A Study Of AG-013736 (Axitinib) Or Bevacizumab (Avastin) In Combination With Paclitaxel And Carboplatin In Patients With Advanced Lung Cancer.: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study Of AG-013736 (Axitinib) Or Bevacizumab (Avastin) In Combination With Paclitaxel And Carboplatin In Patients With Advanced Lung Cancer. (NCT00600821), a randomized phase 2 study. The primary outcome is Progression Free Survival (PFS).

2. Clinical Question

The comparison concerns the added study drug in the context of the same chemotherapy combination; it is not a comparison of either study drug with chemotherapy alone or with no treatment. The registry summary frames the study as evaluating the addition of AG-013736 to chemotherapy in previously untreated advanced lung cancer. It does not establish that every participant received treatment exactly as assigned, and the intent-to-treat definition explicitly retains participants according to initial randomization regardless of treatment received.

The clinical question is therefore about the relative pattern of progression-free survival in these randomized treatment groups, as measured under the registry's stated endpoint definition. Overall survival, objective response, response duration, pharmacokinetics, quality of life, RNA expression, and circulating endothelial cells were also listed as secondary outcomes. Their presence broadens the questions recorded for the trial, but the primary comparison remains progression-free survival.

3. Trial Design

The registry describes a randomized, parallel-group Phase 2 trial with no masking. The two study regimens were axitinib plus paclitaxel and carboplatin, and bevacizumab plus paclitaxel and carboplatin. The registry enrollment is 118. The study status is completed; the recorded start is 2008-04 and primary completion is 2011-04.

The listed population had advanced non-small-cell lung cancer, including adenocarcinoma. The brief summary says the study aimed to determine whether adding AG-013736 to chemotherapy was beneficial in participants with advanced lung cancer who had not previously been treated. The official title describes first-line treatment for advanced non-small-cell lung cancer. These descriptions identify the intended clinical setting, but details such as additional eligibility criteria are not reported in the registry record.

The primary endpoint was progression-free survival, with assessments described as occurring at baseline and every 6 weeks until disease progression or initiation of subsequent anticancer therapy, up to 2.75 years. The stated analysis population was intent-to-treat. Randomized participants remained analyzed according to their initial assignment even if they did not receive study drug or received a different drug from the one assigned.

The registry gives the primary comparison method as Log Rank and reports a hazard ratio with a 95% confidence interval and p-value. The record does not describe the original sample-size justification, allocation ratio, planned interim analyses, or a detailed statistical analysis plan; these are not reported in the registry record. The enrollment figure is a registry entry, and the record does not establish from that figure alone how many participants contributed information to every particular endpoint or analysis.

Secondary outcomes included overall survival, objective response, duration of response, pharmacokinetic analysis for axitinib, quality-of-life measures, RNA expression profiles in whole blood, and circulating endothelial cells. The study's listed outcome schedule and the specific analysis population for each secondary endpoint are not fully described in the record.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Intent-to-treat (ITT) population included all participants who were randomized, with study drug assignment designated according to initial randomization, regardless of whether participants received study drug, or received a different drug from that to which they were randomized.

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: Intent-to-treat (ITT) population included all participants who were randomized, with study drug assignment designated according to initial randomization, regardless of whether participants received study drug, or received a different drug from that to which they were randomized.

6. Primary Endpoint

The registry lists the primary outcome as Progression Free Survival (PFS), measured over Baseline, every 6 weeks until disease progression or initiation of subsequent anticancer therapy up to 2.75 years.

Progression-free survival measures time from the start of study treatment to the first of objective tumor progression or death from any cause. The record specifies that progression was identified from oncologic assessment data meeting the criteria for progressive disease, or from adverse-event data when the outcome was death. The endpoint was calculated as the first event date minus the first randomization date, plus 1, divided by 30.4, and expressed in months.

Time-to-event endpoints retain information about both whether an event occurred and when it occurred. They are useful when the timing of progression or death matters and when participants may have different lengths of follow-up. A median provides a summary point in the observed event-time distribution; it does not describe the experience of every participant or capture the full shape of that distribution.

