This page separates reported trial results from statistical interpretation. Every numerical result presented here comes from the ClinicalTrials.gov record. The registry record is the official source for the trial's registered design and posted analyses.
1. Trial at a Glance
RxPONDER is a randomized, parallel-group, phase 3 treatment trial evaluating systemic chemotherapy with endocrine therapy versus endocrine therapy alone in patients represented by the registry's invasive breast carcinoma and related breast-cancer conditions. The principal registered endpoint is invasive disease-free survival (IDFS), a time-to-event endpoint assessed 5 years after randomization.
| Feature | RxPONDER |
|---|---|
| Phase | Phase 3 |
| Conditions | Breast Ductal Carcinoma In Situ; Invasive Breast Carcinoma; Multicentric Breast Carcinoma; Multifocal Breast Carcinoma; Synchronous Bilateral Breast Carcinoma |
| Design | Randomized, parallel-group |
| Masking | None |
| Primary purpose | Treatment |
| Primary endpoint | Invasive Disease-Free Survival (IDFS) |
| Primary endpoint time frame | 5 years after randomization |
| Enrollment | 5018 |
| Lead sponsor | National Cancer Institute (NCI) |
| Sponsor type | NIH |
| ClinicalTrials.gov | NCT01272037 |
2. Clinical Question
The registered trial compares Chemo and Endocrine Therapy with Endocrine Therapy Alone. The principal statistical question is whether the addition of systemic chemotherapy is associated with a different invasive disease-free survival experience over the registered 5-year time frame, and whether that association differs according to menopausal status.
Population
The registry lists Breast Ductal Carcinoma In Situ, Invasive Breast Carcinoma, Multicentric Breast Carcinoma, Multifocal Breast Carcinoma, and Synchronous Bilateral Breast Carcinoma among the trial conditions.
Intervention
Systemic chemotherapy with endocrine therapy. The registered interventions include systemic chemotherapy and endocrine therapies including tamoxifen citrate, letrozole, anastrozole, and exemestane.
Comparator
Endocrine Therapy Alone, with endocrine therapies represented among the registered interventions.
Primary question
How does randomized treatment assignment relate to invasive disease-free survival, and does the treatment effect depend on menopausal status?
3. Trial Design
Chemo and Endocrine Therapy
- Systemic chemotherapy
- Endocrine therapy
- Registered intervention list includes tamoxifen citrate, letrozole, anastrozole, and exemestane
Endocrine Therapy Alone
- Endocrine therapy
- Registered intervention list includes tamoxifen citrate, letrozole, anastrozole, and exemestane
4. Randomization and Analysis Structure
The allocation field is RANDOMIZED, and the design model is PARALLEL. The posted analyses use Cox regression for the comparison of the randomized treatment groups. The registry analysis notes also identify stratified analysis and covariate adjustment as concepts used in specific analyses.
| Analysis feature | Registry-supported information |
|---|---|
| Allocation | Randomized |
| Design model | Parallel |
| Masking | None |
| Primary endpoint type | Time-to-event |
| Primary statistical method | Cox proportional-hazards model |
| Effect measure | Hazard ratio |
| Analysis concepts identified in posted analyses | Stratified analysis; covariate adjustment |
| Hypothesis types in posted analyses | Other / not stated; Superiority |
5. Primary Endpoint
| Endpoint | Registered definition | Time frame |
|---|---|---|
| Invasive Disease-Free Survival (IDFS) | From date of randomization (2nd Registration) to date of first invasive recurrence (local, regional or distant), second invasive primary cancer (breast or not), or death due to any cause. Patients last known to be alive who have not experienced recurrence or second primary cancer are censored at their last contact date. This is the STEEP definition of invasive disease-free survival. | 5 years after randomization |
This definition combines several possible first events: invasive recurrence, a second invasive primary cancer, or death from any cause. It also explicitly incorporates right censoring for participants who remain alive without recurrence or a second primary cancer at their last contact.
6. Statistical Methodology
Kaplan-Meier estimation
Because IDFS is a time-to-event endpoint, the natural descriptive framework is the Kaplan-Meier estimator. It accounts for participants whose event status is not observed through the complete follow-up period by incorporating their information until censoring.
where di is the number of events at time ti and ni is the number at risk immediately before that time.
Cox proportional-hazards model
The registry reports Cox regression as the method for the posted primary and post-hoc analyses. The resulting effect measure is a hazard ratio (HR).
