This page separates reported trial information from statistical interpretation. The ClinicalTrials.gov record reports the primary endpoint and outcome measures, but does not provide formal statistical analyses for the primary endpoint in the ClinicalTrials.gov record.
1. Trial at a Glance
BYLieve was a completed phase 2, non-randomized, parallel-design study evaluating alpelisib in combination with endocrine therapy in patients with PIK3CA mutant, HR+, HER2- advanced breast cancer who had progressed on or after prior treatments. The registry reports 383 participants across 3 cohorts.
| Feature | BYLieve |
|---|---|
| Trial name | BYLieve |
| NCT identifier | NCT03056755 |
| Phase | Phase 2 |
| Condition | Breast Cancer |
| Population | Patients with PIK3CA mutant, HR+, HER2- advanced breast cancer who have progressed on or after prior treatments |
| Allocation | Non-randomized |
| Design model | Parallel |
| Masking | None |
| Primary purpose | Treatment |
| Enrollment | 383 |
| Arms / cohorts | 3 |
| Results posted | Yes |
| Outcome measures posted | 8 |
| Lead sponsor | Novartis Pharmaceuticals |
| Sponsor type | Industry |
2. Clinical Question
The registry describes a treatment study asking whether alpelisib-based therapy can provide clinical benefit in patients with PIK3CA mutant, HR+, HER2- advanced breast cancer who have progressed on or after prior treatments. The study uses different treatment cohorts according to prior endocrine-treatment exposure.
Population
Patients with PIK3CA mutant, HR+, HER2- advanced breast cancer who have progressed on or after prior treatments.
Intervention
Alpelisib used with endocrine-treatment regimens represented by fulvestrant or letrozole, with goserelin or leuprolide also listed among the study interventions.
Comparator
The ClinicalTrials.gov record does not identify a randomized comparator arm. Allocation is explicitly listed as non-randomized.
Primary question
What percentage of participants are alive without disease progression at 6 months in the three study cohorts?
3. Trial Design
The design is listed as non-randomized, parallel, and unmasked. This is a fundamental distinction from a randomized controlled trial. Without random assignment, differences between cohorts may reflect differences in prior treatment history or other characteristics as well as the treatment received.
Pre-treated With CDK 4/6i + AI
- Prior treatment: CDK 4/6 inhibitor plus aromatase inhibitor
- Alpelisib is part of the study intervention set
- Serious adverse events reported for 37/127 participants
Pre-treated With CDK 4/6i + Fulvestrant
- Prior treatment: CDK 4/6 inhibitor plus fulvestrant
- Serious adverse events reported for 48/126 participants
- Primary endpoint applies to this cohort as well as Cohorts A and C
Pre-treated With Systemic Chemotherapy
- Prior treatment: systemic chemotherapy
- Serious adverse events reported for 38/126 participants
- Primary endpoint applies to this cohort as well as Cohorts A and B
4. Study Interventions
The ClinicalTrials.gov record lists five interventions: alpelisib, fulvestrant, letrozole, goserelin, and leuprolide. The trial's brief title specifies alpelisib plus fulvestrant or letrozole, based on prior endocrine therapy.
| Intervention | Registry role in the ClinicalTrials.gov record |
|---|---|
| Alpelisib | Drug; central study intervention |
| Fulvestrant | Drug; endocrine therapy represented in the study |
| Letrozole | Drug; endocrine therapy represented in the study |
| Goserelin | Drug; listed among study interventions |
| Leuprolide | Drug; listed among study interventions |
The ClinicalTrials.gov record does not provide dosing schedules, treatment durations, treatment discontinuation rules, or a more detailed regimen description. Those details are therefore not reproduced here.
