This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record. Numerical trial results on this page are restricted to the ClinicalTrials.gov record.
1. Trial at a Glance
FinHER is registered as a completed diagnostic study in women with invasive breast cancer. The registry describes a single-group design with no masking and an enrollment of 1010 participants. Tumor material was newly assessed with MammaTyper™ and the primary time-to-event endpoint evaluated 5-year distant disease-free survival according to the resulting molecular subgrouping.
| Feature | FinHER |
|---|---|
| ClinicalTrials.gov identifier | NCT02244580 |
| Brief title | Re-Evaluation of Tumor Samples From Women With Breast Cancer With the in Vitro Diagnostic Kit "MammaTyper™" |
| Condition | Invasive Breast Cancer |
| Primary purpose | Diagnostic |
| Design model | Single group |
| Allocation | NA |
| Masking | None |
| Enrollment | 1010 |
| Intervention | MammaTyper™ (device) |
| Lead sponsor | BioNTech Diagnostics GmbH |
| Sponsor type | Industry |
| Status | Completed |
2. Clinical Question
The registry describes a diagnostic re-evaluation question rather than a conventional randomized treatment-comparison question. Tumor material from women with breast cancer was newly assessed using MammaTyper™, and 5-year distant disease-free survival was calculated according to the resulting subgrouping.
Population
Women with breast cancer; the registered condition is invasive breast cancer.
Intervention
MammaTyper™ (device), used to newly assess tumor material.
Comparator
The primary analysis compares the Luminal A subgroup with the combined subgroup consisting of Luminal B, HER2 positive, and triple negative tumors.
Primary question
How does 5-year distant disease-free survival differ between the Luminal A subgroup and the combined Luminal B, HER2 positive, and triple negative subgroup based on MammaTyper™ subtyping?
3. Trial Design
The ClinicalTrials.gov record does not describe a two-arm randomized treatment intervention. Instead, the analytic comparison arises from molecular subgroup classification after tumor assessment. That distinction is important because the hazard ratio is comparing outcome patterns between classified subgroups rather than estimating the effect of assigning patients to two randomized treatments.
4. Endpoints
| Endpoint | Registry definition / description | Time frame | Analysis |
|---|---|---|---|
| 5 Year Distant Disease Free Survival (DDFS) | Rate of patients without distant metastases in subgroup Luminal A vs. combined subgroup (Luminal B, HER2 positive, triple negative), based on subtyping with MammaTyper™. Tumor material will be newly assessed by MammaTyper™ and 5 year DDFS will be calculated new according to new subgrouping. | 5 year from the date of patient randomisation | Cox proportional-hazards model; hazard ratio |
| Number of Patients With High Ki-67 and Prognosis on Outcome for DDFS and OS | Number of patients with high Ki-67 and prognosis on outcome for DDFS and OS, measured by hazard ratio. | 5 years | Cox proportional-hazards model; hazard ratio |
The registry lists three posted outcome measures and two posted statistical analyses. The ClinicalTrials.gov record identifies one primary analysis and one secondary analysis.
5. Statistical Methodology
Primary analysis: Cox proportional-hazards model
The primary endpoint is a time-to-event endpoint, and the registry reports a Cox regression analysis. The effect measure is a hazard ratio comparing the Luminal A subgroup with the combined Luminal B, HER2 positive, and triple negative subgroup.
The registry reports the method as "Regression, Cox" and normalizes this to a Cox proportional-hazards model. The reported hypothesis type is superiority.
What the model contributes
A Cox model relates the instantaneous event rate to subgroup classification while accounting for differing follow-up times and censoring. This is appropriate to the registered time-to-event framework because not every participant necessarily has to experience the event before analysis.
The important distinction in FinHER is that the model is not presented as estimating the causal effect of a randomized intervention. The subgroup itself is generated by the MammaTyper™ assessment, and the registry describes the design as single group. Thus the reported HR describes an association between subgroup classification and the time-to-event outcome under the specified model.
Hazard ratio as the effect measure
An HR below 1 indicates a lower estimated instantaneous event rate in the Luminal A subgroup relative to the combined subgroup under the fitted Cox model.
The ClinicalTrials.gov record reports a hazard ratio of 0.423. Descriptively, this corresponds to an estimated hazard approximately 42.3% as large as that of the comparison subgroup under the model, or approximately a 57.7% lower estimated instantaneous hazard. That arithmetic interpretation does not mean that 57.7% of patients avoided distant metastases, nor does it represent an absolute difference in 5-year DDFS.
