Study Assessing the Efficacy and Safety of Alpelisib Plus Fulvestrant in Men and Postmenopausal Women With Advanced Breast Cancer Which Progressed on or After Aromatase Inhibitor Treatment.: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Study Assessing the Efficacy and Safety of Alpelisib Plus Fulvestrant in Men and Postmenopausal Women With Advanced Breast Cancer Which Progressed on or After Aromatase Inhibitor Treatment. (NCT02437318), a randomized phase 3 study. The primary outcome is Progression-free Survival (PFS) Per Investigator Assessment in the PIK3CA Mutant Cohort.

2. Clinical Question

The registry describes the study as a phase III, randomized, double-blind, placebo-controlled trial. Its main comparison is specific to the mutation-defined primary cohort, rather than an estimate for every participant with advanced breast cancer. The registry also lists outcomes for the PIK3CA non-mutant cohort and other clinical measures, but those outcomes address additional questions and should not be treated as interchangeable with the primary endpoint. The posted primary result is a time-to-event comparison, so it reflects both the timing of events and censoring, not simply the proportion of participants who experienced progression by a single fixed date.

3. Trial Design

The registry identifies NCT02437318 as a completed, randomized, parallel-group phase III study sponsored by Novartis Pharmaceuticals. The record lists an enrollment of 572 and dates the study from 2015-07-23, with primary completion on 2018-06-12. It describes triple masking and a placebo-controlled comparison. The listed arms received alpelisib plus fulvestrant or placebo plus fulvestrant.

Both groups followed the same fulvestrant regimen: 500 mg by intramuscular injection on Day 1 and Day 15 of Cycle 1, followed by Day 1 of each subsequent 28-day cycle. Alpelisib or placebo was given orally once daily at 300 mg. This design makes the randomized contrast the addition of alpelisib or its placebo to a common fulvestrant regimen. The registry does not report further operational details about how masking was maintained or how treatment allocation was generated.

The primary outcome was investigator-assessed progression-free survival in participants with a PIK3CA mutation based on central tumor-tissue testing. The posted primary analysis includes 169 participants in the alpelisib combination group and 172 in the placebo combination group. These are the reported group sizes for that cohort, not a restatement of the overall enrollment. Secondary outcome results include response and clinical benefit measures in both mutation-defined cohorts. The record lists additional secondary outcomes, including overall survival and quality-of-life measures, but results for those measures are not reported in the registry record.

The primary endpoint timeframe is described as after approximately 243 PFS events in the mutant cohort had been observed, up to 33.3 months. The registry record does not report the trial's original sample-size justification, detailed allocation procedures, or a complete statistical analysis plan. Those omissions limit what can be concluded about the prespecified design choices beyond the features explicitly recorded.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Before assignment: One subject in the PIK3CA mutant cohort, who was randomized to the placebo + fulvestrant arm, did not receive study treatment due to a protocol deviation.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: All subjects with a PIK3CA mutation (based on central testing of hotspot-mutations in tumor tissue) who were randomized to study treatment

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

6. Primary Endpoint

The registry lists the primary outcome as Progression-free Survival (PFS) Per Investigator Assessment in the PIK3CA Mutant Cohort, measured over Once approximately 243 PFS events in the PIK3CA mutant cohort had been observed, up to 33.3 months.

Progression-free survival (PFS) is the time from randomization until the first documented disease progression or death from any cause. In this study, progression was assessed locally by radiology according to RECIST 1.1. A participant who had not experienced either event was censored at the date of the last adequate tumor assessment. The primary analysis estimated the PFS distribution using Kaplan-Meier methodology.

The registry's progression definition required at least a 20% increase in the sum of diameters of measured target lesions, using the smallest recorded sum at or after baseline as the reference. It also required an absolute increase of at least 5 mm. These criteria make the endpoint a defined time-to-event outcome rather than a subjective summary of whether a participant improved or worsened.

As a general statistical point, time-to-event endpoints can use information about when events occur while accounting for participants whose follow-up ends before an event is documented. Censoring does not mean that an event occurred; it means the event time was not observed during the available follow-up. PFS is commonly used in trials where progression is an important clinical event and may be observed before overall survival is fully known. The registry record's chosen endpoint does not, by itself, establish how well it represents every aspect of a participant's health or experience.

7. Results Posted in the Registry

• PIK3CA Mutant Cohort by Tumor Tissue: Alpelisib + Fulvestrant vs PIK3CA Mutant Cohort by Tumor Tissue: Placebo + Fulvestrant: hazard ratio 0.65 (95% CI 0.50–0.85), reported p < 0.001 (as reported in the registry).

