Nivolumab and Epacadostat With Platinum Doublet Chemotherapy Versus Platinum Doublet Chemotherapy in Non-Small Cell Lung Cancer: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Nivolumab and Epacadostat With Platinum Doublet Chemotherapy Versus Platinum Doublet Chemotherapy in Non-Small Cell Lung Cancer (NCT03348904), a randomized phase 3 study. The primary outcome is Overall Survival (OS) of Nivolumab Plus Epacadostat in Combination With Chemotherapy (Arm A) Compared to Chemotherapy (Arm B).

2. Clinical Question

This is therefore a question about the planned comparison, not an estimate of a treatment effect. The record identifies the target population, the intervention and comparator, and the outcome, but it does not provide an observed overall-survival result for the comparison. The listed objective response rate and duration of response outcomes were also intended to compare Arm A with Arm B. Other listed outcomes concerned estimates for Arm C. The absence of efficacy analyses means the record cannot answer those questions from comparative outcome data.

3. Trial Design

The registry describes a randomized, phase 3, global, parallel-group trial with quadruple masking. Its official title identifies the intended setting as first-line treatment of Stage IV or recurrent non-small cell lung cancer. The brief summary describes treatment-naïve participants. The listed arms were Arm A, nivolumab plus epacadostat with platinum doublet chemotherapy; Arm B, platinum doublet chemotherapy; and Arm C, nivolumab plus placebo with platinum doublet chemotherapy.

The primary outcome was overall survival for Arm A compared with Arm B, with a stated timeframe of approximately 38 months. The record defines the outcome as time from randomization to death from any cause. Secondary outcomes included objective response rate and duration of response for the Arm A versus Arm B comparison, as well as estimates of overall survival, progression-free survival, objective response rate, and duration of response for Arm C. The registry lists 8 outcome measures, but reports no analyses.

The record gives an enrollment of 2 and a status of terminated. It lists the start date as 2017-12-27 and primary completion as 2018-05-22. The primary outcome description states that Arm A did not enroll any participants. It also states that no comparisons were performed and that the study was terminated early without efficacy analyses. Thus, the design describes the intended randomized comparison, while the observed record reflects a trial that did not produce the planned efficacy comparison.

The registry identifies the allocation and masking structure, but further operational details are not reported in the registry record. For example, the method used to generate the random sequence, how allocation was concealed, and which roles were included in quadruple masking are not reported in the registry record. The protocol-level rationale for the planned duration and the basis for the enrollment target are also 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.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: All randomized participants; Arm A did not enroll any participants in the study and as a result no comparisons were performed and since the study was terminated early no efficacy analyses were conducted.

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 randomized participants.

6. Primary Endpoint

The registry lists the primary outcome as Overall Survival (OS) of Nivolumab Plus Epacadostat in Combination With Chemotherapy (Arm A) Compared to Chemotherapy (Arm B), measured over Approximately 38 months.

Overall survival is the time from randomization until death from any cause. Because death is an event that can be defined without assigning a cause-specific diagnosis, this endpoint can capture mortality across the follow-up period. In general, overall survival is used in clinical trials to compare the timing of deaths between treatment groups. The time-to-event form retains information about when events occur and can accommodate participants whose event has not been observed by the analysis cutoff through censoring methods, if those methods are specified and used.

For this trial, the primary endpoint was framed as a comparison of overall survival for nivolumab plus epacadostat with chemotherapy versus chemotherapy alone. The registry states a timeframe of approximately 38 months. That timeframe does not itself indicate the length of follow-up achieved for every participant or the timing of an analysis. Details of the planned analysis and actual follow-up are not reported in the registry record.

The endpoint’s definition is available, but its comparative result is not. Arm A did not enroll any participants, and the record says no comparisons were performed and no efficacy analyses were conducted. Consequently, there is no reported hazard ratio, survival estimate, confidence interval, or significance test for the primary comparison in this record. The endpoint description establishes what the study intended to measure; it does not establish an observed difference between the groups.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does randomization contribute?

In general, randomization assigns participants to treatment groups by chance. When implemented as planned, it can reduce systematic differences in measured and unmeasured characteristics between groups, supporting a fair comparison. Randomization does not guarantee balance in every study, and it cannot substitute for outcome data. Here the registry identifies the trial as randomized, but the primary description says Arm A did not enroll any participants. The planned allocation therefore did not yield an analyzable Arm A versus Arm B comparison.

Details of the random-sequence method and allocation concealment are not reported in the registry record. The record identifies quadruple masking, but does not specify which participants, clinicians, assessors, or other personnel were masked. Masking is generally used to reduce the influence of treatment knowledge on care, outcome assessment, or reporting, depending on who is masked. The label alone does not establish how masking operated in practice.

What would an overall-survival analysis estimate?

In general, a time-to-event analysis uses both whether an event occurred and the time until it occurred. Participants who have not had the event by the end of observed follow-up may be censored rather than treated as if they had died. Analyses must account for that incomplete event-time information under stated assumptions. Common approaches can compare event-time distributions or estimate a relative event rate, but the selected method and its assumptions matter.

