A Study Comparing Trametinib and Dabrafenib Combination Therapy to Dabrafenib Monotherapy in Subjects With BRAF-mutant Melanoma: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study Comparing Trametinib and Dabrafenib Combination Therapy to Dabrafenib Monotherapy in Subjects With BRAF-mutant Melanoma (NCT01584648), a randomized phase 3 study. The primary outcome is Progression-Free Survival (PFS) as Assessed by the Investigator.

2. Clinical Question

The trial addresses a comparative question: among eligible participants assigned to one of these two regimens, how did the time from randomization to documented progression or death differ? Progression was assessed by investigators using RECIST, version 1.1. Death from any cause was also counted as an event, so the endpoint incorporates both disease progression and mortality rather than measuring tumor response alone.

The comparison is between treatment strategies as assigned in this randomized study. It is not a comparison with no treatment, and the registry record does not establish how these results apply to people outside the specified disease stage, mutation status, or treatment-history criteria. The posted results include an investigator-assessed endpoint; they do not report a separate blinded independent review of progression in the primary result.

3. Trial Design

This was a Phase III, randomized, double-blinded, parallel-group study with quadruple masking recorded in the registry. The two groups received dabrafenib with trametinib or dabrafenib with trametinib placebo. The registry lists enrollment of 423 participants. For the primary endpoint, the posted analysis included 211 participants in the combination group and 212 in the dabrafenib-plus-placebo group.

Participants were screened for histologically confirmed cutaneous melanoma that was unresectable Stage IIIC or metastatic Stage IV and BRAF V600E/K mutation-positive. People with prior systemic anti-cancer treatment in the advanced or metastatic setting were not eligible; prior systemic treatment in the adjuvant setting was allowed. Randomization was stratified according to baseline lactate dehydrogenase level and BRAF genotype. The registry record does not report the specific allocation ratio or describe the implementation of the stratification procedure.

The primary endpoint was investigator-assessed progression-free survival in the Intent-to-Treat population. Its timeframe ran from randomization until progression or death, with follow-up described as up to approximately 6 years. The trial is recorded as completed. The registry gives a start date of 2012-05-04 and a primary completion date of 2013-08-26; those record dates should not be read as the length of follow-up for every participant.

The trial also posted overall survival, investigator-assessed objective response rate, duration of response, pharmacokinetic outcomes, and adverse-event outcomes among its secondary measures. The full original protocol, the original sample-size justification, and a detailed statistical analysis plan are not reported in the registry record. The posted record therefore supports a description of the design and available analyses, but not a reconstruction of every design decision or analysis procedure.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Before assignment: Approximately, 340 subjects were planned to be randomized in a 1:1 ratio (170 subjects each in combination therapy and monotherapy). A total of 423 subjects with unresectable or metastatic, BRAF V600E or V600K mutation-positive melanoma were randomized to dabrafenib and trametinib (n=211) or dabrafenib and placebo (n=212).

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Intent-to-Treat (ITT) Population was considered.

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

6. Primary Endpoint

The registry lists the primary outcome as Progression-Free Survival (PFS) as Assessed by the Investigator, measured over From randomization until the earliest date of disease progression (PD) or death due to any cause (up to approximately 6 years).

Progression-free survival (PFS) measures the interval from randomization until the earliest of documented disease progression or death from any cause. The registry defines progression using radiographic or photographic evidence assessed by the investigator according to RECIST, version 1.1. Progression included a specified increase in the sum of target-lesion diameters, the appearance of new lesions, or worsening of non-target lesions significant enough to require study-treatment discontinuation.

PFS is used in trials because it combines two clinically relevant events—disease progression and death—into a time-to-event outcome. General statistical explanation: unlike a simple count of participants with an event at one fixed time, a time-to-event analysis can use the timing of events and account for participants whose event has not been observed by the end of their available follow-up. Such participants may be censored, meaning their observed time contributes information up to a stated point without being treated as an event at that point.

