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
PHERGain is a randomized, open-label, parallel phase 2 trial in breast cancer evaluating a response-adapted treatment strategy involving trastuzumab and pertuzumab. The registry lists three arms and an enrollment of 377 participants, with primary endpoints focused on pathologic complete response and 3-year invasive disease-free survival.
| Feature | PHERGain |
|---|---|
| Phase | Phase 2 |
| Condition | Breast Cancer |
| Allocation | Randomized |
| Design | Parallel |
| Masking | None |
| Primary purpose | Treatment |
| Enrollment | 377 |
| Number of arms | 3 |
| Start date | 2017-06-26 |
| Primary completion date | 2023-11-01 |
| Lead sponsor | MedSIR |
| Sponsor type | OTHER |
| ClinicalTrials.gov | NCT03161353 |
2. Clinical Question
The central statistical question is whether a response-adapted strategy involving trastuzumab and pertuzumab can identify patients with HER2-positive breast cancer who achieve sufficient metabolic response to receive the combination as exclusive neoadjuvant treatment, while evaluating subsequent invasive disease-free survival.
Population
Patients with breast cancer enrolled in a randomized phase 2 trial evaluating a response-adapted strategy. The primary endpoint description specifically identifies PET responders in cohort B.
Intervention strategy
The registry title and primary endpoint describe trastuzumab and pertuzumab as the chemotherapy-free combination being evaluated, with endocrine therapy permitted in the described strategy.
Comparator
The trial has three randomized arms. The registry information identifies the study as randomized and lists the interventions, but does not provide enough arm-level detail here to reconstruct each comparator assignment.
Primary questions
What is the rate of pathologic complete response after the specified neoadjuvant treatment period, and what is the 3-year invasive disease-free survival rate in cohort B?
3. Trial Design
The design is randomized, parallel, and unmasked. The registry lists three arms and an enrollment of 377. The response-adapted component is central to the statistical structure because the primary pCR endpoint is specifically defined among PET responders in cohort B.
Interventions listed in the registry
Targeted agents
- Perjeta
- Herceptin
Additional agents
- Docetaxel
- Carboplatin
- Letrozole
- Tamoxifen
The registry lists these interventions for the three-arm trial but does not provide sufficient arm-specific information here to assign each drug or combination to a particular randomized arm. The statistical analysis therefore focuses on the design and endpoints that are explicitly defined rather than reconstructing arm assignments.
4. Trial Timeline
Trial start
The registry lists June 26, 2017 as the study start date.
Randomized treatment evaluation
The study uses randomized allocation, a parallel design, no masking, and three treatment arms.
Primary pCR assessment
The registry defines the pCR endpoint after Cycle 6 or 8, with each cycle lasting 21 days and the corresponding time frame stated as 4.2 or 5.6 months.
Primary iDFS assessment
The second primary endpoint is the 3-year invasive disease-free survival rate, assessed after 36 months.
Primary completion
The registry lists November 1, 2023 as the primary completion date and the study status as active, not recruiting.
5. Primary Endpoints
| Endpoint | Registry definition | Time frame |
|---|---|---|
| Evaluate the rate of pCR | Evaluate the rate of pCR as defined by the absence of invasive disease in the breast and axilla (ypT0/isN0) at the time of surgery achieved with the combination of trastuzumab and pertuzumab (± endocrine therapy) as exclusive neoadjuvant treatment in PET responders patients (cohort B/PET responders). The endpoint also describes the PET/CT positive predictive value (PPV) for a pCR among patients who are PET responders. | After Cycle 6 or 8 (each cycle 21 days — after 4.2 or 5.6 months) |
| 3-year iDFS rate | Time from the first date of no disease (i.e., date of surgery) to invasive recurrence, new invasive disease, or death by any cause. Recurrence will be defined in accordance with the standardized efficacy endpoints (STEEP) criteria. The primary analysis will be to estimate 3-year iDFS rate in cohort B. | After 3 years (36 months) |
Endpoint 1: Pathologic complete response
The pCR endpoint is a binary surgical outcome. The registry defines pCR as absence of invasive disease in the breast and axilla, specifically ypT0/isN0, at the time of surgery. The primary pCR analysis is focused on PET responders in cohort B and evaluates the rate achieved with trastuzumab and pertuzumab, with or without endocrine therapy, as exclusive neoadjuvant treatment.
