An Open-Label, Randomized, Phase 3 Trial of Nivolumab Versus Investigator's Choice Chemotherapy as First-Line Therapy for Stage IV or Recurrent PD-L1+ Non-Small Cell Lung Cancer (CheckMate 026): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for An Open-Label, Randomized, Phase 3 Trial of Nivolumab Versus Investigator's Choice Chemotherapy as First-Line Therapy for Stage IV or Recurrent PD-L1+ Non-Small Cell Lung Cancer (CheckMate 026) (NCT02041533), a randomized phase 3 study. The primary outcome is Progression-Free Survival in Participants With PD-L1 Expression >= 5%.

2. Clinical Question

The study was randomized, open-label, and parallel-group. Its title identifies the setting as first-line therapy for stage IV or recurrent PD-L1-positive non-small cell lung cancer. The primary endpoint was assessed by an Independent Radiology Review Committee using RECIST v1.1 criteria for tumor progression. This distinction matters: the primary question is tied to the specified PD-L1 subgroup and to a defined time-to-event outcome, rather than to tumor response alone or to overall survival.

The clinical question is comparative. It asks how the recorded progression-free survival results in the two assigned groups relate to one another; it does not by itself establish the effect of every chemotherapy option separately. The registry record does not report the distribution of participants among the different chemotherapy regimens.

3. Trial Design

This was a phase 3 randomized trial with a parallel-group design and no masking. The registry record lists Bristol-Myers Squibb as the sponsor and 541 participants enrolled. The trial is marked completed. The recorded study start is 2014-03-27, and primary completion is recorded as 2016-07-01.

Participants were assigned to nivolumab or investigator-choice chemotherapy. Nivolumab was given by intravenous injection every 2 weeks at 3 mg/kg, continuing until one of the stopping conditions in the arm description occurred. Chemotherapy was given in 3-week cycles, up to a maximum of 6 cycles, with stopping conditions that included disease progression, unacceptable toxicity, or completion of the cycles. The record lists several chemotherapy drugs, but does not describe the full regimen options for every participant in the available arm description. It identifies gemcitabine administration on Day 1 and Day 8 for squamous participants.

The primary endpoint population was randomized participants with PD-L1 expression of at least 5%. The primary results list 211 participants in the nivolumab group and 212 in the chemotherapy group. The all-randomized progression-free survival outcome lists 271 and 270 participants, respectively. These outcome-specific counts are not interchangeable: each result pertains to the population identified for that endpoint.

Open-label means that treatment assignment was not masked from participants and study personnel as described in the registry record. It does not mean that every component of outcome assessment was unblinded: the primary endpoint description names an Independent Radiology Review Committee. Further details on the committee’s masking arrangements are not reported in the registry record. Details of the randomization procedure, allocation concealment, and the trial’s original sample-size justification 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.

Pre-Treatment Period

Reasons for not completing this period:

Treatment Period

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: All randomized participants with PD-L1 expression levels \>= 5%

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 Progression-Free Survival in Participants With PD-L1 Expression >= 5%, measured over From date of randomization until date of documented tumor progression (assessed up to August 2016, approximately 28 months).

The primary endpoint was progression-free survival among randomized participants with PD-L1 expression of at least 5%. It measured the time from randomization until the first documented tumor progression, as determined by an Independent Radiology Review Committee using RECIST v1.1, or death from any cause, whichever occurred first. The registry record reports the endpoint in months and gives a median and 95% confidence interval for each group.

The registry’s endpoint definition also specifies how observations without a recorded event were handled. Participants who died without a reported progression were considered to have progressed on the date of death. Participants without progression or death were censored at their last evaluable tumor assessment. Participants with no on-study tumor assessment and no death were censored on the randomization date. If subsequent anti-cancer therapy began before documented progression, censoring occurred at the last evaluable tumor assessment before that therapy began.