For this endpoint, the registry says participants who received anti-cancer therapy before a documented event were censored at their last adequate assessment before starting that therapy. Participants without documented progression or death were censored at their last adequate assessment. Those rules matter because interpretation depends not only on the event definition but also on how follow-up ends for participants without an observed event.

The posted primary result is a median PFS for each group and a hazard ratio comparing groups. A median describes the time by which half of the relevant estimated event-time distribution has occurred; it does not describe every participant's outcome or imply that all participants were followed for that duration.

7. Results Posted in the Registry

• Dabrafenib + Trametinib vs Dabrafenib + Placebo: hazard ratio 0.73 (95% CI 0.59–0.91) (as reported in the registry).

8. Secondary Efficacy Endpoints

The registry also posts results for secondary outcomes. For each outcome below, the endpoint type, the test applied to the posted summary statistics, the result and its interpretation are given in turn. Secondary endpoints are usually not adjusted for multiple testing, so they support rather than confirm the primary finding.

Overall Survival (OS)

• Endpoint type: Time to event: compared as a hazard ratio.
• Test: Hazard ratio as reported by the sponsor; it was not recalculated.
• Result:
  • Dabrafenib + Trametinib vs Dabrafenib + Placebo: hazard ratio 0.81 (95% CI 0.64–1.02) (as reported in the registry).
• Interpretation:
  • The 95% confidence interval includes no difference (a ratio of 1), so these data do not show a statistically significant difference between Dabrafenib + Trametinib vs Dabrafenib + Placebo and the control. The interval gives the range of effects that remain compatible with the data.

Objective Response Rate (ORR) as Assessed by the Investigator

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Dabrafenib + Trametinib: 69.5% (146/210) versus 53.8% (113/210) on control; risk difference 15.7 percentage points (95% CI 6.5 to 24.9), risk ratio 1.29 (1.11–1.51); unadjusted two-proportion z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically higher for Dabrafenib + Trametinib than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

The secondary outcomes extend the primary PFS question in different directions. Overall survival measures time to death from any cause, while objective response rate records whether a participant achieved a documented complete or partial response. Duration of response concerns how long a response lasts. The registry also lists pharmacokinetic concentration measures and adverse-event outcomes. These outcomes provide complementary descriptions, but they do not measure the same construct as investigator-assessed PFS.

The posted OS comparison has a hazard-ratio interval that includes 1, whereas the PFS interval does not. The response-rate recalculation has a positive risk-difference interval and an unadjusted p-value of <0.001. Taken together, these estimates describe a pattern across distinct endpoints, not a single combined effect. The OS estimate is less conclusive relative to its no-difference reference than the PFS estimate, and response rate is based on a separate binary endpoint. None should be substituted for the primary endpoint.

General statistical explanation: when a study tests several endpoints, each additional test creates another opportunity for a result to appear noteworthy by chance. Procedures such as hierarchical testing or multiplicity adjustment can address this, but the registry record does not report whether such a procedure was used or how outcomes were ordered for inferential testing. The unadjusted response-rate result therefore carries less inferential weight than a result supported by a documented prespecified multiplicity plan.

The record lists other secondary outcomes but does not provide numerical results for all of them in the registry record for this summary. Their absence from the posted results is not evidence of no effect. The available secondary findings are best read as supportive context with their own uncertainty and analysis limitations, rather than as independent confirmation of the primary comparison.

Secondary outcomes registered for this trial, for reference (only those above were recalculated):

• Overall Survival (OS)
• Objective Response Rate (ORR) as Assessed by the Investigator
• Duration of Response (DoR)
• Trametinib Pharmacokinetic Concentrations
• Dabrafenib and Dabrafenib Metabolites (Hydroxy-, Carboxy- and Desmethyl-Dabrafenib) Concentrations
• Number of Participants With Adverse Events and Serious Adverse Events

9. Statistical Interpretation

10. Statistical Methods Explained

General statistical explanation: a randomized parallel-group trial assigns participants to different strategies and compares outcomes between those groups. Randomization is intended to make baseline characteristics comparable on average, though it does not guarantee identical groups in every particular. The registry says participants were stratified by baseline lactate dehydrogenase level and BRAF genotype. It does not report the allocation ratio or the method used to generate and conceal the allocation sequence.

