Study of Nivolumab (BMS-936558) Compared With Dacarbazine in Untreated, Unresectable, or Metastatic Melanoma: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Study of Nivolumab (BMS-936558) Compared With Dacarbazine in Untreated, Unresectable, or Metastatic Melanoma (NCT01721772), a randomized phase 3 study. The primary outcome is Overall Survival (OS).

2. Clinical Question

3. Trial Design

The registry describes a phase 3, randomized, double-blind, parallel-group study sponsored by Bristol-Myers Squibb. It was conducted in participants with previously untreated, unresectable or metastatic melanoma. Enrollment was 418. The record lists the study as completed, with a start date of 2013-01-18 and primary completion date of 2014-06-24.

Participants were assigned to nivolumab, 3 mg/kg, plus placebo matching dacarbazine, or dacarbazine, 1000 mg/m^2, plus placebo matching nivolumab. The matching placebos correspond to the double-blind design. The posted arm descriptions specify intravenous administration, with nivolumab given every 2 weeks and dacarbazine every 3 weeks. They also describe possible later treatment with nivolumab for eligible participants in either group. The complete eligibility and treatment-transition details are not reported in the registry record.

The primary outcome was overall survival in all randomized participants. The registry defines the outcome as time from randomization to death or the last date the participant was known to be alive, with assessment described as up to 17 months. The primary results list 210 participants in the nivolumab group and 208 in the dacarbazine group. The reported comparison is a hazard ratio from a stratified log-rank analysis, accompanied by a confidence interval and p-value.

The registry also lists progression-free survival, progression-free survival rate, objective response rate, overall survival by PD-L1 expression level, and change from baseline in health-related quality-of-life scores among the secondary outcomes. Numerical results are posted for progression-free survival and objective response rate. Results for the other listed secondary outcomes are not reported in the registry record. The original trial protocol, detailed allocation procedures, and the trial’s original sample-size justification are not reported in the registry record.

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: 418 were randomized (210 to nivolumab, 208 to dacarbazine) and 411 received treatment (206 with nivolumab, 205 with dacarbazine).

Pre-Treatment

Reasons for not completing this period:

Treatment

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: All randomized participants

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.

Population: All participants randomized to receive treatment

6. Primary Endpoint

The registry lists the primary outcome as Overall Survival (OS), measured over From date of randomization to date of death. For those without documentation of death, to the last date the participant was known to be alive, assessed up to 17 months..

Overall survival measures the time from randomization until death, or, for a participant without documented death, until the last date the participant was known to be alive. It is a time-to-event endpoint: the analysis uses information about both whether an event occurred and how long participants were followed before the event or last known survival date. The registry describes assessment up to 17 months for this primary outcome.

This endpoint is used to compare the timing of deaths between randomized groups, rather than simply counting deaths at a single follow-up point. A time-to-event analysis can account for participants whose follow-up ends before death is documented by using their last known alive date. The reported hazard ratio summarizes the relative event rate over the analyzed follow-up under the model and analysis approach used. It is not the same quantity as a difference in median survival time, and it does not state how many months any individual participant lived.

The registry lists the median overall survival value and confidence interval for the dacarbazine group: 10.84 months, with a 95% confidence interval from 9.33 to 12.09. The corresponding median and confidence interval for the nivolumab group are not reported in the registry record. Consequently, the primary comparison is interpreted from the reported hazard ratio and its interval, not from a comparison of posted medians. The hazard ratio is reported as 0.42, with a 95% confidence interval from 0.30 to 0.60.

7. Results Posted in the Registry

• Nivolumab, 3 mg/kg + Placebo-matching Dacarbazine vs Dacarbazine, 1000 mg/m^2 + Placebo-matching Nivolumab: hazard ratio 0.42 (95% CI 0.30–0.60), reported p < 0.001 (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.

Progression-free Survival (PFS)

• Endpoint type: Time to event: compared as a hazard ratio.
• Test: Hazard ratio as reported by the sponsor (Stratified Log Rank Test); it was not recalculated.
• Result:
  • Nivolumab, 3 mg/kg + Placebo-matching Dacarbazine vs Dacarbazine, 1000 mg/m^2 + Placebo-matching Nivolumab: hazard ratio 0.44 (95% CI 0.34–0.56), reported p < 0.001 (as reported in the registry).
• Interpretation:
  • The 95% confidence interval excludes no difference (a ratio of 1), so the hazard of the event is statistically lower for Nivolumab, 3 mg/kg + Placebo-matching Dacarbazine vs Dacarbazine, 1000 mg/m^2 + Placebo-matching Nivolumab than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Objective Response Rate (ORR)