The hazard ratio compares the modeled instantaneous event rates associated with two treatment groups, conditional on the variables included in the model.
For a treatment comparison, an HR below 1 indicates a lower estimated instantaneous event rate in the first-listed treatment group relative to the comparator, while an HR above 1 indicates a higher estimated instantaneous event rate. This is a relative time-to-event measure; it is not an absolute difference in the proportion of participants experiencing an event.
Covariate adjustment
The registry notes that the premenopausal and postmenopausal IDFS analyses used Cox models with adjustment for the continuous recurrence score. Covariate adjustment can improve precision and account for an important measured prognostic variable while preserving the treatment-group comparison defined by randomization.
Stratified analysis
The posted interaction analysis involving menopausal status identifies stratified analysis as an analysis concept. Stratification is useful when an important factor is expected to influence the baseline event process and the analysis seeks to respect that structure rather than treating all participants as though they had identical underlying hazards.
Right censoring
The registered IDFS definition explicitly states that participants who are last known to be alive without recurrence or a second primary cancer are censored at their last contact date. Censoring therefore forms part of the statistical definition of the endpoint, rather than representing an event itself.
7. Primary IDFS Results: Overall Treatment-by-Recurrence-Score Interaction
The first posted primary analysis evaluates the interaction between treatment arm and the recurrence score in the overall study population. This is an interaction analysis, not simply the overall treatment hazard ratio. Its purpose was to determine whether chemotherapy benefit depended on the recurrence score.
Treatment × recurrence-score interaction
95% CI: 0.98–1.05 · P = 0.35
Cox regression; two-sided 95% confidence interval
The reported interaction estimate of 1.02 concerns whether the association between treatment assignment and IDFS changes with the recurrence score. It should therefore not be read as saying that the chemotherapy-plus-endocrine group had an HR of 1.02 for IDFS overall.
The 95% CI of 0.98–1.05 quantifies uncertainty around the estimated interaction parameter. Because the interval is close to 1, the posted analysis does not provide strong evidence, under this model and test, that the treatment effect changes according to the continuous recurrence score.
The P = 0.35 value addresses the statistical evidence for the interaction hypothesis; it does not measure the magnitude of a treatment effect. A p-value should not be interpreted as the probability that the treatment has no effect, nor does it replace the confidence interval as a measure of precision.
Importantly, an interaction analysis and a treatment-effect analysis answer different questions. A treatment HR describes the relative event rate between treatment groups, whereas an interaction term asks whether that treatment contrast varies with another variable.
8. Primary IDFS Results: Treatment × Menopausal-Status Interaction
A second primary analysis evaluated the interaction between treatment arm and menopausal status, identified in the registry as one of the study stratification factors. The purpose was to determine whether chemotherapy benefit depended on menopausal status.
Treatment × menopausal-status interaction
95% CI: 1.15–2.54 · P = 0.008
Cox regression; two-sided 95% confidence interval
The 1.71 estimate is an interaction result. It should not be interpreted as the hazard ratio for chemotherapy plus endocrine therapy versus endocrine therapy alone in all participants, nor as the hazard ratio for either menopausal group by itself.
The 95% CI of 1.15–2.54 describes uncertainty around the interaction estimate. Unlike a treatment HR, the relevant null value for an interaction on the hazard-ratio scale is also 1: values away from 1 indicate a difference in the treatment contrast across levels of the interacting variable.
The reported P = 0.008 is evidence against the null interaction hypothesis under the posted analysis. It does not quantify how large the clinical treatment benefit is in either menopausal group, and it does not mean that there is a 0.8% probability that the null hypothesis is true.
The important next step is therefore to examine the prespecified or posted treatment comparisons within the relevant menopausal groups. Those analyses are reported below.
9. Primary IDFS Result: Premenopausal Participants
The registry reports a separate IDFS analysis restricted to premenopausal participants. The Cox model compared Chemo and Endocrine Therapy with Endocrine Therapy Alone and adjusted for the continuous recurrence score.
Invasive disease-free survival
95% CI: 0.43–0.83 · P = 0.002
Superiority analysis; Cox model adjusted for continuous recurrence score
An HR of 0.60 means that, under the fitted Cox model, the estimated instantaneous rate of an IDFS event in the Chemo and Endocrine Therapy group was about 40% lower than in the Endocrine Therapy Alone group for the premenopausal analysis. This is a relative hazard interpretation, not a statement that 40% of participants avoided an event.
