← Clinical Trials
Cutaneous Melanoma Phase 3 Time-to-Event Analysis NCT01597908

COMBI-v: Complete Statistical Analysis of Dabrafenib Plus Trametinib in BRAF V600E/K Melanoma

An independent statistical analysis of the randomized phase 3 COMBI-v trial comparing dabrafenib plus trametinib with vemurafenib alone in unresectable or metastatic BRAF V600E/K cutaneous melanoma.

Trial status: COMPLETED  ·  Enrollment: 704  ·  Primary completion: April 17, 2014
Scope of this record

This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record. Numerical trial results on this page are restricted to the ClinicalTrials.gov record.

1. Trial at a Glance

COMBI-v was a randomized, open-label, parallel phase 3 trial evaluating dabrafenib plus trametinib versus vemurafenib in patients with unresectable or metastatic BRAF V600E/K cutaneous melanoma. The registered primary endpoint was overall survival, a time-to-event endpoint analyzed in the intention-to-treat population.

704
Enrollment
2 treatment arms
3
Phase
Randomized
0.70
OS HR
95% CI 0.58–0.83
0.62
PFS HR
95% CI 0.52–0.73
FeatureCOMBI-v
Trial nameCOMBI-v
PhasePhase 3
ConditionMelanoma
PopulationUnresectable or metastatic BRAF V600E/K cutaneous melanoma
DesignRandomized, parallel-group, unmasked
Primary purposeTreatment
Enrollment704
Primary endpointOverall Survival (OS)
Primary endpoint typeTime-to-event
Primary hypothesisSuperiority
Lead sponsorNovartis Pharmaceuticals
ClinicalTrials.govNCT01597908

2. Clinical Question

The primary statistical question was whether dabrafenib plus trametinib produced a different overall-survival experience from vemurafenib in patients with unresectable or metastatic BRAF V600E/K cutaneous melanoma, under a prespecified superiority framework.

Population

Patients with unresectable or metastatic BRAF V600E/K cutaneous melanoma.

Intervention

Dabrafenib plus trametinib.

Comparator

Vemurafenib alone.

Primary question

Does dabrafenib plus trametinib improve overall survival relative to vemurafenib?

3. Trial Design

01
Randomize704 participants
02
Two armsCombination vs single drug
03
FollowTime-to-event outcomes
04
AnalyzeITT population
05
CompareHazard ratios
Allocation
Randomized
Participants were allocated to the two treatment groups.
Design model
Parallel
The registry identifies the study as a parallel-group trial.
Masking
None
The registry identifies the trial as unmasked.
Primary purpose
Treatment
The study was designed to evaluate treatment effects.
TREATMENT ARM

Dabrafenib Plus Trametinib

  • Dabrafenib
  • Trametinib
  • Combination treatment
COMPARATOR ARM

Vemurafenib

  • Vemurafenib
  • Single-drug comparator

The ClinicalTrials.gov record does not provide arm-specific randomized sample sizes, allocation ratios, stratification factors, crossover rules, non-inferiority margins, or details of interim-analysis boundaries. Those design features are therefore not inferred here.

4. Trial Timeline

June 4, 2012

Study start

The registered study start date was 2012-06-04.

April 17, 2014

Primary completion

The registered primary completion date was 2014-04-17.

Completed

Registry status

The trial is listed as COMPLETED, with results posted on ClinicalTrials.gov.

5. Endpoints

EndpointRegistry definition / time frameRole
Overall Survival (OS) From the date of randomization until date of death due to any cause (up to approximately 6 years). For patients who did not die, OS was censored at the date of last contact. Primary
Progression-Free Survival (PFS), as Assessed by the Investigator From randomization until the earliest date of disease progression (PD) or death due to any cause (up to approximately 6 years). Secondary
Duration of Response (DOR), as Assessed by the Investigator From the time of the first documented response (CR or PR) until disease progression (up to approximately 6 years). Secondary

The registry classifies the primary endpoint as a time-to-event outcome. The same general time-to-event framework is used for the posted secondary PFS and DOR analyses.

6. Analysis Populations

PopulationRole in the registry-reported analysis
Intent-to-Treat (ITT) The analysis population specified for the posted OS, PFS, and DOR statistical analyses.
Confirmed responders For the DOR analysis, only participants with confirmed response by RECIST version 1.1 were included.

