This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record.
1. Trial at a Glance
MINDACT is a completed randomized phase 3 trial sponsored by the European Organisation for Research and Treatment of Cancer (EORTC). The registry describes a parallel, unmasked study involving women with breast cancer that involves no more than 3 lymph nodes, with chemotherapy, endocrine treatment, and genetic or clinical assessment incorporated into the trial design.
| Feature | MINDACT |
|---|---|
| Study acronym | MINDACT |
| Phase | Phase 3 |
| Condition | Breast Cancer |
| Population | Women with breast cancer that involves no more than 3 lymph nodes |
| Allocation | Randomized |
| Design model | Parallel |
| Masking | None |
| Primary purpose | Treatment |
| Enrollment | 6600 |
| Number of arms | 6 |
| Lead sponsor | European Organisation for Research and Treatment of Cancer - EORTC |
| Sponsor type | Network |
| Start date | 2007-02 |
| Primary completion date | 2020-03 |
| Status | Completed |
2. Clinical Question
The central statistical question is how treatment decisions based on clinical assessment and genetic testing relate to chemotherapy use and subsequent time-to-event outcomes in women with breast cancer involving no more than 3 lymph nodes.
Population
Women with breast cancer that involves no more than 3 lymph nodes.
Interventions
The registry lists anthracycline-based therapy, docetaxel and capecitabine, tamoxifen, and letrozole among the study interventions.
Comparative structure
The registry identifies chemotherapy randomization (R-C), endocrine randomization (R-E), and a chemotherapy-versus-no-chemotherapy comparison described as R-T.
Primary questions
The registered primary endpoints evaluate 5-year distant metastasis-free survival for the chemotherapy-versus-no-chemotherapy comparison and disease-free survival for both the chemotherapy and endocrine randomizations.
3. Trial Design
MINDACT is registered as a randomized, parallel-group, unmasked phase 3 study. The record describes six study arms and an enrollment of 6600 participants. The primary purpose is treatment.
Registered intervention framework
Systemic chemotherapy
- Anthracycline-based therapy
- Docetaxel and capecitabine
Endocrine therapy
- Tamoxifen
- Letrozole
The intervention list establishes the therapies registered for the study, but the record summarized here does not provide arm-level sample sizes or a complete mapping of every intervention to each of the 6 arms. Those details therefore are not used to construct additional treatment-group estimates.
4. Randomization and Comparative Questions
The primary endpoints reveal three distinct randomized comparisons or decision structures. Keeping these questions separate is important because the endpoint associated with one randomization should not automatically be interpreted as the endpoint for another.
| Randomization / comparison | Registered primary endpoint | Time frame |
|---|---|---|
| R-T: chemotherapy versus no chemotherapy | Distant metastasis-free survival at 5 years | From enrollment/randomization |
| R-C: chemotherapy randomization | Disease-free survival (DFS) | From enrollment/randomization |
| R-E: endocrine randomization | DFS | From enrollment/randomization |
This structure matters statistically because a single trial can contain several related randomized questions without those questions being interchangeable. Each comparison has its own estimand, analysis population, treatment contrast, and interpretation.
5. Primary Endpoints
| Endpoint | Registry definition / description | Time frame |
|---|---|---|
| Distant metastasis-free survival at 5 years | The primary endpoint for chemotherapy versus no chemotherapy (R-T) is that the 5-year distant metastasis free survival (DMFS) of 92% will be tested using a one-sided test. | From enrollment/randomization |
| Disease-free survival (DFS) | The primary endpoint for the chemotherapy randomization (R-C) is disease free survival (DFS). | From enrollment/randomization |
| DFS | The primary test for the endocrine randomization (R-E) is DFS. | From enrollment/randomization |
6. Statistical Methodology
Time-to-event analysis
DMFS and DFS are time-to-event endpoints. In a typical analysis, the time origin would be enrollment or randomization as specified in the registry, and each participant would contribute follow-up until the event of interest or an appropriate censoring time.
Here, S(t) represents the probability of remaining event-free beyond time t. A time-to-event analysis uses the ordering and timing of events rather than reducing every participant to a simple event/no-event indicator.
Kaplan-Meier estimation
For a time-to-event endpoint such as DMFS or DFS, the Kaplan-Meier estimator is a standard nonparametric method for estimating the event-free survival curve while accommodating right-censored follow-up.
At each observed event time, di is the number of events and ni is the number of participants at risk immediately before that time.
Log-rank comparison
A randomized comparison of two or more time-to-event groups is commonly evaluated using a log-rank test. The test compares the observed number of events with the number expected under a null hypothesis of equal survival experience between the groups.
Cox proportional-hazards model
A Cox proportional-hazards model is a common way to estimate a hazard ratio for a time-to-event comparison. If used for the registered endpoints, the hazard ratio would summarize the relative instantaneous event rate between randomized groups under the model.