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Dacarbazine, 1000 mg/m^2 + Placebo-matching Nivolumab: 14.4% (30/208) versus 42.4% (89/210) on control; risk difference -28.0 percentage points (95% CI -36.2 to -19.7), risk ratio 0.34 (0.24–0.49); unadjusted two-proportion z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically lower for Dacarbazine, 1000 mg/m^2 + Placebo-matching Nivolumab than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

The posted secondary results for progression-free survival and objective response rate point in a direction consistent with the primary overall-survival comparison. In broad terms, these endpoints capture different aspects of the randomized groups: progression-free survival concerns the timing of progression or death, while objective response rate concerns tumor response under the registry’s assessment schedule. Their consistency offers context for interpreting the primary result, but it does not make them equivalent measures or independently establish the same clinical question.

The registry lists additional secondary outcomes: progression-free survival rate, overall survival by PD-L1 expression level, and change from baseline in health-related quality-of-life scores. Numerical results for these outcomes are not reported in the registry record. The pattern across all listed secondary endpoints therefore cannot be fully evaluated. In particular, no conclusion about the subgroup results by PD-L1 expression or change in quality-of-life scores can be drawn from the posted numerical results described in the record.

Secondary outcomes generally carry less weight than a prespecified primary endpoint because they may involve additional hypotheses, varying assessment methods, or smaller subgroups. When many outcomes are examined, the chance of at least one apparently notable result can increase unless the analysis plan accounts for multiplicity. The registry record does not report whether a multiplicity procedure was used. Secondary p-values and intervals should therefore be interpreted with that uncertainty in mind, without treating each result as a separate confirmatory conclusion.

The numerical secondary findings can be useful as supporting evidence when considered alongside the primary analysis, but they do not resolve the missing details in the primary record, such as the absent nivolumab median overall survival. Nor do they establish results for outcomes whose numerical data are not reported in the registry record. Their role is complementary: they describe related outcomes and help assess whether the general pattern is coherent, while the primary endpoint remains the central randomized comparison.

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

• Progression-free Survival (PFS)
• Progression-free Survival (PFS) Rate
• Objective Response Rate (ORR)
• Overall Survival by Programmed Cell Death Ligand 1 (PD-L1) Expression Level
• Change From Baseline in Health-related Quality of Life (HRQoL) Scores

9. Statistical Interpretation

10. Statistical Methods Explained

What does the hazard ratio summarize? A hazard ratio compares the event rates between groups over follow-up in a time-to-event analysis. Here the event for overall survival is death, and the reported comparison is nivolumab versus dacarbazine. The reported estimate is 0.42. An estimate below 1 indicates a lower estimated event rate in the first-named group under the analysis, but it is not a direct comparison of median survival times or survival probabilities at a particular time. The hazard ratio also assumes that a single relative summary is meaningful across the analyzed follow-up; the registry record does not report diagnostics or details for assessing that assumption.

Why use a stratified log-rank test? The registry names the stratified log-rank test as the method for the primary comparison. In general, a log-rank procedure compares the event timing between groups while accounting for participants’ observed follow-up and censoring. A stratified version carries out that comparison within strata defined by specified factors and combines the information across them. The factors used for stratification are not reported in the registry record. The associated hazard ratio is reported as 0.42, with a 95% confidence interval from 0.30 to 0.60, and p < 0.001.

What does censoring mean in this setting? In general, a participant is censored when the event of interest has not been documented by the end of the participant’s available follow-up. For overall survival, the registry says participants without documented death were followed to the last date they were known to be alive. Censoring allows such participants to contribute information up to that date rather than treating them as if they had died or as if they had complete follow-up. The registry record does not describe the detailed censoring rules, follow-up completeness, or missing-data procedures.

How is a confidence interval different from a p-value? In general, a confidence interval expresses the precision of an estimate by giving a range of values compatible with the data under the analysis method. A p-value describes how unusual the observed data, or more extreme data, would be under a specified null hypothesis and statistical model. Neither quantity gives the probability that the null hypothesis is true, and neither alone measures the importance of an effect. For the primary endpoint, the registry reports a 95% confidence interval from 0.30 to 0.60 and p < 0.001. These describe different aspects of the same comparison.

What does the median describe? In general, the median survival time is the time by which half of the group has experienced the event, as estimated by the analysis. It is a useful description of the distribution of time to event, but it compresses the survival experience into one point. It does not describe the full pattern of events or capture the relative event rate across follow-up in the same way as a hazard ratio. The registry reports a dacarbazine median of 10.84 months, with a 95% confidence interval from 9.33 to 12.09. The nivolumab median is not reported in the registry record, so a median-based comparison is unavailable.