Confidence interval
The primary analysis reports a two-sided 95% confidence interval of 0.246 to 0.726. A confidence interval communicates statistical precision around the estimated hazard ratio under the model and sampling framework. It does not describe the range of outcomes that individual patients will experience.
P-value
The primary analysis reports a two-sided P-value of 0.0018. In the context of the registered superiority hypothesis, this quantifies the compatibility of the observed analysis with the null hypothesis represented by the statistical test. It is not a measure of the magnitude of the hazard ratio and does not tell us how clinically important an association is.
6. Primary Endpoint Result
The primary endpoint is 5-year distant disease-free survival, assessed as the rate of patients without distant metastases in the Luminal A subgroup versus the combined Luminal B, HER2 positive, and triple negative subgroup, based on MammaTyper™ subtyping.
Primary hazard ratio
95% CI: 0.246–0.726 · P = 0.0018
Two-sided Cox proportional-hazards analysis; superiority hypothesis.
| Primary endpoint component | Reported value |
|---|---|
| Endpoint | 5 Year Distant Disease Free Survival (DDFS) |
| Comparison | Luminal A vs Combined Subtype |
| Time frame | 5 year from the date of patient randomisation |
| Endpoint type | Time-to-event |
| Effect measure | Hazard Ratio (HR) |
| HR | 0.423 |
| 95% CI | 0.246–0.726 |
| P-value | 0.0018 |
| Hypothesis type | Superiority |
What the estimate means: An HR of 0.423 means that, under the fitted Cox model, the estimated instantaneous hazard of the time-to-event outcome for the Luminal A subgroup was approximately 42.3% of that in the combined comparison subgroup. Equivalently, the estimated hazard was approximately 57.7% lower.
What it does not mean: It does not mean that 57.7% more patients were free of distant metastases at five years, that 57.7% of patients benefited, or that an individual patient's probability of an event was reduced by exactly 57.7%.
What the confidence interval says: The two-sided 95% CI of 0.246–0.726 describes uncertainty around the estimated hazard ratio. The interval is entirely below 1, but its width also shows that the exact magnitude of the association is not known with certainty.
Why the P-value is different: The P-value of 0.0018 addresses the statistical test under the registered superiority framework. It does not measure effect size. The HR and its confidence interval are needed to describe the estimated magnitude and its precision.
Important caution: The registry profile describes FinHER as a single-group diagnostic study rather than a randomized treatment comparison. Therefore, the hazard ratio should not automatically be described as a treatment-effect estimate. The association may reflect differences between the molecularly defined subgroups and other factors related to prognosis.
7. Secondary Endpoint Result
The registry also reports a secondary time-to-event analysis concerning patients with high Ki-67 and prognosis for DDFS and OS, measured by hazard ratio.
Secondary hazard ratio
95% CI: 0.25–0.71 · P = 0.001
Two-sided Cox proportional-hazards analysis; superiority hypothesis.
| Secondary endpoint component | Reported value |
|---|---|
| Outcome measure | Number of Patients With High Ki-67 and Prognosis on Outcome for DDFS and OS (Measured by Hazard Ratio) |
| Time frame | 5 years |
| Groups compared | Patients With MKI67 mRNA Determination |
| Endpoint type | Time-to-event |
| Method | Cox proportional-hazards model |
| Effect measure | Hazard Ratio (HR) |
| HR | 0.42 |
| 95% CI | 0.25–0.71 |
| P-value | 0.001 |
| Hypothesis type | Superiority |
What the estimate means: The reported HR of 0.42 indicates a lower estimated instantaneous hazard in the comparison represented by the analysis under the fitted Cox model.
What it does not mean: The HR is not a percentage of patients with high Ki-67, is not a direct five-year risk difference, and should not be interpreted as a causal treatment effect from the ClinicalTrials.gov record.
Precision: The two-sided 95% CI is 0.25–0.71. This interval quantifies uncertainty around the estimated HR rather than uncertainty about an individual patient's outcome.
P-value: The reported P = 0.001 is evidence from the specified statistical test, not a measure of the magnitude or clinical importance of the association.