What does a hazard ratio describe?

A hazard ratio compares the instantaneous event rates between groups among participants who have not yet experienced the event, over the follow-up represented in the analysis. It summarizes a time-to-event comparison and is not the same quantity as a difference between medians, a relative risk at a fixed time, or the fraction of participants who experienced an event. A hazard ratio below 1 indicates a lower estimated event hazard in the first named group relative to the comparator, under the model and analysis used.

For PFS, the posted hazard ratio is 0.73, with a 95% confidence interval from 0.59 to 0.91. For OS, it is 0.81, with a 95% confidence interval from 0.64 to 1.02. The registry does not report the model method, the assumptions checked for the model, or the analysis code. General statistical explanation: proportional-hazards methods are often used to summarize time-to-event comparisons, but the registry record does not establish that a particular model or assumption was used. It would be inappropriate to infer those details from the reported hazard ratio alone.

How should medians and censoring be understood?

A median time-to-event estimate describes the point at which the estimated event-time distribution reaches its midpoint. It is not an arithmetic average and does not say that every participant experienced an event at that time. Participants may have different follow-up lengths. Censoring records that an event was not observed by a specified point, rather than asserting that no event would occur later.

The registry defines the PFS censoring rules for participants who started anti-cancer therapy before a documented event and for those without documented progression or death. It does not provide detailed information about the numbers censored, the timing of censoring, or the sensitivity analyses used to evaluate the effect of censoring. General statistical explanation: time-to-event estimates depend on how censoring relates to future event risk; if censoring is informative, estimates may be affected. The record does not allow that assumption to be evaluated fully.

What do confidence intervals contribute?

A confidence interval gives a range of values compatible with an estimate under the statistical procedure and its assumptions. It conveys precision: a narrower interval indicates less sampling uncertainty than a wider one, while neither width nor inclusion of a reference value resolves every source of bias or uncertainty. The PFS interval does not include 1; the OS interval does include 1. This is a description of the reported intervals, not a substitute for a complete analysis plan or a multiplicity assessment.

The interval is not a range containing a specified proportion of individual participant outcomes, nor does it mean there is a 95% probability that the fixed true value lies in this particular interval. In general, the confidence level describes the long-run performance of the interval procedure under its assumptions. The registry does not report all details of the procedure used to calculate the hazard-ratio intervals.

What is different about the response-rate calculation?

Objective response rate is a binary outcome: a participant is counted as having a response or not according to the endpoint definition. The posted counts are 146 of 210 in the combination group and 113 of 210 in the dabrafenib-plus-placebo group. The unadjusted recalculation gives a risk difference and confidence interval, along with an unadjusted p-value of <0.001. General statistical explanation: a risk difference expresses the difference between two event proportions, while a hazard ratio compares time-varying event rates. These measures answer different questions and should not be compared as if they were interchangeable.

The response-rate comparison is explicitly an unadjusted recalculation from the posted summary counts. The registry does not list a sponsor-reported analysis for this endpoint in the posted results. The sponsor's own reported analyses take precedence when they are available; where none is listed for this endpoint, the recalculation remains a distinct descriptive analysis rather than evidence that the sponsor used the same procedure.

Why can over-dispersion matter?

General statistical explanation: over-dispersion occurs when outcome variability exceeds what a statistical model assumes. If a model underestimates variability, uncertainty intervals may be too narrow and tests may be too optimistic. This is a general consideration, not a reported feature of this trial. The registry record does not report whether over-dispersion was assessed or relevant to any analysis; it should not be assumed to have occurred.

Similarly, the record does not report every model choice, adjustment, diagnostic, or sensitivity analysis. The posted estimates can be described as recorded, but fine-grained evaluation of model adequacy requires information that is not reported in the registry record.

