← Clinical Trials
Melanoma Phase 3 Time-to-Event NCT04099251

CheckMate-76K: Complete Statistical Analysis of Nivolumab in Resected Stage IIB/C Melanoma

An independent statistical analysis of the randomized phase 3 CheckMate-76K trial evaluating nivolumab versus placebo for prevention of recurrent melanoma after complete resection of stage IIB/C melanoma.

Trial start: October 28, 2019  ·  Primary completion: June 28, 2022  ·  Enrollment: 790
Scope of this record

This page separates reported trial results from statistical interpretation. Numerical trial results are restricted to the ClinicalTrials.gov record. The registry provides one formal statistical analysis for the primary recurrence-free survival endpoint.

Registry note: This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record.

1. Trial at a Glance

CheckMate-76K is a randomized, parallel-group, phase 3 trial evaluating nivolumab compared with placebo in participants with melanoma after complete resection of stage IIB/C melanoma. The registered primary endpoint is recurrence-free survival (RFS), a time-to-event endpoint analyzed using a Cox proportional-hazards model.

790
Enrollment
Randomized trial
2
Arms
Nivolumab vs placebo
0.42
RFS HR
95% CI 0.30–0.59
< 0.0001
P-value
Superiority analysis
FeatureCheckMate-76K
PhasePhase 3
ConditionMelanoma
PopulationPrevention of recurrent melanoma after complete resection of stage IIB/C melanoma
DesignRandomized, parallel-group, triple-masked
Primary purposeTreatment
Enrollment790
InterventionsNivolumab and placebo
Primary endpointRecurrence Free Survival (RFS)
Primary endpoint typeTime-to-event
Primary hypothesisSuperiority
Primary analysis methodCox proportional-hazards model
ClinicalTrials.govNCT04099251
Trial statusActive, not recruiting
Lead sponsorBristol-Myers Squibb

2. Clinical Question

The central statistical question is whether nivolumab produces a different recurrence-free survival experience than placebo after complete resection of stage IIB/C melanoma. The registry classifies the primary hypothesis as superiority.

Population

Participants with melanoma after complete resection of stage IIB/C melanoma, as described by the registered brief title.

Intervention

Nivolumab, classified in the registry as a biological intervention.

Comparator

Placebo, classified in the registry as an other intervention.

Primary question

Does nivolumab improve recurrence-free survival compared with placebo after complete resection of stage IIB/C melanoma?

3. Trial Design

The registry describes CheckMate-76K as a randomized, parallel-group, triple-masked phase 3 trial. Its primary purpose is treatment, and the trial contains two intervention arms.

01
Randomize790 participants
02
Two armsNivolumab or placebo
03
Triple maskingRegistry-designated masking
04
FollowTime-to-event outcome
05
AnalyzeCox model for RFS
ARM A · Nivolumab

Nivolumab

  • Nivolumab
  • Biological intervention
  • Compared with placebo in the randomized trial
ARM B · Placebo

Placebo

  • Placebo
  • Other intervention
  • Comparator for the randomized treatment comparison

The registry reports 790 participants enrolled and two arms, but the ClinicalTrials.gov record does not provide arm-specific randomized sample sizes. They therefore are not inferred here.

Trial timing

October 28, 2019

Trial start

The registered study start date was 2019-10-28.

June 28, 2022

Primary completion

The registered primary completion date was 2022-06-28.

Current registry status

Active, not recruiting

The ClinicalTrials.gov record classifies the study as ACTIVE_NOT_RECRUITING.

4. Endpoints

The registry identifies one primary endpoint, and it is a time-to-event endpoint.

EndpointRegistry definition / time frameAnalysis
Recurrence Free Survival (RFS) Time between the date of randomization and the date of first recurrence (local, regional or distant metastasis), new primary melanoma (including melanoma in situ), or death (whatever the cause), whichever occurs first. For participants who remain alive and whose disease has not recurred or did not die, RFS will be censored on the date of last evaluable disease assessment. For those participants who remained alive and had no recorded post-randomization tumor assessment, RFS will be censored on the day of randomization. Cox proportional-hazards model
Why this endpoint is statistically distinctive: RFS combines several possible first events into a single time-to-event outcome. The analysis therefore concerns not simply whether recurrence occurred, but the time from randomization until the first qualifying event or censoring.