5. Primary Endpoint
| Endpoint | Registry definition | Endpoint type | Time frame |
|---|---|---|---|
| Percentage of Participants Who Were Alive Without Disease Progression at 6 Months | Percentage of participants who were alive without disease progression at 6-month follow-up based on local investigator assessment per RECIST v1.1 in Cohort A, Cohort B and Cohort C. Participants who progressed, died, or discontinued study before 6 months were counted as a failure. | Binary | At 6 months |
This endpoint converts follow-up status at a prespecified time point into a binary outcome. A participant is counted as a success if alive without disease progression at 6 months; the registry definition explicitly counts progression, death, or study discontinuation before 6 months as failure.
The primary analysis therefore concerns a proportion: the number of participants meeting the 6-month success definition divided by the relevant cohort population under the prespecified analysis framework.
6. Statistical Methodology
Binary outcome analysis
Because the registered primary endpoint is binary, the natural descriptive quantity is the 6-month success proportion. For a cohort with \(x\) successes among \(n\) evaluable participants, the observed proportion is \(x/n\).
A confidence interval can then be constructed around the estimated proportion using an appropriate binomial method. The ClinicalTrials.gov record does not provide the number of primary-endpoint successes, confidence intervals, or formal hypothesis-test results.
Why the endpoint is not a conventional randomized comparison
In a randomized trial, a binary endpoint can be compared between treatment groups using quantities such as a risk difference, risk ratio, odds ratio, or a model-based treatment effect. BYLieve is listed as non-randomized, however, so the principal statistical question is different: the observed 6-month proportions describe outcomes within the predefined cohorts, but differences between cohorts can be confounded by differences in the populations entering those cohorts.
RECIST v1.1 assessment
The primary endpoint is based on local investigator assessment per RECIST v1.1. Thus, progression status is determined using a standardized tumor-response framework, but the assessment is not described in the ClinicalTrials.gov record as blinded central review.
Failure definition
The registry's definition is particularly important statistically. Participants who progressed, died, or discontinued study before 6 months were counted as failures. This prevents those outcomes from simply disappearing from the denominator and makes the endpoint a composite operational definition of being alive and progression-free at the specified time point.
7. Planned Analysis
The ClinicalTrials.gov record identifies the primary endpoint and report that results are posted, but no formal statistical analyses were posted. Consequently, there is no reported primary-endpoint estimate, confidence interval, or P-value in the ClinicalTrials.gov record that can be reproduced on this page.
What is actually reported in the ClinicalTrials.gov record?
The registry identifies a binary 6-month endpoint across Cohorts A, B, and C, but the ClinicalTrials.gov record contains no treatment-effect estimate, confidence interval, or P-value.
How this endpoint would typically be analysed
A conventional analysis would first calculate the proportion of participants who are alive without disease progression at 6 months within each cohort. Because the endpoint is binary, a binomial confidence interval would normally accompany each estimated proportion.
If the scientific objective were to compare two cohorts, investigators could use a comparison of proportions or an appropriate regression model. However, because this study is non-randomized, such a comparison would require careful consideration of differences in prior treatment exposure and other prognostic factors. A numerical difference alone would not establish that alpelisib caused the difference.
8. Statistical Methods Explained
What does a binary 6-month endpoint measure?
It asks a yes-or-no question at a fixed time point: was the participant alive without disease progression at 6 months? This differs from a time-to-event endpoint such as progression-free survival, which uses the timing of progression or death throughout follow-up.
Why are participants who discontinue before 6 months counted as failures?
That rule is part of the registered endpoint definition. Without it, early discontinuation could create ambiguity about whether a participant should remain a success, be excluded, or be treated as missing. Counting such participants as failures creates a definite binary classification for the primary endpoint.
Why would a confidence interval be useful for a proportion?
An observed proportion is estimated from a finite number of participants. A confidence interval communicates uncertainty around that estimate. It is not a prediction interval for future patients and does not describe the range of individual patient outcomes.
Why is non-randomization important when comparing cohorts?
Randomization balances measured and unmeasured factors probabilistically between treatment groups. BYLieve is listed as non-randomized, and the cohorts are defined by different prior-treatment histories. Therefore, differences in observed outcomes can reflect both treatment and differences in the populations entering the cohorts.