General: a time-to-event endpoint uses both the timing of events and the follow-up information from participants who have not yet had an event. Progression-free survival is used to summarize the time until progression or death, rather than simply counting how many participants experienced either event by a single fixed date. The censoring rules are important because they define how incomplete follow-up and treatment changes enter the analysis. The registry record states these rules, but does not report how many participants were censored for each reason.

7. Results Posted in the Registry

• Nivolumab vs Investigator Choice of Chemotherapy: hazard ratio 1.15 (95% CI 0.91–1.45), reported p = 0.251 (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 in All Randomized Participants

• Endpoint type: Time to event: compared as a hazard ratio.
• Test: Hazard ratio as reported by the sponsor; it was not recalculated.
• Result:
  • Nivolumab vs Investigator Choice of Chemotherapy: hazard ratio 1.17 (95% CI 0.95–1.43) (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 Nivolumab vs Investigator Choice of Chemotherapy and the control. The interval gives the range of effects that remain compatible with the data.

Overall Survival in Participants With PD-L1 Expression >= 5%

• Endpoint type: Time to event: compared as a hazard ratio.
• Test: Hazard ratio as reported by the sponsor; it was not recalculated.
• Result:
  • Nivolumab vs Investigator Choice of Chemotherapy: hazard ratio 0.97 (95% CI 0.79–1.20) (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 Nivolumab vs Investigator Choice of Chemotherapy and the control. The interval gives the range of effects that remain compatible with the data.

Overall Survival in All Randomized Participants

• Endpoint type: Time to event: compared as a hazard ratio.
• Test: Hazard ratio as reported by the sponsor; it was not recalculated.
• Result:
  • Nivolumab vs Investigator Choice of Chemotherapy: hazard ratio 1.03 (95% CI 0.86–1.24) (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 Nivolumab vs Investigator Choice of Chemotherapy and the control. The interval gives the range of effects that remain compatible with the data.

The secondary outcomes extend the assessment beyond the primary PD-L1 subgroup, including progression-free survival and overall survival among all randomized participants, overall survival in the PD-L1 subgroup, objective response rate in the PD-L1 subgroup, and disease-related symptom improvement by Week 12. The posted time-to-event comparisons do not show a consistent, clearly separated pattern between groups: their hazard-ratio confidence intervals include 1. This is broadly compatible with the uncertainty in the primary comparison rather than a sharply different secondary signal.

Secondary outcomes help describe whether the primary result is accompanied by similar or contrasting patterns in broader populations and other outcome domains. They do not automatically provide independent confirmation. Outcomes may involve different populations, definitions, follow-up periods, and statistical properties. For example, an all-randomized outcome does not answer exactly the same question as an outcome limited to participants with PD-L1 expression of at least 5%.

The registry record lists objective response rate and disease-related symptom improvement among the secondary outcomes, but numerical results for those outcomes are not reported in the registry record. The time-to-event secondary results have posted hazard-ratio estimates and confidence intervals, while their p-values are not reported in the registry record. This limits how far the overall secondary evidence can be characterized from the record alone.

General: when multiple outcomes or subgroups are examined, the chance of finding an apparently notable result by chance can increase. Unless a prespecified multiplicity strategy controls that risk, individual secondary findings are usually less conclusive than a primary result. The multiplicity strategy is not reported in the registry record. Secondary results therefore provide context, but should not be treated as a set of equally definitive tests or as a substitute for the primary comparison.

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

• Progression-Free Survival in All Randomized Participants
• Overall Survival in Participants With PD-L1 Expression >= 5%
• Overall Survival in All Randomized Participants
• Objective Response Rate (ORR) in Participants With PD-L1 Expression >= 5%
• Disease-related Symptom Improvement Rate by Week 12

9. Statistical Interpretation

10. Statistical Methods Explained

The primary analysis is a time-to-event comparison. The registry record reports a median progression-free survival for each group and a hazard ratio for nivolumab versus investigator-choice chemotherapy. It identifies a log-rank method for the reported primary comparison. The posted hazard ratio and confidence interval are the sponsor’s reported comparison; the registry record does not provide enough detail to reconstruct the full analysis model or all analysis procedures.