The HR does not say that every premenopausal participant experienced a 40% reduction in personal risk. It also does not give an absolute risk reduction, a number needed to treat, or a difference in 5-year IDFS percentages.
The 95% CI of 0.43–0.83 provides a range of values reflecting statistical uncertainty around the estimated hazard ratio under the specified model. Because the interval does not include 1, the estimate is statistically separated from the null value in this posted analysis.
The P = 0.002 value addresses evidence against the null treatment comparison under the stated superiority analysis. It does not measure the size or clinical importance of the effect. Interpretation also depends on the analysis population, censoring, model specification, and the proportional-hazards assumption underlying the Cox model.
10. Primary IDFS Result: Postmenopausal Participants
The registry also reports a separate IDFS analysis restricted to postmenopausal participants. As in the premenopausal analysis, the Cox model adjusted for the continuous recurrence score.
Invasive disease-free survival
95% CI: 0.82–1.26 · P = 0.89
Superiority analysis; Cox model adjusted for continuous recurrence score
An HR of 1.02 is close to the null value of 1. Under the fitted Cox model, the estimated instantaneous IDFS event rate in the Chemo and Endocrine Therapy group was approximately similar to that in the Endocrine Therapy Alone group in this postmenopausal analysis.
The HR does not mean that the two groups had identical outcomes at every time point. A single HR summarizes a model-based relative comparison over follow-up; it is not a complete description of the survival curves.
The 95% CI of 0.82–1.26 indicates uncertainty around the estimate and includes 1. Thus, the interval is compatible with both a lower and a higher hazard under the model. The P = 0.89 value likewise does not provide evidence against the null treatment comparison in this posted analysis.
A nonsignificant result should not be translated into proof that the treatment has exactly no effect. It means that this analysis did not produce sufficiently strong statistical evidence against the null hypothesis, given its data, model, and assumptions.
11. Reading the Primary IDFS Analyses Together
| Analysis | Role | HR | 95% CI | P-value |
|---|---|---|---|---|
| Overall treatment × recurrence-score interaction | Primary | 1.02 | 0.98–1.05 | 0.35 |
| Treatment × menopausal-status interaction | Primary | 1.71 | 1.15–2.54 | 0.008 |
| Premenopausal IDFS | Primary; superiority | 0.60 | 0.43–0.83 | 0.002 |
| Postmenopausal IDFS | Primary; superiority | 1.02 | 0.82–1.26 | 0.89 |
The statistical story is therefore more specific than simply saying that "the treatment worked" or "the treatment did not work." The registry reports an interaction between treatment and menopausal status, followed by separate Cox analyses in premenopausal and postmenopausal participants. The treatment HRs in those analyses differ substantially, while the corresponding confidence intervals quantify their uncertainty.
The recurrence-score interaction addresses a separate question. Its HR of 1.02 and P = 0.35 concern whether treatment effect varies with recurrence score, rather than representing the treatment effect at a particular recurrence-score value.
12. Post-Hoc Distant Relapse-Free Survival Results
The registry also posts two post-hoc analyses of Distant Relapse-Free Survival (DRFS) at 5 years after randomization. These analyses compare Chemo and Endocrine Therapy with Endocrine Therapy Alone and use Cox regression with adjustment for the continuous recurrence score.
Premenopausal participants
Distant relapse-free survival
95% CI: 0.39–0.87 · P = 0.009
Post-hoc superiority analysis; eligible and evaluable participants
An HR of 0.58 corresponds to an approximately 42% lower estimated instantaneous event rate in the Chemo and Endocrine Therapy group under the fitted model. This is a relative hazard interpretation and does not provide an absolute difference in the probability of distant relapse-free survival.
The 95% CI of 0.39–0.87 reflects uncertainty around the estimated hazard ratio. The P = 0.009 value addresses the statistical comparison under the posted analysis; it does not measure the magnitude of the effect.
Because this analysis is explicitly identified as post-hoc, its inferential role should be distinguished from the primary endpoint analyses. A post-hoc analysis can provide useful evidence and context, but it should not automatically be interpreted as though it were a separately prespecified confirmatory endpoint.
Postmenopausal participants
Distant relapse-free survival
95% CI: 0.81–1.37 · P = 0.70
Post-hoc superiority analysis; eligible and evaluable participants
An HR of 1.05 is close to 1, so the fitted model estimates a similar instantaneous DRFS event rate between the two randomized groups in this postmenopausal post-hoc analysis.
The 95% CI of 0.81–1.37 includes the null value of 1 and expresses uncertainty that extends in both directions. The P = 0.70 result does not provide statistical evidence against the null treatment comparison in this analysis.
