Study of Pemetrexed + Carboplatin + Bevacizumab + Enzastaurin Versus Pemetrexed + Carboplatin + Bevacizumab + Placebo in Participants With Non-Small Cell Lung Cancer Who Have Not Been Previously Treated With Chemotherapy: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Study of Pemetrexed + Carboplatin + Bevacizumab + Enzastaurin Versus Pemetrexed + Carboplatin + Bevacizumab + Placebo in Participants With Non-Small Cell Lung Cancer Who Have Not Been Previously Treated With Chemotherapy (NCT00533429), a randomized phase 2 study. The primary outcome is Progression-Free Survival (PFS).

2. Clinical Question

3. Trial Design

This was a randomized, double-blind, phase 2, parallel-group study sponsored by Eli Lilly and Company. Its registered title describes a comparison of pemetrexed, carboplatin and bevacizumab combined with either enzastaurin or placebo in participants with non-small cell lung cancer who had not previously been treated with chemotherapy. The official title specifies chemonaive participants with stage IIIB or IV disease.

The registry records enrollment of 40 and a completed status. The recorded start was 2007-10, and primary completion was 2009-02. The two assigned regimens shared pemetrexed, carboplatin and bevacizumab; the distinguishing treatment was enzastaurin versus placebo. The study summary describes the experimental regimen as followed by maintenance bevacizumab plus enzastaurin.

Randomization supports a comparison between assigned groups, while double masking is intended to limit the influence of treatment knowledge on participant care and outcome assessment. The record does not report the randomization procedure, allocation ratio, masking roles, or details of how adherence to masking was assessed. It also does not report the planned sample-size justification or a target number of participants distinct from the recorded enrollment.

The primary outcome was progression-free survival. The primary analysis population is described as all randomized participants identified by assigned treatment group, although the reported group counts for the primary outcome are 15 and 18. The registry separately reports 20 participants in each group for response rate. The reasons for these differences in analysis counts are not reported in the registry record, so the outcome-specific denominators should be kept in view when comparing results.

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: Participant flow reports those participants who discontinued from study drug.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: All randomized participants identified by assigned treatment group. Number of participants censored: Pemetrexed + Carboplatin + Bevacizumab + Enzastaurin = 5; Pemetrexed + Carboplatin + Bevacizumab + Placebo = 2.

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 who received at least 1 dose of study drug.

6. Primary Endpoint

The registry lists the primary outcome as Progression-Free Survival (PFS), measured over Randomization to measured PD or death from any cause up to 12.2 months.

Progression-free survival measures the time from randomization until the first recorded progressive disease or death from any cause. It is a time-to-event outcome: each participant contributes information about the time until an event, or about the period during which they are known to remain free of the event. The endpoint can therefore use information from participants whose follow-up ends before an event is observed, through censoring, rather than treating them as if an event had occurred.

The registry defines censoring for participants not known to have died by the data cutoff and without objective progressive disease as occurring at their last objective progression-free assessment. It also describes censoring when a participant began subsequent anticancer therapy after study treatment ended but before objective progression or death: in that circumstance, censoring was at the last objective progression-free assessment before the subsequent therapy began.

Progression-free survival is used to capture both disease progression and death in a single time-to-event measure. It does not, by itself, describe overall survival, response quality, symptoms, or treatment burden. The reported timeframe is randomization to measured progressive disease or death from any cause up to 12.2 months. The registry does not provide further detail about the assessment schedule or the criteria used to determine progression.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does a time-to-event analysis use?

Progression-free survival is measured from randomization until progressive disease or death. Unlike a simple count of whether an event occurred, a time-to-event analysis uses both event timing and follow-up time. Participants whose event is not observed contribute information up to the time they were last known to be progression-free, subject to the analysis's censoring rules. This allows the analysis to retain partial follow-up information rather than discarding every participant without an observed event.

Censoring is not the same as a favorable outcome or a missing event. It indicates that the event time is not fully observed within the relevant follow-up. Valid interpretation depends on assumptions about censoring, including that censoring does not systematically hide a different event pattern in one group in a way that distorts the comparison. The registry describes when censoring was applied, but does not report detailed diagnostics of the censoring assumptions or the full distribution of follow-up times.

