Safety and Efficacy of Capmatinib (INC280) Plus Pembrolizumab vs Pembrolizumab Alone in NSCLC With PD-L1≥ 50%: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Safety and Efficacy of Capmatinib (INC280) Plus Pembrolizumab vs Pembrolizumab Alone in NSCLC With PD-L1≥ 50% (NCT04139317), a randomized phase 2 study. The primary outcome is Progression-Free Survival (PFS) by Investigator Assessment as Per RECIST 1.1.

2. Clinical Question

The trial asked whether adding capmatinib to pembrolizumab was associated with different efficacy and safety results than pembrolizumab alone when used as first-line treatment for locally advanced or metastatic non-small cell lung cancer. The intended population had PD-L1 expression at least 50% and no EGFR mutation or ALK rearrangement. The primary outcome was progression-free survival assessed by investigators using RECIST 1.1.

3. Trial Design

This was a randomized, open-label, multicenter Phase II study with a parallel-group design. Participants were assigned to capmatinib plus pembrolizumab or to pembrolizumab alone. The registry reports 76 participants enrolled: 51 in the combination arm and 25 in the pembrolizumab-alone arm. Because masking was absent, investigators and participants knew the assigned treatment. Randomization can help limit baseline allocation bias, but an open-label design may still influence assessment or care decisions; the record does not describe any safeguards beyond investigator assessment according to RECIST 1.1.

The study began on 2020-01-22, and primary completion was recorded as 2022-01-14. Its status is terminated. The registry does not report the reason for termination, so the status alone does not establish why the study ended or what motivated the decision. The record describes the study as first-line treatment for the specified NSCLC population.

The combination regimen was capmatinib 400 mg orally twice daily plus pembrolizumab 200 mg intravenously every 3 weeks. The comparator regimen was pembrolizumab 200 mg intravenously every 3 weeks. According to the primary-outcome description, capmatinib was discontinued in all participants in the combination arm, and progression-free survival in that arm was censored on 21-Jan-2021 or at the last adequate tumor assessment before that date. This is central to interpreting the primary result: after that censoring point, later progression or death was not incorporated into the combination-arm progression-free survival estimate in the usual way.

The registry does not report the intended sample-size justification, recruitment targets by arm, randomization procedure, allocation concealment, or a detailed statistical analysis plan. The enrollment and arm counts describe who entered the posted results, not the assumptions used to design the study.

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: The screening period began once patients had signed the study informed consent. Screening evaluations were performed within 28 days prior to the first dose of study treatment. After screening, the treatment period started on Cycle 1 Day 1.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: All patients to whom study treatment had been assigned by randomization.

5. Baseline Characteristics

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

6. Primary Endpoint

The registry lists the primary outcome as Progression-Free Survival (PFS) by Investigator Assessment as Per RECIST 1.1, measured over Up to 1.3 years.

Progression-free survival measures the time from randomization until the first documented disease progression or death from any cause. Tumor response and progression were assessed by investigators using RECIST 1.1. A time-to-event endpoint is useful when the timing of progression or death matters, rather than only whether an event occurred by a single fixed visit. Kaplan-Meier estimation can use information from participants with different follow-up durations while accounting for censoring.

Censoring means that an exact event time is not counted after a participant’s last usable assessment or another specified cutoff, even though the participant may remain alive or may later experience progression. In this study, participants without an event could be censored at the last adequate tumor assessment, the start of subsequent anti-neoplastic therapy, or the sponsor’s decision to discontinue capmatinib, as applicable. The registry specifically states that all participants in the combination arm had capmatinib discontinued and that this arm’s progression-free survival was censored on 21-Jan-2021 or at the last adequate assessment before that date.

The median is the estimated time by which half of the relevant progression-free survival distribution has experienced an event, under the assumptions of the time-to-event analysis. It is not the average follow-up time and does not describe every participant’s experience. Confidence intervals communicate uncertainty around an estimate. The posted upper confidence limits for the primary medians are not reported in the registry record, limiting how fully their precision can be assessed.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is a time-to-event measure, so the registry reports median progression-free survival estimated with the Kaplan-Meier method. The endpoint begins at randomization and ends at first documented progression or death due to any cause. This definition combines progression and death into a single event outcome; it does not require a participant to experience both. Tumor assessment was by investigators under RECIST 1.1.

