Study of Cisplatin/Vinorelbine +/- Cetuximab as First-line Treatment of Advanced Non Small Cell Lung Cancer (FLEX): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Study of Cisplatin/Vinorelbine +/- Cetuximab as First-line Treatment of Advanced Non Small Cell Lung Cancer (FLEX) (NCT00148798), a randomized phase 3 study. The primary outcome is Overall Survival Time (OS).

2. Clinical Question

The clinical question is whether adding cetuximab to the chemotherapy regimen was associated with a different overall-survival distribution in the trial population. Overall survival is a time-to-event outcome: it reflects not only whether death was observed, but also when it occurred during follow-up. Patients who were not known to have died by the relevant observation limit were censored at the last date known to be alive or the clinical cut-off, whichever came first. The posted record gives median survival estimates and confidence intervals by group, alongside a stratified log-rank comparison. It does not provide a hazard ratio for the primary comparison. That distinction matters: the reported medians describe the middle of each group’s observed survival distribution, while the log-rank test compares event-time information across follow-up. Neither measure alone describes every feature of the survival curves.

3. Trial Design

FLEX was a randomized, open-label, controlled, multicenter Phase III study with parallel treatment groups. It enrolled 1861 participants and compared cetuximab plus cisplatin and vinorelbine with cisplatin and vinorelbine alone. The registry lists the study status as completed. The record gives the start as 2004-10 and primary completion as 2007-07.

The trial description identifies the intended population as patients with advanced NSCLC expressing EGFR who had not received prior chemotherapy. The registry lists overall survival as the primary outcome and gives the primary analysis population as ITT. It defines the outcome as time from randomisation to death, with participants without a recorded event censored at the last date known to be alive or at the clinical cut-off, whichever was earlier. The reported period runs from the first patient’s randomisation in Oct 2004 to the cut-off date of 18 Jul 2007.

Random assignment supports a comparative interpretation by allocating participants to treatment groups rather than allowing treatment choice alone to determine group membership. The open-label design means masking was absent; the registry does not describe any masking of outcome assessment. The primary endpoint is death, a relatively direct event, but the record does not report details about independent event adjudication or other procedures that could affect outcome ascertainment.

The registry’s enrollment figure and the group denominators reported for the primary outcome are not the same set of numbers. The reasons for the difference are not reported in the registry record, so it should not be attributed to a particular form of exclusion, missingness or analysis decision. The sample-size justification, target number of primary events, and any interim-analysis plan are not reported in the registry record. The planning illustrations associated with secondary comparisons are not evidence of the trial’s original design assumptions.

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: Enrolled: 1,861 after consent to epidermal growth factor receptor (EGFR) assessment; 603 excluded (mainly non-fulfillment of inclusion or exclusion criteria). 1,258 screened for eligibility after consent for study procedures; 143 excluded (mainly non-fulfillment of inclusion or exclusion criteria). 1,125 subjects randomized.

Overall Study

Analysis population for the primary outcome, as defined in the registry: ITT

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 Overall Survival Time (OS), measured over Time from randomisation to death or last day known to be alive, reported between day of first patient randomised, Oct 2004, until cut-off date 18 Jul 2007.

Overall survival is the time from randomisation to death. A participant who had not experienced the event by the applicable observation limit was censored at the last date known to be alive or at the clinical cut-off, whichever came first. The registry reports the endpoint in months, using the median and a 95% confidence interval, for the ITT population.

The median is the point in the estimated survival distribution by which half of the group’s observed survival time lies on either side. It offers a concise description of the central part of a time-to-event distribution, but it does not convey the full shape of the survival experience or the timing of every event. The confidence interval expresses uncertainty in the estimated median; it is not a range containing individual survival times.

Overall survival is used in trials because it records death from any cause and does not require researchers to decide whether a particular progression or symptom event qualifies. It also needs follow-up: participants who are alive at the cut-off contribute information up to their censoring time, rather than being treated as if they had died. The registry reports a stratified log-rank comparison between groups, but does not report a hazard ratio for this primary outcome. A median comparison and a log-rank test answer related but non-identical statistical questions.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary endpoint is a time-to-event measure, analyzed using a stratified log-rank test according to the registry. General explanation: time-to-event methods use both whether an event occurred and the time until it occurred. They can also retain partial follow-up information from participants who have not had an event by the cut-off, through censoring. The registry does not describe the stratification factors or give further details of the analysis implementation.

