Carboxyamidotriazole in Treating Patients With Stage III or Stage IV Non-small Cell Lung Cancer: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Carboxyamidotriazole in Treating Patients With Stage III or Stage IV Non-small Cell Lung Cancer (NCT00003869), a randomized phase 3 study. The primary outcome is Overall Survival (OS).

2. Clinical Question

The clinical context in the registry record is advanced lung cancer after standard chemotherapy. The trial summary notes that chemotherapy may be followed by tumor relapse and growth, and states that it was not known whether carboxyamidotriazole was more effective than no further treatment. The randomized placebo comparison was designed to address that uncertainty in this study population; it does not, by itself, answer questions about other populations or treatment settings.

3. Trial Design

This was a randomized, double-blind, parallel-group phase III trial sponsored by the National Cancer Institute. The registry record lists 186 participants and describes the trial as completed. The start date is 1999-04 and the primary completion date is 2008-05. The record does not report the reasons for the length of the study period or any details about the sequence used to assign participants.

The two groups were carboxyamidotriazole and placebo. The intervention description states that participants in each group received their assigned oral treatment daily. The posted overall survival results include 94.0 participants in the carboxyamidotriazole group and 92.0 in the placebo group. These are the reported arm counts for that outcome; they need not match the number available for every secondary measure.

The primary endpoint was overall survival, defined as time from randomization to death from any cause. Follow-up for this endpoint was for a maximum of 5 years from randomization. Participants who did not die, or who were lost to follow-up, were censored at their last evaluation or follow-up date. The registry describes the analysis population as all randomized participants on an intent-to-treat basis.

Listed secondary outcomes include time to disease progression, severe non-hematologic adverse events, confirmed tumor responses, and clinically significant decreases in two quality-of-life measures from baseline to week 8. Quality-of-life assessment and laboratory biomarker analysis are also listed among the trial activities. Detailed protocol information about the frequency of assessment, allocation concealment, masking procedures, adherence, and the planned analysis sequence is not reported in the registry record.

4. Participant Flow and Accounting

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

Before assignment: 186 patients were randomized to either arm

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Overall survival was analyzed on all randomized participants on an intent to treat basis.

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 (OS), measured over up to 5 years.

Overall survival measures the time from randomization until death from any cause. It is a time-to-event endpoint: the analysis uses both whether an event occurred and how long participants were followed before an event or censoring. This allows information from participants with different follow-up durations to contribute without treating a participant who has not died by the end of observation as though their eventual survival time were known.

The registry reports the median overall survival estimated using the Kaplan-Meier method, with a 95% confidence interval for each group. The median is the estimated time by which half of the group has experienced the event. A confidence interval conveys uncertainty in that estimate; it is not a range guaranteed to contain the true value. The reported follow-up was for a maximum of 5 years from randomization.

Overall survival is clinically interpretable because it concerns death from any cause rather than a surrogate measure. Its definition also avoids requiring investigators to classify the cause of death. However, a difference between group medians alone does not describe the full time-to-event pattern. The registry also reports a Log Rank comparison, which assesses the evidence for a difference in survival experience across follow-up. Neither the median nor the test result establishes why any group difference might occur.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

General statistical explanation: time-to-event methods are useful when participants can have different follow-up durations and some have not experienced the event by the end of observation. Censoring records the last point at which a participant's event-free status is known, rather than assigning an unobserved event time. The registry specifies censoring at the last evaluation or follow-up date for participants who did not die or were lost to follow-up.

What does the Kaplan-Meier method estimate?

The Kaplan-Meier method estimates survival over time while accounting for censored observations. The reported median is the time at which the estimated survival curve reaches the halfway point. It is an estimate of the middle of the event-time distribution, not a measure of the average follow-up or a summary of every feature of the survival curves. Medians can be useful for communication, but they do not fully describe how event rates may vary during follow-up.

The registry gives a 95% confidence interval for each median. General statistical explanation: a confidence interval expresses the uncertainty associated with an estimate under the method's assumptions. A wider interval indicates less precision than a narrower interval. Overlap between two arm-specific intervals is not, by itself, the formal test of the between-group difference. The registry's reported comparison uses a Log Rank test rather than a test based only on the two medians.