Endpoint 2: 3-year invasive disease-free survival
The iDFS endpoint is a time-to-event outcome. The clock begins at the first date of no disease, identified in the registry as the date of surgery, and ends at invasive recurrence, new invasive disease, or death from any cause. The primary analysis is intended to estimate the 3-year iDFS rate in cohort B.
6. Planned Analysis
No results have been posted on ClinicalTrials.gov for the registered primary endpoints. The registry therefore defines what will be measured without providing numerical estimates, confidence intervals, p-values, event counts, or formal posted statistical analyses.
Primary endpoint 1
Binary endpoint evaluated after Cycle 6 or 8, corresponding to 4.2 or 5.6 months.
The registry defines pCR as ypT0/isN0 at surgery among the specified PET responders in cohort B.
For a binary endpoint such as pCR, the natural statistical summary is the proportion of evaluable patients meeting the pCR definition. An appropriate confidence interval would quantify uncertainty around that proportion. Because this is a response-adapted randomized design, the analysis population is particularly important: the registry explicitly identifies PET responders in cohort B as the population for the primary pCR endpoint.
The proportion is descriptive of the specified analysis population. It should not automatically be interpreted as the pCR probability for all 377 enrolled participants.
Primary endpoint 2
Time-to-event endpoint assessed after 36 months.
The registry defines the event as invasive recurrence, new invasive disease, or death by any cause.
For a time-to-event endpoint such as iDFS, the usual analysis estimates the survival function over time, accounts for right-censoring, and reports the estimated survival probability at the prespecified 36-month landmark. Kaplan-Meier estimation is a standard framework for this purpose.
Here di represents events at an event time and ni represents the number at risk immediately before that time. The 3-year iDFS rate is the estimated survival probability at 36 months.
Why the response-adapted population matters
The pCR endpoint is explicitly defined in PET responders in cohort B. That means the statistical interpretation depends on the treatment-selection pathway leading to that population. A response-selected estimate answers a different question from an all-randomized-population estimate because PET response is itself part of the trial strategy.
7. Statistical Methodology
Analysis of a binary pCR endpoint
The pCR endpoint is fundamentally a binary outcome: a patient either satisfies the registry definition of ypT0/isN0 at surgery or does not. The primary descriptive quantity is therefore a proportion. Confidence intervals around that proportion quantify statistical uncertainty and should be presented alongside the point estimate rather than relying on the percentage alone.
For a response-adapted study, the analysis should preserve the prespecified population definition. In this trial, the registry explicitly identifies cohort B/PET responders for the primary pCR analysis.
Kaplan-Meier analysis of iDFS
iDFS is a time-to-event endpoint because patients can experience an event at different times after surgery, while others may remain event-free at their last assessment. Kaplan-Meier estimation is designed to use the available follow-up from both types of observations without treating censored patients as though they experienced the event.
Censored → contributes up to the last eligible follow-up time
The distinction allows an analysis to use partial follow-up information without incorrectly assigning an event to a patient whose event status is not observed.
Estimating the 3-year iDFS rate
The registry specifies a 36-month endpoint. The statistical target is therefore a survival probability at a fixed time point rather than simply a median survival time. This distinction is important: a 3-year iDFS rate asks what proportion is estimated to remain free of the defined invasive disease event through 36 months, while a median iDFS asks when the estimated survival function crosses 50%.
Potential role of Cox regression
A Cox proportional-hazards model is a standard complementary approach for time-to-event endpoints when investigators want a relative hazard measure in addition to the survival probability. However, the ClinicalTrials.gov record does not report a posted Cox model, hazard ratio, stratification factors, or model specification for PHERGain. Those details should therefore not be attributed to the trial.