Progression-free survival is used to assess the period during which participants remain alive without documented progression. The registry's definition includes death from any cause, so the endpoint is not limited to radiographic tumor progression. The recorded schedule was baseline and every 6 weeks until progression or initiation of subsequent anticancer therapy, up to 2.75 years. How assessments after treatment discontinuation, missed assessments, or other specific data situations were handled is not reported in the registry record.

The primary result should be read as a comparison of time-to-event patterns under this endpoint definition. The median values and hazard ratio answer related but distinct questions: the medians describe the middle of each group's observed event-time distribution, while the hazard ratio summarizes a relative event-rate contrast over follow-up under the model and analysis method used.

7. Results Posted in the Registry

• Axitinib + Paclitaxel + Carboplatin vs Bevacizumab + Paclitaxel + Carboplatin: hazard ratio 1.09 (95% CI 0.68–1.76), reported p = 0.639 (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.

Overall Survival (OS)

• Endpoint type: Time to event: compared as a hazard ratio.
• Test: Hazard ratio as reported by the sponsor (Log Rank); it was not recalculated.
• Result:
  • Axitinib + Paclitaxel + Carboplatin vs Bevacizumab + Paclitaxel + Carboplatin: hazard ratio 1.12 (95% CI 0.74–1.69), reported p = 0.699 (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 Axitinib + Paclitaxel + Carboplatin vs Bevacizumab + Paclitaxel + Carboplatin and the control. The interval gives the range of effects that remain compatible with the data.

Percentage of Participants With Objective Response (OR)

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Bevacizumab + Paclitaxel + Carboplatin: 43.3% (26/60) versus 29.3% (17/58) on control; risk difference 14.0 percentage points (95% CI -3.1 to 31.2), risk ratio 1.48 (0.90–2.42); unadjusted two-proportion z-test p = 0.114.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Bevacizumab + Paclitaxel + Carboplatin and the control. The interval gives the range of effects that remain compatible with the data.

The secondary outcomes span time-to-event measures, response, treatment exposure, quality of life, and biological measurements. Overall survival and duration of response address clinical event timing; objective response concerns tumor response classification. The listed pharmacokinetic, questionnaire, RNA-expression, and circulating-endothelial-cell outcomes address drug exposure, reported quality of life, and biological features. This breadth means that secondary outcomes answer different questions and should not be collapsed into a single overall conclusion.

For overall survival, the posted estimates show a median of 10.6 months for the axitinib combination and 13.3 months for the bevacizumab combination; the sponsor-reported hazard ratio is 1.117, with a 95% confidence interval from 0.739 to 1.689 and p = 0.699. For objective response, the registry reports group percentages and a sponsor analysis using a Cochran-Mantel-Haenszel risk ratio. These results do not change the interpretation of the primary progression-free survival result: they are separate outcomes, with their own definitions and uncertainty.

The registry lists additional secondary outcomes, but detailed results for every listed measure are not included in the registry record for this analysis. Results, analysis populations, missing-data procedures, and confidence intervals for those outcomes are not reported in the registry record. Their inclusion in the outcome list should not be mistaken for evidence that a particular pattern was observed.

Secondary analyses generally carry less confirmatory weight when many outcomes or comparisons are examined without a stated multiplicity strategy. Repeated testing can increase the chance of at least one apparently notable result even when no real effect exists. The registry does not report a multiplicity adjustment or a prespecified hierarchy for the secondary outcomes. Their results are therefore best interpreted as supporting or exploratory context, not as independent confirmation of the primary result.

These endpoints also differ in what they measure. Overall survival is not interchangeable with progression-free survival; response percentage does not describe how long response lasts; and quality-of-life or biomarker measures cannot be treated as direct substitutes for clinical event outcomes. A coherent interpretation considers each endpoint's question and analysis separately, while recognizing that the record does not supply all details needed for a complete synthesis.

