This page separates information reported in the ClinicalTrials.gov record from statistical interpretation. The registry reports that results have been posted, but the ClinicalTrials.gov record does not contain formal statistical-analysis estimates, confidence intervals, or p-values for the registered primary endpoint. Accordingly, this page does not reconstruct or infer an unreported treatment effect.
1. Trial at a Glance
NSABP B-36 was a randomized, parallel-group, phase 3 treatment trial comparing two combination chemotherapy regimens in women with breast cancer. The ClinicalTrials.gov record identifies disease-free survival at 9 years as the single registered primary endpoint.
| Feature | NSABP B-36 |
|---|---|
| Trial name | NSABP B-36 |
| Brief title | Comparison of Two Combination Chemotherapy Regimens in Treating Women With Breast Cancer |
| Phase | Phase 3 |
| Condition | Breast Cancer |
| Allocation | Randomized |
| Design model | Parallel |
| Masking | None |
| Primary purpose | Treatment |
| Enrollment | 2722 |
| Arms | 2 |
| Primary endpoints registered | 1 |
| Primary endpoint type | Time-to-event |
| Results posted | Yes |
| Outcome measures posted | 9 |
| Lead sponsor | NSABP Foundation Inc |
| Sponsor type | Network |
2. Clinical Question
The ClinicalTrials.gov record describes a phase 3 randomized comparison of two combination chemotherapy regimens in women with breast cancer. The primary statistical question is framed through disease-free survival over a 9-year time frame.
Population
Women with breast cancer enrolled in the NSABP B-36 phase 3 treatment trial.
Interventions
The registry lists cyclophosphamide, adriamycin, epirubicin, and fluorouracil among the trial interventions.
Comparator structure
The trial used two randomized parallel treatment arms. The ClinicalTrials.gov record identifies the regimen components but do not provide a more detailed protocol description of dosing or sequencing.
Primary question
How does disease-free survival differ between the two randomized chemotherapy regimens over the registered 9-year time frame?
3. Trial Design
Adriamycin + Cyclophosphamide
- Adriamycin
- Cyclophosphamide
- Serious adverse events: 12/1351 affected/at risk
Fluorouracil + Epirubicin + Cyclophosphamide
- Fluorouracil
- Epirubicin
- Cyclophosphamide
- Serious adverse events: 26/1337 affected/at risk
The registry identifies the overall design as randomized, parallel, unmasked, and intended for treatment. The ClinicalTrials.gov record does not specify the randomization ratio, stratification variables, block structure, allocation concealment procedures, or detailed treatment schedule.
4. Endpoints
| Endpoint | Time frame | Registry definition | Endpoint type |
|---|---|---|---|
| Disease Free Survival | 9 years | Percentage of patients free from DFS event. DFS events include local, regional, or distant recurrence, second primary cancer or death from any cause prior to recurrence or second primary cancer | Time-to-event |
Disease-free survival is therefore a composite time-to-event endpoint. An event can occur through recurrence at a local, regional, or distant site, development of a second primary cancer, or death from any cause before recurrence or a second primary cancer.
The wording is important statistically: the endpoint is not simply the percentage of patients who remain recurrence-free. Death is included as an event under the registered definition, and the definition also includes second primary cancer.
5. Statistical Methodology
Why disease-free survival is a time-to-event endpoint
The primary endpoint is explicitly classified as time-to-event. Rather than recording only whether an event occurred, a time-to-event analysis uses the follow-up time available for each participant. This is important because participants may have different lengths of observation, and some participants may reach the end of available follow-up without experiencing a DFS event.
Typical analysis of disease-free survival
For a randomized trial with disease-free survival as the primary endpoint, the standard analytical framework would ordinarily include Kaplan-Meier estimation of the disease-free survival function, a comparison between randomized treatment groups using a log-rank test, and estimation of a treatment-effect measure such as a hazard ratio using a Cox proportional-hazards model.
Here, di represents the number of observed events at time ti, while ni represents the number of participants at risk immediately before that event time.
Why censoring matters
A participant who has not experienced a DFS event by the last observation does not simply disappear from the analysis. Under a standard right-censoring framework, that participant contributes information through the time at which follow-up ends. The validity of the resulting estimator depends on assumptions about the censoring mechanism and the way the endpoint is defined.