Response-adapted analysis
The response-adapted structure creates a distinction between randomization at enrollment and subsequent classification according to FDG-PET response. For statistical interpretation, this means that analyses restricted to PET responders are conditional analyses rather than simple comparisons of all randomized patients.
8. Statistical Methods Explained
What does the pCR endpoint measure?
pCR is a binary pathologic outcome assessed at surgery. In PHERGain, the registry defines it as absence of invasive disease in the breast and axilla, ypT0/isN0. The resulting rate is the proportion of patients in the specified analysis population who meet that definition.
Why is the PET-responder population important?
The primary pCR description specifically refers to PET responders in cohort B. Conditioning on PET response changes the population being summarized. Consequently, a pCR rate in PET responders should not be interpreted as the unconditional probability of pCR among everyone entering the trial.
Why is iDFS analyzed differently from pCR?
pCR is determined at a particular clinical milestone: surgery. iDFS records whether and when a defined event occurs during follow-up. Because event times vary and some patients may not have an event during observed follow-up, time-to-event methods such as Kaplan-Meier estimation are appropriate for iDFS.
What does a 3-year iDFS rate represent?
It is the estimated probability of remaining free of the registry-defined iDFS event through 36 months from the specified starting point, which is the first date of no disease or date of surgery. It is not the same as a response rate and does not describe pathologic response at surgery.
Why does the iDFS definition include death from any cause?
Including death from any cause makes death an iDFS event even when the death is not preceded by a documented invasive recurrence. This is important because competing clinical outcomes can otherwise complicate interpretation of a disease-free endpoint.
Why does the definition begin at surgery?
The registry defines the time origin for iDFS as the first date of no disease, identified as the date of surgery. This makes the iDFS estimand different from an endpoint measured from randomization because the follow-up clock begins later in the treatment pathway.
What is the statistical consequence of a response-adapted strategy?
The strategy creates a treatment pathway in which PET response influences subsequent treatment. Analyses of PET responders therefore describe outcomes in a selected subgroup defined after treatment begins. The distinction between randomized groups and post-randomization response groups is essential when interpreting causal comparisons.
9. Confidence Intervals and Statistical Precision
No numerical confidence intervals for the PHERGain primary endpoints are posted in the registry information summarized here. Nevertheless, confidence intervals are an important component of the planned statistical interpretation.
For pCR
A confidence interval around the pCR proportion describes statistical uncertainty around the estimated response rate in the specified analysis population.
For iDFS
A confidence interval around the 3-year survival estimate describes uncertainty around the estimated probability of remaining free of the defined event through 36 months.
A point estimate alone can appear more precise than the underlying evidence warrants. In particular, response-adapted analyses may involve fewer patients than the full randomized population, so the effective sample size for a particular endpoint can be smaller than the overall enrollment of 377.
10. Multiplicity and the Three-Arm Design
PHERGain has three randomized arms and two registered primary endpoints. Those features create a potentially complex multiplicity structure. However, the ClinicalTrials.gov record does not specify an alpha-allocation strategy, hierarchical testing procedure, adjustment method, or other formal multiplicity procedure.
| Feature | Statistical implication |
|---|---|
| Three randomized arms | Multiple treatment comparisons may arise depending on the prespecified comparison structure. |
| Two primary endpoints | More than one primary hypothesis can require coordinated type I error control when formal hypothesis testing is performed. |
| PET-response selection | Post-randomization response status must be distinguished from the original randomized population. |
| 36-month iDFS target | The primary time-to-event estimand is a fixed-time survival probability rather than simply a median. |
Without a specified multiplicity strategy, a numerical p-value or significance conclusion cannot be reconstructed from the registry fields alone. The statistically important point is that the number of arms and endpoints determines the structure of the inferential problem, while the prespecified analysis plan determines how that problem should be handled.