What a Kaplan-Meier estimate represents

Kaplan-Meier estimation uses the timing of observed events and the timing at which participants are censored. A participant who has not had a recorded progression or death by the last usable follow-up contributes information up to that point, but does not contribute an observed event after it. The method can accommodate unequal observation times, but it cannot recover unobserved event times without assumptions. Its interpretation relies on a suitable censoring process: broadly, censoring should not make the remaining observed participants unrepresentative of the future event experience that the estimate is meant to describe.

The median is read from the estimated event-time distribution. It identifies the point at which the estimated proportion still free of progression or death reaches the median threshold. It is a summary of the distribution, not a comparison test. Two medians can look similar while the curves differ at other times, and similar medians do not by themselves demonstrate equivalence. A comparative hazard ratio or another prespecified contrast would address a different question, but a sponsor-reported hazard ratio is not reported in the registry record.

How to read the confidence limits

A confidence interval describes statistical uncertainty in an estimate under the method’s assumptions. A wider interval generally signals less precision, while a narrower interval signals greater precision. The primary outcome is described as having 95% confidence intervals, but only the lower limits are posted for the two medians. Their upper limits are not reported in the registry record. Consequently, the full interval width cannot be assessed, and the posted lower limits should not be mistaken for complete confidence intervals.

Confidence intervals around separate arm estimates are not a substitute for an interval around a between-arm effect. Nor does the presence or absence of overlap between separate intervals serve as a formal test of the treatment contrast. The record does not post a primary comparative estimate, uncertainty interval, or p-value, so a formal inferential conclusion about the randomized treatment contrast cannot be reconstructed from the medians alone.

Why the censoring rule matters

The registry states that participants without a progression or death event could be censored at the last adequate tumor assessment, the start of subsequent anti-neoplastic therapy, or the sponsor’s decision to discontinue capmatinib, as applicable. It further states that capmatinib was discontinued in all participants in the combination arm and that progression-free survival for that arm was censored on 21-Jan-2021 or at the last adequate tumor assessment before that date. The cutoff limits which later events can contribute to that arm’s estimate. If censoring is related to prognosis, disease status, or treatment experience, the usual interpretation of the Kaplan-Meier estimate may be less secure. The record does not report analyses that test the effect of this censoring decision.

The distinction between an observed event and a censored observation is important. Censoring does not mean that a participant had no later progression or death; it means that the analysis did not count an event time beyond the applicable censoring point. The record does not provide participant-level timelines, numbers at risk over time, event counts for progression-free survival, or additional censoring diagnostics. Those details are not reported in the registry record.

Comparisons and planning illustrations

For the secondary binary outcomes, posted unadjusted comparisons are based on the reported counts and do not represent sponsor-reported analyses. Such comparisons summarize an observed difference in proportions and can be sensitive to small group sizes. The listed p-values are unadjusted, meaning they do not account for the number of outcomes examined. They should be read as exploratory rather than as definitive confirmatory evidence. No sponsor-reported analyses are posted for those outcomes.

The planning illustrations associated with binary outcomes use a two-sided significance level of 0.05, equal allocation, and a normal approximation, with target effects expressed as fractions of the observed effect. These are hypothetical planning calculations based on the observed result, not the trial’s own design assumptions and not observed power. They do not indicate the probability that the study detected a true effect. The registry record does not report the original sample-size justification, target effect, or planned power calculation.

Over-dispersion is a general statistical concern when variation in observed data exceeds what a simple model assumes. For a time-to-event outcome, the record’s key interpretive issue is the stated censoring and lack of comparative analysis; it does not report a model or diagnostics from which over-dispersion can be evaluated. No additional numerical estimates should be inferred from the posted medians.

10. Multiplicity, Missing Data and Analysis Populations

The record lists multiple secondary outcomes spanning efficacy, safety, survival, and pharmacokinetics. Testing or interpreting many outcomes creates a multiplicity problem: even when no true treatment differences exist, a sufficiently large set of comparisons can produce apparently notable results by chance. Statistical plans may address this through a prespecified hierarchy, adjustment, or explicit designation of exploratory outcomes. No multiplicity procedure or outcome hierarchy is reported in the registry record. The secondary comparisons should therefore be treated as unadjusted and exploratory, not as a collection of confirmatory tests.