The ITT designation is particularly important for the primary OS comparison. Participants are analyzed according to their randomized treatment assignment rather than being reassigned according to subsequent treatment exposure. This preserves the comparison created by randomization.

7. Primary Result: Overall Survival

The registered primary endpoint was overall survival, defined as the interval between randomization and death due to any cause, with censoring at last contact for participants who did not die. The registry-reported statistical analysis compares dabrafenib plus trametinib with vemurafenib in the ITT population.

Hazard ratio for overall survival

0.70

95% CI: 0.58–0.83   ·   Two-sided 95% CI

Superiority analysis; p-value not reported in the ClinicalTrials.gov record.

EndpointComparisonEffect measureEstimate95% CIHypothesis
Overall Survival Dabrafenib + Trametinib vs Vemurafenib Hazard Ratio 0.70 0.58–0.83 Superiority
Clinical Biostats interpretation

What the estimate means: An HR of 0.70 means that, under the time-to-event model represented by the reported hazard ratio, the estimated instantaneous hazard of death in the dabrafenib-plus-trametinib group was 70% of that in the vemurafenib group. Expressed as a simple relative-hazard interpretation, this corresponds to an estimated 30% lower hazard.

What it does not mean: It does not mean that 30% of patients were saved, that every patient experienced a 30% reduction in risk, or that the probability of death was exactly 30% lower at every point in time.

What the confidence interval says: The two-sided 95% confidence interval of 0.58–0.83 describes uncertainty around the estimated hazard ratio under the analysis framework. It does not describe the range of individual patient outcomes.

Why the p-value is different: The ClinicalTrials.gov record does not report a p-value for this comparison. In general, a p-value addresses compatibility with a null hypothesis; it does not measure the size or clinical importance of an observed effect.

Important time-to-event caution: A hazard ratio is a relative measure of event rates over follow-up. Its interpretation relies on the underlying survival-analysis framework, and a single HR can be less descriptive if the relative hazards vary materially over time. The registry analysis does not report a separate assessment of the proportional-hazards assumption.

Analysis method reported by the registry

Pike estimator
Hazard Ratio = estimated relative event hazard for dabrafenib + trametinib versus vemurafenib

The registry states that hazard ratios were estimated using a Pike estimator. The ClinicalTrials.gov record does not report a more detailed normalized statistical-method description.

8. Secondary Result: Progression-Free Survival

Investigator-assessed progression-free survival was a secondary time-to-event endpoint. The registry definition measures time from randomization until the earliest date of disease progression or death due to any cause, with a time frame of up to approximately 6 years.

Hazard ratio for progression or death

0.62

95% CI: 0.52–0.73   ·   Two-sided 95% CI

Superiority analysis; p-value not reported in the ClinicalTrials.gov record.

EndpointComparisonEffect measureEstimate95% CIHypothesis
Progression-Free Survival Dabrafenib + Trametinib vs Vemurafenib Hazard Ratio 0.62 0.52–0.73 Superiority
Clinical Biostats interpretation

What the estimate means: An HR of 0.62 corresponds to an estimated instantaneous hazard of progression or death that is 62% of the comparator hazard under the reported time-to-event analysis. In simple relative terms, this corresponds to an estimated 38% lower hazard.

What it does not mean: It is not a 38% absolute reduction in the probability of progression, and it does not mean that each participant experienced the same proportional reduction in risk.

Precision: The 95% CI of 0.52–0.73 quantifies uncertainty around the estimated hazard ratio. The interval is entirely below 1, but the ClinicalTrials.gov record does not provide a formal p-value.

Interpretation: PFS combines two possible events—disease progression and death. Consequently, an HR for PFS is not an estimate of the hazard of death alone.

Censoring: Time-to-event methods can incorporate participants whose event has not occurred by their last evaluable follow-up. The registry text does not provide enough detail to reconstruct the censoring rules beyond the endpoint definition.

9. Secondary Result: Duration of Response

Duration of response was assessed by the investigator and defined from the first documented response, complete response or partial response, until disease progression, with a time frame of up to approximately 6 years. The analysis population was the ITT population with the additional requirement that participants have a confirmed response by RECIST version 1.1.

Hazard ratio for duration of response

0.64

95% CI: 0.51–0.81   ·   Two-sided 95% CI

Superiority analysis; p-value not reported in the ClinicalTrials.gov record.