Why the randomized comparison is important
Randomization is the principal design feature that creates comparable treatment groups in expectation with respect to both measured and unmeasured baseline characteristics. For the primary efficacy question, analysis according to randomized assignment is generally the natural framework for preserving the benefit of randomization.
What the ClinicalTrials.gov record does not establish
The ClinicalTrials.gov record does not specify the actual statistical model used for the posted results. They do not identify a hazard ratio, confidence interval, p-value, Kaplan-Meier estimate, log-rank result, Cox-model specification, censoring rules, stratification variables, or analysis population. Those details therefore cannot be represented as observed NSABP B-36 analysis results here.
6. Planned Analysis
Registered primary endpoint
Time frame: 9 years
The registry defines a DFS event as local, regional, or distant recurrence, second primary cancer, or death from any cause prior to recurrence or second primary cancer.
ClinicalTrials.gov indicates that results have been posted for the trial and that 9 outcome measures are posted. However, the ClinicalTrials.gov recordset contains no formal statistical analyses for the registered primary endpoint. There is therefore no reported hazard ratio, confidence interval, p-value, event count, or treatment-group DFS estimate available in the ClinicalTrials.gov record.
How this endpoint would normally be analyzed
For a randomized disease-free-survival endpoint, a conventional analysis would estimate DFS curves with the Kaplan-Meier method and compare randomized groups with a log-rank test. A Cox proportional-hazards model could then provide a hazard ratio with a confidence interval. If the analysis were stratified, the corresponding stratification factors would be incorporated into the comparison and model.
| Analytical component | Typical role for DFS | Status in reported NSABP B-36 data |
|---|---|---|
| Kaplan-Meier estimation | Estimate the probability of remaining free from a DFS event over time | Not in the posted results |
| Log-rank test | Compare DFS distributions between randomized groups | Not in the posted results |
| Cox proportional-hazards model | Estimate a relative hazard measure between treatment groups | Not in the posted results |
| Hazard ratio | Summarize the relative event rate under a proportional-hazards model | Not reported |
| 95% confidence interval | Quantify statistical uncertainty around an estimated treatment effect | Not reported |
| P-value | Assess evidence against a specified null hypothesis under the chosen test | Not reported |
7. Results Reporting Status
The registry information posted on ClinicalTrials.gov for NSABP B-36 states Results Posted: Yes and identifies 9 posted outcome measures. At the same time, no formal statistical analyses were posted.
What is available
The trial has posted results and identifies disease-free survival as the registered primary endpoint with a 9-year time frame.
What is not available here
No formal statistical-analysis estimate, confidence interval, or p-value for the primary endpoint is contained in the ClinicalTrials.gov record.
What should not be inferred
Why the distinction matters
Descriptive registry information and formal inferential statistics answer different questions and should not be silently substituted for one another.
8. Safety Results
The ClinicalTrials.gov record reports serious adverse events by randomized arm using affected/at-risk counts. These are the only safety results provided in the trial data.
| Treatment arm | Serious adverse events affected | At risk | Reported affected/at risk |
|---|---|---|---|
| Adriamycin + Cyclophosphamide | 12 | 1351 | 12/1351 |
| Fluorouracil + Epirubicin + Cyclophospha | 26 | 1337 | 26/1337 |
The registry wording posted on ClinicalTrials.gov for the second arm abbreviates the regimen as “Fluorouracil + Epirubicin + Cyclophospha”; that wording is retained here rather than silently changing the reported label.
These serious-adverse-event counts should not be treated as an efficacy result. They describe a safety outcome and are distinct from the disease-free-survival endpoint. The ClinicalTrials.gov record also do not provide a formal statistical comparison of serious adverse events, definitions of the safety-analysis population beyond the reported at-risk denominators, exposure duration, or confidence intervals.
9. Statistical Methods Explained
Why is disease-free survival analyzed as time-to-event?
Because participants can experience a DFS event at different times, a simple proportion at a single point does not use all of the available follow-up information. A time-to-event analysis incorporates both the timing of events and the follow-up contributed by participants who have not yet experienced an event.
What counts as a DFS event in this trial?
The registered definition includes local, regional, or distant recurrence; a second primary cancer; or death from any cause prior to recurrence or second primary cancer. This makes DFS a composite endpoint rather than a measure limited to recurrence alone.
Why would Kaplan-Meier estimation be useful?