What do the median and confidence interval describe?

The median survival estimate locates the midpoint of the estimated survival distribution. It is not an arithmetic average and does not imply that every participant survived for that amount of time. The 95% confidence interval describes uncertainty in estimating the group median under the statistical method used. In general, a confidence interval is a range generated by a procedure that, under its assumptions, would capture the target parameter at the stated rate across repeated applications. It is not a statement that a particular participant’s outcome lies within the interval.

The intervals for the two reported medians overlap. General explanation: overlap or non-overlap of separate confidence intervals is not a substitute for a direct test of the between-group contrast. The registry’s primary comparison is the stratified log-rank test, which uses event information over follow-up. The record does not provide the primary hazard ratio, so the relative rate of death over time and its precision cannot be summarized from a hazard-ratio estimate.

What does the log-rank test compare?

General explanation: a log-rank test compares observed and expected event patterns between groups across event times, accounting for censoring under its assumptions. Stratification allows comparisons to be made within specified strata and then combined. The record identifies the method as stratified but does not report which variables defined the strata. A p-value indicates how compatible the observed data are with a specified null model under the test; it does not give the probability that the null model is true, and it does not express the size of the treatment contrast.

The posted primary p-value is below 0.05. The statistical significance threshold used in the sponsor’s primary analysis is not reported in the registry record, so the planning illustration’s two-sided significance level should not be mistaken for the trial’s stated analysis threshold. The primary test result should be interpreted alongside the group estimates and the absence of a posted hazard ratio, rather than in isolation.

How should the binary secondary outcomes be read?

Best overall response and disease control are reported as percentages of participants with 95% confidence intervals. The registry identifies a Cochran-Mantel-Haenszel analysis for each of these comparisons. General explanation: a risk difference expresses the absolute difference between group proportions; a confidence interval around it indicates the uncertainty in that contrast. A confidence interval that includes no difference is compatible with a range of effects in either direction under the method used. The registry also includes simple unadjusted recalculations from posted summary statistics. Those recalculations are not replacements for the sponsor’s analyses, which use the reported Cochran-Mantel-Haenszel method.

The posted recalculated p-values for these binary outcomes are unadjusted. The sponsor’s reported values are based on the named stratified method. Small differences between the recalculated and sponsor-reported values can arise because the procedures differ; the registry does not flag a discrepancy. Such p-values do not account for the number of outcomes examined unless a multiplicity procedure is applied, and no such procedure is reported in the registry record.

What does the pharmacokinetic comparison represent?

The pharmacokinetic outcome reports mean cetuximab concentrations with standard deviations at two different sampling occasions: end of infusion in Week 1 and before infusion in Week 7. The listed groups have different sample sizes and represent different measurement times. A numerical comparison between these summaries is therefore not a randomized comparison of treatment arms. General explanation: a standard deviation describes variation among measurements around a mean; it is not the same as a confidence interval for that mean. The registry does not report a sponsor comparative analysis for this outcome.

What are the planning illustrations?

The registry’s planning illustrations use the observed secondary-outcome effects as inputs, with equal allocation, a two-sided significance level of 0.05, and a normal approximation. They show hypothetical participants per arm for stated power levels and effect fractions. General explanation: a planning calculation asks how many participants might be needed under specified assumptions to detect a chosen effect. It is not observed power, does not establish the power of the completed trial, and is not the trial’s original sample-size justification. The original justification is not reported in the registry record. For a time-to-event endpoint, the event count rather than only the number enrolled is central to many power calculations; the posted record does not give a primary event target or a planning illustration for the primary result.

10. Analyses Reported by the Sponsor

For overall survival, the registry reports group medians and 95% confidence intervals, plus a stratified log-rank test with a p-value below 0.05. It does not report a sponsor hazard ratio for this primary comparison. Because a primary time-to-event contrast cannot be recreated from the posted medians and confidence intervals alone, no simple recalculated hazard ratio or log-rank result is available. The sponsor’s reported analysis is therefore the appropriate result to describe; it should not be replaced by a calculation that the posted summaries cannot support.

