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
PHEREXA was a randomized, parallel, phase 3 treatment trial in HER2-positive metastatic breast cancer. The registered primary endpoint was progression-free survival assessed independently by an Independent Review Facility, with the primary comparison analyzed using a two-sided stratified log-rank test and summarized with a hazard ratio.
| Feature | PHEREXA |
|---|---|
| Phase | Phase 3 |
| Condition | Breast Cancer |
| Population | Patients with HER2-positive metastatic breast cancer |
| Design | Randomized, parallel |
| Masking | None |
| Primary purpose | Treatment |
| Enrollment | 452.0 |
| Arms | 2 |
| Primary endpoint | Progression Free Survival (Independent Assessment) |
| Primary endpoint type | Time-to-event |
| Primary hypothesis | Superiority |
| Primary analysis | Two-sided stratified log-rank test with hazard ratio |
| ClinicalTrials.gov | NCT01026142 |
| Lead sponsor | Hoffmann-La Roche |
2. Clinical Question
The central statistical question was whether adding pertuzumab to capecitabine plus trastuzumab changed the distribution of independent-assessment progression-free survival compared with capecitabine plus trastuzumab alone in patients with HER2-positive metastatic breast cancer.
Population
Patients with HER2-positive metastatic breast cancer.
Intervention
Capecitabine + trastuzumab + pertuzumab.
Comparator
Capecitabine + trastuzumab.
Primary question
Do the two treatment groups have different survival distributions for independently assessed progression-free survival?
3. Trial Design
Capecitabine + Trastuzumab
- Capecitabine
- Trastuzumab
Capecitabine + Trastuzumab + Pertuzumab
- Capecitabine
- Trastuzumab
- Pertuzumab
4. Trial Timeline and Registry Status
Trial start
The registered start date for PHEREXA was 2010-01-26.
Primary completion
The registered primary completion date was 2015-05-29.
Registry status
The ClinicalTrials.gov record identifies the study as completed and reports that results were posted.
5. Primary Endpoint
| Endpoint | Registry definition / assessment | Analysis |
|---|---|---|
| Progression Free Survival (Independent Assessment) | Progression Free Survival (PFS) was defined as the time from randomization to first documented disease progression (PD), as determined by an Independent Review Facility (IRF) using Response Evaluation Criteria in Solid Tumors (RECIST) version 1.0, or death from any cause, whichever occurred first. | Stratified log-rank test; hazard ratio |
| Time frame | Tumor assessments every 9 weeks from randomization until Week 27, then every 12 weeks thereafter, until IRF-determined PD, initiation of alternative anticancer medication, or death (up to 5.5 years). | Time-to-event endpoint |
6. Results: Primary Progression-Free Survival Analysis
The registry contains one formal statistical analysis for the primary endpoint. The analysis population was all randomized participants. The comparison was between capecitabine plus trastuzumab and capecitabine plus trastuzumab plus pertuzumab.
Hazard ratio for independent-assessment PFS
95% CI: 0.65–1.02 · P = 0.0731
Two-sided, stratified log-rank analysis · Superiority hypothesis
| Analysis characteristic | Reported result |
|---|---|
| Endpoint | Progression Free Survival (Independent Assessment) |
| Analysis population | All randomized participants. |
| Group A | Capecitabine + Trastuzumab |
| Group B | Capecitabine + Trastuzumab + Pertuzumab |
| Method | Log Rank |
| Other analysis concept | Stratified analysis |
| Effect measure | Hazard Ratio (HR) |
| Estimate | 0.82 |
| Confidence interval | 95% two-sided CI: 0.65–1.02 |
| P-value | 0.0731 |
| Hypothesis type | Superiority |
The estimated hazard ratio of 0.82 means that the estimated instantaneous rate of the PFS event in the pertuzumab-containing group was approximately 82% of the corresponding rate in the capecitabine-plus-trastuzumab group, under the time-to-event comparison represented by the hazard ratio. Equivalently, an HR of 0.82 corresponds to an estimated 18% lower instantaneous event rate.
This is not the same as saying that 18% of patients had their risk reduced by 18%, that 18% more patients remained progression-free, or that the absolute difference in PFS probability was 18 percentage points. A hazard ratio is a relative time-to-event measure, not an absolute risk difference.
