative breast-cancer trial used CBR at 16 weeks as its primary endpoint and also reported a 24-week endpoint. Those clocks are not interchangeable. Gucalp et al., 2013, Traina et al., 2018.
The latter reported both all-enrolled and a specified evaluable subgroup, with different biomarker and assessment requirements. Removing people without follow-up assessments can raise a rate. Keep the analysis population and handling of missing assessments attached; neither denominator should quietly replace the other.
The readout card
The unit is usually responding participants / specified participants, expressed as a percentage with a confidence interval (CI). Record measurable-disease requirements, response criteria, scan schedule, reviewer, confirmation, data cutoff and missing-data rules. A lesion-level response rate has a lesion denominator and cannot be relabeled as a patient-level rate.
Duration of response asks how long response persists under a defined start, event and censoring rule. It adds information that ORR alone lacks. It is not the same as progression-free survival in the full study population.
Common confusions
- A 30% ORR means 30% met the response definition, not that every tumor shrank by 30%.
- Adding SD changes the endpoint; it does not increase ORR.
- Best response over follow-up differs from disease status at one fixed visit.
- A response signal does not establish comparative survival benefit. Surrogate endpoints explains that bridge.
Try it
In a fictional 20-person study, two have confirmed CR, four confirmed PR and eight SD at week 8. Only three of those eight retain qualifying SD through week 24. Everyone has complete assessments. Its protocol defines DCR at week 8 as CR + PR + SD, and CBR as confirmed CR + PR or SD lasting at least 24 weeks.
What are the three rates?
Answer: ORR is 6/20 = 30%; week-8 DCR is 14/20 = 70%; protocol-defined CBR is 9/20 = 45%. None alone establishes survival benefit. These definitions belong to this invented protocol.
Explain it back
“This percentage counts ___ out of ___ using ___.” One answer: “confirmed complete or partial responses; all 20 participants; the stated criteria and confirmation rules.”
Takeaway
Read the numerator, denominator, time and confirmation rule before the headline percentage.
Related concepts
Response criteria explains the lesion-level measurements behind overall categories. Cross-trial comparison explains why percentages from different studies do not create a direct comparison.
Sources and scope
Source check: October 10, 2026. Fictional arithmetic; named studies illustrate endpoint and population definitions, not current access or treatment recommendations. Expert and learner review pending.
- FDA, December 2018: oncology clinical-trial endpoint guidance — ORR, SD and response durability.
- Eisenhauer et al., 2009: primary RECIST 1.1 guideline — overall response, confirmation and protocol rules.
- Gucalp et al., 2013: bicalutamide phase II study — the abstract's explicit CBR definition in selected advanced breast cancer.
- Traina et al., 2018: enzalutamide phase II study — distinct timepoints and analysis populations in advanced disease.