5. Statistical Methodology

Primary analysis population

The posted statistical analysis specifies all randomized participants as the analysis population. This is important because the primary efficacy comparison remains anchored to treatment assignment rather than being restricted to participants who completed treatment.

Time-to-event framework

RFS is measured over time rather than as a simple binary endpoint. Participants who experience recurrence, a new primary melanoma, or death contribute an event time. Participants who remain alive without a qualifying event contribute follow-up until censoring.

Cox proportional-hazards model

The registry reports the method as Regression, Cox, normalized here as a Cox proportional-hazards model. The effect measure is the hazard ratio.

Conceptual Cox model
h(t | X) = h0(t) exp(βX)

The hazard ratio for a treatment indicator compares the modeled instantaneous event rates between treatment groups, conditional on the model structure.

Hazard ratio

The hazard ratio is a relative measure of the event rate at a given point in follow-up under the fitted model. It is not a ratio of median RFS times, not a relative risk at a fixed time point, and not a statement that the same percentage of individuals experience the same reduction in recurrence risk.

Interpretation of the hazard ratio
HR < 1  →  lower estimated instantaneous event rate in the nivolumab group

For this trial, the estimated HR is below 1, so the fitted model estimates a lower instantaneous rate of an RFS event for nivolumab relative to placebo over the analyzed follow-up.

Confidence interval

The reported confidence interval quantifies uncertainty around the estimated hazard ratio under the statistical model and sampling framework. It does not describe the range of effects experienced by individual participants.

Superiority hypothesis

The registry classifies the primary hypothesis as superiority. In a superiority analysis, the treatment comparison asks whether the observed data provide evidence that the treatment groups differ in the prespecified direction of interest rather than asking whether a treatment is merely no worse than a predefined non-inferiority margin.

6. Results: Recurrence-Free Survival

ClinicalTrials.gov reports one formal statistical analysis for the primary endpoint. It compares nivolumab with placebo among all randomized participants using a Cox proportional-hazards model.

Primary endpoint: Recurrence-Free Survival

HR 0.42

95% CI: 0.30–0.59   ·   P < 0.0001

Analysis population: all randomized participants  ·  Hypothesis: superiority

Primary endpointNivolumab vs placebo95% CIP-valueMethod
Recurrence Free Survival (RFS) HR 0.42 0.30–0.59 < 0.0001 Cox proportional-hazards model
Clinical Biostats interpretation

The estimated hazard ratio of 0.42 means that, under the fitted Cox model, the estimated instantaneous rate of an RFS event in the nivolumab group was approximately 42% of the rate in the placebo group over the analyzed follow-up. Equivalently, 1 − 0.42 = 0.58, so the estimate corresponds to an approximately 58% lower estimated hazard of an RFS event.

This does not mean that 58% of participants were protected from recurrence, that 58% of participants were cured, or that every individual participant had exactly a 58% reduction in risk. A hazard ratio is a model-based relative measure of event rates over time.

The 95% confidence interval of 0.30–0.59 describes the statistical uncertainty around the estimated hazard ratio. The interval is entirely below 1, which is consistent with a treatment effect favoring nivolumab under the prespecified superiority framework.

The P < 0.0001 result addresses evidence against the relevant null hypothesis under the statistical testing framework. It does not measure the size of the treatment effect. Effect size is described by the hazard ratio, while the confidence interval communicates its precision.

Because the analysis uses a Cox proportional-hazards model, interpretation of a single hazard ratio should be considered in light of the model's proportional-hazards assumption. The ClinicalTrials.gov record does not report a formal assessment of that assumption, nor do they provide a Kaplan-Meier curve, median RFS, event counts, or time-specific RFS estimates.

Educational note: a Kaplan-Meier curve cannot be reconstructed responsibly from the reported HR, confidence interval, and P-value alone. A valid reconstruction requires the underlying event and censoring information or sufficiently detailed source data.

7. How to Read the Primary Result

Relative effect

The HR of 0.42 is the primary reported relative treatment-effect measure. It compares the modeled instantaneous rates of the RFS event between the randomized groups.

Direction

An HR below 1 indicates a lower estimated event hazard for nivolumab relative to placebo under the fitted model.

Precision

The 95% CI of 0.30–0.59 shows the uncertainty around the point estimate. It should not be interpreted as an interval containing individual patient treatment effects.

Statistical evidence

The P-value is reported as < 0.0001. It describes statistical evidence against the null hypothesis; it is not a measure of clinical importance or effect magnitude.