Missing-data handling is only partly described. For progression-free survival, the record specifies censoring at the last adequate tumor assessment, the start of subsequent anti-neoplastic therapy, or the sponsor’s decision to discontinue capmatinib, as applicable. It also specifies the 21-Jan-2021 cutoff for the combination arm after capmatinib discontinuation. These rules do not explain how all missing assessments were handled, why individual participants had no adequate assessment, or whether sensitivity analyses examined alternative assumptions. Those details are not reported in the registry record.

The primary population is described as all patients to whom treatment had been assigned by randomization. The record does not report whether this population is also called intention-to-treat, how protocol deviations were handled, or whether any participants were excluded from particular analyses. It does not report analysis-population definitions for the secondary outcomes, either. Such details matter because exclusions after randomization can affect how closely an analysis preserves the protection offered by random assignment.

The registry does not report the original sample-size justification, planned power, interim analyses, stopping boundaries, or the detailed statistical analysis plan. It also does not report whether the study’s termination was connected to efficacy, safety, feasibility, or another reason. These omissions prevent a full assessment of how the observed results relate to the trial’s planned inferential framework. The record’s planning illustrations based on observed effects are not a substitute for the original design assumptions.

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: pneumonia (8.0%), chronic obstructive pulmonary disease (8.0%), dyspnoea (8.0%), pyrexia (5.9%), cardiac failure (4.0%).

The safety record distinguishes on-treatment periods from post-treatment survival follow-up. In the combination on-treatment group, 26 of 51 participants were recorded as having a serious adverse event, and 41 of 51 had other adverse events. In the pembrolizumab-after-combination on-treatment group, 9 of 51 had a serious adverse event and 29 of 51 had other adverse events. In the pembrolizumab-alone on-treatment group, 13 of 25 had a serious adverse event and 24 of 25 had other adverse events. These are group-specific counts and should not be added across treatment periods: the same participants may appear in more than one period.

The registry lists pneumonia, chronic obstructive pulmonary disease, and dyspnoea at 8.0% each among the top serious adverse event terms; pyrexia at 5.9%; and cardiac failure and ileus at 4.0% each. The record does not provide enough context in these entries to establish causal attribution or a comparative treatment effect for each term. These figures describe events recorded in the registry, not a complete clinical interpretation of risk.

Deaths are listed separately for on-treatment and post-treatment survival follow-up. They are not interchangeable with serious adverse event counts, and the post-treatment follow-up groups have zero participants at risk for the adverse-event counts shown. The registry record does not report a detailed causal review of deaths or a statistical comparison of safety between arms. Differences in exposure duration and treatment sequence also limit simple comparisons of the period-specific counts.

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

The posted results describe a randomized comparison of capmatinib plus pembrolizumab with pembrolizumab alone in a defined first-line NSCLC population. They report arm-specific median progression-free survival estimates and selected secondary results. The primary medians are close in magnitude, but that observation by itself is not evidence of equivalence, non-inferiority, or superiority. No sponsor-reported primary comparative estimate or p-value is posted.

Precision and completeness matter as much as the point estimates. Only the lower 95% confidence limits for the primary medians are reported; the upper limits are not reported in the registry record. This prevents assessment of the full uncertainty interval. The absence of a reported hazard ratio or another comparative effect estimate also means the posted primary results do not quantify the relative time-to-event contrast between treatment groups.

The capmatinib discontinuation and censoring rule further restrict interpretation. The combination-arm progression-free survival data were censored on 21-Jan-2021 or at the last adequate tumor assessment before that date. This means the reported estimate reflects follow-up under that rule, and cannot be read as a complete account of later progression or death in that arm. The record does not establish how the censoring may have influenced the comparison.

The secondary outcomes can add descriptive context but do not resolve these limitations. Their unadjusted comparisons carry less inferential weight when many outcomes are examined and no multiplicity strategy is reported. Safety counts likewise describe recorded events over particular treatment periods; they are not, on their own, a causal or fully adjusted comparison of risk.