8. Secondary Efficacy Endpoints

The registry also posts results for secondary outcomes. For each outcome below, the endpoint type, the test applied to the posted summary statistics, the result and its interpretation are given in turn. Secondary endpoints are usually not adjusted for multiple testing, so they support rather than confirm the primary finding.

Overall Response Rate (ORR) Per Investigator Assessment

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • PIK3CA Mutant Cohort by Tumor Tissue: Alpelisib + Fulvestrant: 26.6% (45/169) versus 13.4% (23/172) on control; risk difference 13.3 percentage points (95% CI 4.9 to 21.6), risk ratio 1.99 (1.26–3.14); unadjusted two-proportion z-test p = 0.002.
  • PIK3CA Non-mutant Cohort by Tumor Tissue: Alpelisib + Fulvestrant: 20.9% (24/115) versus 13.4% (23/172) on control; risk difference 7.5 percentage points (95% CI -1.5 to 16.5), risk ratio 1.56 (0.93–2.63); unadjusted two-proportion z-test p = 0.093.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically higher for PIK3CA Mutant Cohort by Tumor Tissue: Alpelisib + Fulvestrant than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between PIK3CA Non-mutant Cohort by Tumor Tissue: Alpelisib + Fulvestrant and the control. The interval gives the range of effects that remain compatible with the data.

Clinical Benefit Rate (CBR) Per Investigator Assessment

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • PIK3CA Mutant Cohort by Tumor Tissue: Alpelisib + Fulvestrant: 61.5% (104/169) versus 44.8% (77/172) on control; risk difference 16.8 percentage points (95% CI 6.3 to 27.2), risk ratio 1.37 (1.12–1.69); unadjusted two-proportion z-test p = 0.002.
  • PIK3CA Non-mutant Cohort by Tumor Tissue: Alpelisib + Fulvestrant: 53.9% (62/115) versus 44.8% (77/172) on control; risk difference 9.1 percentage points (95% CI -2.6 to 20.9), risk ratio 1.20 (0.95–1.53); unadjusted two-proportion z-test p = 0.129.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically higher for PIK3CA Mutant Cohort by Tumor Tissue: Alpelisib + Fulvestrant than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between PIK3CA Non-mutant Cohort by Tumor Tissue: Alpelisib + Fulvestrant and the control. The interval gives the range of effects that remain compatible with the data.

The available secondary summaries cover investigator-assessed overall response rate and clinical benefit rate in both PIK3CA mutation cohorts. In the mutant cohort, the posted percentages are higher in the alpelisib combination group for both measures. In the non-mutant cohort, the point estimates also favor the combination group, but the unadjusted recalculated intervals include no difference. This makes the pattern less conclusive for that cohort than the pattern in the mutant cohort. These secondary findings do not replace or redefine the primary PFS result.

The registry lists additional secondary outcomes, including overall survival in both mutation cohorts, investigator-assessed PFS in the non-mutant cohort, and performance status and quality-of-life measures. Numerical results for these outcomes are not reported in the registry record. Their inclusion in the outcome list alone does not indicate the direction, magnitude, or precision of any treatment contrast.

Unadjusted secondary results carry less weight than the primary comparison for several general reasons. A trial may examine multiple endpoints and populations, creating multiple opportunities for apparently notable results by chance. The registry record does not report whether or how multiplicity was controlled. In addition, the posted secondary comparisons are recalculations from summary statistics rather than sponsor-reported inferential analyses. They can describe a pattern worth considering, but should not be treated as confirmatory evidence equal to the primary analysis. Their interpretation is also limited by the absence of a complete outcome-by-outcome analysis plan and reporting of missing-data handling.

Secondary outcomes registered for this trial, for reference (only those above were recalculated):

• Overall Survival (OS) in the PIK3CA Mutant Cohort
• PFS Per Investigator Assessment in the PIK3CA Non-mutant Cohort
• OS in the PIK3CA Non-mutant Cohort
• Overall Response Rate (ORR) Per Investigator Assessment
• Clinical Benefit Rate (CBR) Per Investigator Assessment
• Time to Definitive Deterioration of Eastern Cooperative Oncology Group (ECOG) Performance Status (PS) Score From Baseline
• Time to 10% Deterioration in the Global Health Status (GHS) /Quality of Life (QOL) Scale Score of the European Organization for Research and Treatment of Cancer's Core Quality of Life Questionnaire (EORTC QLQ-C30)
• Change From Baseline in the GHS/QOL Scale Score of the EORTC QLQ-C30

9. Statistical Interpretation

10. Statistical Methods Explained

The primary outcome is a time-to-event endpoint. Rather than reducing follow-up to whether each participant had an event by an arbitrary cutoff, the analysis uses the event times and the available follow-up. Participants without documented progression or death at their last adequate tumor assessment contribute information up to that time and are then censored under the stated endpoint definition.