The 95% confidence interval of 0.65–1.02 expresses statistical uncertainty around the estimated hazard ratio. Because the interval extends above 1, the registry-reported confidence interval includes a hazard ratio corresponding to no difference in the event rate. The interval therefore matters as much as the point estimate when interpreting the precision of the result.
The P-value of 0.0731 is a measure of compatibility between the observed data and the stated null hypothesis under the specified testing framework. It does not measure the magnitude or clinical importance of the hazard ratio. A P-value should not be read as the probability that the null hypothesis is true, nor as the probability that the treatment has no effect.
The analysis was two-sided and based on a superiority hypothesis. The registry also identifies the analysis as stratified, but the ClinicalTrials.gov record does not identify the individual stratification factors. Those factors therefore should not be inferred here.
7. The Primary Hypothesis in Statistical Terms
The registry analysis notes define the primary hypothesis in terms of the survival distributions of independently assessed PFS. The null hypothesis is that the survival distributions are the same between the two treatment groups; the alternative hypothesis is that they are different.
H1: IRF PFSpertuzumab ≠ IRF PFScontrol
The registry-reported analysis notes specify a two-sided and stratified comparison. The formal reported method is the log-rank test, with the hazard ratio used as the effect measure.
This distinction is important because the hazard ratio and log-rank P-value answer related but different questions. The hazard ratio summarizes the relative event rate under the survival-analysis model, while the log-rank test evaluates whether the observed survival distributions differ under the specified hypothesis-testing framework.
8. Statistical Methodology
Time-to-event analysis
Progression-free survival is a time-to-event endpoint because the analysis concerns the time from randomization until the first qualifying event: disease progression or death, whichever occurs first. Unlike a simple binary outcome, the timing of the event contributes directly to the analysis.
The registry identifies PFS as a time-to-event endpoint measured in months and reports tumor assessments at scheduled intervals. This structure creates the possibility that the exact time of progression is not observed continuously; instead, progression is identified through scheduled tumor assessments. The ClinicalTrials.gov record does not provide additional information about censoring rules, missing assessments, or imputation, so those elements are not added to this analysis.
Kaplan-Meier estimation
Kaplan-Meier estimation is the standard descriptive framework for displaying and estimating a time-to-event survival distribution. It accounts for patients whose event has not yet occurred by the time of their last usable follow-up through right censoring.
Here, di is the number of events at event time ti, while ni is the number of participants at risk immediately before that event time.
The ClinicalTrials.gov record does not provide Kaplan-Meier event-time data or median PFS estimates. A Kaplan-Meier curve therefore cannot be reconstructed from the ClinicalTrials.gov record without introducing unsupported data.
Stratified log-rank test
The reported primary comparison used a stratified log-rank test. The ordinary log-rank test compares observed and expected event patterns between treatment groups over follow-up. A stratified version performs that comparison while accounting for prespecified strata.
The registry explicitly identifies the analysis as stratified, but the ClinicalTrials.gov record does not list the individual stratification variables. Consequently, this page describes the method without assigning specific clinical variables to the strata.
Hazard ratio
The hazard ratio provides a relative measure of the event rate between treatment groups. For this analysis, the reported HR was 0.82.
An HR of 0.82 is compatible with an estimated instantaneous event rate approximately 18% lower in the pertuzumab-containing group, but the HR is not an absolute probability and does not indicate how many individual patients benefited.
Confidence intervals
The reported 95% two-sided confidence interval was 0.65–1.02. Confidence intervals provide a range of values reflecting statistical uncertainty around an estimated effect under the specified analysis framework.
For a hazard ratio, the value 1 represents equal estimated hazard between the groups. Because the reported interval reaches above 1, the precision of the estimate is important when interpreting the point estimate of 0.82.
P-values
The reported P-value was 0.0731. A P-value does not quantify the size of the treatment effect. Its interpretation depends on the prespecified null hypothesis, the analysis method, and the testing framework.
For this trial, the registry-reported analysis notes specify a two-sided superiority hypothesis concerning equality versus difference in the PFS survival distributions. The P-value should therefore be read in that context rather than as a standalone measure of treatment magnitude.