Accordingly, the results do not establish that either regimen is better or worse, and they do not justify a treatment recommendation. They describe the posted observations and the limitations of the available analysis. The reason for termination, the original sample-size rationale, and several details needed for a fuller statistical assessment are not reported in the registry record.

13. Why This Trial Matters Statistically

This trial is statistically instructive because a time-to-event estimate can be difficult to interpret when treatment exposure changes and censoring is tied to that change. The primary outcome was defined in a conventional way, but the record states that capmatinib was discontinued in all participants in the combination arm and specifies a common censoring cutoff for that arm. That information is essential context, not a minor administrative detail: it determines which observations contribute to the estimated progression-free survival distribution.

The study also illustrates why arm-specific medians are not enough to evaluate a randomized comparison. A median compresses an event-time distribution into a single summary. It does not convey the full shape of the curve, the amount of follow-up, the pattern of censoring, or a direct between-arm effect. Comparative estimates and their uncertainty are needed for a more complete inferential interpretation, but sponsor-reported comparative analyses are not posted for the primary outcome.

Open-label allocation, unequal observed arm sizes, missing upper confidence limits, and the terminated status all provide context for assessing completeness and precision. None of these features alone determines the validity of the findings. Together with the limited information about the analysis plan and censoring assumptions, however, they leave important questions unanswered.

The record is also a useful reminder that numerous secondary outcomes may generate interesting descriptive signals without providing confirmatory evidence. Unadjusted comparisons across multiple outcomes require restraint, particularly when the record does not report a multiplicity strategy or sponsor analyses. Statistical interpretation depends not only on the result that appears most notable, but also on how outcomes were selected, analyzed, and reported.

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 the censoring of progression-free survival in the combination arm after capmatinib was discontinued in all participants. The registry states that censoring occurred on 21-Jan-2021 or at the last adequate tumor assessment before that date. The posted result is therefore tied to this cutoff, and the record does not report sensitivity analyses evaluating other censoring approaches or the consequences of stopping observation at that point.

The primary outcome record does not include upper 95% confidence limits, event counts, numbers at risk over time, or a sponsor-reported comparative effect estimate. Those omissions make it difficult to assess precision, the contribution of follow-up at later times, and the magnitude of a direct treatment contrast. The registry does not report the distribution of follow-up or the number and reasons for censored observations.

The trial is marked terminated, but the reason is not reported in the registry record. The original sample-size justification, planned power, interim monitoring, stopping criteria, and detailed analysis plan are also not reported. As a result, the posted results cannot be fully evaluated against the study’s intended design assumptions or stopping framework.

The study was open-label, and the primary endpoint relied on investigator assessment. RECIST 1.1 provides stated assessment criteria, but the registry does not report independent central review or other procedures that might limit assessment bias. The observed randomized arm sizes were unequal, and the record does not explain the allocation process or whether any participants were excluded from particular analyses.

Finally, the registry lists many secondary outcomes but does not post analyses for all of them. The presented secondary comparisons are unadjusted, and no multiplicity strategy is reported. Safety figures are divided across treatment periods and cannot be combined as if they represented separate participant groups. These limits call for a descriptive reading rather than a definitive comparative conclusion.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Safety and Efficacy of Capmatinib (INC280) Plus Pembrolizumab vs Pembrolizumab Alone in NSCLC With PD-L1≥ 50% — NCT04139317.
• Registry record last updated 2024-10-09; status last verified 2024-10.
• Results first posted 2024-02-08.
• 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

NCT04139317 was a randomized, open-label, multicenter Phase II parallel-group study of first-line capmatinib plus pembrolizumab versus pembrolizumab alone for locally advanced or metastatic NSCLC with PD-L1 expression at least 50% and no EGFR mutation or ALK rearrangement. The registry reports 76 enrolled participants, with 51 assigned to the combination and 25 to pembrolizumab alone. The study began on 2020-01-22, had primary completion on 2022-01-14, and is listed as terminated.

The primary endpoint was investigator-assessed progression-free survival by RECIST 1.1, analyzed with Kaplan-Meier estimates. The registry reports arm-specific medians but does not post upper confidence limits or sponsor-reported comparative analyses. The combination-arm observations were censored after capmatinib discontinuation according to the stated cutoff. Selected response and safety results are also posted; other listed secondary analyses and several design and analysis details are not reported in the registry record.