What does the median represent?

General: a median time-to-event estimate is the point in time by which an estimated half of the analyzed group has experienced the event, under the survival-analysis estimate. It differs from an arithmetic average and does not imply that every participant experienced the event near that time. Some participants may have an event earlier, while others may remain event-free longer or be censored before an event is recorded.

The reported medians are accompanied by 95% confidence intervals. Those intervals describe uncertainty around the median estimates. They are not ranges containing 95% of individual participants’ event times. The endpoint-specific counts identify how many participants were included in the registry’s displayed results, but do not by themselves show how many progression events occurred or how follow-up was distributed.

What does the hazard ratio describe?

General: a hazard is the instantaneous event rate among participants still under observation and event-free at a given time. A hazard ratio compares the hazards between groups over follow-up, subject to the assumptions and model used for the analysis. A value of 1 represents no relative difference in hazard under that comparison; values on either side represent different relative directions. Interpretation also depends on the ordering of the groups and the event definition. The reported comparison is labeled nivolumab versus investigator-choice chemotherapy.

A hazard ratio is not a ratio of median times, a probability that a participant will experience progression, or a measure of how many months a participant gains or loses. It also compresses the time-to-event information into a single estimate. If the relative hazards change over time, a single hazard ratio may not capture that changing pattern. Whether proportional hazards were assessed, and what diagnostic results were obtained, are not reported in the registry record.

Why use a log-rank comparison?

General: the log-rank test compares event-time patterns between groups while accounting for censoring. It uses information from participants who experience the event and from those whose follow-up ends without a recorded event. Its interpretation depends on the event and censoring definitions, and it does not replace the effect estimate or its confidence interval. The registry record identifies the log-rank method for the primary result, but does not report further implementation details or whether additional adjusted analyses were performed.

The p-value of 0.251 is the reported test result for the primary comparison. General: a p-value describes how unusual data at least as incompatible with a specified null hypothesis would be under that hypothesis and the test assumptions. It is not the probability that the null hypothesis is true, and it does not measure the size or clinical importance of an effect. The registry record does not report a full prespecified statistical analysis plan, so the reported value should not be read as answering questions about unreported analyses or decision rules.

How should censoring be understood?

General: censoring records that an event time is not fully observed, while retaining the available time under observation. It is not the same as recording that no event will ever occur. For this endpoint, the registry specifies censoring at the last evaluable tumor assessment for participants without progression or death, and at the last assessment before subsequent anti-cancer therapy if that therapy began before progression. Participants with no on-study assessment and no death were censored on the randomization date.

These rules make the endpoint definition more precise, but they do not reveal the amount or pattern of censoring. The number censored, the duration of follow-up by group, and sensitivity analyses for alternative censoring assumptions are not reported in the registry record. General: time-to-event methods rely on assumptions about the relationship between censoring and the event process. The posted information does not permit assessment of those assumptions.

The primary analysis therefore has a clear stated endpoint, population, and reported comparison method, alongside limits on what can be evaluated from the registry entry. The record does not report a complete analysis plan, details of model diagnostics, or whether interim analyses were conducted. Those omissions should be treated as unknown rather than filled in by assumption.

11. Confidence Intervals and Precision

The primary median estimates have 95% confidence intervals of 2.96 to 5.55 months for nivolumab and 5.42 to 6.93 months for investigator-choice chemotherapy. The comparison hazard ratio is 1.15, with a 95% confidence interval from 0.91 to 1.45. That interval includes 1, so the data are compatible with a range of relative hazard values around the no-difference reference, including values in either direction.

General: a confidence interval expresses uncertainty in an estimate under the statistical method used. It is not a guarantee that the true value lies inside that particular interval, and it does not show the spread of individual outcomes. The width of an interval reflects both the amount of information and the variability in the data, along with the analysis method. A confidence interval should be read together with the estimate and endpoint definition, not as a stand-alone conclusion.