9. Analysis Population and What It Means
The statistical analysis used all randomized participants. This is important because randomization creates the basis for the treatment-group comparison. An analysis based on randomized assignment preserves that allocation framework rather than restricting the primary comparison to participants who completed treatment or remained evaluable for some later assessment.
Why randomization matters
Randomization provides the structural basis for comparing the treatment groups without relying solely on adjustment for measured baseline characteristics.
Why the analysis population matters
The primary analysis was explicitly reported in all randomized participants, so the hazard ratio and P-value should be interpreted as results from that analysis population.
The ClinicalTrials.gov record does not specify a separate per-protocol population, treated safety population, or response-evaluable population for the primary analysis. Those populations are therefore not described as if they were part of the reported primary analysis.
10. Safety Results
The ClinicalTrials.gov record reports serious adverse events by treatment arm as affected participants divided by participants at risk.
| Treatment group | Serious adverse events |
|---|---|
| Capecitabine + Trastuzumab | 53/218 |
| Capecitabine + Trastuzumab + Pertuzumab | 58/228 |
These are the serious-adverse-event counts and denominators reported in the ClinicalTrials.gov record. The page does not convert them into percentages because the instruction for this analysis is to preserve reported numbers and not introduce additional calculated results when they are not needed.
11. What the Primary Result Does — and Does Not — Mean
The reported HR of 0.82 represents a relative comparison of the estimated instantaneous PFS event rate between the pertuzumab-containing group and the capecitabine-plus-trastuzumab group. The point estimate is below 1, corresponding to an estimated 18% lower instantaneous event rate.
It does not mean that 18% of patients benefited, that every patient experienced an 18% reduction in risk, or that the probability of remaining progression-free differed by exactly 18 percentage points.
The 95% CI of 0.65–1.02 shows the uncertainty around the estimated hazard ratio. Because the interval crosses 1, the data posted on ClinicalTrials.gov for the primary analysis do not identify a narrow range entirely below the no-difference value.
The two-sided P-value of 0.0731 summarizes the evidence against the stated null hypothesis under the reported testing procedure. It does not tell us how large the treatment effect is. The effect size is described by the hazard ratio, while its statistical precision is described by the confidence interval.
12. Statistical Methods Explained
Why was a log-rank test used?
The primary endpoint is time-to-event PFS rather than a simple binary endpoint. The log-rank framework compares the event experience of the randomized groups over the follow-up period and therefore uses information about when events occur rather than reducing every participant to a single yes/no outcome at one fixed time.
Why was the analysis stratified?
The registry explicitly identifies the analysis as stratified. Stratification allows the survival comparison to account for predefined strata when testing the treatment groups. The ClinicalTrials.gov record does not identify those strata, so no specific stratification factors are attributed to PHEREXA here.
What does a hazard ratio of 0.82 mean?
An HR of 0.82 indicates an estimated instantaneous PFS event rate approximately 82% as large in the pertuzumab-containing group as in the comparator group, corresponding to an 18% lower estimated instantaneous event rate. It is a relative time-to-event measure, not an absolute probability.
Why does the confidence interval extend above 1?
The reported 95% CI is 0.65–1.02. A hazard ratio of 1 represents equal estimated event rates. Because the interval includes values above 1 as well as values below 1, the point estimate of 0.82 has uncertainty that includes the no-difference value and a small range favoring a higher event rate in the pertuzumab-containing group.
Why is the P-value not a measure of effect size?
The P-value depends on both the observed data and the testing framework. It addresses compatibility with the null hypothesis; it does not measure how large the treatment effect is. In PHEREXA, the effect estimate is the HR of 0.82 and the associated precision is summarized by the 95% CI of 0.65–1.02.
Why is all-randomized analysis important?
The registry reports that the analysis population consisted of all randomized participants. This keeps the primary comparison anchored to the treatment assignment generated by randomization, which is the key design feature supporting a comparative treatment analysis.
Why should a hazard ratio not be treated as a relative risk?