An HR below 1 indicates a lower estimated hazard in the numerator group at the times represented by the model. It is not the same quantity as a risk ratio, an absolute risk difference, or a probability of individual benefit.
One-sided testing for the R-T endpoint
The registry explicitly states that the 5-year DMFS target of 92% for chemotherapy versus no chemotherapy (R-T) will be tested using a one-sided test. A one-sided test places the rejection region in the direction specified by the alternative hypothesis.
The important statistical implication is that the direction of the hypothesis is part of the prespecified design. A one-sided test should not be interpreted as simply a smaller version of a two-sided p-value; it represents a different hypothesis-testing framework.
7. Planned Analysis
No formal statistical analyses were posted to ClinicalTrials.gov. The registry nevertheless specifies the principal time-to-event outcomes and, for the R-T comparison, the 5-year DMFS target and one-sided testing framework.
R-T: 5-year distant metastasis-free survival
The registered primary endpoint is distant metastasis-free survival at 5 years, measured from enrollment/randomization. The registry states that the 5-year DMFS of 92% will be tested using a one-sided test.
Prespecified statistical target
5-year distant metastasis-free survival
One-sided test for chemotherapy versus no chemotherapy (R-T)
The natural analysis of a 5-year time-to-event endpoint would use the survival function at the specified time point, with censoring handled through a survival-analysis framework. The key estimand is the probability of remaining free of the defined distant-metastasis event through 5 years.
R-C: Disease-free survival
The primary endpoint for the chemotherapy randomization is DFS from enrollment/randomization. A standard analysis would estimate DFS over time and compare the randomized groups using a time-to-event method such as a Kaplan-Meier analysis with an appropriate between-group test and, where prespecified, a Cox model for a hazard ratio.
R-E: Disease-free survival
The primary test for the endocrine randomization is also DFS from enrollment/randomization. The same general time-to-event framework would normally be appropriate, while preserving the specific randomized comparison defined by R-E.
8. Statistical Methods Explained
Why is distant metastasis-free survival a time-to-event endpoint?
Because the outcome concerns whether and when a distant metastasis event occurs during follow-up. Participants who remain free of the event when follow-up ends do not simply become equivalent to participants who experienced the event; their partial follow-up still contributes information. Survival-analysis methods preserve that timing information.
Why is the 5-year point important for R-T?
The registry defines the R-T primary endpoint specifically as distant metastasis-free survival at 5 years. This means the statistical question is tied to the event-free probability at a prespecified time point rather than only to an unspecified comparison of entire survival curves.
What does the 92% target represent?
The registry states that a 5-year DMFS of 92% will be tested for the chemotherapy-versus-no-chemotherapy comparison. Statistically, this is a prespecified benchmark for the survival probability at 5 years. It is not itself a hazard ratio and should not be converted into one without additional data and a defined model.
Why use a one-sided test?
The registry explicitly specifies a one-sided test for the R-T endpoint. A one-sided hypothesis test asks whether the result is sufficiently consistent with an alternative in one prespecified direction. The direction and testing framework need to be established before examining the outcome data.
Why would Kaplan-Meier analysis be useful?
Kaplan-Meier estimation provides an estimate of the probability of remaining event-free over time while accommodating censoring. It also allows the analyst to examine whether the pattern of event-free survival differs between randomized groups throughout follow-up rather than only at one time point.
What would a Cox hazard ratio add?
A Cox model would provide a relative measure of the event hazard between randomized groups. It is useful for summarizing a time-to-event comparison, but it does not directly state the absolute probability of being event-free at 5 years. The proportional-hazards assumption also matters when a single hazard ratio is used to summarize follow-up.
9. Confidence Intervals and P-values in the Planned Analysis
No numerical confidence intervals or p-values are posted for the registered primary endpoints in the ClinicalTrials.gov record.
Confidence intervals
For an estimated survival probability, hazard ratio, or other effect measure, a confidence interval quantifies statistical uncertainty around the estimate under the chosen analysis framework. It does not describe the range of outcomes for individual patients.
P-values
A p-value measures the compatibility of the observed data with a specified null hypothesis under the statistical model and testing procedure. It does not measure the size or clinical importance of an effect.
For the R-T endpoint, the interpretation of any eventual p-value would also depend on the prespecified one-sided testing framework. For the DFS endpoints, the appropriate interpretation would depend on the statistical testing strategy specified for the respective randomization.