EndpointComparisonEffect measureEstimate95% CIHypothesis
Duration of Response Dabrafenib + Trametinib vs Vemurafenib Hazard Ratio 0.64 0.51–0.81 Superiority
Clinical Biostats interpretation

What the estimate means: An HR of 0.64 corresponds to an estimated instantaneous hazard of disease progression of 64% of the comparator hazard within the analyzed responder population. In simple relative terms, this corresponds to an estimated 36% lower hazard.

Why this endpoint differs from PFS: PFS begins at randomization and includes progression or death. DOR begins later, at the first documented response, and therefore addresses durability among participants who achieved a confirmed response.

Population matters: Because only participants with confirmed response by RECIST version 1.1 were included, the DOR analysis should not be interpreted as an estimate of the probability that an unselected randomized participant will respond or remain progression-free.

Precision and p-value: The two-sided 95% CI is 0.51–0.81. The registry statistical-analysis record does not report a p-value for this comparison.

10. Results Overview

EndpointRoleAnalysis populationHR95% CIMethod note
Overall Survival Primary ITT 0.70 0.58–0.83 Pike estimator
Progression-Free Survival Secondary ITT 0.62 0.52–0.73 Pike estimator
Duration of Response Secondary ITT; confirmed responders by RECIST version 1.1 0.64 0.51–0.81 Hazard ratio reported; formal method not normalized in registry record
Reading the three hazard ratios together: all three reported estimates are below 1, but they answer different questions. OS concerns death from any cause after randomization; PFS concerns progression or death after randomization; DOR concerns progression after a confirmed response. They should therefore not be treated as interchangeable measures.

11. Statistical Methodology

Intention-to-treat analysis

The statistical analyses posted on ClinicalTrials.gov use the Intent-to-Treat (ITT) Population. In an ITT analysis, participants are analyzed according to their randomized treatment assignment. This is important because randomization is the mechanism that creates the basis for a causal comparison between treatment groups.

For an overall-survival endpoint, maintaining the randomized assignment means that later treatment changes or differences in treatment exposure do not automatically redefine the original comparison. That makes the ITT analysis particularly useful for preserving the interpretation of the randomized trial.

Time-to-event analysis

OS, PFS, and DOR are time-to-event endpoints. Unlike a simple binary outcome measured at one fixed date, these analyses use both whether an event occurred and information about when the event occurred. Participants who have not experienced the specified event during available follow-up may contribute censored observations.

Generic survival-function concept
S(t) = P(T > t)

Here, S(t) represents the probability that the event time T exceeds time t. The ClinicalTrials.gov record does not state that a particular survival estimator such as Kaplan-Meier was used, so the page does not attribute that specific estimator to COMBI-v.

Hazard ratio

The principal reported effect measure for all three posted statistical analyses was the hazard ratio. A hazard ratio compares event hazards between two groups within a time-to-event framework.

Hazard-ratio interpretation
HR < 1  →  lower estimated event hazard in the first-named treatment group

For COMBI-v, the first-named group is dabrafenib plus trametinib and the comparator is vemurafenib. An HR below 1 therefore corresponds to a lower estimated event hazard for the combination under the reported analysis.

Pike estimator

The registry analysis notes that hazard ratios are estimated using a Pike estimator. The Pike estimator is a method used to estimate a relative risk or hazard ratio from survival-event information. The registry does not provide the additional computational details needed to reconstruct the estimator from individual-level data.

Confidence intervals

Each posted effect estimate is accompanied by a two-sided 95% confidence interval. The interval provides a measure of statistical uncertainty around the point estimate under the relevant sampling and analysis framework.

A confidence interval should not be interpreted as the probability that the true hazard ratio lies inside the particular interval after the data have been observed. It is better understood as a procedure that, under repeated sampling and the model assumptions, produces intervals with the stated coverage property.

12. Statistical Methods Explained

Why was an intention-to-treat population used?

The ITT principle keeps participants associated with their randomized treatment assignment. This protects the comparison generated by randomization and avoids redefining treatment groups according to events that occurred after randomization. For COMBI-v, the registry specifically identifies the ITT population for OS, PFS, and DOR analyses.

What does an OS hazard ratio of 0.70 mean?

An OS HR of 0.70 indicates that the estimated instantaneous hazard of death for dabrafenib plus trametinib was 70% of the estimated hazard for vemurafenib under the reported analysis. The complementary interpretation is a 30% lower estimated hazard. It is not a 30-percentage-point difference in survival probability.

Why is a confidence interval needed in addition to the hazard ratio?