What does the log-rank test assess?

The registry reports a log-rank comparison for the primary outcome. In general, the log-rank test compares the event experience of groups across follow-up, accounting for when events occur and who remains at risk at different times. Its p-value measures how surprising the observed separation, or a more extreme separation, would be under a statistical model in which there is no group difference and the test assumptions hold. It is not the probability that the null hypothesis is true, nor does it quantify the size of a treatment effect.

The reported p-value is 0.469. It should be interpreted alongside the group estimates, confidence intervals, sample sizes and censoring. A p-value alone does not show whether the estimate is precise or whether a clinically important difference can be ruled out. The registry does not report a hazard ratio or its confidence interval for the primary comparison, and no such estimate should be inferred from the log-rank p-value.

How should the confidence intervals be read?

The primary group values are accompanied by 95% confidence intervals. In general, a confidence interval communicates uncertainty around an estimate under the method used to construct it. A wide interval indicates that the data are compatible with a broad range of underlying values. It is not a range containing a fixed proportion of individual participants, and it does not express the probability that a particular treatment effect is true.

The intervals reported for the two groups overlap, and both extend across a range of month values. This reinforces that the displayed group values are not precise enough to support a sharp distinction between them. Group-specific intervals are not a substitute for a confidence interval for the between-group contrast. The registry does not report a confidence interval for the log-rank comparison or a direct estimate of the primary group difference.

What is the response-rate comparison?

Response rate is a binary outcome: each participant is classified according to whether a complete or partial response was recorded. The registry reports percentages and 95% confidence intervals for each assigned group. Fisher's exact test, the sponsor's reported method, is a test for association between group and a binary outcome. It is often suitable when counts are small because it calculates a probability using the observed margins rather than relying on a large-sample approximation.

The sponsor reports p = 0.462 for response rate. An unadjusted recalculation from the posted summary statistics reports p = 0.465. The recalculation is not an adjusted analysis and does not replace the sponsor's analysis. The small difference between the two p-values does not change the broad interpretation: the data do not establish a clear group difference, and the confidence intervals are wide. The registry does not report an adjusted response analysis or a prespecified threshold for interpreting the response-rate difference.

Why does over-dispersion matter in general?

Over-dispersion is a general issue for count outcomes when the variability in the data exceeds what a chosen model assumes. If a model underestimates variability, its standard errors can be too small and intervals too narrow. A suitable approach depends on the outcome distribution and the analysis design. The posted primary endpoint is time-to-event, and the response endpoint is binary; the registry does not identify an over-dispersion model for these comparisons. No over-dispersion conclusion should be attached to these results based on the information reported.

How should the planning illustrations be used?

The planning illustrations use the observed response effect and specify a two-sided significance level of 0.05, equal allocation between groups, and a normal approximation. They present participants per arm for 80% power and participants per arm for 90% power at 100% of the observed effect, 75% of the observed effect, and 50% of the observed effect. The corresponding values are 294 and 392; 540 and 722; and 1,251 and 1,674, respectively.

These are planning illustrations, not observed power and not the trial's own design justification. They show how the required sample size can change when the effect chosen for planning is smaller than the observed effect. The registry does not report that these assumptions were used to design the trial, nor does it report the original sample-size calculation. The illustration is therefore useful for understanding planning sensitivity, but it cannot establish whether the original trial was adequately sized for its intended question.

10. Analyses Reported by the Sponsor

For progression-free survival, the registry reports the sponsor's log-rank p-value of 0.469. No recalculated primary comparison is posted, so there is no recalculated primary p-value or contrast to set alongside the sponsor's analysis. The group values and their 95% confidence intervals remain useful for describing the observed data, but they do not supply a separate test of the difference.

For response rate, the sponsor reports Fisher Exact p = 0.462, while the unadjusted recalculation from the posted summary statistics is p = 0.465. These values are close and lead to the same broad reading of substantial uncertainty. The recalculation is a simple, unadjusted check using the posted group summaries; it does not incorporate adjustment for other outcomes, covariates, or any analysis details that may not appear in the registry. The registry does not report further details about the response analysis that would explain the small difference between the p-values.