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

• Overall Survival (OS)
• Percentage of Participants With Objective Response (OR)
• Duration of Response (DR)
• Population Pharmacokinetic (PK) Analysis for Axitinib (AG-013736)
• European Organization for Research and Treatment of Cancer, Quality of Life Questionnaire Core-30 (EORTC QLQ-C30) Score
• European Organization for Research and Treatment of Cancer, Quality of Life Questionnaire Lung Cancer-13 (QLQ- LC13) Score
• Percentage of Participants by Ribonucleic Acid (RNA) Expression Profile in Whole Blood
• Circulating Endothelial Cells (CEC) in Blood: Total CEC

9. Statistical Interpretation

10. Statistical Methods Explained

The primary endpoint is a time-to-event outcome. Each participant contributes information about the time to progression or death, or about the period for which the participant was known to remain free of either event. When follow-up ends before an event is observed, time-to-event methods can retain the observed follow-up rather than treating the participant as if an event occurred at the end of observation. The registry does not provide the detailed censoring rules used for this analysis, so the handling of individual follow-up situations cannot be reconstructed from the record.

What does the median describe?

The median progression-free survival is the point in the event-time distribution at which half the observed distribution lies on either side, subject to the survival-analysis estimation used. It is a compact summary, not a statement that every participant experienced progression or death at that time. It also does not show whether the groups' event-time patterns were similar throughout follow-up. The registry reports the medians and their confidence intervals, but does not provide the full survival curves as numeric values in the record.

What does a hazard ratio describe?

A hazard is the event rate at a given time among participants who have not yet experienced the event. A hazard ratio compares those rates between groups over the analysis period, under the assumptions of the method. The reported comparison is axitinib combination versus bevacizumab combination, with a hazard ratio of 1.093. A value of 1 is the reference for no relative difference in hazards; a value on either side indicates a relative direction according to the stated group order. A hazard ratio is not a ratio of median survival times and does not directly give the probability that a participant will experience an event.

The hazard-ratio interpretation is strongest when the relative event rate is reasonably stable over time. If the relative rate changes substantially, one overall hazard ratio can compress different time patterns into a single summary. The registry record does not report a test or assessment of the proportional-hazards assumption, nor does it present a time-varying hazard-ratio analysis. That assumption therefore cannot be evaluated from the posted result alone.

Why use the Log Rank method?

The Log Rank method compares time-to-event experience across groups while incorporating event timing and censored follow-up. It is commonly used for comparing survival distributions. Its result depends on the event and follow-up information available, not just the median values. The registry reports a Log Rank analysis and an associated hazard ratio, but does not give further details about model specifications or adjustments. It is not reported in the registry record whether any covariates were included in the primary comparison.

The p-value of 0.639 describes how compatible the test statistic is with a null comparison under the test's assumptions. It is not the probability that the null hypothesis is true, nor the probability that the observed result occurred by chance in a general sense. A large p-value does not establish that the groups are equivalent; the confidence interval and the design question must also be considered.

What information is in a confidence interval?

A confidence interval expresses uncertainty in an estimated quantity. In repeated use of the same procedure under its assumptions, intervals constructed at the stated confidence level would cover the target parameter at that rate. For an individual interval, it is more useful to view the endpoints as a range of effects reasonably compatible with the data and method than as a probability statement about the fixed true value. The primary hazard-ratio interval, 0.679 to 1.761, crosses the no-relative-difference reference of 1 and includes a range of possible relative effects.

Confidence intervals also help distinguish a point estimate from its precision. A point estimate can look close to the no-difference reference while uncertainty still permits effects in either direction. Conversely, an interval that excludes a reference does not by itself establish clinical importance. Interpretation needs the endpoint, comparison direction, uncertainty, and clinical context, and the registry does not specify a threshold for clinical importance.

Why can over-dispersion matter for other outcomes?

For count outcomes, over-dispersion means the observed variability exceeds what a simple count model would expect. If it is ignored, standard errors can be too small and confidence intervals too narrow. This is a general consideration for count data, not a claim that over-dispersion occurred in this trial. The primary endpoint is time-to-event, and the registry does not report an over-dispersion analysis for the secondary outcomes.

For binary outcomes such as objective response, estimates depend on the number classified as responders and the population used as the denominator. Different methods can estimate different measures, such as a risk ratio or risk difference, and can use different procedures for inference. The registry includes a sponsor-reported Cochran-Mantel-Haenszel risk ratio for objective response, alongside an unadjusted comparison in the posted comparison data. Those quantities should not be treated as the same analysis or as interchangeable results.

Statistical methods can address sampling uncertainty but cannot remove limitations in measurement, follow-up, missing data, or design. The registry lists an ITT definition for the primary endpoint, but detailed methods for missing observations, censoring, multiplicity, and covariate adjustment are not reported in the registry record. These omissions limit how fully the analyses can be independently evaluated.