11. Confidence Intervals and Precision

The PFS median estimate is 10.2 months in the combination group, with a 95% confidence interval from 8.1 to 12.8 months. In the comparator group, the median is 8.8 months, with a 95% confidence interval from 5.9 to 9.3 months. The group-specific intervals describe uncertainty in each median estimate; they are not the interval for the comparison between groups. The reported comparison is the hazard ratio of 0.73, with a 95% confidence interval from 0.59 to 0.91.

The PFS hazard-ratio interval lies below 1. This indicates that the reported interval does not include the no-difference value for a hazard ratio, given the procedure used. The interval still permits a range of possible relative event-rate comparisons, so the point estimate should not be treated as exact. The registry does not provide a p-value for this comparison.

For OS, the group medians are 25.8 months, with a 95% confidence interval from 19.2 to 38.2 months, and 18.7 months, with a 95% confidence interval from 15.2 to 23.1 months. The OS hazard ratio is 0.81, with a 95% confidence interval from 0.64 to 1.02. Because that interval includes 1, it includes values on both sides of the no-difference reference. This reflects uncertainty in the comparison, not proof that the groups are identical.

For the unadjusted response-rate recalculation, the risk-difference interval is from 0.065 to 0.249. Its width indicates that the observed counts are compatible with a range of effect magnitudes. General statistical explanation: confidence intervals reflect sampling uncertainty under the analysis method, but do not automatically account for selection bias, outcome measurement limitations, multiplicity, or departures from model assumptions. The registry record does not report enough detail to assess all of those sources of uncertainty.

12. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The planning figures are illustrations based on observed effects, not the original trial design. For time-to-event outcomes, the figures represent total events required; for the other outcome, they represent participants per arm. The displayed planning assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. Each scenario considers an effect equal to 100%, 75%, or 50% of the observed effect. The original sample-size assumptions and justification are not reported in the registry record.

The 80% power column describes the planned event count or per-arm participant count that would, under those assumptions, give an 80% chance of detecting the stated effect if that effect were the true one. The 90% power column gives the corresponding planning target for a 90% chance. These are design concepts, not guarantees about the outcome of a future or completed trial.

The time-to-event planning calculations are based on the observed hazard-ratio effects and express required total events. The response-rate planning calculations express participants per arm. Because the response-rate outcome has a fixed binary comparison rather than a time-to-event event count, its planning quantity is not a total number of events across follow-up.

The scenarios illustrate that smaller effects need more participants or events to detect at the same power and significance level. In practice, the required number also depends on design and analysis assumptions, including those used to specify the effect and event accrual. These calculations are not observed power: they do not describe the probability that the completed trial detected an effect after the results were known. The registry record does not report whether the trial's original design used these same assumptions or targets.

13. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

d is the total number of events needed (Schoenfeld's formula) with 1:1 allocation and proportional hazards.

Worked calculation for this trial

1. Hazard ratio HR = 0.730, so ln(HR) = -0.3147.
2. d = 4 × (1.96 + 0.842)² / [ln(0.730)]² = 317.0, rounded up to 317 events in total for 80% power.

The smaller-effect rows in the table above repeat the same calculation with the effect multiplied by 0.75 and 0.5 (for ratios, applied to the logarithm of the ratio).

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
hr <- 0.73                              # observed hazard ratio
events_needed <- function(power, frac = 1) ceiling(round(4 * (z_a + qnorm(power))^2 / log(hr^frac)^2, 6))   # Schoenfeld, 1:1 allocation
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: events = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(events_needed(0.80, f), big.mark = ","), format(events_needed(0.90, f), big.mark = ",")))

# Simulation: given d total events under proportional hazards, events in the treatment arm are Binomial(d, hr / (1 + hr)).
# The score (log-rank type) statistic is (D1 - d / 2) / sqrt(d / 4).
set.seed(1584648)
d    <- events_needed(0.80)
reps <- 20000
D1 <- rbinom(reps, d, hr / (1 + hr))
z  <- (D1 - d / 2) / sqrt(d / 4)
cat(sprintf("Simulated power with %s events: %.3f (%d replicates)\n", format(d, big.mark = ","), mean(abs(z) > z_a), reps))