8. Statistical Methods Explained

Why was a Cox proportional-hazards model used?

RFS is a time-to-event endpoint. Some participants experience an event during follow-up, while others may be censored before an event is observed. A Cox model is designed for this setting because it compares event hazards over time while accommodating censored observations.

What does an HR of 0.42 mean?

An HR of 0.42 means that the estimated instantaneous event rate under the fitted model is 0.42 times the comparator rate. Expressed as a relative reduction, 1 − 0.42 = 0.58, corresponding to an approximately 58% lower estimated hazard. This is not equivalent to saying that 58% of patients avoid recurrence.

Why is RFS a time-to-event endpoint rather than a simple recurrence rate?

The registered definition incorporates the time from randomization to the first recurrence, new primary melanoma, or death. The time dimension therefore carries information that would be lost if the analysis simply classified every participant as having or not having an event.

What does the 95% confidence interval tell us?

The interval of 0.30–0.59 expresses uncertainty around the estimated HR under the model and statistical framework. A narrower interval would generally indicate greater precision, while a wider interval would indicate more uncertainty. The interval should not be read as the range of effects among individual patients.

Why doesn't the P-value measure effect size?

The P-value evaluates the compatibility of the observed data with a specified null hypothesis under the statistical testing framework. It depends on both the observed effect and the amount of information in the analysis. The HR communicates the estimated relative effect, while the confidence interval communicates uncertainty around that estimate.

What does the proportional-hazards assumption mean?

The Cox model expresses treatment effects through a hazard ratio that is commonly interpreted as a relative hazard over time. If the proportional-hazards assumption does not adequately describe the underlying event process, a single HR may provide an incomplete summary of how treatment effects vary during follow-up. The ClinicalTrials.gov record does not report a formal proportional-hazards diagnostic.

9. What the Hazard Ratio Does — and Does Not — Mean

The reported estimate

The reported RFS hazard ratio is 0.42. In the context of the Cox model, this corresponds to an estimated instantaneous RFS event rate that is 42% of the placebo rate, or approximately a 58% lower estimated hazard for nivolumab.

What it does not mean

The HR is not a 58% probability of avoiding recurrence, not a 58% reduction in every patient's individual risk, and not a 58% increase in median RFS. Those interpretations require different quantities and, where applicable, different statistical summaries.

Why absolute measures matter

A hazard ratio provides a relative treatment effect but does not directly state the absolute probability of remaining recurrence-free at a specified time. The ClinicalTrials.gov record does not provide time-specific RFS estimates or median RFS, so those quantities are not inferred here.

10. Safety: Serious Adverse Events

The ClinicalTrials.gov record reports serious adverse events by arm as affected participants divided by participants at risk. These figures should be kept separate from the efficacy analysis because safety and efficacy answer different questions and can use different analysis populations or definitions.

GroupParticipants with serious adverse eventsParticipants at risk
Nivolumab95524
Placebo37264
Open-Label Nivolumab530
Safety interpretation: the registry-reported serious-adverse-event data include an Open-Label Nivolumab group in addition to the randomized nivolumab and placebo groups. Because the ClinicalTrials.gov record does not describe the circumstances defining that group, these counts should not be folded into the randomized efficacy comparison or interpreted as if all three groups were randomized arms.

11. Analysis Population and Randomization

The formal RFS analysis specifies all randomized participants as the analysis population. This is an important feature of randomized clinical-trial inference because treatment assignment is determined before the outcome is observed.

FeatureRegistered information
AllocationRandomized
Design modelParallel
MaskingTriple
Primary analysis populationAll randomized participants
Groups comparedNivolumab vs placebo
Primary hypothesisSuperiority
Primary endpoint typeTime-to-event

Randomization helps make the treatment groups comparable in expectation with respect to measured and unmeasured baseline factors. Triple masking can additionally reduce opportunities for treatment knowledge to influence aspects of trial conduct or assessment, although the ClinicalTrials.gov record does not specify the identities of the three masked parties.

12. Multiplicity, Interim Analysis, and Other Design Topics

The ClinicalTrials.gov record identifies one registered primary endpoint and one posted formal statistical analysis. They do not provide a multiplicity-adjustment procedure, an interim-analysis schedule, an alpha-spending plan, a non-inferiority margin, a crossover strategy, an imputation method, or a Bayesian analysis.