11. Non-Inferiority, Crossover, and Bayesian Methods
The registry information provided for PHERGain does not specify a non-inferiority margin, crossover procedure, factorial design, Bayesian analysis, or interim-analysis framework. The trial is described as randomized and parallel, not factorial, and its masking status is none.
| Design topic | What the registry reports |
|---|---|
| Non-inferiority margin | Not reported in the available registry information. |
| Crossover | Not reported in the available registry information. |
| Factorial design | The registry identifies a parallel design with three arms. |
| Bayesian methods | Not reported in the available registry information. |
| Interim analysis | Not reported in the available registry information. |
| Multiplicity adjustment | Not reported in the available registry information. |
12. Missing Data and Analysis Populations
The registry defines the primary endpoints but does not provide the missing-data or imputation rules for those endpoints in the information summarized here. That distinction is important because the pCR and iDFS endpoints have different data structures.
pCR missingness
pCR is determined at surgery. Patients without an evaluable pathologic assessment require a prespecified rule for defining the analysis population and handling missing assessments.
iDFS censoring
iDFS naturally incorporates right-censoring when patients remain without an event at the last available follow-up. The censoring rules determine how those observations contribute to the survival estimate.
For a response-adapted endpoint, it is also important to distinguish missing PET-response information from missing surgical pathology. These are not necessarily interchangeable missing-data problems because they occur at different stages of the treatment pathway.
13. Interpretation of the Primary Endpoints
The pCR endpoint measures the frequency with which the prespecified pathologic definition is achieved at surgery among the population identified by the registry as PET responders in cohort B. It does not measure long-term disease control by itself, and it should not be interpreted as equivalent to invasive disease-free survival.
The 3-year iDFS endpoint summarizes the probability of remaining free from invasive recurrence, new invasive disease, or death through 36 months from the first date of no disease. It incorporates both recurrence and mortality and therefore represents a broader post-surgical outcome than pCR.
pCR is an early pathologic measure obtained at surgery, whereas iDFS follows patients over a substantially longer period. A response-adapted trial can therefore connect an early treatment-response classification with a later clinical outcome, but the two endpoints remain statistically distinct estimands.
14. Important Limitations and Interpretation Issues
- No posted numerical results: the registry does not provide numerical results, confidence intervals, p-values, or event counts for the registered primary endpoints.
- Arm-level detail: the registry information identifies three randomized arms and lists six interventions, but the available fields do not specify enough arm-level detail to reconstruct every treatment assignment.
- Response selection: the primary pCR endpoint is defined among PET responders in cohort B, making the analysis population different from the full randomized population.
- Endpoint timing: pCR is assessed after Cycle 6 or 8, while iDFS begins from the first date of no disease or surgery and is evaluated at 36 months.
- Missing-data rules: the available registry information does not specify the imputation or missing-assessment rules for the primary endpoints.
- Multiplicity: three randomized arms and two primary endpoints can create multiple comparison issues, but the available registry information does not specify the formal type I error-control strategy.
- Survival assumptions: time-to-event analyses depend on appropriate handling of censoring and, if a Cox model is used, assumptions concerning the hazard structure.
- Generalizability: the trial concerns breast cancer and a specific response-adapted strategy, so its findings should be interpreted in the context of the enrolled study population and treatment pathway.
15. Why This Trial Matters Statistically
PHERGain is statistically interesting because it links randomized treatment allocation with an intermediate response assessment and then evaluates a longer-term clinical endpoint. That structure illustrates why the definition of the analysis population can be as important as the choice of statistical test.