The hazard ratio is a point estimate, so it provides only one value summarizing the treatment comparison. The 95% confidence interval shows how much statistical uncertainty surrounds that estimate. For OS, the estimate is 0.70 and the corresponding interval is 0.58–0.83.

Why does PFS use a different event definition from OS?

OS counts death from any cause. PFS counts the earliest of disease progression or death. A participant can therefore experience a PFS event without having died. This means that an HR for PFS and an HR for OS are estimates for different event processes.

Why is DOR analyzed in a responder population?

DOR begins at the first documented response. A participant cannot have a response duration without first experiencing a confirmed response. The registry-reported analysis therefore includes only participants with confirmed response by RECIST version 1.1. This makes DOR a conditional measure of durability rather than an endpoint defined from randomization for every participant.

What does the Pike estimator contribute?

The registry states that the hazard ratios were estimated using a Pike estimator. Its role is to produce a relative treatment-effect estimate from time-to-event information. The ClinicalTrials.gov record does not provide enough detail to reconstruct the underlying event table or calculation, so the reported hazard ratios should be treated as registry results rather than independently recomputed estimates.

Why does a p-value not measure effect size?

A p-value addresses evidence against a specified null hypothesis under a statistical model. It depends on both the magnitude of an observed effect and the amount of information available. A hazard ratio and its confidence interval are therefore more directly informative about the estimated size and precision of a time-to-event effect. For the three registry-reported statistical analyses, the ClinicalTrials.gov record does not report p-values.

13. Primary Endpoint and Statistical Inference

Superiority framework

The primary OS analysis is explicitly classified as a superiority analysis. This differs fundamentally from a non-inferiority analysis, where the question is whether an effect remains within a prespecified acceptable margin.

Two-sided confidence interval

The registry-reported OS analysis reports a two-sided 95% confidence interval of 0.58–0.83 around the hazard ratio of 0.70.

Effect measure

The registry reports the hazard ratio as the effect measure for the primary OS comparison.

Formal p-value

No p-value is included in the ClinicalTrials.gov record, so no p-value is attributed to the COMBI-v result on this page.

Do not substitute a p-value that is not in the record. A common reporting error is to fill an incomplete registry result with a value remembered from a publication or another data cutoff. This page deliberately does not do that. The numerical statistical evidence is limited to the ClinicalTrials.gov record.

14. Safety Results

The ClinicalTrials.gov record reports serious adverse events by arm using affected participants over participants at risk. These counts are distinct from the efficacy analyses and should not be interpreted as estimates of treatment effect on overall survival or progression-free survival.

Safety groupSerious adverse eventsAffected / at risk
Dabrafenib Plus Trametinib Serious adverse events 172/350
Vemurafenib Serious adverse events 139/349
Crossover Dabrafenib Plus Trametinib Serious adverse events 15/34
All Patients Serious adverse events 319/699
Registry-count interpretation: The ClinicalTrials.gov record reports 704 participants enrolled but give a serious-adverse-event denominator of 699 for "All Patients." This page preserves the registry-reported denominators rather than attempting to reconcile or replace them. The ClinicalTrials.gov record also identify a crossover dabrafenib-plus-trametinib subgroup for safety reporting but do not provide the crossover rules or timing needed to characterize the crossover design further.

15. What the Hazard Ratios Do — and Do Not — Mean

Overall survival

An OS HR of 0.70 means that the estimated instantaneous hazard of death was 70% as high in the dabrafenib-plus-trametinib group as in the vemurafenib group under the reported time-to-event analysis. This corresponds to an estimated 30% lower hazard.

It does not mean that 30% of participants survived, that 30% of deaths were prevented, or that each individual experienced exactly a 30% reduction in their probability of death.

Progression-free survival

An HR of 0.62 means that the estimated instantaneous hazard of progression or death was 62% as high in the dabrafenib-plus-trametinib group as in the vemurafenib group. This corresponds to an estimated 38% lower hazard.

Because PFS includes both progression and death, it is not equivalent to an estimate of mortality alone.

Duration of response

An HR of 0.64 means that the estimated instantaneous hazard of disease progression during the response-duration analysis was 64% as high in the combination group as in the comparator group within the analyzed confirmed-responder population.

This does not establish how many randomized participants responded in the first place; that is a separate question from how long a confirmed response lasted.