Interpretive caution: The registry wording identifies the analysis as concerning high Ki-67 and prognosis for DDFS and OS. The ClinicalTrials.gov record does not provide additional details about covariate adjustment, the precise threshold defining high Ki-67, or the number of participants included in this analysis. Those details should therefore not be inferred.
8. Statistical Methods Explained
Why was a Cox proportional-hazards model used?
The registered primary endpoint is a time-to-event endpoint measured over five years. Cox regression is designed for this setting because it models the instantaneous event rate while allowing participants to contribute information for different lengths of follow-up and to be censored before an event occurs.
What does an HR of 0.423 mean?
Under the fitted model, the estimated instantaneous hazard for the Luminal A subgroup is approximately 42.3% of that for the combined comparison subgroup. It is a relative time-to-event measure, not a probability and not an absolute difference in five-year disease-free survival.
Why is the confidence interval important?
The estimate 0.423 is only one estimate from the study. The 95% confidence interval of 0.246–0.726 communicates the uncertainty surrounding that estimate. A reader should consider the interval alongside the point estimate rather than focusing on the point estimate alone.
Why does the P-value not measure effect size?
The P-value answers a hypothesis-testing question. It depends on both the magnitude of the observed effect and the amount of statistical information available. Two analyses can have the same P-value while having different effect sizes, and a large effect estimate can have an uninformative P-value when information is limited.
What is the proportional-hazards assumption?
A standard Cox proportional-hazards interpretation assumes that the relative hazard represented by the hazard ratio is reasonably stable over the relevant follow-up. The ClinicalTrials.gov record reports the Cox model but do not provide an assessment of whether the proportional-hazards assumption was checked. That assessment therefore cannot be added here.
Why is this not automatically a randomized treatment effect?
the ClinicalTrials.gov record posted on ClinicalTrials.gov for FinHER specifies a single-group design, NA allocation, no masking, and diagnostic primary purpose. The primary analysis compares molecular subgroups generated through MammaTyper™ classification. Without a randomized assignment between those subgroups, the observed hazard ratio should be interpreted as an association between subgroup classification and outcome rather than automatically as a causal treatment effect.
9. Reading the Primary Result Carefully
Relative effect
HR 0.423 is a model-based relative measure of the instantaneous event rate between the two molecularly defined groups.
Absolute outcome
The registry describes the primary endpoint as the rate of patients without distant metastases at five years, but the ClinicalTrials.gov record does not provide the corresponding subgroup-specific percentages.
Uncertainty
The 95% CI of 0.246–0.726 shows uncertainty around the estimated hazard ratio and should accompany the point estimate.
Statistical evidence
The two-sided P-value of 0.0018 is the reported hypothesis-test result and should not be substituted for an effect estimate or confidence interval.
This distinction is particularly important for time-to-event outcomes. A hazard ratio summarizes relative event dynamics over follow-up; it does not directly provide the probability of remaining event-free at a specified time. A five-year survival or disease-free survival percentage answers an absolute time-point question, while the hazard ratio answers a relative hazard question.
10. What the Registry Does Not Report in the Supplied Data
The registry-reported FinHER record contains enough information to describe the registered endpoint and its two posted Cox analyses, but it does not provide several details that are often included in a full clinical-trial statistical report.
| Potential analysis detail | Supplied information |
|---|---|
| Subgroup-specific 5-year DDFS percentages | Not provided in the ClinicalTrials.gov record. |
| Median DDFS | Not provided. |
| Number of DDFS events | Not provided. |
| Kaplan-Meier estimates | Not provided. |
| Baseline characteristics | Not provided. |
| Serious adverse events by arm | Not applicable to the registry-reported single-group design, and no safety results are provided. |
| Randomization scheme | the ClinicalTrials.gov record specifies allocation as NA and design model as SINGLE_GROUP. |
| Interim analysis | Not provided. |
| Multiplicity procedure | Not provided. |
| Missing-data or imputation method | Not provided. |
| Bayesian methods | Not reported. |
| Stratification factors | Not provided. |
These omissions matter because a hazard ratio alone does not reconstruct the full statistical analysis. In particular, the ClinicalTrials.gov record does not identify whether additional covariates were included in the Cox model, how missing observations were handled, or whether any prespecified multiplicity adjustment was used.