The registry defines overall survival as time from randomization to death from any cause and states a timeframe of approximately 38 months. It does not report the survival-analysis method, censoring rules, analysis cutoff, or an estimate. Since the planned Arm A group had no participants, there is no basis in this record for an Arm A versus Arm B time-to-event estimate. The endpoint definition should not be mistaken for a reported analysis.

How should estimates and uncertainty be read?

In general, an estimate summarizes the observed data in a form relevant to the research question, while a confidence interval communicates uncertainty around that estimate under the model and analysis assumptions. A p-value addresses compatibility between the data and a specified null hypothesis under an analysis model; it is not the probability that a treatment works or that the null hypothesis is true. Estimates, intervals, and tests are most informative when presented together with the analysis population and methods.

No efficacy estimate, confidence interval, or p-value is reported for the primary comparison. There is therefore no effect size to interpret and no statistical test result to use as evidence for or against a difference. A result that is not estimable because the comparison was not performed is distinct from a null result in a completed analysis.

What do the planning illustrations mean?

The registry’s sample-size assumptions describe planning illustrations based on a stated fraction of an observed effect, a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. For time-to-event tables, the stated quantity is the total number of events; other tables are described in participants per arm. These are planning conventions, not the trial’s own design justification. The registry record does not report a sample-size justification.

In general, a planning illustration asks what sample size or event total would be needed under chosen assumptions to detect a specified effect with a chosen power. Such a calculation depends on the assumed effect, variability or event process, allocation, statistical test, and significance level. It does not show that a trial actually had that power, nor does it establish what effect would occur in a new study. Sample sizes in such illustrations are planning quantities, never observed power. Because the efficacy comparison was not performed and no observed effect is reported in the registry record, the illustrations cannot be treated as evidence about the treatment effect in this trial.

Over-dispersion is a general issue when outcome variation exceeds what a model assumes. If a model ignores extra variation, its uncertainty estimates may be too small and its tests may appear more decisive than warranted. This is relevant to some count outcomes, but it does not create an analysis where none was conducted. The registry record does not report an over-dispersion assessment or a statistical model for efficacy outcomes.

What is not described?

The registry record does not report the detailed statistical analysis plan, the planned estimand, model covariates, handling of protocol deviations, or the approach to missing outcomes. It also does not report any interim-analysis plan. These details can affect interpretation, but they should not be reconstructed from standard practice or from the randomized and masked design labels. The record’s direct statement that no efficacy analyses were conducted governs interpretation of the primary and secondary efficacy questions.

10. Multiplicity, Missing Data and Analysis Populations

The record lists a primary overall-survival comparison and several secondary outcomes, including objective response rate and duration of response for Arm A versus Arm B, and estimates for Arm C. When a trial tests multiple outcomes or comparisons, multiplicity can increase the chance of at least one apparently unusual result under a collection of null hypotheses. In general, a statistical plan may address this through a prespecified hierarchy, adjusted significance thresholds, or other procedures. The registry record does not report a multiplicity strategy.

That omission does not imply that unadjusted tests were performed. The record lists 0 analyses and states that no efficacy analyses were conducted because the study was terminated early. There are no reported efficacy test results to interpret as multiplicity-adjusted or unadjusted. The planned outcome list describes intended questions, not a set of completed tests.

The registry identifies the primary outcome population as all randomized participants, but details of the analysis population are not reported beyond that description. The record does not specify how missing outcome data, loss to follow-up, protocol deviations, or treatment discontinuation would have been handled. Those are important elements of an analysis plan, especially for time-to-event outcomes, where censoring and follow-up practices can affect estimates. No assumptions about these procedures should be inferred.

The reason for early termination, any planned interim looks, and whether missing data were assessed are not reported in the registry record. Since there were no efficacy analyses, the record cannot show how multiplicity or missingness affected an efficacy estimate. These are limits on what can be concluded from the registry entry, rather than evidence that a particular statistical procedure was used or omitted in an unreported analysis.

11. 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: dehydration (50.0%), back pain (50.0%), osteomyelitis (50.0%).

The adverse-event entries are descriptive counts by listed group. For the nivolumab plus epacadostat and platinum doublet group, the registry lists 0 participants at risk for serious adverse events, 0 affected, 0 at risk for other adverse events, 0 affected, and 0 deaths. For platinum doublet chemotherapy, it lists 2 at risk for serious adverse events and 1 affected; it lists 2 at risk for other adverse events and 2 affected, with 1 death. For nivolumab plus placebo with platinum doublet chemotherapy, the corresponding listed counts are 0, and deaths are 0.

The serious-event terms listed at 50.0% are dehydration, back pain, and osteomyelitis. These percentages are registry entries for the terms; the record excerpt does not provide the term-level counts or further detail needed to interpret their denominators. They should not be treated as rates that can be compared across treatment groups from this information alone.

Safety counts are not efficacy results. With the small recorded safety population, these figures are descriptive and do not establish comparative risks or identify a causal relationship. The registry record does not provide additional context such as event timing, severity grading details, exposure duration, or adjudication methods. The entries should therefore be read as the reported counts, not as a complete characterization of safety.