There is no reported discrepancy flag for these results. The sponsor's own analysis is therefore presented as reported, with the recalculation treated as a secondary check rather than a replacement. Neither p-value should be interpreted as a measure of the probability that the treatment effect is zero.

11. Multiplicity, Missing Data and Analysis Populations

The registry identifies progression-free survival as the primary outcome and lists several secondary outcomes, including response rate, overall survival, time to progressive disease, duration of response, and pharmacology, toxicity and adverse events. It records six outcome measures and three analyses. However, a statistical plan describing how the primary and secondary comparisons were organized is not reported in the registry record. In particular, it does not report a multiplicity adjustment, a hierarchy among secondary outcomes, or whether secondary tests were intended to be confirmatory or exploratory.

Multiplicity matters because repeated testing can make chance findings more likely across a collection of outcomes. In general, an analysis plan can address this by defining which comparisons are primary, setting an order for testing, or adjusting error rates. Without a reported plan, individual secondary p-values should be interpreted cautiously and should not be read as independent confirmation. The posted primary log-rank result remains the central inferential result, while unadjusted secondary comparisons are supportive or exploratory at most.

The record reports censoring rules for the primary endpoint. Participants without known death or objective progression by the cutoff were censored at their last objective progression-free assessment. Those beginning subsequent anticancer therapy after discontinuation but before progression or death were censored at their last objective progression-free assessment before that therapy. The primary outcome population is described as all randomized participants by assigned group. The numbers censored were 5 in the enzastaurin group and 2 in the placebo group.

The registry does not report the amount of missing outcome data by reason, detailed follow-up completeness, sensitivity analyses for censoring, or a missing-data analysis plan. It also does not report a separate per-protocol population or other analysis populations beyond the stated randomized population for the primary endpoint. The primary endpoint counts are 15 and 18, while the response-rate counts are 20 in each group; reasons for the differing counts are not reported. This limits assessment of how the analysis denominators relate to all enrolled or randomized participants.

The registry does not report interim analyses, stopping rules, or a detailed statistical analysis plan. Their absence from the record should not be interpreted as evidence that they did or did not occur; their status is not reported in the registry record.

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: constipation (10.0%), dehydration (10.0%), dyspnoea (10.0%), thrombosis (10.0%), abdominal pain (5.0%).

The posted safety counts distinguish serious adverse events from other adverse events. Serious adverse events affected 6 of 20 participants in the enzastaurin group and 7 of 20 in the placebo group. Other adverse events affected 19 of 20 and 20 of 20 participants, respectively. These are participant counts in the registry's safety groups; they do not describe event duration, severity within each category, or whether an individual experienced multiple events.

The registry also lists serious adverse event terms with percentages: constipation, dehydration, dyspnoea and thrombosis at 10.0% each, and abdominal pain and ascites at 5.0% each. These entries should be read as the reported term-specific percentages, not as a complete accounting of every adverse event or as evidence that each term was caused by a particular treatment. The record does not report death counts in the safety summary.

Safety counts are descriptive and do not establish comparative risk by themselves. The small group sizes and the lack of further detail about event timing, recurrence and severity constrain interpretation. A difference in counts should not be treated as proof of a treatment-related effect.

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

The trial provides a randomized comparison of two regimens that differed by enzastaurin versus placebo, with progression-free survival identified as the primary outcome. The posted group values are 3.5 months and 4.3 months, and the sponsor's log-rank comparison reports p = 0.469. These findings describe the observed data and indicate that the primary comparison did not establish a statistically clear difference in this trial.

That result does not show that the regimens are equivalent. A non-significant comparison can arise when the true difference is small, when estimates are imprecise, or when the study has limited information to distinguish groups. Equivalence or non-inferiority requires a prespecified design, margin and analysis; none is reported in the registry record. The result also does not establish that the treatments have no effect. The reported confidence intervals and the modest outcome counts leave considerable uncertainty about the underlying values.

The findings do not, by themselves, settle the secondary questions. Response-rate percentages are also imprecise, and several other secondary outcomes are listed without numerical results in the registry record. The safety counts offer a descriptive account of affected participants, but do not establish causation or comparative safety. No clinical recommendation follows from these statistical results alone.