11. Limitations and Interpretation Issues
- Single-group design: the ClinicalTrials.gov record identifies FinHER as a single-group diagnostic study. The molecular subgroup comparison therefore does not have the same causal interpretation as a randomized treatment comparison.
- Registry endpoint wording: the primary time frame says "5 year from the date of patient randomisation," even though the registry-reported design fields specify single group and allocation NA. This page preserves both pieces of registry information rather than resolving the discrepancy by inference.
- Hazard-ratio interpretation: the HR is a relative model-based measure and should not be interpreted as a five-year absolute risk difference.
- Proportional-hazards assumption: the ClinicalTrials.gov record identifies a Cox model but do not report an assessment of the proportional-hazards assumption.
- Incomplete numerical reporting: the ClinicalTrials.gov record gives HRs, confidence intervals, and P-values but does not give subgroup-specific five-year DDFS percentages or event counts.
- Unreported model details: the ClinicalTrials.gov record does not specify covariate adjustment, stratification, missing-data handling, or other model implementation details.
- Secondary endpoint interpretation: the secondary analysis has its own reported HR, confidence interval, and P-value, but the ClinicalTrials.gov record does not describe a multiplicity strategy connecting the secondary analysis to the primary hypothesis.
- Prognostic versus predictive interpretation: the registry describes prognosis on DDFS and OS in relation to high Ki-67. The ClinicalTrials.gov record does not establish that Ki-67 predicts differential benefit from a treatment.
12. Why This Trial Matters Statistically
FinHER is useful as a statistical teaching case because it illustrates how a time-to-event analysis can be used in a diagnostic and molecular-classification setting rather than a conventional randomized treatment comparison. The most important lesson is to interpret the statistical method together with the study design.
| Concept | How it appears in FinHER |
|---|---|
| Time-to-event endpoint | 5-year distant disease-free survival is the registered primary endpoint. |
| Cox proportional-hazards model | The registry reports "Regression, Cox" for the primary and secondary analyses. |
| Hazard ratio | The primary analysis reports HR 0.423; the secondary analysis reports HR 0.42. |
| Confidence interval | The primary 95% CI is 0.246–0.726; the secondary 95% CI is 0.25–0.71. |
| P-value | The primary P-value is 0.0018; the secondary P-value is 0.001. |
| Subgroup classification | The primary comparison is Luminal A versus the combined Luminal B, HER2 positive, and triple negative subgroup. |
| Molecular assessment | Tumor material is newly assessed by MammaTyper™. |
| Diagnostic purpose | The registry identifies the primary purpose as DIAGNOSTIC. |
| Single-group design | the ClinicalTrials.gov record identifies the design model as SINGLE_GROUP and allocation as NA. |
| Interpretation of association | The molecular subgroup comparison should not automatically be interpreted as a randomized causal treatment effect. |
13. Related Tutorials
Learn more about the methods used in this trial:
14. Related Calculators
15. Sources
- ClinicalTrials.gov: FinHER (NCT02244580), the official registry record used as the source for the trial design, endpoints, statistical analyses, estimates, confidence intervals, and P-values presented on this page.
Continue through the Clinical Biostats statistical pathway
Use the related tutorials and calculators to explore the time-to-event methods represented in this trial, including hazard ratios, Cox models, confidence intervals, and survival analysis.
16. Record Summary
FinHER provides a focused example of time-to-event statistical analysis in a diagnostic molecular-classification setting. The ClinicalTrials.gov record describes a completed single-group study of 1010 participants with invasive breast cancer in which tumor material was newly assessed using MammaTyper™. The primary endpoint was 5-year distant disease-free survival, comparing the Luminal A subgroup with a combined Luminal B, HER2 positive, and triple negative subgroup.
The primary Cox proportional-hazards analysis reported a hazard ratio of 0.423, with a two-sided 95% confidence interval of 0.246–0.726 and a two-sided P-value of 0.0018. A secondary Cox analysis concerning high Ki-67 and prognosis on DDFS and OS reported an HR of 0.42, a 95% CI of 0.25–0.71, and a P-value of 0.001.
The central statistical lesson is that these numbers must be interpreted in the context of the design. A Cox hazard ratio can quantify the relative event rate between two observed subgroups, but the presence of a hazard ratio does not by itself make an analysis randomized or causal. For FinHER, the registry design fields identify a single-group diagnostic study, making careful distinction between association, prognostic classification, and treatment effect especially important.