Output from running the code

100% of observed effect: events = 317 (80% power), 425 (90% power)
 75% of observed effect: events = 564 (80% power), 755 (90% power)
 50% of observed effect: events = 1,268 (80% power), 1,698 (90% power)
Simulated power with 317 events: 0.812 (20000 replicates)

The simulation uses the fact that, with d events in total and proportional hazards, the number of events in the treatment arm is binomial. It is an approximation that conditions on the total number of events, and it should land close to the target power.

14. Group-Sequential Design Analysis

Many time-to-event trials, especially in oncology, stop or report early when interim data are overwhelming, so a single fixed-sample number understates the real design problem. The calculation below is an illustrative group-sequential design built from this trial's posted control median and hazard ratio. The registry record does not state the planned number of looks or the spending function, so the design is a planning illustration, not a reconstruction of the protocol. Assumed: one-sided α = 0.025, 90% power, three equally spaced looks on information, Lan-DeMets O'Brien-Fleming spending, 24 months of accrual, 60 months in total and a 1% monthly dropout hazard.

Under these assumptions the design needs about 508 participants in total, with the boundaries below.

Fixed design versus group-sequential alternatives

Group-sequential designs cost a small increase in maximum sample size and events compared with the fixed design, in exchange for the chance to stop early for overwhelming efficacy. O'Brien-Fleming spending keeps early boundaries very conservative and the final boundary close to the fixed-design value, whereas Pocock spending spends more alpha early.

Sensitivity to a smaller true effect

Planning for exactly the observed effect is optimistic, because observed effects in published trials tend to be somewhat larger than the true effect. The table repeats the main design (three looks, O'Brien-Fleming spending, 90% power) for smaller true hazard ratios.

For context, this trial enrolled 423 participants.

Cross-check, upper boundaries: gsDesign 3.71, 2.511, 1.993 | rpact 3.71, 2.511, 1.993. Agreement between the two packages confirms the boundary calculation.

R code and output

suppressWarnings(suppressMessages({ library(gsDesign); library(rpact) }))
med_control <- 8.8          # control-arm median, months (converted from the posted unit)
hr          <- 0.73          # hazard ratio used for planning
lambdaC     <- log(2) / med_control
eta         <- 0.01               # dropout hazard per month (assumption)
acc_months  <- 24                 # accrual duration (assumption)
study_months<- 60                 # total study duration (assumption)

design <- function(k, sfu, beta, h = hr) gsSurv(k = k, test.type = 1, alpha = 0.025, beta = beta, sfu = sfu,
  lambdaC = lambdaC, hr = h, hr0 = 1, eta = eta, gamma = 1, R = acc_months, T = study_months, minfup = study_months - acc_months)

# Main design: 3 looks, Lan-DeMets O'Brien-Fleming spending, one-sided alpha 0.025, 90% power
x <- design(3, sfLDOF, 0.10)
cat(sprintf("N_TOTAL|%s\n", format(ceiling(sum(x$gamma * x$R)), big.mark = ",")))
for (i in 1:3) cat(sprintf("GS|%d|%.1f|%.0f|%.3f|%.5f|%.5f\n", i, x$T[i], x$n.I[i], x$upper$bound[i], pnorm(-x$upper$bound[i]), cumsum(x$upper$spend)[i]))
cat("\n"); print(gsBoundSummary(x))

# Comparison of designs
cmp <- list(list("Fixed design, one analysis", 1, sfLDOF), list("Two looks, O'Brien-Fleming", 2, sfLDOF),
            list("Three looks, O'Brien-Fleming", 3, sfLDOF), list("Three looks, Pocock", 3, sfLDPocock))
cat("\n")
for (b in c(0.20, 0.10)) for (d in cmp) {
  y <- design(d[[2]], d[[3]], b)
  cat(sprintf("DES|%s|%d|%.0f|%s|%.1f\n", d[[1]], round(100 * (1 - b)), tail(y$n.I, 1), format(ceiling(sum(y$gamma * y$R)), big.mark = ","), tail(y$T, 1)))
}