What does a Kaplan-Meier estimate describe?

As a general statistical explanation, the Kaplan-Meier method estimates the probability of remaining free of the event over time while accommodating censored follow-up. The median is the time by which the estimated event-free proportion has reached its midpoint. It is therefore a feature of the estimated distribution, not the arithmetic average of observed follow-up times. In this trial, the reported median PFS values summarize the separate estimated distributions in the two mutant-cohort groups.

The registry defines progression using RECIST 1.1 and specifies a required relative increase in the sum of target-lesion diameters along with an absolute increase of at least 5 mm. Death from any cause is also an event. These definitions matter because endpoint classification depends on a stated rule: changes that do not meet the progression criteria are not, on that definition alone, progression events. The registry says assessment was local radiology assessment by investigators; central review of progression is not reported in the registry record.

What does the hazard ratio add?

The hazard ratio compares the event hazard between randomized groups across follow-up. In general, the hazard is the instantaneous rate of an event among participants who have not yet had one. A hazard ratio is not a ratio of median times, and it is not the percentage of participants who progressed. The sponsor reported a hazard ratio of 0.65 with a 95% confidence interval from 0.50 to 0.85. The registry identifies the log-rank method for the comparison.

A single hazard ratio gives a compact summary of the relative event experience, but its interpretation is strongest when the relative contrast is reasonably stable over time. The registry record does not report an assessment of whether that assumption held, nor does it provide a time-varying treatment-effect analysis. The posted value should therefore be understood as the reported summary, not as proof that the relative hazard was identical at every point during follow-up.

Why does censoring matter?

In general, survival methods rely on censoring being handled appropriately. Censoring is not an event and does not establish that a participant remained event-free after the recorded follow-up ended. Kaplan-Meier estimation can accommodate censoring, but interpretation depends on whether censored participants' future event experience would be informative about the observed risk. The registry states that PFS was censored at the last adequate tumor assessment for participants without an event. It does not report the amount or pattern of censoring by group or detailed sensitivity analyses for alternative assumptions.

Over-dispersion is a concern for some count outcomes when the observed variability exceeds what a simple model assumes. It is not the central issue for the posted primary analysis, which is a time-to-event analysis. General statistical reasoning still emphasizes matching methods to the outcome: event timing, censoring, and the definition of progression are central. The registry record does not report further model diagnostics or a detailed specification beyond the Kaplan-Meier estimation and log-rank comparison.

How should the secondary recalculations be read?

The posted ORR and CBR comparisons include simple unadjusted recalculations from summary statistics. These are not identified as sponsor-reported analyses in the registry record. A recalculated comparison can help readers understand the size and uncertainty of a contrast using the posted group summaries, but it does not replace a trial's prespecified analysis. For binary outcomes, confidence intervals describe uncertainty in the estimated group contrast; wide intervals indicate less precision. The registry record does not report sponsor analyses for these posted secondary comparisons.

The record lists 14 outcome measures and 11 posted secondary outcomes, but the results represented in the available secondary summaries cover ORR and CBR. Other listed outcomes, including overall survival and quality-of-life measures, are not reported numerically in the registry record. The analysis population and missing-data handling for each unreported outcome cannot be inferred from the outcome titles. In general, interpreting a trial requires distinguishing the endpoint definition, population, estimation method, and uncertainty rather than treating every outcome as though it came from one identical analysis.

11. Confidence Intervals and Precision

The primary PFS estimates are accompanied by 95% confidence intervals: 7.49 to 14.52 months for alpelisib plus fulvestrant and 3.65 to 7.36 months for placebo plus fulvestrant. The hazard ratio has a 95% confidence interval from 0.50 to 0.85. These intervals communicate uncertainty about the estimates under the analysis, rather than a range containing 95% of individual participants' outcomes.

As a general explanation, a confidence interval reflects both the observed data and the method used to estimate uncertainty. A narrower interval indicates a more precise estimate under that method; a wider interval indicates greater uncertainty. Interval width should be considered alongside the estimate and endpoint scale. For a hazard ratio, the interval is interpreted relative to the no-difference reference. The reported interval remains on one side of that reference, which is consistent with the direction of the reported comparison.