The median intervals also overlap, but overlap between two separate confidence intervals is not itself a formal test of the difference between groups. The reported hazard-ratio interval directly describes uncertainty for the relative comparison. Its breadth means the estimate is not highly precise: a range of hazard ratios remains compatible with the posted data under the stated method.

Precision is also population-specific. The primary confidence intervals apply to participants with PD-L1 expression of at least 5%, while the all-randomized analysis has its own estimates and intervals. Neither set should be transferred to a different population without evidence. The registry record does not report interval estimates for every secondary outcome listed, so precision cannot be assessed from the record for outcomes whose numerical results are not posted.

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 the observed effect, not a report of the trial’s original design calculations. For the time-to-event outcomes, the entries represent total events, not participants per arm. The planning assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation; the calculation uses the stated fraction of the observed effect. The trial’s own original sample-size assumptions and justification are not reported in the registry record.

The 80% power column describes the total number of events the planning calculation estimates would be needed to have an 80% chance of detecting the effect specified for that row, if the assumptions hold. The 90% power column describes the corresponding event total for a 90% chance. These are design-planning quantities, not numbers of participants who actually entered the trial. Because the outcome is time-to-event, the calculation is expressed in events; translating event totals into a required enrollment would also depend on how many enrolled participants experience an event during follow-up.

Rows that use smaller fractions of the observed effect represent smaller target effects. Smaller effects are harder to distinguish from random variation and therefore require more information, in this case more events, under the stated assumptions. The planning entries should not be treated as forecasts of trial results or as a retrospective measure of the trial’s ability to detect its own result.

This is not observed power. Observed power is not a useful substitute for reading the estimate and its confidence interval: the reported interval gives a more direct account of the uncertainty in the observed comparison. The figures are best read as conditional planning illustrations that vary the target effect and requested power while holding the stated assumptions fixed.

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 = 1.150, so ln(HR) = 0.1398.
2. d = 4 × (1.96 + 0.842)² / [ln(1.150)]² = 1,607.3, rounded up to 1,608 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 <- 1.15                              # 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(2041533)
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 = 1,608 (80% power), 2,152 (90% power)
 75% of observed effect: events = 2,858 (80% power), 3,826 (90% power)
 50% of observed effect: events = 6,430 (80% power), 8,607 (90% power)
Simulated power with 1,608 events: 0.800 (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

The registry record reports the primary hazard ratio as 1.15, with a 95% confidence interval of 0.91 to 1.45 and p = 0.251 by the log-rank method. The reported analysis identifies the same hazard-ratio estimate, interval, and p-value. The posted median estimates are 4.21 months for nivolumab and 5.88 months for investigator-choice chemotherapy, with their respective confidence intervals.

The registry record’s comparison is the relevant sponsor-reported analysis. The recalculation is not a replacement for that analysis: a simple re-calculation from summary statistics cannot reproduce all participant-level time-to-event information or the sponsor’s full analysis procedures. The record does not indicate a discrepancy between the reported primary comparison and the posted comparison, so no conflicting estimate is identified in the registry entry.

The record does not report enough detail to verify every analytical choice, including the full model specification or analysis-plan decisions. Accordingly, the reported hazard ratio, confidence interval, and p-value should be attributed to the registry’s reported analysis, while the median estimates should be understood as the endpoint summaries posted for each group. Neither a planning illustration nor a calculation from summaries should be confused with the sponsor’s inferential result.

15. Multiplicity, Missing Data and Analysis Populations

The registry record identifies a primary progression-free survival outcome and lists several secondary outcomes, but does not report a multiplicity strategy. It is not reported in the registry record whether the testing sequence was hierarchical, whether significance levels were divided among outcomes, or whether any adjustment was applied to account for multiple testing. Without this information, individual secondary p-values, if available elsewhere in the record, could not automatically be interpreted as confirmatory evidence. The secondary time-to-event p-values are not reported in the registry record.