# Sensitivity: what if the true effect is smaller than the observed one? (3 looks, O'Brien-Fleming, 90% power)
cat("\n")
for (fr in c(1, 0.75, 0.5)) {
  y <- design(3, sfLDOF, 0.10, exp(log(hr) * fr))
  cat(sprintf("SENS|%.2f|%.3f|%.0f|%s\n", fr, exp(log(hr) * fr), tail(y$n.I, 1), format(ceiling(sum(y$gamma * y$R)), big.mark = ",")))
}

# Independent cross-check of the boundaries with rpact
r <- getDesignGroupSequential(kMax = 3, alpha = 0.025, beta = 0.10, sided = 1, typeOfDesign = "asOF", informationRates = x$timing)
cat(sprintf("\nCross-check, upper boundaries: gsDesign %s | rpact %s\n", paste(round(x$upper$bound, 3), collapse = ", "), paste(round(r$criticalValues, 3), collapse = ", ")))

Method: LachinFoulkes 
    Analysis               Value Efficacy
   IA 1: 33%                   Z   3.7103
      N: 366         p (1-sided)   0.0001
 Events: 143        ~HR at bound   0.5373
   Month: 17    P(Cross) if HR=1   0.0001
             P(Cross) if HR=0.73   0.0338
   IA 2: 67%                   Z   2.5114
      N: 508         p (1-sided)   0.0060
 Events: 286        ~HR at bound   0.7428
   Month: 27    P(Cross) if HR=1   0.0060
             P(Cross) if HR=0.73   0.5603
       Final                   Z   1.9930
      N: 508         p (1-sided)   0.0231
 Events: 429        ~HR at bound   0.8248
   Month: 60    P(Cross) if HR=1   0.0250
             P(Cross) if HR=0.73   0.9000



Cross-check, upper boundaries: gsDesign 3.71, 2.511, 1.993 | rpact 3.71, 2.511, 1.993

15. Analyses Reported by the Sponsor

For PFS, the registry reports the sponsor analysis as a hazard ratio of 0.73 with a 95% confidence interval from 0.59 to 0.91. The posted OS analysis reports a hazard ratio of 0.81 with a 95% confidence interval from 0.64 to 1.02. These are the reported hazard ratios used for the time-to-event comparisons; there is no recalculated comparison replacing them. No p-value or analysis method is reported for either comparison in the registry record.

The objective response-rate comparison is different. The registry lists counts by group but no sponsor-reported analysis for that endpoint. The posted risk difference, confidence interval, and unadjusted p-value are recalculated from the summary counts. They are unadjusted and should not be presented as the sponsor's analysis or as evidence of the sponsor's prespecified method.

When a sponsor analysis is reported, it takes precedence over a simple recalculation from summary statistics. Recalculation may not reproduce adjustments, censoring rules, stratification, or other decisions in the sponsor's full analysis. For this record, the time-to-event results are described using the sponsor-reported hazard ratios, while the response-rate comparison is identified as an unadjusted recalculation. The registry does not report enough detail to reconcile that response-rate calculation with a sponsor analysis that is not posted.

16. Multiplicity, Missing Data and Analysis Populations

Multiplicity arises when several outcomes, comparisons, time points, or analysis choices are considered. General statistical explanation: testing many hypotheses without an appropriate plan can increase the chance of at least one apparently significant result, even when no true difference exists. Strategies such as a prespecified hierarchy, family-wise error control, or other adjustment can manage this issue. The registry record does not report which, if any, multiplicity procedure was planned or applied to the primary and secondary outcomes.

The primary endpoint is identified as investigator-assessed PFS. OS, objective response rate, duration of response, pharmacokinetic measures, and adverse events are also listed as secondary outcomes. However, the registry does not state how inferential priority was assigned among these outcomes, whether any secondary endpoint was formally tested, or whether the reported response-rate recalculation was part of a prespecified confirmatory testing sequence. Its p-value is explicitly unadjusted, so it should not be treated as controlling the error rate across the set of outcomes.