Why does RECIST v1.1 matter?
The endpoint depends on determining whether disease has progressed. RECIST v1.1 supplies standardized criteria for tumor-response assessment, reducing ambiguity in how radiographic disease status is classified. The ClinicalTrials.gov record specifically identify local investigator assessment.
Would a P-value measure the size of the treatment effect?
No. A P-value addresses compatibility with a specified statistical null hypothesis under the assumptions of the test. It does not quantify the magnitude of an effect. For a binary endpoint, the estimated proportion and its confidence interval are needed to describe the observed outcome and its precision.
9. Interpreting a 6-Month Progression-Free Proportion
The primary endpoint is a time-point proportion, not a hazard ratio. If an analysis reports a 6-month progression-free proportion, that number describes the fraction of participants meeting the registered success definition at 6 months. It does not describe the instantaneous hazard of progression or death over follow-up.
A 6-month progression-free proportion should not be interpreted as the probability that a particular future patient will remain progression-free for exactly 6 months. Trial eligibility, treatment history, sample size, censoring or discontinuation rules, and the analysis population all affect how the observed proportion should be interpreted.
A proportion based on a smaller cohort is generally more statistically uncertain than the same proportion based on a larger cohort. The appropriate confidence interval communicates that uncertainty. The ClinicalTrials.gov record does not provide confidence intervals for the primary endpoint.
Even a precisely estimated cohort-specific proportion would not by itself establish a causal effect relative to another cohort. Precision and causal identification are separate statistical concepts.
10. Secondary Outcome Measures
the ClinicalTrials.gov record states that 8 outcome measures were posted. However, the ClinicalTrials.gov record identifies only the registered primary endpoint in sufficient detail to reproduce its definition and time frame. No secondary-endpoint estimates or formal statistical analyses are included in the ClinicalTrials.gov record.
11. Safety Results
The ClinicalTrials.gov record reports serious adverse events by cohort. These are presented as affected participants divided by participants at risk.
| Cohort | Serious adverse events, affected / at risk |
|---|---|
| Cohort A: Pre-treated With CDK 4/6i + AI | 37/127 |
| Cohort B: Pre-treated With CDK 4/6i + Fu | 48/126 |
| Cohort C: Pre-treated With Systemic Chem | 38/126 |
The registry extract also contains two additional safety entries: 0/1 for Cohort A: Pre-treated With CDK 4/6i + AI and 3/10 for Cohort C: Pre-treated With Systemic Chem. The registry-reported field does not provide an accompanying description that identifies how these additional entries differ from the principal serious-adverse-event counts, so they should not be assigned a more specific interpretation.
12. Cohort Size and Safety Context
The principal safety denominators are 127, 126, and 126 for Cohorts A, B, and C, respectively. These values sum to 379, whereas total trial enrollment is reported as 383. That difference is important: the registry-reported safety field should not be treated as though it represents every enrolled participant without knowing the specific safety-analysis population and denominator definitions used for the posted result.
13. Trial Timeline
Study start
The BYLieve study began on August 29, 2017, according to the ClinicalTrials.gov record.
Non-randomized parallel study
The study used a phase 2, non-randomized, parallel design with 3 cohorts and a total enrollment of 383 participants.
6-month disease-control status
The registered primary endpoint was the percentage of participants alive without disease progression at 6 months, assessed locally according to RECIST v1.1.
Primary completion
The registry profile lists June 14, 2021 as the primary completion date.
Current trial status in the ClinicalTrials.gov record
the ClinicalTrials.gov record identifies BYLieve as completed and states that results have been posted.
14. What This Trial Can — and Cannot — Establish Statistically
What the design supports
The study can describe the 6-month alive-without-progression outcome within its predefined cohorts and characterize safety outcomes among the reported analysis populations.
What the design does not establish by itself
Because allocation was non-randomized, differences between cohorts do not automatically isolate the causal effect of treatment from differences in the populations receiving the treatment.