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

The record establishes what the trial set out to compare: nivolumab plus epacadostat with platinum doublet chemotherapy against platinum doublet chemotherapy alone, with overall survival as the primary endpoint. It also records randomized allocation, quadruple masking, a parallel design, and the intended first-line population. These are features of the planned study, not evidence that the main comparison was completed.

The key limitation is explicit: Arm A did not enroll any participants. The registry says no comparisons were performed and that no efficacy analyses were conducted after early termination. Consequently, the record does not estimate whether the intervention changed overall survival, response, or duration of response relative to chemotherapy. It also does not show that the treatments were equivalent, that there was no effect, or that one approach was better. Those conclusions would require outcome data and an analysis that are not reported.

The enrollment entry of 2 and the safety counts provide limited descriptive information, but they cannot repair the absence of the planned primary comparison. A study’s intended endpoint and a study’s observed result are different kinds of information. The endpoint definition tells readers what was to be measured; only reported outcome analyses can quantify the observed contrast and its uncertainty.

The trial’s terminated status is important context, but the reason for termination is not reported in the registry record. It should not be interpreted as proof of an efficacy or safety finding. Nor should the listed planning assumptions be mistaken for observed power or for an estimate of the treatment effect. The defensible conclusion is narrow: this registry record describes a planned phase 3 trial but supplies no efficacy comparison with which to answer its primary question.

That conclusion applies to this record. It is not a claim about other evidence, other studies, or clinical decisions. No medical recommendation follows from the absence of a reported analysis.

13. Why This Trial Matters Statistically

This trial is statistically notable because the planned design and the available evidence diverge sharply. A randomized phase 3 comparison can, in principle, provide a structured estimate of treatment differences. But that inferential value depends on having data in the groups being compared and completing the planned analysis. The registry states that Arm A did not enroll, so the central contrast could not be made.

The distinction between a missing comparison and a negative result is fundamental. A completed analysis might yield an estimate close to no difference, an estimate favoring either group, or an imprecise result. None of those possibilities is reported. Treating no analysis as a null finding would substitute an unobserved result for evidence.

The record also illustrates why outcome listings should be read as plans rather than results. Overall survival, response, and duration of response appear among the intended outcomes, but the registry reports 0 analyses. The presence of outcome definitions does not mean that data were collected, evaluated, or compared. This matters when readers assess registry records, abstracts, or other summaries: design features and planned endpoints do not stand in for effect estimates.

Finally, the entry underscores the limits of safety counts in a very small record. The adverse-event figures describe what was recorded by group, but they do not support stable comparative estimates or a broad account of risk. Statistical interpretation begins by identifying which quantities were actually observed and analyzed. In this case, the most important result is that the planned efficacy comparison is unavailable, not a numerical estimate of treatment benefit or harm.

14. 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 principal trial-specific limitation is that Arm A did not enroll any participants, preventing the intended primary comparison with Arm B. The registry states that no comparisons were performed and that no efficacy analyses were conducted. As a result, there is no reported treatment-effect estimate, confidence interval, p-value, or efficacy outcome summary for the planned comparison.

The record reports enrollment of 2, and the safety entries are limited to the listed groups and counts. This small recorded population cannot provide a robust basis for comparative safety conclusions. The serious adverse-event terms are presented with percentages, but term-level counts and fuller denominator information are not reported in the registry record. The safety entries should not be generalized beyond their descriptive role.

The study was terminated, but the reason for termination is not reported in the registry record. The record does not explain why Arm A did not enroll, whether the termination was related to recruitment, operations, safety, or another issue, or whether any interim review informed the decision. It would be speculative to attribute a cause.

Important methods details are also not reported in the registry record, including the analysis plan, missing-data procedures, random-sequence generation, allocation concealment, and the specific roles covered by quadruple masking. Although the primary outcome gives a timeframe of approximately 38 months, the achieved follow-up and analysis cutoff are not described. These omissions limit assessment of the intended methods, but they do not change the central fact that the efficacy comparison was not conducted.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Nivolumab and Epacadostat With Platinum Doublet Chemotherapy Versus Platinum Doublet Chemotherapy in Non-Small Cell Lung Cancer — NCT03348904.
• Registry record last updated 2019-06-13; status last verified 2019-06.
• Results first posted 2019-06-13.
• 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.

18. Record Summary

NCT03348904 is a randomized, phase 3, global, parallel trial with quadruple masking, intended to compare nivolumab plus epacadostat and platinum doublet chemotherapy with platinum doublet chemotherapy in treatment-naïve Stage IV or recurrent non-small cell lung cancer. Arm C was nivolumab plus placebo with platinum doublet chemotherapy. The primary endpoint was overall survival, defined as time from randomization to death from any cause, with a stated timeframe of approximately 38 months.

The record lists enrollment of 2, a start date of 2017-12-27, primary completion on 2018-05-22, and terminated status. Arm A did not enroll any participants. The registry reports no comparisons and no efficacy analyses, with 0 analyses listed. Safety counts are descriptive and limited. The reason for termination and further analysis details are not reported in the registry record. The central statistical conclusion is that this record does not provide an efficacy estimate for its primary comparison.