Kaplan-Meier estimation provides a way to estimate the probability of remaining free from the defined DFS event over time while accommodating right-censored observations. It produces a curve that can show when events accumulate rather than reducing the entire follow-up experience to a single percentage.
Why would a log-rank test be used?
A log-rank test is commonly used to compare two time-to-event distributions in randomized trials. It evaluates whether the observed pattern of event times differs between groups under the test's assumptions and weighting structure.
What would a Cox hazard ratio mean?
A Cox-model hazard ratio would describe the relative instantaneous event rate between the two treatment groups under the fitted model. A value below 1 would indicate a lower estimated event hazard in the numerator treatment group, while a value above 1 would indicate a higher estimated event hazard. It would not mean that the same percentage of patients experienced or avoided an event.
Why does the proportional-hazards assumption matter?
A conventional Cox hazard ratio is most straightforward to interpret as a common relative hazard when the proportional-hazards assumption is reasonable. If the relative hazard changes substantially over time, one summary hazard ratio may obscure important features of the survival curves. The registry-reported NSABP B-36 data do not report whether proportional hazards were assessed or what model diagnostics were performed.
Why can't the safety counts establish the efficacy result?
Serious-adverse-event counts and disease-free-survival events represent different clinical outcomes. Even though both arise from the same randomized trial, safety counts cannot be used to calculate or infer a DFS hazard ratio, DFS p-value, or disease-free-survival curve.
10. Randomization and Analysis Populations
The trial allocation is explicitly identified as randomized, with a parallel design and two arms. Randomization is central to the causal interpretation of the treatment comparison because treatment assignment is determined by the trial design rather than by a participant's subsequent outcome.
| Design feature | Supplied information | Statistical significance |
|---|---|---|
| Allocation | Randomized | Supports a treatment-group comparison based on randomized assignment |
| Design model | Parallel | Participants are assigned to one of two treatment pathways rather than sequentially receiving both randomized regimens |
| Masking | None | The trial was not masked according to the ClinicalTrials.gov record |
| Primary purpose | Treatment | The trial is designed to evaluate treatment rather than prevention, diagnostic performance, or another primary purpose |
| Arms | 2 | Defines the primary randomized comparison |
The ClinicalTrials.gov record does not state whether the primary analysis used intention-to-treat, per-protocol, or another analysis population. Consequently, this page does not assign a specific analysis population to the reported DFS endpoint.
11. Censoring and Composite Endpoint Structure
Two features of the registered DFS definition are especially important statistically: the endpoint is time-to-event, and the event itself is composite.
Time-to-event component
Participants contribute follow-up until a qualifying DFS event or the end of their usable observation under the prespecified analysis rules.
Composite component
Local, regional, and distant recurrence, second primary cancer, and death from any cause are all included in the registered event definition.
A composite endpoint can increase the number of observed events and therefore potentially provide more information for a time-to-event comparison. But the interpretation of the endpoint depends on the clinical meaning and frequency of its individual components. The ClinicalTrials.gov record does not provide component-specific event counts, so this page does not attribute the DFS endpoint to any particular component.
12. Interim Analysis, Multiplicity, and Other Design Features
The ClinicalTrials.gov record does not report an interim-analysis plan, alpha-spending method, multiplicity strategy, non-inferiority margin, Bayesian analysis, crossover procedure, missing-data or imputation method, or prespecified stratification variables.
| Design topic | What the ClinicalTrials.gov record establishes |
|---|---|
| Non-inferiority margin | Not reported in the ClinicalTrials.gov record |
| Crossover | Not reported in the ClinicalTrials.gov record |
| Factorial design | Not reported; the registry-reported design model is parallel with 2 arms |
| Multiplicity | Not reported in the ClinicalTrials.gov record |
| Interim analysis | Not reported in the ClinicalTrials.gov record |
| Missing-data / imputation method | Not reported in the ClinicalTrials.gov record |
| Stratification | Not reported in the ClinicalTrials.gov record |
| Bayesian methods | Not reported in the ClinicalTrials.gov record |
13. Interpreting a Disease-Free Survival Analysis
A hazard ratio would summarize the relative rate of experiencing a qualifying DFS event between the randomized treatment groups under the fitted survival model. For example, an HR below 1 would correspond to a lower estimated event hazard in the treatment group defined as the numerator of the comparison.