The record also does not report whether interim analyses were planned or conducted, or whether any decision boundaries were used. These details should not be inferred from the phase of the trial, its completion status, or the fact that it reports a primary result. The original statistical analysis plan, including its rules for handling multiple outcomes and any interim looks, is not reported in the registry record.

Some missing-data handling is specified for the primary endpoint through the censoring rules. Participants without progression or death were censored at their last evaluable tumor assessment; other censoring situations are also described in the endpoint definition. However, the record does not give the amount of censoring by group or reason, the number of participants with missing tumor assessments, or sensitivity analyses addressing possible departures from the censoring assumptions. Missingness beyond the stated event-time rules is not characterized in the registry record.

The analysis populations are identified for the primary subgroup and all-randomized progression-free survival outcomes, but a full account of analysis-population rules is not reported in the registry record. Details such as exclusions after randomization, handling of protocol deviations, and the approach to participants with unavailable PD-L1 status are not reported. These unknowns affect how completely the results can be independently assessed. They are limits of the registry entry, not evidence that any particular procedure was or was not used.

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 (36.1%), pneumonia (9.9%), anaemia (6.1%), pulmonary embolism (4.2%), pyrexia (3.8%).

The registry safety summary lists serious adverse events affecting 195 of 267 participants in the nivolumab group and 203 of 263 participants in the investigator-choice chemotherapy group. It also lists a post-chemotherapy optional nivolumab group, in which 110 of 159 participants were affected. Other adverse events affected 245 of 267, 252 of 263, and 137 of 159 participants, respectively. These are descriptive counts for the safety groups identified in the record.

The listed serious adverse event terms include malignant neoplasm progression at 36.1%, pneumonia at 9.9%, anaemia at 6.1%, pulmonary embolism at 4.2%, pyrexia at 3.8%, and metastases to the central nervous system at 3.8%. The registry record does not describe the denominator or detailed calculation for these term percentages in the posted summary.

The death counts listed are 233, 92, and 136 for the three safety groups, respectively. These counts do not identify cause or establish that deaths were treatment-related. The post-chemotherapy optional nivolumab group is a distinct group in the safety record and should not be read as a direct randomized comparison with the original treatment groups. Exposure duration, attribution procedures, and detailed event rates are not reported in the registry record. These safety figures are descriptive and should not be used alone to make treatment decisions.

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

The primary result describes progression-free survival in randomized participants with PD-L1 expression of at least 5%. The posted median was 4.21 months in the nivolumab group and 5.88 months in the investigator-choice chemotherapy group. The reported hazard ratio was 1.15, with a 95% confidence interval of 0.91 to 1.45 and p = 0.251. The interval includes 1, and the reported test does not establish a statistically clear difference under a conventional reading of the posted result.

That finding does not demonstrate equivalence or prove that the treatments have identical effects. An equivalence or non-inferiority conclusion requires a design and analysis that specify an acceptable margin and test it; such a margin is not reported in the registry record. Conversely, the result should not be described as proving that one treatment has no effect. The estimate remains uncertain, and its interval permits more than one interpretation of the relative hazard.

The endpoint is not overall survival. It combines documented tumor progression and death, and it is subject to the stated censoring rules. The trial also reports secondary overall-survival outcomes, but those are distinct outcomes with their own populations and uncertainty. A result for progression-free survival cannot be assumed to determine overall survival, response, symptoms, or other outcomes.

General: statistical evidence is one part of interpreting a clinical comparison. The estimate indicates the direction and size of the result under the chosen measure, while the confidence interval shows its precision. Neither alone establishes clinical importance. The registry record does not report a threshold for a clinically important difference in progression-free survival, so the clinical importance of a particular estimate cannot be assessed against such a threshold from this record.