What the endpoint captures
The primary endpoint captures a clinically meaningful binary state at a fixed time: alive and without disease progression at 6 months.
What remains unavailable here
The ClinicalTrials.gov record contains no formal primary-endpoint estimates, confidence intervals, or P-values. Those quantities are therefore not reproduced or reconstructed.
15. Limitations
- Non-randomized allocation: the absence of randomization limits causal comparisons between cohorts because treatment and prior-treatment characteristics are not assigned randomly.
- Cohort heterogeneity: the three cohorts are defined by different prior-treatment histories, which can be associated with different underlying patient populations.
- Fixed-time endpoint: a 6-month binary endpoint does not use the full timing information contained in disease progression and death after or before the time point.
- Local investigator assessment: the registered primary endpoint uses local investigator assessment per RECIST v1.1 rather than a blinded central assessment in the registry-reported definition.
- Discontinuation handling: participants who discontinued before 6 months were counted as failures under the registered endpoint definition. This rule should be retained when interpreting the endpoint.
- Incomplete statistical-analysis detail: the ClinicalTrials.gov record contains no formal statistical analyses for the primary endpoint, preventing reproduction of a registry-reported estimate, confidence interval, or P-value.
- Safety denominators: the principal serious-adverse-event denominators total 379, while overall enrollment is 383. The ClinicalTrials.gov record does not provide enough information to reconcile those populations.
- Additional safety entries: two additional affected/at-risk entries are listed without sufficient accompanying description to establish their precise statistical role.
- Secondary outcomes: although 8 outcome measures are reported as posted, the ClinicalTrials.gov record does not provide sufficient numerical detail for secondary-endpoint analysis.
16. Why This Trial Matters Statistically
BYLieve is a useful teaching example because it highlights an important distinction between estimating an outcome within cohorts and estimating a causal treatment effect between randomized groups. The trial's primary endpoint is binary and fixed at 6 months, while the study itself is explicitly non-randomized.
| Statistical concept | How it appears in BYLieve |
|---|---|
| Binary endpoint | Alive without disease progression at 6 months is classified as a success or failure. |
| Fixed time point | The registered primary endpoint is assessed at 6 months. |
| RECIST v1.1 | Progression status is based on local investigator assessment per RECIST v1.1. |
| Non-randomized design | The allocation field is explicitly listed as non-randomized. |
| Cohort analysis | The primary endpoint is evaluated in Cohort A, Cohort B, and Cohort C. |
| Confidence intervals | Would ordinarily be used to communicate uncertainty around estimated binary proportions, although none are reported in the ClinicalTrials.gov record. |
| Safety denominators | Serious adverse events are reported as affected participants divided by participants at risk. |
| Causal inference | Non-randomization requires caution when interpreting differences between cohorts. |
17. Related Statistical Tutorials
Learn more about the methods used in this trial:
18. Related Statistical Calculators
19. Sources
- ClinicalTrials.gov: BYLieve, NCT03056755.
- PubMed: PubMed record for PMID 39637900.
- PubMed: PubMed record for PMID 39177931.
- PubMed: PubMed record for PMID 38439079.
- PubMed: PubMed record for PMID 33794206.
Continue through the Clinical Biostats statistical library
Connect this trial's binary endpoint, cohort design, confidence intervals, and safety analysis to broader statistical tutorials and calculation tools.
20. Record Summary
BYLieve is a completed phase 2, non-randomized, parallel study enrolling 383 participants across 3 cohorts. Its registered primary endpoint is the percentage of participants who were alive without disease progression at 6 months, based on local investigator assessment per RECIST v1.1, with participants who progressed, died, or discontinued before 6 months counted as failures.
Statistically, the central teaching point is the distinction between a binary outcome measured at a fixed time and a causal comparison generated by randomization. The ClinicalTrials.gov record reports serious adverse events for the three principal cohorts as 37/127, 48/126, and 38/126, while additional entries of 0/1 and 3/10 are present without enough context to establish their precise role.