A hazard ratio would not directly state the percentage of patients who are disease-free at 9 years, the percentage cured, or the absolute reduction in DFS events. Those are different estimands and require corresponding absolute-risk or survival estimates.
A confidence interval would describe statistical uncertainty around the estimated treatment effect under the specified model and sampling framework. A narrow interval generally indicates greater precision than a wide interval, but the interval does not describe the range of effects experienced by individual patients.
A p-value would quantify the compatibility of the observed data with a specified null hypothesis under the chosen test. It would not measure how large or clinically important a treatment effect is. Effect size and uncertainty require the estimated treatment effect and its confidence interval.
14. Limitations
- No formal primary statistical estimate reported: the trial data identify disease-free survival as the registered primary endpoint, but no formal statistical analyses were posted.
- No confidence interval or p-value: the ClinicalTrials.gov record does not provide inferential statistics for the primary endpoint.
- No event counts for DFS: the number of disease-free-survival events by treatment arm is not reported.
- No survival estimates: the ClinicalTrials.gov record does not provide Kaplan-Meier estimates, median DFS, or a 9-year DFS percentage.
- No subgroup results: the ClinicalTrials.gov record contains no subgroup treatment-effect estimates, so subgroup interpretation is not possible from the provided record.
- No detailed statistical-analysis plan: stratification, model specification, multiplicity, interim monitoring, censoring rules, and missing-data procedures are not reported.
- Safety information is limited: only serious adverse-event affected/at-risk counts by arm are provided, without a formal comparative analysis.
- Registry and publication information are not interchangeable: the registry states that results are posted, but the absence of formal statistical-analysis entries in the ClinicalTrials.gov record prevents reconstruction of a complete inferential results section.
15. Why This Trial Matters Statistically
NSABP B-36 is a useful teaching example because the registry information combines a straightforward randomized phase 3 structure with a clinically meaningful composite time-to-event endpoint. The statistical challenge is not simply to compare two percentages. The primary endpoint requires attention to event timing, censoring, the definition of an event, and the distinction between descriptive results and formal inference.
| Concept | How it appears in NSABP B-36 |
|---|---|
| Randomization | The trial is identified as randomized. |
| Parallel-group design | The design model is parallel with 2 arms. |
| Time-to-event analysis | Disease-free survival is the registered primary endpoint and is classified as time-to-event. |
| Composite endpoint | DFS includes recurrence, second primary cancer, and death under the registered definition. |
| Kaplan-Meier estimation | A standard framework for estimating disease-free survival over time. |
| Log-rank comparison | A standard method for comparing randomized time-to-event distributions. |
| Cox model | A standard framework for estimating a hazard ratio when appropriate. |
| Censoring | Important because follow-up can end before a qualifying event. |
| Analysis population | The ClinicalTrials.gov record does not specify the actual primary efficacy analysis population. |
| Safety analysis | Serious adverse events are reported by arm using affected/at-risk counts. |
16. Related Tutorials
Learn more about the methods used in this trial:
17. Related Calculators
18. Sources
- ClinicalTrials.gov: NCT00087178 — NSABP B-36.
- Linked publication: PubMed record — PMID 35230585.
- Linked publication: PubMed record — PMID 35112166.
Continue through the Clinical Biostats statistical pathway
Explore the underlying concepts of randomization, disease-free survival, Kaplan-Meier estimation, hazard ratios, Cox models, and clinical-trial time-to-event analysis.
19. Record Summary
NSABP B-36 is a completed randomized phase 3 trial with 2722 enrolled participants, 2 parallel treatment arms, and disease-free survival registered as the single primary endpoint over 9 years. The endpoint is explicitly defined as the percentage of patients free from a DFS event, where events include local, regional, or distant recurrence, second primary cancer, or death from any cause prior to recurrence or second primary cancer.
The statistical structure is therefore centered on time-to-event methodology. Kaplan-Meier estimation, log-rank comparison, and Cox proportional-hazards modeling are natural analytical frameworks for this endpoint, but the ClinicalTrials.gov record does not identify which formal statistical procedures were actually used or provide the resulting hazard ratio, confidence interval, p-value, event counts, or DFS estimates. The registry does report that results are posted, but that fact should not be confused with the presence of a formal statistical-analysis record in the ClinicalTrials.gov recordset.