For best overall response and disease control, the registry reports Cochran-Mantel-Haenszel analyses. The simple recalculations are unadjusted and use the posted group counts and denominators. Their p-values are close to, but not identical to, the sponsor-reported values. The record does not flag a discrepancy. The two methods do not have to produce identical results, because one is an unadjusted calculation and the other is the named sponsor analysis.

For the pharmacokinetic endpoint, the registry reports no sponsor comparative analysis. Its posted summaries refer to concentrations at different sampling times, so the recalculation is not a randomized treatment contrast. In general, recalculated comparisons are descriptive checks based on available summaries; when a sponsor analysis is reported, it retains priority for describing that analysis.

11. Multiplicity, Missing Data and Analysis Populations

The registry identifies overall survival as the primary outcome and lists several secondary outcomes, including progression-free survival, response, disease control, quality of life, pharmacokinetics and safety. It also lists 8 outcome measures and 4 analyses. The analysis plan’s outcome hierarchy, any multiplicity adjustment, and any rules for confirmatory interpretation are not reported in the registry record.

Multiplicity matters because repeated testing across outcomes can increase the chance of at least one apparently unusual result even when there is no underlying effect. General explanation: a prespecified adjustment or testing hierarchy can help control that risk, but whether a particular method is appropriate depends on the trial’s aims and analysis plan. Without the relevant plan, secondary p-values should be treated as unadjusted evidence unless the registry explicitly identifies an adjustment. The posted recalculated p-values for binary comparisons are unadjusted, and the record does not report a correction for the number of outcomes examined.

The primary outcome population is listed as ITT. The registry does not provide detailed rules for inclusion in that population or explain the difference between enrollment of 1861 and the primary group denominators of 557 and 568. Missing-data methods for the primary and secondary outcomes are not reported in the registry record. The description states how participants without a primary event were censored, but it does not give details on follow-up completeness, withdrawals, or sensitivity analyses for censoring.

The record also does not report an interim-analysis plan, stopping rules, or the original sample-size justification. These omissions do not establish that such procedures were absent; they mean that they cannot be assessed from the registry record. The planning illustrations based on observed secondary effects are not a substitute for the original design assumptions and should not be read as observed power.

12. Safety Results

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

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: febrile neutropenia (17.5%), neutropenia (8.6%), general physical health deterioration (4.0%), pulmonary embolism (3.6%), pneumonia (3.5%).

The registry lists participants at risk for serious adverse events as 548 in the cetuximab-plus-chemotherapy group and 562 in the chemotherapy-alone group. It reports 325 and 244 participants, respectively, as affected by serious adverse events. These are participant counts, not counts of all individual adverse-event episodes. The registry also reports 532 and 537 participants with any adverse event in the corresponding groups.

The listed serious-event terms include febrile neutropenia, neutropenia, general physical health deterioration, pulmonary embolism, pneumonia and dyspnoea, with percentages reported for those terms. The record does not provide an arm attribution alongside this list in the posted information, so the percentages should not be assigned to a particular group. The registry does not report deaths in the adverse-event summary.

General explanation: safety counts describe events observed and recorded during the trial, but comparisons depend on exposure, follow-up, definitions and reporting procedures. The posted figures alone do not establish causality or allow a complete comparison of event timing or duration. They should be read as descriptive safety data, not as a standalone measure of overall benefit or harm.

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

The registry records a randomized comparison of cetuximab plus cisplatin and vinorelbine against cisplatin and vinorelbine alone in the stated first-line advanced NSCLC population. For the primary endpoint, it reports median overall survival estimates of 11.3 months and 10.1 months, with the corresponding 95% confidence intervals, and a stratified log-rank p-value below 0.05. These findings describe the posted trial analysis; they do not establish how an individual patient would fare.

A p-value below 0.05 does not measure the size or importance of a difference, and the median estimates do not describe the complete survival distributions. The registry does not post a primary hazard ratio, which limits description of the relative event rate over time. Nor does the record report the sponsor’s significance threshold for the primary analysis, the original sample-size justification, or details of any interim analyses.