The primary endpoint description provides censoring rules for participants who initiated anti-cancer therapy before documented events and for those without documented progression or death. This helps explain how some observations were handled, but it does not report the number of censored participants, the distribution of censoring times, or sensitivity analyses. The registry record also does not provide a comprehensive missing-data plan for all outcomes.

The primary population is the Intent-to-Treat population. General statistical explanation: an intention-to-treat approach generally analyzes participants according to randomized assignment, which preserves the comparison created by randomization. The registry identifies the population but does not fully document how every participant's outcome data were handled, whether any participants were excluded from particular analyses, or how protocol deviations were treated. The posted PFS arm counts are 211 and 212; the response-rate counts use 210 participants in each group. The record does not explain the difference in analysis counts.

These omissions limit assessment of multiplicity, missingness, and analysis-population implementation. They do not invalidate the reported estimates by themselves, but they mean the record cannot establish all the assumptions needed for a complete audit of the analyses.

17. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 5% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: pyrexia (17.2%), squamous cell carcinoma of skin (7.1%), basal cell carcinoma (6.6%), ejection fraction decreased (6.2%), chills (4.8%).

The serious adverse-event summary lists 100 affected participants among 209 at risk in the dabrafenib-plus-trametinib group and 80 among 211 at risk in the dabrafenib-plus-placebo group. The all-patients listing gives 183 affected among 420 at risk. These are counts of participants with serious adverse events, not a comparison adjusted for follow-up time or a measure of event severity beyond the registry's serious-event classification.

The registry also reports other adverse events affecting 194 of 209 participants in the combination group and 200 of 211 in the comparator group. Deaths are listed as 135 and 146 for those respective groups. These are descriptive safety counts; the record does not provide a causal attribution for deaths in this summary. A crossover combination group is listed separately, with 8 of 28 participants affected by serious adverse events and 23 of 28 affected by other adverse events, along with 5 deaths. Those participants should not be combined casually with the original randomized groups when interpreting the randomized comparison.

Among the listed serious adverse-event terms, pyrexia is reported at 17.2%, followed by squamous cell carcinoma of skin at 7.1%, basal cell carcinoma at 6.6%, decreased ejection fraction at 6.2%, chills at 4.8%, and cardiac failure at 3.6%. The registry summary does not provide event-term counts by randomized group here, nor does it establish that these terms account for all serious events. Descriptive safety counts alone do not establish causation or comparative risk.

18. What the Result Does, and Does Not, Mean

The primary result describes a randomized comparison of two regimens among participants meeting the trial's advanced melanoma and mutation-status criteria. The registry reports longer median PFS in the dabrafenib-plus-trametinib group than in the dabrafenib-plus-placebo group, and a hazard-ratio interval that lies below 1. These statistics describe the observed estimates and their precision; they are not a guarantee of an individual participant's outcome.

The result does not establish that every participant in one group had a better outcome, nor does it predict a particular person's response or time without progression. Median values summarize group-level event-time distributions, and hazard ratios summarize relative event rates over follow-up under the analysis. Both conceal variation among participants and depend on how events and censoring were defined and handled.

The record also does not show that the regimens are equivalent for outcomes whose intervals include the no-difference value, or that a secondary finding is definitive after considering multiple outcomes. The OS interval includes 1, and the response-rate p-value is an unadjusted recalculation. Those results require the distinctions between endpoint, analysis method, and statistical uncertainty to remain clear.

General statistical explanation: a confidence interval addresses sampling uncertainty under a procedure's assumptions. It does not by itself account for every possible source of bias, nor does it describe the probability that a treatment will help a particular person. The registry record does not report all protocol and analysis details needed to assess every source of uncertainty.