What does the Log Rank test address?

The Log Rank test compares the observed and expected event patterns between groups across follow-up, taking account of censoring. Its p-value summarizes how compatible the observed data are with the test's no-difference hypothesis, under the assumptions of the test. The reported p-value is 0.54. A p-value is not the probability that the no-difference hypothesis is true, and it does not measure the size or clinical importance of a difference. The registry record does not provide a hazard ratio or a confidence interval for a between-group time-to-event effect.

The registry identifies the primary population as all randomized participants analyzed on an intent-to-treat basis. General statistical explanation: an intent-to-treat approach retains participants in their randomized groups for analysis, regardless of later treatment adherence, which helps preserve the purpose of randomization. The registry does not give enough detail to assess how every participant's follow-up was handled beyond the stated censoring rule.

How should the secondary binary outcomes be read?

The week 8 quality-of-life outcomes are reported as participant counts meeting a clinically significant decrease definition. The registry's reported comparison method is Fisher Exact. The accompanying recalculated risk differences are unadjusted comparisons derived from posted summary counts. A risk difference describes the difference in the observed proportions between groups, with the sign depending on the stated comparison direction. Its confidence interval expresses uncertainty about that difference.

General statistical explanation: for a risk-difference interval, an interval that includes zero is compatible with no difference under the interval method; an interval that excludes zero indicates evidence of a nonzero unadjusted difference under that method. Neither interpretation resolves issues such as missing observations, the choice of analysis population, or multiplicity from considering several outcomes. A risk difference also does not convey the same information as a relative risk or odds ratio, even though these measures can be calculated from the same binary data.

The registry reports Fisher Exact p-values for the two quality-of-life outcomes. The recalculated p-values are simple unadjusted re-calculations from posted summary statistics, while the sponsor's reported tests are the registry's stated analyses. Their results should be distinguished rather than treated as if they came from an identical calculation. The registry does not describe a multiplicity adjustment for these secondary comparisons.

What do the planning illustrations mean?

General statistical explanation: a sample-size calculation asks how many participants might be needed to detect a specified effect under stated assumptions. It is not a measure of the probability that the completed trial has found an effect, and it is not observed power. The planning illustrations in the posted material are based on fractions of observed secondary effects, a two-sided significance level 0.05, equal allocation between the groups, and a normal approximation. They are not the trial's original design justification; that justification is not reported in the registry record. Because the illustrations use observed effects, they should be read as hypothetical planning exercises, not as confirmation that those effects are reproducible.

For time-to-event planning, event counts and participant counts are not interchangeable. The stated planning description distinguishes total events for time-to-event outcomes from participants per arm for other outcomes. General statistical explanation: event-driven calculations depend on the number of observed events, while binary-outcome calculations depend on the number of participants and the assumed event proportions. These assumptions shape a planning result and do not change the evidence observed in the completed trial.

The registry does not report the full statistical analysis plan, sample-size justification, interim analyses, or a comprehensive approach to missing data and multiplicity. The reported methods therefore allow a reader to understand the main endpoint definition and named tests, but not to audit every analysis decision made during the trial.

10. Analyses Reported by the Sponsor

For overall survival, the registry reports the sponsor's Log Rank analysis with p = 0.54. The posted arm medians and confidence intervals are not sufficient to reconstruct that time-to-event test or a hazard ratio. No recalculated hazard ratio is reported in the registry record, and the median-time comparison is not replaced by an unadjusted calculation from the arm summaries.

For the two posted quality-of-life outcomes, the registry reports Fisher Exact tests. The unadjusted recalculation for the UNISCALE decrease gives p = 0.036, compared with the reported p = 0.04. For the FACT-L decrease, the unadjusted recalculation gives p = 0.164, compared with the reported p = 0.18. The registry does not report a reason for these numerical differences. The recalculations are simple comparisons from the posted summary counts and should not be presented as adjusted or as a replacement for the named sponsor analyses.

The primary sponsor analysis remains the basis for describing the primary result. For the secondary outcomes, the recalculated estimates are useful as unadjusted summaries, but their interpretation is limited by uncertainty, incomplete outcome information, and the absence of a reported multiplicity strategy. Neither set of p-values alone describes the magnitude or precision of an effect; those judgments also depend on the endpoint estimates and their confidence intervals.