11. Confidence Intervals and Precision

The primary medians are 5.7 months for the axitinib combination, with a 95% confidence interval from 4.1 to 7.5 months, and 6.1 months for the bevacizumab combination, with a 95% confidence interval from 4.2 to 8.7 months. These intervals overlap. Overlap is not itself a formal test of the difference between groups; the registry's between-group analysis is the Log Rank comparison with hazard ratio 1.093, 95% confidence interval 0.679 to 1.761, and p = 0.639.

The hazard-ratio interval crosses 1, the reference corresponding to equal hazards under this measure. Its endpoints leave room for relative effects in either direction, so the point estimate should not be read as a precise estimate of the treatment contrast. The p-value also does not establish equivalence or rule out a clinically important effect. The registry does not report a threshold for clinical importance or an equivalence margin.

A confidence interval describes uncertainty under the analysis method and its assumptions. It does not account automatically for every source of uncertainty, including unreported choices about missing data, censoring, or endpoint assessment. The registry specifies the progression and death events that define the endpoint, but detailed handling of incomplete assessments is not reported in the registry record.

Precision is distinct from whether a result crosses a conventional significance threshold. The posted estimates provide a point summary and an uncertainty range; the range is broad enough that conclusions about the direction and magnitude of a true relative effect should remain cautious. The trial's original precision target or sample-size justification is not reported in the registry record.

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 the trial's original design calculations. The assumptions listed for the illustration are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect to detect is specified as a fraction of the observed effect. For time-to-event outcomes, the figures are total events calculated using the Schoenfeld approach; for other outcomes, they are participants per arm. The original trial design assumptions are not reported in the registry record.

For the primary time-to-event result, planning for 100% of the observed effect gives 3,971 total events for 80% power and 5,315 total events for 90% power. Planning for 75% of the observed effect gives 7,059 total events for 80% power and 9,449 total events for 90% power. Planning for 50% of the observed effect gives 15,881 total events for 80% power and 21,260 total events for 90% power. These are total event counts, not participants per arm.

In plain terms, the 80% and 90% power columns describe different planning targets for the chance of detecting the specified effect if that effect were present and the assumptions held. The higher power target requires more information. Smaller effects require more participants or events to detect reliably than larger effects. These figures are not observed power; they are planning illustrations and do not describe the probability that the completed trial's result is correct.

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 = 1.093, so ln(HR) = 0.0889.
2. d = 4 × (1.96 + 0.842)² / [ln(1.093)]² = 3,970.2, rounded up to 3,971 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 <- 1.093                              # 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(600821)
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 = 3,971 (80% power), 5,315 (90% power)
 75% of observed effect: events = 7,059 (80% power), 9,449 (90% power)
 50% of observed effect: events = 15,881 (80% power), 21,260 (90% power)
Simulated power with 3,971 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

For progression-free survival, the registry reports the sponsor's Log Rank hazard ratio as 1.093, with a 95% confidence interval from 0.679 to 1.761 and p = 0.639. Those are the appropriate reported figures for describing the primary comparison. The registry lists the same estimate and inferential values in its analysis record.

The sponsor-reported hazard ratio should be distinguished from the group medians. Medians summarize each group's event-time distribution separately, while the hazard ratio is a relative time-to-event comparison. Neither is a substitute for the other. The reported analysis is not a recalculation from the median values, and no discrepancy is identified in the record for this primary result.

For objective response, the record gives a sponsor-reported Cochran-Mantel-Haenszel risk ratio of 0.676, with a 95% confidence interval from 0.412 to 1.107 and p = 0.942. The posted comparison data also include an unadjusted risk-difference calculation with p = 0.114. These are different estimands and methods, so their numerical results need not match. The sponsor's reported analysis should be kept distinct from the unadjusted comparison rather than treated as a direct recalculation of it.

The registry does not provide enough detail to assess the full specifications behind every sponsor analysis, including any stratification factors or adjustment variables. Where the record gives a sponsor-reported estimate, that result is the basis for reporting the sponsor's analysis; it should not be silently replaced by a different method's estimate.