Design topicWhat can be concluded from the ClinicalTrials.gov record
Primary endpoint multiplicityOne primary endpoint is identified: RFS.
Formal primary analysesOne statistical analysis is posted for the primary endpoint.
Hypothesis typeSuperiority.
Non-inferiority marginNot applicable to the reported superiority hypothesis; no margin is reported.
Interim analysisNo interim-analysis details are reported.
CrossoverNo crossover details are reported.
Missing-data / imputation methodNo specific imputation method is posted on ClinicalTrials.gov for the primary analysis.
Bayesian methodsNo Bayesian method is reported in the ClinicalTrials.gov record.

This distinction is important. The absence of a registry-reported design detail should not be converted into an assumption about how the protocol or statistical analysis plan handled it.

13. What Is and Is Not Reported in the Supplied Results

The ClinicalTrials.gov record contains 12 posted outcome measures but only one posted statistical analysis. That formal analysis concerns the primary RFS endpoint.

Reported formally

RFS, analyzed among all randomized participants using a Cox proportional-hazards model, with HR 0.42, 95% CI 0.30–0.59, and P < 0.0001.

Not interpreted as formal comparisons

The ClinicalTrials.gov record does not provide formal statistical analyses for secondary endpoints, so no secondary efficacy comparison is constructed here.

Not inferred

Median RFS, event counts, time-specific RFS rates, subgroup estimates, and Kaplan-Meier estimates are not reported and are therefore not calculated.

Safety information available

Serious adverse-event counts and denominators are posted on ClinicalTrials.gov for nivolumab, placebo, and open-label nivolumab.

14. Limitations and Interpretation Issues

15. Why This Trial Matters Statistically

CheckMate-76K is a useful teaching example because the ClinicalTrials.gov record illustrates the central statistical structure of a randomized time-to-event trial without requiring unsupported assumptions about additional analyses.

ConceptHow it appears in CheckMate-76K
RandomizationThe allocation is explicitly randomized.
Parallel-group designThe registry identifies a parallel design model with two intervention groups.
BlindingThe trial is classified as triple-masked.
Time-to-event endpointRFS is the registered primary endpoint.
Kaplan-Meier frameworkRFS is a time-to-event endpoint for which Kaplan-Meier estimation is a standard descriptive approach.
Cox modelThe formal posted analysis uses a Cox proportional-hazards model.
Hazard ratioThe primary treatment effect is reported as an HR of 0.42.
Confidence intervalThe 95% CI is 0.30–0.59.
P-valueThe primary analysis reports P < 0.0001.
Superiority testingThe registered hypothesis type is superiority.
ITT-style efficacy principleThe primary analysis population is all randomized participants.
Safety analysisSerious adverse events are reported by group with affected and at-risk counts.

16. Statistical Concepts in This Trial

Learn more about the methods used in this trial:

17. Related Statistical Calculators

18. Sources

Continue with the statistical methods

Explore the underlying survival-analysis concepts, confidence intervals, hypothesis testing, and clinical-trial methods used to interpret randomized time-to-event studies.

19. Record Summary

CheckMate-76K provides a clear example of randomized time-to-event analysis. The trial enrolled 790 participants in a randomized, parallel, triple-masked phase 3 design comparing nivolumab with placebo. Its registered primary endpoint was recurrence-free survival, defined from randomization to the first recurrence, new primary melanoma, or death, with censoring for participants remaining alive without a qualifying event. The posted formal analysis used a Cox proportional-hazards model in all randomized participants and reported an HR of 0.42 with a 95% CI of 0.30–0.59 and P < 0.0001.

Statistically, the most important lesson is that these three quantities answer different questions. The hazard ratio describes the estimated relative event rate under the Cox model, the confidence interval describes uncertainty around that estimate, and the P-value describes statistical evidence against the relevant null hypothesis. None of these quantities alone describes the absolute probability of remaining recurrence-free at a particular time.

The ClinicalTrials.gov record does not provide median RFS, event counts, time-specific RFS estimates, subgroup estimates, or detailed interim and multiplicity procedures. Those quantities are therefore not reconstructed here. Keeping the distinction between reported evidence and statistical interpretation is essential when analyzing a clinical trial from registry data.

Clinical Biostats methodology: A trial-results page should not merely repeat the registry. The goal is to explain what the reported statistical quantities mean, what they do not mean, and which conclusions are supported by the available analysis without filling gaps in the record with unsupported assumptions.