How should the secondary endpoint analyses be understood? Progression-free survival is also a time-to-event outcome, with a registry definition based on progression or death. Its reported comparison uses a stratified log-rank test and a hazard ratio. Objective response rate is a binary outcome that counts whether a participant met the response criterion during assessment. Its reported sponsor analysis uses a Cochran-Mantel-Haenszel method and reports an odds ratio. In general, the odds ratio compares the odds of response between groups; it is not the same measure as a risk difference or a response percentage. The response assessment schedule is specified in the registry, but the full response criteria are not reported in the registry record.

What can be concluded from the methods listed? The registry gives the named analysis methods and selected results, but not a complete statistical analysis plan. It does not report the complete model specification, stratification variables, diagnostic checks, interim-analysis procedures, or the original design assumptions. These omissions limit independent reconstruction and assessment of the analyses. General explanations of the methods describe their usual meaning; they do not establish which unreported implementation details applied in this trial.

11. Confidence Intervals and Precision

The primary hazard ratio is 0.42, with a 95% confidence interval from 0.30 to 0.60. This interval shows the uncertainty around the estimated relative event rate under the reported analysis. It is not a prediction interval for individual participants and does not mean that 95% of participants’ outcomes lie within those limits. In general, a confidence interval reflects both the estimate and the amount of information available under the analysis method.

The interval is comparatively informative about the direction of the estimated relative comparison: its lower and upper bounds are both below the reference value for equal event rates. That interpretation applies to the hazard ratio scale, not to absolute survival probabilities or median times. The registry does not report a median overall survival estimate for the nivolumab group, so precision for a median difference cannot be assessed from the primary result. The dacarbazine median is 10.84 months, with a 95% confidence interval from 9.33 to 12.09.

The reported p-value, < 0.001, addresses the test of the null hypothesis used by the analysis. It is not a measure of precision or effect size and should be read alongside the hazard ratio and its interval. The primary result concerns all randomized participants according to the registry outcome description; follow-up is described as assessed up to 17 months. The available record does not provide the number of deaths used in the primary analysis, the follow-up distribution, or additional details needed to judge how the estimate changes over time.

Precision also depends on the analysis and its assumptions. The stratification factors, model diagnostics, and detailed censoring rules are not reported in the registry record. These limits do not alter the posted interval, but they restrict independent evaluation of its construction and robustness.

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 sample-size planning figures are illustrations based on the observed effect, not the original trial design justification. The registry’s original sample-size assumptions are not reported in the registry record. The listed planning assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect to detect is set at the stated fraction of the observed effect.

For time-to-event outcomes, the planning figures are total events across the groups, not participants per arm. For other outcomes, the figures are participants per arm. Thus, the units depend on the endpoint: overall survival and progression-free survival use total event counts, while the objective response rate planning figures use participants per arm.

The 80% power column means the planning illustration’s chance of detecting an effect of the specified size would be 80% under its assumptions, if that effect and the planning model were correct. The 90% power column represents the corresponding 90% target. These are planning targets, not guarantees about a future study or descriptions of the completed trial’s probability of finding a result.

The rows describe planning for 100%, 75%, and 50% of the observed effect. A smaller target effect is harder to distinguish from random variation, so it generally requires more events or participants under the same planning assumptions. The posted figures are not observed power. They do not retrospectively establish whether the completed trial was adequately powered, and they should not be mistaken for the enrollment or event counts actually used in the trial analysis.

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 = 0.420, so ln(HR) = -0.8675.
2. d = 4 × (1.96 + 0.842)² / [ln(0.420)]² = 41.7, rounded up to 42 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 <- 0.42                              # 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(1721772)
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 = 42 (80% power), 56 (90% power)
 75% of observed effect: events = 75 (80% power), 100 (90% power)
 50% of observed effect: events = 167 (80% power), 224 (90% power)
Simulated power with 42 events: 0.765 (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. Analyses Reported by the Sponsor

For the primary time-to-event outcome, the reported hazard ratio is the sponsor’s analysis result used in the statistical interpretation. The registry names a stratified log-rank test and reports a hazard ratio of 0.42, a 95% confidence interval from 0.30 to 0.60, and p < 0.001. The time-to-event presentation does not substitute a newly calculated hazard ratio for the sponsor’s reported estimate.