15. Multiplicity, Missing Data and Analysis Populations

The registry identifies progression-free survival as the primary outcome and lists multiple secondary outcomes, including overall survival, objective response, duration of response, pharmacokinetic analysis, quality-of-life scores, RNA expression, and circulating endothelial cells. Testing multiple endpoints can raise the chance of a nominally positive result occurring by chance across the family of analyses. A prespecified multiplicity strategy can control or describe that risk, but one is not reported in the registry record.

The registry does not report whether secondary endpoints were organized into a testing hierarchy, whether any adjustment was planned, or whether analyses were designated as confirmatory or exploratory. Without those details, individual secondary p-values should not be treated as though each were the sole prespecified test. The primary result remains the clearest basis for the central comparison, with secondary findings interpreted cautiously and in their separate clinical and statistical contexts.

The primary population is defined as intent-to-treat: randomized participants were analyzed according to initial assignment regardless of whether they received study drug or received a different drug. This is an important part of the endpoint definition. The record does not specify analysis populations for every secondary outcome, nor does it provide participant flow, endpoint-specific missingness, or the number contributing to each analysis beyond the group information recorded for particular outcomes.

Missing-data handling is not reported in the registry record. For a time-to-event outcome, censoring rules determine how incomplete follow-up contributes information. The record describes event definitions and assessment timing, but does not detail censoring, missed assessment handling, or sensitivity analyses. For questionnaire, biomarker, and binary outcomes, the consequences of missing measurements depend on the analysis population and missing-data method; those methods are not reported in the registry record.

The original sample-size justification, planned interim analyses, detailed statistical analysis plan, and covariate-adjustment strategy are also not reported in the registry record. These omissions do not invalidate the posted estimates, but they limit assessment of whether all analyses followed prespecified plans and how robust the results are to alternative assumptions. The planning illustrations based on observed effects do not supply the trial's original design rationale.

16. 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: disease progression (12.1%), dehydration (6.9%), asthenia (5.2%), myocardial infarction (3.4%), sepsis (3.4%).

The safety figures describe participants recorded as affected by serious adverse events and by other adverse events in each group. For axitinib plus paclitaxel and carboplatin, serious adverse events affected 31 of 58 participants at risk, and other adverse events affected 54 of 58 at risk. For bevacizumab plus paclitaxel and carboplatin, serious adverse events affected 19 of 59 participants at risk, and other adverse events affected 58 of 59 at risk. These are participant counts with the recorded denominators, not event counts.

The registry's listed serious adverse-event terms include disease progression, dehydration, asthenia, myocardial infarction, sepsis, and neutropenia, with percentages reported for each term. Those percentages should be read as the record's term-specific figures; the record does not provide the denominator details needed to independently reconstruct them. Disease progression appears among the listed serious terms, so the terms should not be interpreted as a list of effects caused by study treatment.

These safety summaries do not establish comparative risk by themselves. The record does not give a causal attribution for each event, detailed exposure time, or a complete account of severity and timing. Death counts are not reported in the registry record. Differences in recorded counts should be interpreted in light of the study's size, follow-up, and the available reporting details, rather than as definitive evidence of a treatment-related safety difference.

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

The posted primary result describes a randomized comparison of two chemotherapy combinations, one with axitinib and one with bevacizumab, for advanced non-small-cell lung cancer. Median progression-free survival was 5.7 months in the axitinib group and 6.1 months in the bevacizumab group. The sponsor-reported hazard ratio was 1.093, with a 95% confidence interval from 0.679 to 1.761 and p = 0.639. These estimates do not provide precise evidence of a difference between the regimens.

That conclusion is not the same as saying the regimens are equal. The confidence interval includes the no-difference reference and a range of values on both sides, leaving uncertainty about the relative effect. The registry does not report an equivalence or non-inferiority design or margin, so equivalence cannot be inferred. Nor does a non-significant p-value establish that no difference exists.

The result also does not establish that either regimen works or does not work in an absolute sense. The record describes a head-to-head comparison, not a comparison with no treatment or a control without the chemotherapy combination. The interpretation is limited to the posted outcomes and analysis, and should not be extended to populations or treatment settings not described in the record.

The primary endpoint incorporates progression and death, while secondary endpoints address other outcomes. Their estimates do not erase the uncertainty in the primary comparison, and their different definitions prevent treating them as interchangeable. The record's secondary analysis details and multiplicity strategy are incomplete, which further limits their confirmatory interpretation.