The result applies to the enrolled trial population and specified analysis populations, not automatically to people outside them. The registry entry does not provide enough information to assess all aspects of generalizability, such as detailed participant characteristics or chemotherapy-regimen distribution. The result is therefore best understood as a report of this randomized comparison and its uncertainty, rather than a universal statement about all treatment settings.

18. Why This Trial Matters Statistically

Statistically, this trial illustrates the difference between a group-specific median and a comparative time-to-event estimate. The medians describe the estimated midpoint of progression-free survival in each group, while the hazard ratio summarizes the relative event hazard across follow-up. Those measures answer related but non-identical questions. Reading the hazard ratio together with its confidence interval helps avoid treating a numerical difference between medians as a complete account of the comparison.

The registry entry also makes the analysis population important. The primary result is limited to participants with PD-L1 expression of at least 5%, whereas another progression-free survival result includes all randomized participants. The distinction is consequential: a subgroup estimate should not be presented as if it applied to everyone randomized, and the broader analysis should not be substituted for the stated primary question.

Another statistical lesson is the role of endpoint definitions. The primary outcome combines tumor progression assessed by an independent review committee with death, and the registry specifies how certain incomplete observations were censored. These details determine what event is being compared and how follow-up contributes to the analysis. Without those definitions, a hazard ratio can be difficult to interpret.

Finally, the posted result is imprecise enough that its interval includes the no-relative-difference reference. That makes clear why an estimate alone is insufficient: uncertainty, the analysis population, the endpoint, and the reported test all need to be considered. The record’s missing details about multiplicity, interim analyses, and the original design assumptions also show why conclusions should remain within what the posted statistical information can support.

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 registry entry does not report the original sample-size justification or the full statistical analysis plan. It therefore does not establish whether the trial’s planned enrollment or event targets were met, what significance threshold governed the primary test, or which analyses were prespecified beyond the outcome descriptions. Planning illustrations based on observed effects cannot fill these gaps.

Although the primary endpoint definition is detailed, the record does not report numbers of progression events, deaths, or censored observations by reason. The follow-up distribution and the number of participants who began subsequent anti-cancer therapy before documented progression are also not reported. These omissions limit independent evaluation of the time-to-event analysis and its censoring assumptions.

The study was open-label, and details about masking of the radiology review committee are not reported in the registry record. The primary endpoint description identifies independent review, but does not describe the committee’s procedures. Randomization procedures, allocation concealment, and the distribution of participants across the chemotherapy choices are also not reported in the registry entry.

Numerical results for some listed secondary outcomes, including objective response rate and disease-related symptom improvement, are not reported in the registry record. The multiplicity strategy, interim analyses, detailed participant characteristics, and analysis-population rules are not reported either. These limits do not negate the posted results; they define the additional information that would be needed for a fuller assessment of robustness, applicability, and the relationship among the trial’s outcomes.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: An Open-Label, Randomized, Phase 3 Trial of Nivolumab Versus Investigator's Choice Chemotherapy as First-Line Therapy for Stage IV or Recurrent PD-L1+ Non-Small Cell Lung Cancer (CheckMate 026) — NCT02041533.
• Registry record last updated 2023-03-02; status last verified 2023-02.
• Results first posted 2017-07-26.
• 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 record describes a completed, randomized, open-label, parallel-group phase 3 trial of nivolumab versus investigator-choice chemotherapy for stage IV or recurrent PD-L1-positive non-small cell lung cancer. It lists 541 participants enrolled. The primary endpoint was progression-free survival among randomized participants with PD-L1 expression of at least 5%, defined as time to documented tumor progression or death.

The posted primary comparison reports a hazard ratio of 1.15, a 95% confidence interval of 0.91 to 1.45, and p = 0.251 by the log-rank method. Median progression-free survival was 4.21 months for nivolumab and 5.88 months for investigator-choice chemotherapy. The confidence interval includes 1, and the registry entry leaves important design and analysis details unreported, including multiplicity handling and the original sample-size justification.