The secondary outcomes offer only partial context. Some results are posted, while results for other listed outcomes are not reported in the registry record. The presence of a secondary outcome or a statistically notable p-value is not, by itself, evidence that a finding is reliable or independently confirmatory. The analysis plan and multiplicity handling are not reported, and the pharmacokinetic summaries compare different sampling times rather than randomized groups.

The data do not establish that the same results would apply outside the enrolled trial population, nor do they answer every question about treatment experience. The open-label design, unexplained difference between enrollment and primary analysis denominators, and incomplete reporting of missing-data procedures are relevant limitations. The interpretation should remain tied to the population, endpoint, follow-up period and analyses recorded in the registry.

14. Why This Trial Matters Statistically

This trial illustrates why a clinical result should be read through both its endpoint and its analysis method. Overall survival is a time-to-event outcome, and the registry pairs median estimates with a stratified log-rank test. Those statistics summarize different features of the data: the medians locate the centers of the estimated group distributions, while the test uses event timing across follow-up. One cannot reconstruct the full time-to-event comparison from the medians alone.

It also demonstrates the limits of interpreting a p-value without an effect measure that describes the comparison over time. A primary hazard ratio is not reported, so the posted primary evidence does not include that relative-effect summary or its precision. The group confidence intervals provide uncertainty around each median, but they are not a confidence interval for a between-group hazard ratio.

The record provides a useful contrast between sponsor analyses and simple unadjusted recalculations for selected binary outcomes. These analyses can differ because they use different methods. The comparison reinforces that a recalculation from summary counts is not a replacement for the sponsor’s specified analysis. Meanwhile, the pharmacokinetic summaries show why not every numerical contrast is a randomized arm comparison: measurements taken at different times answer a different question.

Finally, the case highlights the importance of transparent reporting. Information about multiplicity, missing-data handling, sample-size justification and interim analyses is not reported in the registry record. Those details affect how readers assess precision and the strength of evidence, but their absence from the record cannot be taken as proof that a procedure was or was not used. Statistical interpretation depends both on the results and on the design and analysis details available to evaluate them.

15. Limitations

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

The primary endpoint results include group medians and confidence intervals, but no hazard ratio or confidence interval for the primary contrast. This limits the available description of the relative event rate across follow-up. The registry identifies a stratified log-rank test but does not name the stratification variables or provide further analysis details.

The record reports enrollment of 1861, while the primary outcome denominators are 557 and 568. The reason for this difference is not reported in the registry record. It would be inappropriate to infer a particular exclusion, missing-data pattern or analysis decision from the counts alone. The handling of missing data beyond the stated censoring rule is also not reported.

The study was open-label. The record does not state whether outcome assessment or other trial procedures were masked. It also does not report the original sample-size justification, interim-analysis plan, stopping rules, or a primary event target. These are reporting limitations, not evidence that such procedures were absent.

Several secondary outcomes are listed, but results for some are not reported in the registry record. The multiplicity strategy and the hierarchy for secondary outcomes are also not reported, so the confirmatory status of secondary findings cannot be assessed. The simple recalculations are unadjusted and should not be confused with the sponsor’s reported analyses. Lastly, the pharmacokinetic summaries refer to different sampling occasions and do not provide a randomized comparison of treatment arms.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Study of Cisplatin/Vinorelbine +/- Cetuximab as First-line Treatment of Advanced Non Small Cell Lung Cancer (FLEX) — NCT00148798.
• Registry record last updated 2014-06-25; status last verified 2014-06.
• Results first posted 2011-10-04.
• 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.

19. Record Summary

FLEX was a randomized, open-label, multicenter Phase III trial comparing cetuximab plus cisplatin and vinorelbine with chemotherapy alone in patients described as having advanced EGFR-expressing NSCLC without prior chemotherapy. The registry reports enrollment of 1861 and identifies overall survival as the primary outcome for the ITT population.

Median overall survival was 11.3 months in the cetuximab-plus-chemotherapy group and 10.1 months in the chemotherapy-alone group. The sponsor reported a stratified log-rank comparison with a p-value below 0.05. A primary hazard ratio is not reported in the registry record. Secondary outcomes provide a mixed and incomplete picture, and details on multiplicity, missing-data procedures and the original sample-size justification are not reported.