11. Multiplicity, Missing Data and Analysis Populations

The trial includes a primary overall survival endpoint and several secondary outcomes. The registry record does not report a statistical strategy for controlling the chance of false-positive findings across the secondary comparisons, such as a prespecified hierarchy or adjustment procedure. Without that information, secondary p-values should be regarded as unadjusted evidence for their individual comparisons, not as results that account for all the outcomes considered. This matters particularly when interpreting an isolated secondary result alongside other outcomes.

The primary endpoint's missing-data handling is partly described. Participants who did not die or were lost to follow-up were censored at the last evaluation or follow-up date, and follow-up was for a maximum of 5 years from randomization. The registry does not report additional details about censoring patterns, reasons for loss to follow-up, or sensitivity analyses assessing how departures from assumptions might affect the results. Those details are relevant because censoring methods rely on assumptions about the information available after censoring.

The overall survival population is described as all randomized participants on an intent-to-treat basis. The quality-of-life outcomes have smaller posted arm counts than the overall survival outcome. The registry does not report a full participant flow for these assessments, reasons for missing quality-of-life data, or the method used to handle missing assessments. Numerical results for time to disease progression and confirmed tumor responses are not reported in the registry record. It is therefore not possible to assess the completeness of those outcome analyses from the posted information.

The registry names the Kaplan-Meier method and Log Rank test for overall survival and Fisher Exact tests for the quality-of-life comparisons. The full statistical analysis plan, the sample-size justification, any interim analyses, and a broader prespecified approach to multiplicity are not reported in the registry record. These omissions limit assessment of whether all analysis choices followed plans established before results were known. They do not change the reported results, but they constrain how comprehensively those results can be evaluated.

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 0% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: dyspnea (2.2%), arrythmia (1.1%), edema (1.1%), supraventricular arrhythmias (svt/atrial fibrillation/flutter) (1.1%), nausea (1.1%).

The posted serious adverse-event figures list 90 participants at risk in the carboxyamidotriazole group, of whom 9 were affected, and 92 at risk in the placebo group, of whom 6 were affected. The registry also lists other adverse events for 90 participants at risk in the carboxyamidotriazole group, with 87 affected, and 92 at risk in the placebo group, with 84 affected. These are counts within the reported safety groups and should not be confused with the overall survival population counts.

The registry lists dyspnea at 2.2% among the top serious terms. Arrythmia, edema, supraventricular arrhythmias (SVT/atrial fibrillation/flutter), nausea, and stomatitis are each listed at 1.1%. These term-specific percentages do not provide a complete account of event timing, duration, severity, or whether events were judged related to treatment. Deaths in the adverse-event data are not reported in the registry record.

These figures describe reported safety events; they do not establish comparative risk on their own. The record's listed secondary outcome of severe non-hematologic adverse events is distinct from the overall survival endpoint, and the available summary does not provide all details needed to evaluate event ascertainment or attribution.

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

The primary result is a comparison of overall survival in randomized participants assigned to daily oral carboxyamidotriazole or daily oral placebo. The registry reports arm-specific Kaplan-Meier median estimates and a Log Rank test. The test result does not provide strong statistical evidence of a difference in survival experience, while the interval estimates show uncertainty around each median. The available summaries do not establish that the groups are equivalent or rule out every potentially important difference.

The result applies to the enrolled population with stage IIIA, stage IIIB, or stage IV non-small cell lung cancer in this trial context. It does not automatically generalize to other stages, other populations, different treatment settings, or different regimens. The registry describes the trial after standard chemotherapy, but does not give enough protocol detail to characterize all prior treatments or participant characteristics.

The secondary quality-of-life findings concern clinically significant decreases over the period from baseline to week 8. Their interpretation is limited by smaller outcome-specific participant counts, uncertainty, and the lack of a reported multiplicity strategy. The registry's adverse-event figures are descriptive safety data. Numerical results for time to disease progression and confirmed tumor responses are not reported in the registry record. These limitations mean the secondary outcomes cannot be used to fill gaps in the primary analysis.