The medians describe different points on the estimated PFS distributions, while the hazard ratio summarizes relative event hazards over follow-up. Their intervals answer different statistical questions, so they should not be treated as interchangeable measures. In particular, comparing median estimates alone does not reproduce the log-rank analysis or establish how the whole event-time distributions compare.

Precision is also specific to the mutation-defined cohort, the observed events, and follow-up in this study. The registry gives a timeframe of approximately 243 PFS events in the mutant cohort, up to 33.3 months, but does not report additional detail about censoring patterns or subgroup precision. The confidence intervals quantify statistical uncertainty under the reported analysis; they do not account for every limitation in study design, measurement, or generalizability.

12. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The sample-size illustrations are planning aids based on observed effects, not descriptions of the trial's original design. For the time-to-event outcome, the figures are total events needed across groups. For the binary secondary outcomes, the figures are participants per arm. The original design assumptions and sample-size justification are not reported in the registry record.

The illustrations assume a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. They show planning for 80% or 90% power to detect the stated fraction of the observed effect. In plain terms, the 80% column gives the planning amount associated with an 80% chance of detecting an effect of the specified size under those assumptions if that effect were present. The 90% column uses a higher target chance of detection and consequently requires more information under otherwise similar assumptions.

Rows for 100% of the observed effect use the observed contrast as the planning target; the 75% and 50% rows consider smaller effects. Smaller effects are harder to distinguish from random variation, so the illustrations require more participants or events to detect them at the same target power. These are not statements about how many participants the trial actually had to enroll to achieve its primary result; nor do they establish that the trial had the stated probability of success. This is not observed power.

The binary-outcome illustrations relate to the posted ORR and CBR contrasts, while the PFS illustration is expressed in total events. They depend on planning assumptions and the observed effect used as a reference, so they should not be read as a recommendation or as a reconstruction of the sponsor's protocol. The registry record does not report whether these particular assumptions were used in the original study design.

13. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

d is the total number of events needed (Schoenfeld's formula) with 1:1 allocation and proportional hazards.

Worked calculation for this trial

1. Hazard ratio HR = 0.650, so ln(HR) = -0.4308.
2. d = 4 × (1.96 + 0.842)² / [ln(0.650)]² = 169.2, rounded up to 170 events in total for 80% power.

The smaller-effect rows in the table above repeat the same calculation with the effect multiplied by 0.75 and 0.5 (for ratios, applied to the logarithm of the ratio).

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
hr <- 0.65                              # observed hazard ratio
events_needed <- function(power, frac = 1) ceiling(round(4 * (z_a + qnorm(power))^2 / log(hr^frac)^2, 6))   # Schoenfeld, 1:1 allocation
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: events = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(events_needed(0.80, f), big.mark = ","), format(events_needed(0.90, f), big.mark = ",")))

# Simulation: given d total events under proportional hazards, events in the treatment arm are Binomial(d, hr / (1 + hr)).
# The score (log-rank type) statistic is (D1 - d / 2) / sqrt(d / 4).
set.seed(2437318)
d    <- events_needed(0.80)
reps <- 20000
D1 <- rbinom(reps, d, hr / (1 + hr))
z  <- (D1 - d / 2) / sqrt(d / 4)
cat(sprintf("Simulated power with %s events: %.3f (%d replicates)\n", format(d, big.mark = ","), mean(abs(z) > z_a), reps))

Output from running the code

100% of observed effect: events = 170 (80% power), 227 (90% power)
 75% of observed effect: events = 301 (80% power), 403 (90% power)
 50% of observed effect: events = 677 (80% power), 906 (90% power)
Simulated power with 170 events: 0.803 (20000 replicates)

The simulation uses the fact that, with d events in total and proportional hazards, the number of events in the treatment arm is binomial. It is an approximation that conditions on the total number of events, and it should land close to the target power.

14. Group-Sequential Design Analysis

Many time-to-event trials, especially in oncology, stop or report early when interim data are overwhelming, so a single fixed-sample number understates the real design problem. The calculation below is an illustrative group-sequential design built from this trial's posted control median and hazard ratio. The registry record does not state the planned number of looks or the spending function, so the design is a planning illustration, not a reconstruction of the protocol. Assumed: one-sided α = 0.025, 90% power, three equally spaced looks on information, Lan-DeMets O'Brien-Fleming spending, 24 months of accrual, 60 months in total and a 1% monthly dropout hazard.