In general, clinical meaning depends on more than a statistical test. The size and precision of an effect, how an endpoint maps to patient experience, the quality of follow-up, and the study design all matter. The registry does not specify a threshold for a clinically important difference. The defensible reading is therefore limited: the trial reports no clear statistical evidence of a difference in its primary time-to-event comparison, while uncertainty remains about the magnitude and direction of the underlying effect.

18. Why This Trial Matters Statistically

This trial illustrates why a randomized comparison needs to be interpreted through both its point estimates and their uncertainty. The two reported median progression-free survival values are close, but the medians alone do not quantify the full time-to-event contrast. The hazard ratio supplies a relative comparison, and its confidence interval shows that the estimate remains imprecise. Reading the point estimate without the interval would give an incomplete account of the evidence.

It also illustrates the difference between failure to show a difference and evidence of equivalence. The reported p-value is not evidence that the treatment combinations have identical effects. Equivalence requires a design and margin suited to that question; neither is reported in the registry. Statistical interpretation should follow the question the design can answer, not substitute a different claim after seeing the result.

The trial's outcomes also demonstrate the challenge of interpreting a broad endpoint set. Progression-free survival, overall survival, response, quality of life, and biological measures answer distinct questions. Without a reported multiplicity strategy, secondary findings require restraint, particularly when analyses use different estimands and methods.

Finally, registry records distinguish what is reported from what cannot be assessed. The randomized design and ITT definition are useful information, while missing detail about sample-size justification, missing data, censoring, interim analyses, and analysis planning limits independent appraisal. Recognizing those boundaries is part of sound statistical reading: it avoids turning an uncertain estimate into a stronger conclusion than the record supports.

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 primary hazard-ratio confidence interval is broad and includes values on both sides of 1, limiting precision about the direction and size of the relative effect. The record does not report a clinically important difference threshold, so the clinical importance of the estimates cannot be judged against a stated margin. It also does not report an equivalence or non-inferiority design.

The registry does not give the original sample-size justification, detailed statistical analysis plan, allocation ratio, interim-analysis plan, covariate-adjustment details, or assessment of the proportional-hazards assumption. Consequently, it is not possible from the record to evaluate all design and modeling choices behind the primary comparison.

Although the primary population is defined as intent-to-treat, detailed participant flow and endpoint-specific analysis populations are not reported in the registry record. The record also does not describe censoring rules, missing-data handling, sensitivity analyses, or how incomplete tumor assessments were treated. These details are important for interpreting time-to-event and secondary outcomes.

Multiple secondary outcomes are listed, but a multiplicity plan and testing hierarchy are not reported. The available secondary analysis details are incomplete, so the full pattern of findings cannot be assessed from the record alone. Safety reporting also has limits: event attribution and exposure-time detail are not reported, and death counts are not reported in the registry record. These constraints call for cautious interpretation and restrict the conclusions to the posted information.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: A Study Of AG-013736 (Axitinib) Or Bevacizumab (Avastin) In Combination With Paclitaxel And Carboplatin In Patients With Advanced Lung Cancer. — NCT00600821.
• Registry record last updated 2013-11-08; status last verified 2013-10.
• Results first posted 2012-05-22.
• 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

NCT00600821 is a completed randomized Phase 2 parallel-group trial without masking. It compared axitinib, also identified as AG-013736, plus paclitaxel and carboplatin with bevacizumab plus paclitaxel and carboplatin in advanced non-small-cell lung cancer. The registry enrollment is 118; the recorded study start is 2008-04 and primary completion is 2011-04.

The primary outcome was progression-free survival, defined from study treatment start to objective tumor progression or death from any cause. Median progression-free survival was 5.7 months for the axitinib combination and 6.1 months for the bevacizumab combination. The sponsor-reported Log Rank hazard ratio was 1.093, with a 95% confidence interval from 0.679 to 1.761 and p = 0.639.

The result does not establish equivalence, and its confidence interval leaves uncertainty about the relative effect. The registry defines an ITT population for the primary endpoint, but does not report several details needed for fuller assessment, including missing-data procedures, multiplicity strategy, and the original sample-size justification.