Again, "not statistically significant" is not synonymous with "proven identical." The confidence interval is essential because it shows the range of treatment contrasts that remain compatible with the observed estimate under the model.
13. Safety Results
The ClinicalTrials.gov record reports serious adverse events by randomized arm as affected participants divided by participants at risk. These are the only arm-specific safety counts provided in the trial data.
| Safety measure | Chemo and Endocrine Therapy | Endocrine Therapy Alone |
|---|---|---|
| Serious adverse events, affected / at risk | 14 / 2062 | 9 / 2342 |
Chemo + endocrine
14 participants were affected among 2062 at risk according to the registry safety field.
Endocrine alone
9 participants were affected among 2342 at risk according to the registry safety field.
The reported safety counts should not be treated as a formal comparative hypothesis test because the ClinicalTrials.gov record does not provide a statistical analysis for this safety measure. In addition, the denominator is explicitly an "at risk" count in the registry field, so this page does not recalculate rates or construct a relative risk from the displayed numbers.
14. Statistical Methods Explained
Why was a Cox proportional-hazards model used?
IDFS and DRFS are time-to-event endpoints. A Cox model is suited to these data because it uses both event timing and censoring information while estimating a relative hazard between treatment groups. The model produces the hazard ratios reported throughout the registry analyses.
What does an HR of 0.60 mean in the premenopausal IDFS analysis?
It means that the fitted model estimates the instantaneous IDFS event rate to be about 40% lower in the Chemo and Endocrine Therapy group than in the Endocrine Therapy Alone group. It does not mean that 40% of participants benefited, nor does it give an absolute 5-year risk difference.
Why is the HR of 1.71 not the treatment effect?
The 1.71 estimate belongs to the treatment × menopausal-status interaction. An interaction asks whether treatment effects differ according to another variable. It is therefore different from the treatment HR estimated within the premenopausal or postmenopausal groups.
Why was recurrence score included as a covariate?
The registry explicitly states that the premenopausal and postmenopausal IDFS analyses adjusted for the continuous recurrence score. Including that measured variable in the Cox model allows the treatment comparison to be estimated conditional on recurrence score rather than leaving that prognostic information unused.
What does the confidence interval tell us?
A 95% confidence interval describes statistical uncertainty around the estimated parameter under the specified model and sampling framework. For the premenopausal IDFS HR of 0.60, the interval is 0.43–0.83. It does not describe the range of individual treatment effects among patients.
Why does the p-value not measure effect size?
The p-value quantifies how compatible the observed data are with a specified null hypothesis under the statistical model. It depends on both the magnitude of the observed effect and the amount of information in the data. Effect magnitude is better conveyed by the hazard ratio and its confidence interval.
Why does the registered IDFS definition include censoring?
Not every participant will necessarily experience an invasive recurrence, second invasive primary cancer, or death during observed follow-up. The endpoint definition therefore specifies that participants last known to be alive without those events are censored at their last contact date. This allows their observed follow-up to contribute information without incorrectly treating their last observation as an event.
15. Understanding the Hazard Ratio More Carefully
The hazard ratio is a relative measure. For example, an HR of 0.60 can be described as an approximately 40% lower estimated instantaneous event rate. It does not reveal the absolute probability of an IDFS event at a particular time unless accompanied by survival estimates or other absolute-risk information.
An instantaneous hazard is a rate-like quantity describing the occurrence of events at a given time among participants who remain event-free immediately before that time. It is not the same quantity as the probability that an individual participant will experience an event.
The Cox proportional-hazards model conventionally interprets the treatment contrast through a hazard ratio that is stable over the relevant model structure. If hazards are not approximately proportional over time, a single HR can become an incomplete summary of the underlying time-to-event distributions. The ClinicalTrials.gov record does not report a formal proportional-hazards diagnostic, so this page does not assume that such a diagnostic was performed or passed.
16. Interaction: Recurrence Score Versus Menopausal Status
RxPONDER's posted analyses illustrate two distinct interaction questions. One examines whether treatment benefit depends on the recurrence score. The other examines whether treatment benefit depends on menopausal status.
| Interaction question | HR | 95% CI | P-value | Interpretation target |
|---|---|---|---|---|
| Treatment × recurrence score | 1.02 | 0.98–1.05 | 0.35 | Whether the treatment contrast varies with recurrence score |
| Treatment × menopausal status | 1.71 | 1.15–2.54 | 0.008 | Whether the treatment contrast varies with menopausal status |
These are important distinctions in clinical-trial statistics. A treatment effect estimated in one subgroup is not itself an interaction test. Conversely, an interaction estimate is not the treatment effect in either subgroup. The interaction analysis provides evidence about whether treatment contrasts differ across levels of the modifying variable; the subgroup-specific Cox analyses describe those treatment contrasts.