A hazard ratio compares instantaneous event rates over time. A relative risk compares probabilities of an event over a specified period. Those quantities are related but are not interchangeable. The PHEREXA registry result is explicitly reported as a hazard ratio for the time-to-event endpoint.
13. Important Limitations and Interpretation Issues
- Limited result set: The ClinicalTrials.gov record contains one formal statistical analysis for the primary endpoint. It does not provide median PFS, Kaplan-Meier probabilities at specified time points, subgroup estimates, or additional formal efficacy analyses.
- Confidence-interval interpretation: The 95% CI of 0.65–1.02 includes the hazard-ratio value of 1, so the point estimate should not be interpreted without its uncertainty interval.
- Two-sided testing: The reported analysis was two-sided. The P-value should therefore be interpreted according to the stated two-sided superiority hypothesis rather than a one-sided threshold.
- Stratification details: The registry identifies a stratified analysis but the ClinicalTrials.gov record does not identify the stratification factors.
- Proportional-hazards considerations: A hazard ratio is a time-to-event summary and is most straightforward to interpret when the relative hazard is reasonably stable over time. The ClinicalTrials.gov record does not report an assessment of the proportional-hazards assumption, so no such conclusion is made here.
- Censoring information: The ClinicalTrials.gov record does not provide detailed censoring rules or missing-assessment handling beyond the registered PFS definition and tumor-assessment schedule.
- Incomplete secondary reporting: Although the registry reports 9 outcome measures, the ClinicalTrials.gov record contains only one formal statistical analysis. No additional efficacy estimates are added from outside sources.
- Safety interpretation: Serious adverse events are reported as affected/at-risk counts by arm. These figures describe safety events and should not be interpreted as efficacy measures.
14. Why This Trial Matters Statistically
PHEREXA is a useful teaching example because its the ClinicalTrials.gov record brings together the central elements of randomized time-to-event analysis: randomization, an independently assessed progression endpoint, stratified log-rank testing, a hazard-ratio effect measure, a two-sided confidence interval, and a formal superiority hypothesis.
| Concept | How it appears in PHEREXA |
|---|---|
| Randomization | The trial is registered as randomized. |
| Parallel design | The two treatment groups are evaluated in a parallel design. |
| Time-to-event endpoint | The primary endpoint is Progression Free Survival (Independent Assessment). |
| Independent assessment | PFS is determined by an Independent Review Facility using RECIST version 1.0. |
| Scheduled assessment | Tumor assessments occur every 9 weeks through Week 27 and every 12 weeks thereafter under the registry time frame. |
| Log-rank testing | The primary analysis uses a log-rank test. |
| Stratified analysis | The registry identifies the primary comparison as stratified. |
| Hazard ratio | The primary effect measure is a hazard ratio of 0.82. |
| Confidence interval | The reported 95% two-sided CI is 0.65–1.02. |
| P-value | The reported two-sided P-value is 0.0731. |
| Superiority hypothesis | The registered hypothesis type is superiority. |
| All-randomized analysis | The primary analysis population is all randomized participants. |
15. Primary Result in Context
The statistical story of PHEREXA is more informative when the point estimate, uncertainty interval, and hypothesis test are read together.
Point estimate
The HR of 0.82 is below 1 and corresponds to an 18% lower estimated instantaneous PFS event rate.
Precision
The 95% CI of 0.65–1.02 spans the no-difference value of 1.
Hypothesis test
The two-sided P-value is 0.0731 under the registered superiority hypothesis.
Analysis framework
The comparison used a stratified log-rank method in all randomized participants.
These components should not be collapsed into a single numerical judgment. The hazard ratio describes the estimated relative event rate, the confidence interval describes uncertainty around that estimate, and the P-value describes the evidence against the stated null hypothesis under the specified test.
16. What Is and Is Not Reported in the Supplied Data
17. Related Tutorials
Learn more about the methods used in this trial:
18. Related Calculators
19. Sources
- ClinicalTrials.gov: PHEREXA, NCT01026142. The ClinicalTrials.gov record and statistical analysis fields used throughout this page are derived from the ClinicalTrials.gov record.
Continue with Clinical Biostats
Explore additional tutorials and statistical calculators covering survival analysis, confidence intervals, hypothesis testing, and clinical-trial methods.