These results are specific to the study population and comparison recorded: first-line therapy in the advanced or metastatic setting, with the stated disease and mutation criteria and restrictions on prior treatment. They should not be extended automatically to other melanoma populations, other treatment settings, or other comparisons. This statistical interpretation is not medical advice.

19. Why This Trial Matters Statistically

Statistically, this trial illustrates why a primary time-to-event endpoint needs to be read through more than one number. The PFS medians give a familiar description of group event timing, while the hazard ratio provides a comparison across follow-up and the confidence interval shows uncertainty around that relative estimate. These measures are related but not interchangeable.

The trial also illustrates the importance of distinguishing primary from secondary outcomes. The registry identifies PFS as primary, while OS and response rate are among several secondary outcomes. Their estimates answer different questions and have different analysis forms: time-to-event comparisons use hazard ratios, while the posted response-rate comparison is an unadjusted risk difference recalculated from counts. A small unadjusted p-value for a secondary endpoint does not erase the need to consider the endpoint hierarchy and multiplicity plan.

Sample-size illustrations provide another statistical lesson. Planning needs depend on the effect size one aims to detect, the chosen power, significance level, allocation, and outcome type. Smaller target effects require more events or participants under the displayed assumptions. These illustrations are distinct from the original trial design, which is not documented in the registry record, and they are not a retrospective measure of the trial's power.

Finally, the record demonstrates the limits of registry summaries for methodological appraisal. It reports endpoint definitions, group estimates, and some censoring rules, but not a full analysis plan, detailed model specifications, or multiplicity strategy. Statistical interpretation should therefore be precise about what is reported and equally clear about what remains unknown.

20. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

The primary endpoint was assessed by investigators using radiographic or photographic evidence. The registry record does not report a separate independent review result for the primary endpoint, so the posted estimate should be understood as investigator-assessed. The record also does not provide the number of progression events, censoring details, or sensitivity analyses that could help evaluate how follow-up and censoring affected the time-to-event estimates.

The record does not report the original sample-size justification, the planned effect size, or the original statistical analysis plan. It also omits the model method for the reported hazard ratios, the approach to multiplicity across outcomes, and detailed handling of missing data and protocol deviations. These gaps limit independent assessment of the prespecification and robustness of the findings.

The PFS and OS estimates use the randomized groups, while the posted response-rate counts include 210 participants in each group. The registry record does not explain why the analysis counts differ across outcomes. The response-rate comparison is an unadjusted recalculation, not a posted sponsor analysis, and its p-value should not be treated as adjusted for multiple outcomes.

The trial population was restricted by disease stage, mutation status, and treatment history. The results therefore do not directly address populations outside those criteria. The record includes a crossover group in the adverse-event summary, but the figures alone do not establish how crossover affected later outcome comparisons. Finally, registry summaries cannot provide the participant-level detail needed to examine the timing and patterns of outcomes or assess all model assumptions.

21. Related Tutorials

22. Related Statistical Calculators

23. Sources

• ClinicalTrials.gov: A Study Comparing Trametinib and Dabrafenib Combination Therapy to Dabrafenib Monotherapy in Subjects With BRAF-mutant Melanoma — NCT01584648.
• Registry record last updated 2021-02-17; status last verified 2021-02.
• Results first posted 2014-08-15.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

24. Record Summary

This completed, randomized Phase III study compared dabrafenib plus trametinib with dabrafenib plus placebo in participants with unresectable Stage IIIC or metastatic Stage IV BRAF V600E/K mutation-positive cutaneous melanoma. Enrollment was 423. The primary endpoint was investigator-assessed PFS in the Intent-to-Treat population, defined as time from randomization to progression or death.

The registry reports median PFS of 10.2 months and 8.8 months, respectively, with a hazard ratio of 0.73 and a 95% confidence interval from 0.59 to 0.91. Secondary results include OS and investigator-assessed response rate. The registry record does not report the original sample-size rationale, full statistical analysis plan, or multiplicity procedure. The response-rate comparison is an unadjusted recalculation from posted counts; the reported time-to-event hazard ratios are used for PFS and OS.