17. Primary and Post-Hoc Evidence
| Endpoint / analysis | Registry role | Method | Effect measure |
|---|---|---|---|
| IDFS, treatment × recurrence-score interaction | Primary | Cox regression | HR |
| IDFS, treatment × menopausal-status interaction | Primary | Cox regression | HR |
| IDFS, premenopausal participants | Primary | Cox regression with recurrence-score adjustment | HR |
| IDFS, postmenopausal participants | Primary | Cox regression with recurrence-score adjustment | HR |
| DRFS, premenopausal participants | Post-hoc | Cox regression with recurrence-score adjustment | HR |
| DRFS, postmenopausal participants | Post-hoc | Cox regression with recurrence-score adjustment | HR |
This distinction matters because statistical evidence has different inferential roles depending on whether an analysis was designated as primary or post-hoc. The ClinicalTrials.gov record identifies the DRFS analyses as post-hoc, so they are presented as additional analyses rather than being merged into the primary IDFS evidence.
18. Multiplicity and Multiple Analyses
The ClinicalTrials.gov record shows six statistical analyses: four primary analyses of IDFS and two post-hoc analyses of DRFS. They include interaction testing as well as treatment comparisons within menopausal-status groups.
When several hypotheses are examined, the interpretation of individual p-values depends on the prespecified statistical testing strategy. The ClinicalTrials.gov record does not provide an alpha-allocation scheme, multiplicity-adjustment procedure, or hierarchical testing strategy. This page therefore does not assign a familywise error interpretation to the posted p-values beyond the hypothesis descriptions explicitly reported by the registry.
Primary analyses
Four posted analyses concern the registered primary endpoint, IDFS. They include two interaction analyses and two menopausal-status-specific treatment comparisons.
Post-hoc analyses
Two additional analyses concern DRFS and are explicitly labeled post-hoc in the ClinicalTrials.gov record.
19. What the Confidence Intervals Add
| Analysis | HR | 95% CI | What the interval contributes |
|---|---|---|---|
| Recurrence-score interaction | 1.02 | 0.98–1.05 | Shows uncertainty around the estimated treatment-by-score interaction. |
| Menopausal-status interaction | 1.71 | 1.15–2.54 | Shows uncertainty around the estimated treatment-by-menopausal-status interaction. |
| Premenopausal IDFS | 0.60 | 0.43–0.83 | Shows uncertainty around the estimated treatment hazard ratio. |
| Postmenopausal IDFS | 1.02 | 0.82–1.26 | Shows uncertainty around the estimated treatment hazard ratio. |
| Premenopausal DRFS | 0.58 | 0.39–0.87 | Shows uncertainty around the post-hoc treatment hazard ratio. |
| Postmenopausal DRFS | 1.05 | 0.81–1.37 | Shows uncertainty around the post-hoc treatment hazard ratio. |
The widths of these intervals also illustrate why the point estimate alone is insufficient. For example, the postmenopausal IDFS estimate is 1.02, but its confidence interval spans 0.82–1.26. The interval communicates substantially more information than the point estimate alone about the precision of the comparison.
20. Time-to-Event Endpoints and Censoring
IDFS is not simply a binary outcome assessed at exactly 5 years. Its registered definition starts at randomization and follows participants until the first qualifying event or censoring. The 5-year time frame specifies the period at which the endpoint is assessed, while the underlying statistical object is the time from randomization to the first qualifying event.
For RxPONDER IDFS, qualifying events include invasive recurrence, second invasive primary cancer, or death from any cause.
This structure explains why Kaplan-Meier estimation and Cox regression are appropriate tools for the registered endpoint. Both methods can incorporate different follow-up durations and right-censored observations rather than requiring every participant to have identical observation time.
21. Limitations and Interpretation Issues
- Open-label design: masking is listed as none. Knowledge of treatment assignment can affect some aspects of participant or investigator behavior and therefore should be considered when interpreting outcomes susceptible to assessment or management effects.
- Endpoint complexity: IDFS includes invasive recurrence, second invasive primary cancer, and death from any cause. These are clinically distinct events combined into a composite time-to-event endpoint.