The distinction matters because a hazard ratio is not recoverable from the group medians alone. The nivolumab median overall survival is not reported in the registry record, and even two medians would not contain all the event-time and censoring information required to reproduce a stratified time-to-event analysis. The reported estimate should therefore be understood as the registry’s sponsor-reported comparison rather than as a simple recalculation from the median values.

For secondary outcomes, the registry also lists sponsor-reported analyses: a stratified log-rank analysis for progression-free survival and a Cochran-Mantel-Haenszel analysis for objective response rate. The latter reports an odds ratio of 4.43, with a 95% confidence interval from 2.75 to 7.13 and p < 0.001. These method-specific estimates are not interchangeable with unadjusted contrasts. No discrepancy is flagged in the registry record; the sponsor-reported results remain the appropriate reported analyses to cite.

15. Multiplicity, Missing Data and Analysis Populations

The registry lists one primary outcome and several secondary outcomes, including progression-free survival, progression-free survival rate, objective response rate, overall survival by PD-L1 expression level, and change from baseline in health-related quality-of-life scores. It also records six analyses. The multiplicity strategy for testing across outcomes or analyses is not reported in the registry record. It is therefore not possible to determine whether the reported p-values were adjusted, whether particular outcomes were organized into a testing hierarchy, or how confirmatory claims were protected against multiple comparisons.

Multiplicity matters because repeated testing creates more opportunities for chance findings. In general, a multiplicity adjustment controls the overall false-positive risk across a defined family of hypotheses, while a hierarchical procedure can specify an order for interpreting results. Without a reported plan, nominal p-values for secondary outcomes should not automatically be read as if each were the sole primary test. The registry’s reported primary p-value remains the result of its named primary comparison, but the broader testing context is not available.

The registry identifies all randomized participants as the primary outcome population. The analysis population definitions for the secondary outcomes, and whether any exclusions were applied, are not reported in the registry record. The record does not report detailed handling of missing outcome data, loss to follow-up, protocol deviations, or censoring beyond the general overall-survival definition. Those details are important for assessing whether the analysis population and observed data match the intended randomized comparison.

Interim analyses and any associated procedures for controlling statistical error are not reported in the registry record. Nor are the original sample-size justification, full statistical analysis plan, or the timing of any analysis decisions reported. These omissions prevent a complete audit of the testing strategy. They do not by themselves show that an analysis was inappropriate; they delimit what can be concluded from the registry record.

16. 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: malignant neoplasm progression (33.7%), basal cell carcinoma (3.9%), pleural effusion (3.4%), pulmonary embolism (3.4%), general physical health deterioration (2.9%).

The registry reports serious adverse events in 125 of 206 participants at risk in the nivolumab group and 119 of 205 participants at risk in the dacarbazine group. It reports other adverse events in 191 of 206 and 187 of 205 participants at risk, respectively. Death counts are listed as 126 for the nivolumab group and 165 for the dacarbazine group. These are registry counts; the record does not establish that each event or death was caused by study treatment.

The listed serious adverse event terms include malignant neoplasm progression (33.7%), basal cell carcinoma (3.9%), pleural effusion (3.4%), pulmonary embolism (3.4%), general physical health deterioration (2.9%), and pain (2.9%). The registry listing does not assign these term percentages to a particular randomized group in the information, so they should not be interpreted as group-specific rates.

Serious adverse events, other adverse events, and deaths are distinct summaries and should not be combined into a single safety measure. The registry record does not provide enough detail to compare event timing, severity distributions, attribution, or exposure-adjusted rates. The figures describe what is posted, not a causal assessment or individualized risk estimate.

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

The primary result is a randomized comparison of overall survival between nivolumab and dacarbazine in participants with previously untreated, unresectable or metastatic melanoma. The reported hazard ratio is 0.42, with a 95% confidence interval from 0.30 to 0.60 and p < 0.001. Under the reported stratified log-rank analysis, the estimated event rate over follow-up is lower in the nivolumab group. This is the central statistical finding in the registry record.

The estimate does not mean that every participant experienced the same benefit, nor does it give an individual participant’s expected survival time. A hazard ratio is a relative time-to-event measure, not a fixed-time survival probability, absolute risk difference, or number of months gained. The registry does not report the nivolumab median overall survival, so the posted primary results do not support a direct comparison of the groups’ medians. The dacarbazine median alone cannot fill that gap.

The result also does not establish outcomes for every population with melanoma. The registry describes a specific group: participants with previously untreated, unresectable or metastatic disease. Generalizing beyond that population requires evidence not contained in this record. The record also notes possible switching or crossover to nivolumab for eligible participants, while full details are not reported; the degree to which treatment transitions affected the comparison cannot be assessed.