Interpretation should account for the outcome-specific analysis counts, censoring, unreported details of follow-up and the absence of a reported multiplicity plan. The sample-size illustrations are planning aids based on the observed response effect, not evidence of the study's observed power and not confirmation that the trial was designed around those assumptions. The overall conclusion is bounded: the registry records the trial's comparison and estimates, but those estimates are uncertain and the record does not support a definitive claim of benefit, equivalence, or absence of effect.

14. Why This Trial Matters Statistically

This trial illustrates why a randomized comparison is not interpreted from the numerical group estimates alone. The primary endpoint is time-to-event, so the analysis incorporates both when progression or death occurred and the information contributed by participants whose outcomes were censored. Understanding the censoring rules and analysis population is essential to understanding what the reported group values represent.

It also illustrates the distinction between an estimated value and a test of a between-group difference. The registry reports group-level progression-free survival values with confidence intervals and a log-rank p-value. These answer related but different questions: the values describe each group, while the log-rank test assesses evidence of a difference across follow-up. A p-value does not convey the size or precision of a difference, and the registry does not report a direct primary contrast with a confidence interval.

The response-rate comparison shows why secondary findings need context. Its percentages and Fisher Exact p-value add information about a different outcome, but they are not a substitute for the primary endpoint. When multiple outcomes are considered, the absence of a reported multiplicity plan makes isolated unadjusted findings less persuasive as confirmatory evidence.

Finally, the planning illustrations show how sample-size requirements depend on the effect chosen for planning. They are not observed power, do not explain the original design, and should not be confused with the trial's enrollment. The registry does not report the original sample-size justification, detailed statistical plan or interim-analysis procedures. Recognizing these gaps is part of sound statistical interpretation: the conclusions should remain within the evidence and reporting available.

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 registered enrollment was 40, but the reported primary outcome group counts are 15 and 18, while response-rate counts are 20 in each group. The registry describes the primary population as all randomized participants by assigned group but does not explain the outcome-specific differences in counts. This leaves uncertainty about which participants contribute to each outcome estimate and why the denominators differ.

The primary group values have broad confidence intervals, and the registry does not report a direct estimate and confidence interval for the between-group primary contrast. The sponsor reports a log-rank p-value, but the registry does not provide a hazard ratio or further details about the model and assumptions used. Censoring rules are described, yet the distribution and reasons for censoring, completeness of follow-up, and sensitivity analyses are not reported.

The registry does not report the original sample-size rationale, a detailed statistical analysis plan, interim analyses, stopping rules, or a multiplicity strategy. These omissions limit assessment of whether the reported analyses correspond to prespecified methods and how secondary comparisons should be weighted. The sample-size illustrations are based on the observed response effect and should not be mistaken for the trial's design calculation.

Numerical results are not reported in the registry record for several listed secondary outcomes in the information available for this analysis. Safety reporting is limited to affected-participant counts and selected serious adverse event terms; detail on event timing, recurrence, severity and causal assessment is not reported. These limitations constrain conclusions about both efficacy-related outcomes and safety and reinforce that the results are descriptive evidence from a small phase 2 trial.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Study of Pemetrexed + Carboplatin + Bevacizumab + Enzastaurin Versus Pemetrexed + Carboplatin + Bevacizumab + Placebo in Participants With Non-Small Cell Lung Cancer Who Have Not Been Previously Treated With Chemotherapy — NCT00533429.
• Registry record last updated 2021-05-13; status last verified 2021-04.
• Results first posted 2021-05-13.
• 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.

19. Record Summary

The registry describes a completed randomized, double-blind, phase 2, parallel-group trial sponsored by Eli Lilly and Company. It enrolled 40 participants with previously untreated stage IIIB or IV non-small cell lung cancer and compared pemetrexed, carboplatin and bevacizumab with enzastaurin against the same regimen with placebo.

Progression-free survival was the primary endpoint. The reported group values were 3.5 months and 4.3 months, with a sponsor-reported log-rank p-value of 0.469. Response rate was also reported; the sponsor's Fisher Exact p-value was 0.462, and an unadjusted recalculation was 0.465. The record includes censoring information and selected safety counts, but leaves key design and analysis details, including the sample-size rationale and multiplicity plan, not reported in the registry record.