- Interaction interpretation: the treatment-by-menopausal-status HR of 1.71 is an interaction estimate, not the treatment HR in either subgroup.
- Model dependence: the reported hazard ratios come from Cox regression and therefore depend on the fitted model and its assumptions.
- Proportional hazards: the ClinicalTrials.gov record does not report a formal assessment of the proportional-hazards assumption.
- Post-hoc analyses: DRFS analyses are explicitly identified as post-hoc and should be distinguished from the primary IDFS analyses.
- Multiplicity: the ClinicalTrials.gov record does not specify an alpha-allocation or multiplicity-adjustment strategy for the six posted analyses.
- Safety interpretation: only affected/at-risk counts for serious adverse events are reported; no formal comparative safety analysis is provided in the ClinicalTrials.gov record.
- Missing absolute estimates: the statistical analyses posted on ClinicalTrials.gov report hazard ratios and confidence intervals but do not provide 5-year IDFS or DRFS percentages. This page therefore does not infer absolute survival probabilities from the hazard ratios.
22. Why This Trial Matters Statistically
RxPONDER is a useful teaching example because it shows how a randomized clinical trial can move beyond a single overall treatment comparison. The statistical structure includes a registered time-to-event endpoint, Cox modeling, covariate adjustment, interaction testing, subgroup-specific analyses, and separate post-hoc analyses.
| Concept | How it appears in RxPONDER |
|---|---|
| Randomization | Participants were allocated randomly to two treatment arms. |
| Parallel-group design | The registry specifies a parallel design model with two arms. |
| Time-to-event endpoint | IDFS is the registered primary endpoint, assessed 5 years after randomization. |
| Kaplan-Meier estimation | Appropriate framework for describing the registered time-to-event endpoint with censoring. |
| Cox regression | Used for all six posted statistical analyses. |
| Hazard ratio | The reported effect measure for IDFS and DRFS analyses. |
| Covariate adjustment | Premenopausal and postmenopausal analyses adjusted for continuous recurrence score. |
| Interaction testing | Analyses examined treatment interaction with recurrence score and menopausal status. |
| Stratified analysis | Identified as an analysis concept in the treatment-by-menopausal-status analysis. |
| Post-hoc analysis | DRFS was analyzed post-hoc in premenopausal and postmenopausal participants. |
| Confidence intervals | All six posted statistical analyses provide two-sided 95% confidence intervals. |
| Safety by randomized arm | Serious adverse events are reported as affected participants divided by participants at risk. |
23. Statistical Concepts in This Trial
Learn more about the methods used in this trial:
24. Related Statistical Calculators
Apply the same statistical concepts with Clinical Biostats calculators:
25. Sources
- ClinicalTrials.gov: RxPONDER, NCT01272037.
- PubMed record: PMID 42458791.
- PubMed record: PMID 41379459.
- PubMed record: PMID 39656951.
- PubMed record: PMID 34914339.
Continue through the Clinical Biostats statistical library
Explore the underlying survival-analysis methods, confidence intervals, regression models, and clinical-trial statistical tools used to understand randomized time-to-event studies.
26. Record Summary
RxPONDER provides a detailed example of how a randomized phase 3 trial can use a time-to-event endpoint to investigate treatment effects and effect modification. The registered primary endpoint is invasive disease-free survival at 5 years after randomization, with the endpoint defined by invasive recurrence, second invasive primary cancer, or death, together with explicit censoring rules. The posted analyses use Cox regression and hazard ratios, with recurrence-score adjustment in the menopausal-status-specific analyses.
The primary analyses distinguish between treatment effects and interaction effects. The treatment-by-recurrence-score interaction was reported as HR 1.02 (95% CI 0.98–1.05; P = 0.35), while the treatment-by-menopausal-status interaction was HR 1.71 (95% CI 1.15–2.54; P = 0.008). The subgroup-specific IDFS analyses reported HR 0.60 (95% CI 0.43–0.83; P = 0.002) in premenopausal participants and HR 1.02 (95% CI 0.82–1.26; P = 0.89) in postmenopausal participants.
The two additional DRFS analyses were explicitly identified as post-hoc. They reported HR 0.58 (95% CI 0.39–0.87; P = 0.009) in premenopausal participants and HR 1.05 (95% CI 0.81–1.37; P = 0.70) in postmenopausal participants. These results should be interpreted in their stated post-hoc context rather than combined indiscriminately with the primary endpoint evidence.