Statistical results describe evidence within a defined analysis and do not, by themselves, determine clinical importance. A p-value does not state the probability that a treatment is effective or ineffective. A confidence interval describes uncertainty under a method; it does not eliminate uncertainty. The registry record supports a careful account of what was measured and reported, but leaves important details about missing data, analysis planning, and some secondary outcomes unavailable. No treatment recommendation follows from these statistical summaries alone.

14. Why This Trial Matters Statistically

This trial is statistically informative because it uses random assignment and a placebo comparison to examine a time-to-event primary endpoint in advanced non-small cell lung cancer. Randomization provides the framework for comparing outcomes between groups, while the double-blind design is intended to reduce the influence of knowledge of assignment. The registry does not describe the specific allocation or masking procedures, so their implementation cannot be assessed from the record.

The example also illustrates why an endpoint's summary and its comparison test are not interchangeable. Arm-specific medians offer an accessible description of survival time, but the Log Rank test uses the survival experience across follow-up. Interpreting both the estimates and their uncertainty avoids reducing a time-to-event result to a single p-value or to a comparison of medians alone.

It also shows the difference between primary and secondary evidence. Overall survival is identified as the primary endpoint, while the quality-of-life and other listed outcomes address additional questions. Secondary outcomes may be useful, but unadjusted comparisons across multiple outcomes require caution. The incomplete numerical reporting for some secondary measures and limited detail about missing observations further restrict the strength of conclusions that can be drawn from those results.

Finally, planning calculations based on observed secondary effects are not evidence of the trial's original design rationale or of observed power. Keeping design assumptions, analysis methods, estimates, and uncertainty distinct is important when translating a registry record into a statistical interpretation. The available details support a bounded reading of the primary result while making clear where the record does not permit a fuller audit.

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 registry record does not report a hazard ratio or a confidence interval for the relative time-to-event effect. The primary results therefore offer medians by arm and a Log Rank p-value, but not a direct estimate of the relative event rate over follow-up. The survival curves and their detailed features are not available in the record summary, so changes in the pattern of events during follow-up cannot be evaluated from the medians alone.

Although the primary analysis is described as intent-to-treat, the registry gives limited information about participant flow, follow-up completeness, and censoring patterns. The quality-of-life measures have smaller analysis counts than overall survival, and reasons for missing assessments are not reported. The numerical results for time to disease progression and confirmed tumor responses are not reported in the registry record. These gaps limit comparisons across outcomes and assessment of potential missing-data effects.

The record does not report the full statistical analysis plan, sample-size justification, interim analyses, or a multiplicity strategy. It also lacks details about allocation concealment, masking procedures, adherence, and the extent of prior treatment beyond the stated context after standard chemotherapy. These are trial-specific gaps; they should not be filled by assumptions based on general practices.

The posted serious adverse-event counts and term percentages do not include all information needed to evaluate duration, timing, attribution, or the complete safety profile. In addition, the secondary quality-of-life calculations are unadjusted and should not be treated as confirmatory evidence. Together, these limitations call for a narrow interpretation centered on the reported primary analysis rather than a broad claim about every possible outcome.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Carboxyamidotriazole in Treating Patients With Stage III or Stage IV Non-small Cell Lung Cancer — NCT00003869.
• Registry record last updated 2014-05-20; status last verified 2012-12.
• Results first posted 2010-12-22.
• 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.

19. Record Summary

NCT00003869 is a completed randomized, double-blind, parallel-group phase III trial sponsored by the National Cancer Institute. It enrolled 186 participants with stage IIIA, stage IIIB, or stage IV non-small cell lung cancer and compared daily oral carboxyamidotriazole with daily oral placebo. The start date was 1999-04 and the primary completion date was 2008-05.

The primary endpoint was overall survival, measured from randomization to death from any cause and analyzed by the Kaplan-Meier method. The registry reports medians and 95% confidence intervals by arm and a Log Rank comparison. The posted test result does not provide strong statistical evidence of a difference between groups. Several secondary outcomes are listed, but numerical details are incomplete in the registry record, and the broader statistical plan and missing-data approach are not reported.