The secondary results for progression-free survival and objective response rate are directionally consistent with the primary result, but they address different outcomes and are not replacements for overall survival. Results for the other listed secondary outcomes are not reported in the registry record. The data therefore support interpretation of the posted endpoints only, not claims about unreported subgroup or quality-of-life outcomes.

Finally, statistical evidence and clinical importance are not identical. The p-value indicates that the observed comparison is difficult to reconcile with the specified null hypothesis under the analysis assumptions; it does not quantify benefit, harm, or patient priorities. Those judgments require information beyond the hazard ratio and p-value, including absolute outcome patterns and fuller context that is not all reported in the registry record.

18. Why This Trial Matters Statistically

This trial is statistically notable because it combines random assignment and double masking with a time-to-event primary endpoint. Randomization provides a basis for comparing the groups, while masking can reduce some forms of bias in treatment delivery and outcome assessment. The parallel design assigns participants to one of the two study groups, and the primary analysis compares overall-survival event timing rather than relying only on a single follow-up snapshot.

The reported hazard ratio and confidence interval make the relative comparison interpretable while also communicating uncertainty. A point estimate alone would not show how precisely the event-rate comparison was estimated. The reported 95% confidence interval from 0.30 to 0.60 gives that precision context, and the stratified log-rank method identifies the approach used for the sponsor’s primary comparison.

The record also illustrates why endpoint summaries should not be conflated. A median survival value describes a point in the event-time distribution, whereas a hazard ratio summarizes relative event rates over follow-up. The nivolumab median is not reported, so the central posted comparison relies on the hazard-ratio analysis rather than a comparison of medians. This distinction is important when interpreting registry results that include both kinds of summaries.

Finally, the secondary outcomes provide related but not interchangeable evidence. Progression-free survival and objective response rate broadly align with the primary comparison, while other listed outcomes lack numerical results in the registry record. The incomplete reporting of multiplicity procedures, missing-data handling, and original design assumptions demonstrates why a statistically clear primary result can still have limits in independent review. Careful interpretation distinguishes the reported evidence from unreported analysis details.

19. 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 results do not include the nivolumab median overall survival or its confidence interval. The registry does report the dacarbazine median, but that single value cannot support a direct median comparison. The primary comparison is therefore available through the hazard ratio and its confidence interval, while absolute survival summaries for both groups cannot be assessed from the posted primary outcome.

The registry describes the primary endpoint as assessed up to 17 months and reports enrollment and group sizes, but it does not report the number of deaths contributing to the primary analysis or the distribution of follow-up. The time-to-event estimate can consequently be read as posted, but its information base and changes over follow-up cannot be independently examined from the record. Detailed censoring rules, model diagnostics, and stratification factors are also not reported in the registry record.

Possible treatment transitions complicate interpretation of the treatment contrast. The arm descriptions indicate that eligible participants in the nivolumab group could switch to a different nivolumab schedule and that eligible participants in the dacarbazine group could cross over to open-label nivolumab. The full conditions, timing, and extent of those transitions are not reported in the registry record, so their effect on the overall-survival comparison cannot be assessed.

Numerical results are not reported for all listed secondary outcomes. The record also does not report the multiplicity plan, missing-data procedures, interim analyses, original sample-size justification, or a full analysis plan. These are reporting limits specific to the evidence available in the registry record; they do not by themselves establish that the underlying trial analyses were improperly conducted.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Study of Nivolumab (BMS-936558) Compared With Dacarbazine in Untreated, Unresectable, or Metastatic Melanoma — NCT01721772.
• Registry record last updated 2022-07-12; status last verified 2022-06.
• Results first posted 2016-02-25.
• 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.

23. Record Summary

The registry describes a completed, phase 3, randomized, double-blind, parallel-group study of nivolumab versus dacarbazine in previously untreated, unresectable or metastatic melanoma. Enrollment was 418. The primary endpoint was overall survival in all randomized participants, defined from randomization to death or the last date known alive.

The primary sponsor-reported comparison used a stratified log-rank test and gave a hazard ratio of 0.42, a 95% confidence interval from 0.30 to 0.60, and p < 0.001. The nivolumab median overall survival is not reported in the registry record. Posted secondary results include progression-free survival and objective response rate; numerical results for other listed secondary outcomes are not reported in the registry record.

The record does not report the original sample-size justification, multiplicity strategy, detailed missing-data procedures, or full analysis plan. Those omissions limit independent review of design and analysis details without changing what the registry reports as the primary result.