10. Analysis Population and Censoring
The registry record summarized here identifies randomized allocation but does not provide the detailed analysis-population definitions, censoring rules, or missing-data procedures needed to reproduce a formal survival analysis.
| Statistical component | What is reported | Why it matters |
|---|---|---|
| Randomization | Randomized allocation | Provides the basis for unbiased comparison of randomized groups under the trial design. |
| Time origin | From enrollment/randomization | Defines when follow-up begins for the registered primary endpoints. |
| DMFS endpoint | 5-year distant metastasis-free survival for R-T | Defines the primary time-specific estimand for the R-T comparison. |
| DFS endpoint | DFS for R-C and R-E | Defines the primary time-to-event outcome for the two additional randomized questions. |
| Censoring rules | Not reported in the record summarized here | Censoring assumptions can affect Kaplan-Meier and Cox estimates. |
| Missing-data procedures | Not reported in the record summarized here | Handling incomplete follow-up can influence time-to-event analyses. |
| Analysis populations | Detailed definitions not reported in the record summarized here | The population analyzed determines how the randomized comparison is interpreted. |
11. Multiplicity and Multiple Primary Questions
MINDACT contains three registered primary endpoint statements: one for the R-T chemotherapy-versus-no-chemotherapy comparison and two DFS endpoints corresponding to R-C and R-E. These multiple questions create an important statistical distinction between the existence of several primary analyses and the error-control strategy used across them.
The record summarized here does not report a multiplicity-adjustment procedure, alpha allocation across the three primary questions, or a hierarchical testing strategy. Therefore, the three endpoint statements should be treated as three registered primary questions rather than assuming a particular familywise-error procedure.
| Primary question | Endpoint | Statistical issue |
|---|---|---|
| R-T | 5-year DMFS | One-sided test against the registered 92% target |
| R-C | DFS | Time-to-event comparison following chemotherapy randomization |
| R-E | DFS | Time-to-event comparison following endocrine randomization |
12. Non-Inferiority, Equivalence, and Superiority Logic
The registry description for the R-T primary endpoint states that a 5-year DMFS of 92% will be tested using a one-sided test. That language defines a directional hypothesis-testing framework, but the record summarized here does not provide a conventional non-inferiority margin or an equivalence margin.
This distinction is important. A one-sided test is a feature of hypothesis testing; a non-inferiority trial additionally requires a prespecified margin defining how much loss of efficacy would still be considered acceptable. Without a reported non-inferiority margin, the R-T endpoint should not automatically be described as a non-inferiority analysis.
13. Interim Analysis and Adaptive Features
The registry record summarized here does not report an interim-analysis schedule, stopping boundaries, alpha-spending method, Lan-DeMets procedure, O'Brien-Fleming boundary, or other formal sequential-monitoring method.
Similarly, the record does not provide sufficient statistical detail to characterize the trial as having a particular adaptive design beyond its randomized, parallel structure and the multiple randomized questions described for R-T, R-C, and R-E.
14. Safety and Adverse Events
The record summarized here lists the interventions but does not provide arm-specific serious adverse-event counts, adverse-event rates, treatment discontinuations, or other numerical safety results.
| Safety component | Registry reporting in the record summarized here |
|---|---|
| Serious adverse events by arm | Not reported |
| Adverse-event rates by arm | Not reported |
| Treatment discontinuations due to adverse events | Not reported |
| Safety analysis population | Not reported |
For a randomized clinical trial, safety is generally summarized separately from the primary efficacy estimands. The absence of numerical safety results in the registry record means that no arm-level safety comparison can be characterized from this record.
15. What the Primary Endpoints Do — and Do Not — Measure
The R-T endpoint asks about remaining free of distant metastasis through the specified 5-year time point. It is a time-to-event outcome, so both whether the event occurs and when it occurs are relevant to the analysis.
It does not by itself measure overall survival, quality of life, treatment toxicity, or every form of disease recurrence.
DFS is the registered primary endpoint for both the chemotherapy randomization and the endocrine randomization. It is likewise a time-to-event endpoint, with follow-up beginning at enrollment/randomization according to the registry.
DFS should not automatically be treated as interchangeable with DMFS. They are distinct endpoint definitions and correspond to different registered statistical questions in this trial.
The 92% value is a prespecified 5-year DMFS target in the R-T description. It is a survival-probability benchmark, not an observed treatment effect and not a hazard ratio.
16. Statistical Interpretation of the Trial Structure
MINDACT is statistically interesting because the study is not represented by a single treatment-versus-control question. Instead, the registry identifies multiple randomized questions that connect clinical assessment, genetic testing, chemotherapy, and endocrine treatment with different primary time-to-event endpoints.
Separate estimands
R-T, R-C, and R-E have different primary questions. Their endpoint estimates should therefore be interpreted in the context of the corresponding randomization.
Time-to-event information
Both DMFS and DFS incorporate the timing of events and censoring rather than reducing follow-up to a simple binary outcome.