| Statistical concept | How it appears in PHERGain |
|---|---|
| Randomization | The study uses randomized allocation across three arms. |
| Parallel design | The registry identifies a parallel trial design. |
| Response-adapted treatment | FDG-PET response is incorporated into the treatment strategy. |
| Binary endpoint | pCR is defined as achievement or non-achievement of ypT0/isN0 at surgery. |
| Time-to-event endpoint | iDFS measures time from surgery to invasive recurrence, new invasive disease, or death. |
| Fixed-time survival estimate | The primary iDFS analysis targets the 36-month rate. |
| Post-randomization selection | The primary pCR endpoint focuses on PET responders in cohort B. |
| Multiple treatment arms | Three randomized arms create a potentially multi-comparison analysis structure. |
| Endpoint multiplicity | Two primary endpoints require a prespecified inferential framework when formal testing is performed. |
| Censoring | iDFS requires time-to-event methods that account for patients without an observed event at last follow-up. |
16. The Statistical Story of PHERGain
The trial can be understood as a sequence of distinct statistical questions rather than as one single outcome.
Question 1: Who responds metabolically?
FDG-PET response is incorporated into the response-adapted strategy. This creates a clinically defined subgroup that becomes important for the primary pCR analysis.
Question 2: Who achieves pCR?
The primary pCR endpoint evaluates whether the surgical pathology meets the ypT0/isN0 definition after the specified treatment period.
Question 3: What happens after surgery?
The iDFS endpoint follows patients from the first date of no disease and records invasive recurrence, new invasive disease, or death.
Question 4: What is the 3-year disease-free probability?
The primary iDFS analysis targets the estimated probability of remaining free of the defined event at 36 months.
This sequence illustrates a central principle of clinical-trial statistics: an endpoint is inseparable from its population, time origin, event definition, and analysis method. A percentage without its denominator, a survival rate without its time point, or a subgroup estimate without its selection mechanism can all be misleading.
17. Clinical Interpretation vs Statistical Interpretation
Statistical interpretation
The trial uses a randomized three-arm parallel design with two registered primary endpoints. One is a binary pathologic endpoint among PET responders in cohort B; the other is a 36-month time-to-event endpoint in cohort B.
Clinical interpretation
The study evaluates whether a response-adapted strategy can use FDG-PET response to identify patients for a chemotherapy-free neoadjuvant approach involving trastuzumab and pertuzumab, while tracking subsequent invasive disease-free survival.
18. What Is and Is Not Established by the Registry Record
| Question | Registry information available |
|---|---|
| Was the study randomized? | Yes. |
| How many participants were enrolled? | 377. |
| How many arms were planned? | 3. |
| Was the study masked? | No; masking is listed as none. |
| What are the primary endpoints? | pCR rate and 3-year iDFS rate. |
| What is the pCR definition? | Absence of invasive disease in the breast and axilla, ypT0/isN0, at surgery. |
| What is the iDFS event definition? | Invasive recurrence, new invasive disease, or death by any cause. |
| What is the iDFS time origin? | First date of no disease, identified as the date of surgery. |
| Are numerical primary results posted? | No numerical primary endpoint results are posted in the registry information summarized here. |
| Are formal statistical analyses posted? | No posted statistical analyses are listed for the primary endpoints. |
| Are arm-specific treatment assignments fully specified here? | No. |
19. Sources
- ClinicalTrials.gov: NCT03161353 — PHERGain.
- PubMed: PubMed record.
- PubMed: PubMed record.
- PubMed: PubMed record.
- PubMed: PubMed record.
- PubMed: PubMed record.
Continue with Clinical Biostats
Explore additional clinical-trial analyses and statistical methods for randomized studies, binary endpoints, and time-to-event outcomes.
20. Record Summary
PHERGain is a randomized phase 2, three-arm, parallel trial with an enrollment of 377 participants. Its statistical structure combines an early pathologic endpoint with a longer-term time-to-event endpoint: pCR is defined as ypT0/isN0 at surgery after Cycle 6 or 8, while 3-year iDFS is defined from the first date of no disease to invasive recurrence, new invasive disease, or death by any cause. The response-adapted design makes the distinction between the randomized population and the PET-responder population particularly important. Because the registry does not post numerical primary endpoint results or formal statistical analyses, the appropriate statistical focus is the definition of the estimands, the analysis populations, the handling of time-to-event data, and the implications of response-based treatment selection.