16. Confidence Intervals Across Endpoints

Point estimates and 95% confidence intervals
Overall Survival
0.70
Progression-Free Survival
0.62
Duration of Response
0.64

The visual above is a conceptual comparison of the reported point estimates, not a forest plot. The page does not fabricate interval graphics from insufficient underlying data. The exact reported intervals are shown in the table below.

EndpointHR95% CIInterpretive question
Overall Survival 0.70 0.58–0.83 Relative hazard of death after randomization
Progression-Free Survival 0.62 0.52–0.73 Relative hazard of progression or death after randomization
Duration of Response 0.64 0.51–0.81 Relative hazard of progression after confirmed response

17. Statistical Design Topics Not Reported in the Supplied Data

Several design topics are important in clinical-trial statistics but are not described sufficiently in the registry-reported COMBI-v registry data. Rather than infer them from general knowledge or external publications, this page identifies the limits of the available record.

TopicWhat the ClinicalTrials.gov record supports
Non-inferiority margin Not applicable to the registry-reported primary hypothesis description; the primary OS analysis is classified as superiority.
Crossover A crossover dabrafenib-plus-trametinib safety subgroup of 15/34 is reported, but crossover rules and timing are not reported.
Factorial design No factorial design is reported; the registry identifies a parallel design.
Multiplicity The ClinicalTrials.gov record identifies one primary endpoint and three posted statistical analyses, but do not provide a multiplicity-adjustment procedure.
Interim analysis No interim-analysis boundary, alpha-spending procedure, or information fraction is reported.
Missing data / imputation No missing-data or imputation procedure is reported.
Stratification No stratification factors are reported in the ClinicalTrials.gov record.
Bayesian methods No Bayesian method is reported in the statistical analyses posted on ClinicalTrials.gov.
Why this matters: Statistical trial pages should distinguish between a method that is documented for a particular trial and a method that is merely common in the field. The absence of a registry-reported method description is not evidence that the method was or was not used.

18. Limitations

19. Why This Trial Matters Statistically

COMBI-v is a useful teaching case because the ClinicalTrials.gov record illustrates how a randomized oncology trial can be represented through several related but distinct time-to-event estimands.

ConceptHow it appears in COMBI-v
Randomization The study is registered as randomized with a parallel design.
ITT analysis OS, PFS, and DOR analyses are reported in the ITT population, with the DOR analysis additionally restricted to confirmed responders.
Time-to-event endpoints OS, investigator-assessed PFS, and investigator-assessed DOR are all time-to-event outcomes.
Hazard ratio HR is the reported effect measure for all three statistical analyses.
Confidence intervals Each reported HR is accompanied by a two-sided 95% CI.
Superiority The primary OS analysis is classified as a superiority hypothesis.
Pike estimator The registry states that hazard ratios are estimated using a Pike estimator.
Endpoint conditioning DOR starts at first documented response, unlike OS and PFS, which begin at randomization.
Safety analysis Serious adverse events are reported separately using affected/at-risk counts.

20. Statistical Concepts in This Trial

Learn more about the methods used in this trial:

21. Related Statistical Calculators

22. Sources

Continue with the statistical methods behind clinical trials

Explore tutorials on survival analysis, hazard ratios, confidence intervals, randomization, and intention-to-treat analysis, then apply the concepts with statistical calculators.

23. Record Summary

COMBI-v provides a clear example of a randomized phase 3 oncology trial centered on a time-to-event primary endpoint. The registry analysis reports an OS hazard ratio of 0.70 with a two-sided 95% CI of 0.58–0.83 for dabrafenib plus trametinib versus vemurafenib in the ITT population. Secondary analyses report HRs of 0.62 for investigator-assessed PFS and 0.64 for investigator-assessed DOR, with corresponding 95% CIs of 0.52–0.73 and 0.51–0.81.

The most important statistical distinction is that these three estimates describe different event processes and populations. OS begins at randomization and measures death from any cause. PFS begins at randomization and measures progression or death. DOR begins only after a confirmed response and measures progression thereafter. The hazard ratios quantify relative event hazards, while the confidence intervals describe uncertainty around those estimates.

Clinical Biostats methodology: A trial-results page should distinguish documented trial methods from general statistical conventions. For COMBI-v, this means reporting the registry's ITT populations, superiority framework, hazard-ratio estimates, confidence intervals, and Pike-estimator statement while avoiding unsupported claims about p-values, subgroup analyses, stratification, interim monitoring, multiplicity, or other methods not contained in the ClinicalTrials.gov record.