Directional testing
The R-T description explicitly specifies a one-sided test, making the prespecified direction part of the inferential framework.
Multiple questions
Three primary endpoint statements require careful separation of the individual hypotheses and whatever multiplicity strategy was prespecified.
17. Limitations
- No posted statistical analyses: the ClinicalTrials.gov record does not provide numerical estimates, confidence intervals, p-values, or event counts for the registered primary endpoints.
- Limited arm-level detail: the record identifies 6 arms and lists four intervention types, but the summarized registry data do not provide arm-specific sample sizes or a complete intervention-to-arm mapping.
- Limited analysis-plan detail: the record does not report the complete statistical analysis plan, including detailed censoring rules, analysis populations, model specifications, or missing-data procedures.
- Multiplicity information: the record identifies three primary endpoint statements but does not report how type I error was allocated across the corresponding hypotheses.
- No subgroup estimates: the record summarized here does not provide numerical subgroup treatment effects or interaction tests.
- No safety estimates: arm-specific serious adverse-event and adverse-event results are not reported in the record summarized here.
- No outcome reconstruction: without event-level or sufficient summary results, a formal survival estimate, hazard ratio, confidence interval, or p-value cannot be established from the registry record alone.
18. Why This Trial Matters Statistically
MINDACT is a useful teaching case because it illustrates how a single randomized clinical trial can contain several linked but distinct statistical questions. It also demonstrates why the definition of an endpoint, the randomization that generates the comparison, and the hypothesis-testing framework must be considered together.
| Concept | How it appears in MINDACT |
|---|---|
| Randomization | The study is registered as randomized and contains distinct chemotherapy, endocrine, and chemotherapy-versus-no-chemotherapy questions. |
| Parallel design | The registered design model is parallel. |
| Time-to-event analysis | DMFS and DFS are the registered primary endpoints. |
| Fixed time-point survival | R-T specifically uses 5-year DMFS. |
| One-sided testing | The R-T endpoint description explicitly specifies a one-sided test. |
| Multiple primary questions | Three primary endpoint statements correspond to R-T, R-C, and R-E. |
| Kaplan-Meier estimation | A standard framework for estimating DMFS and DFS over time. |
| Cox regression | A standard model for estimating relative time-to-event effects when its assumptions are appropriate. |
| Censoring | Relevant to any analysis of DMFS or DFS because follow-up can end before an event is observed. |
| Hypothesis testing | The R-T endpoint provides an explicit example of a prespecified one-sided test. |
| Multiplicity | Multiple registered primary questions require attention to the overall testing framework. |
19. Record Timeline
Study start
The registry lists 2007-02 as the study start date.
Randomized clinical trial
The study is registered as a phase 3 randomized, parallel-group trial with no masking and a primary purpose of treatment.
Registered enrollment
The registry lists enrollment of 6600 and 6 study arms.
Primary completion
The registry lists 2020-03 as the primary completion date.
Current registered status
The trial is listed as completed.
20. Clinical Biostats Interpretation
The most important feature of MINDACT is the separation of the clinical question from the statistical question. The clinical question concerns whether chemotherapy or endocrine treatment should be used within a genetic-testing and clinical-assessment framework. The statistical questions are defined by the specific randomized comparison and its prespecified endpoint.
DMFS at 5 years and DFS are both time-to-event outcomes, but they are not interchangeable. The R-T comparison has an explicit 5-year target of 92% and a one-sided test, while the R-C and R-E primary endpoints are DFS from enrollment/randomization.
A trial registration can define an important statistical question without containing the numerical evidence needed to answer it. Estimates, confidence intervals, event counts, and p-values are necessary for evaluating the magnitude and precision of an observed treatment effect. They are not reported by the registry record summarized here.
21. Sources
- ClinicalTrials.gov: MINDACT — NCT00433589.
Continue through the Clinical Biostats trial-analysis library
Explore additional clinical trial analyses and statistical methods across the Clinical Biostats site.
22. Record Summary
MINDACT is a completed phase 3 randomized parallel-group trial with 6600 participants and 6 registered arms, evaluating questions involving chemotherapy, endocrine treatment, genetic testing, and clinical assessment in women with breast cancer involving no more than 3 lymph nodes. Its registered primary endpoints are 5-year distant metastasis-free survival for the chemotherapy-versus-no-chemotherapy comparison and DFS for the chemotherapy and endocrine randomizations.
From a statistical perspective, the trial is particularly useful for understanding how a complex randomized design can contain several distinct estimands. The R-T endpoint specifies a 5-year DMFS target of 92% and a one-sided test, while the R-C and R-E questions use DFS from enrollment/randomization. Appropriate analysis of these endpoints would ordinarily use time-to-event methods that account for censoring, with the exact inferential framework determined by the prespecified statistical analysis plan.