Under these assumptions the design needs about 255 participants in total, with the boundaries below.

Fixed design versus group-sequential alternatives

Group-sequential designs cost a small increase in maximum sample size and events compared with the fixed design, in exchange for the chance to stop early for overwhelming efficacy. O'Brien-Fleming spending keeps early boundaries very conservative and the final boundary close to the fixed-design value, whereas Pocock spending spends more alpha early.

Sensitivity to a smaller true effect

Planning for exactly the observed effect is optimistic, because observed effects in published trials tend to be somewhat larger than the true effect. The table repeats the main design (three looks, O'Brien-Fleming spending, 90% power) for smaller true hazard ratios.

For context, this trial enrolled 572 participants.

Cross-check, upper boundaries: gsDesign 3.71, 2.511, 1.993 | rpact 3.71, 2.511, 1.993. Agreement between the two packages confirms the boundary calculation.

R code and output

suppressWarnings(suppressMessages({ library(gsDesign); library(rpact) }))
med_control <- 5.7          # control-arm median, months (converted from the posted unit)
hr          <- 0.65          # hazard ratio used for planning
lambdaC     <- log(2) / med_control
eta         <- 0.01               # dropout hazard per month (assumption)
acc_months  <- 24                 # accrual duration (assumption)
study_months<- 60                 # total study duration (assumption)

design <- function(k, sfu, beta, h = hr) gsSurv(k = k, test.type = 1, alpha = 0.025, beta = beta, sfu = sfu,
  lambdaC = lambdaC, hr = h, hr0 = 1, eta = eta, gamma = 1, R = acc_months, T = study_months, minfup = study_months - acc_months)

# Main design: 3 looks, Lan-DeMets O'Brien-Fleming spending, one-sided alpha 0.025, 90% power
x <- design(3, sfLDOF, 0.10)
cat(sprintf("N_TOTAL|%s\n", format(ceiling(sum(x$gamma * x$R)), big.mark = ",")))
for (i in 1:3) cat(sprintf("GS|%d|%.1f|%.0f|%.3f|%.5f|%.5f\n", i, x$T[i], x$n.I[i], x$upper$bound[i], pnorm(-x$upper$bound[i]), cumsum(x$upper$spend)[i]))
cat("\n"); print(gsBoundSummary(x))

# Comparison of designs
cmp <- list(list("Fixed design, one analysis", 1, sfLDOF), list("Two looks, O'Brien-Fleming", 2, sfLDOF),
            list("Three looks, O'Brien-Fleming", 3, sfLDOF), list("Three looks, Pocock", 3, sfLDPocock))
cat("\n")
for (b in c(0.20, 0.10)) for (d in cmp) {
  y <- design(d[[2]], d[[3]], b)
  cat(sprintf("DES|%s|%d|%.0f|%s|%.1f\n", d[[1]], round(100 * (1 - b)), tail(y$n.I, 1), format(ceiling(sum(y$gamma * y$R)), big.mark = ","), tail(y$T, 1)))
}

# Sensitivity: what if the true effect is smaller than the observed one? (3 looks, O'Brien-Fleming, 90% power)
cat("\n")
for (fr in c(1, 0.75, 0.5)) {
  y <- design(3, sfLDOF, 0.10, exp(log(hr) * fr))
  cat(sprintf("SENS|%.2f|%.3f|%.0f|%s\n", fr, exp(log(hr) * fr), tail(y$n.I, 1), format(ceiling(sum(y$gamma * y$R)), big.mark = ",")))
}

# Independent cross-check of the boundaries with rpact
r <- getDesignGroupSequential(kMax = 3, alpha = 0.025, beta = 0.10, sided = 1, typeOfDesign = "asOF", informationRates = x$timing)
cat(sprintf("\nCross-check, upper boundaries: gsDesign %s | rpact %s\n", paste(round(x$upper$bound, 3), collapse = ", "), paste(round(r$criticalValues, 3), collapse = ", ")))

Method: LachinFoulkes 
    Analysis               Value Efficacy
   IA 1: 33%                   Z   3.7103
      N: 164         p (1-sided)   0.0001
  Events: 77        ~HR at bound   0.4271
   Month: 15    P(Cross) if HR=1   0.0001
             P(Cross) if HR=0.65   0.0338
   IA 2: 67%                   Z   2.5114
      N: 256         p (1-sided)   0.0060
 Events: 153        ~HR at bound   0.6655
   Month: 24    P(Cross) if HR=1   0.0060
             P(Cross) if HR=0.65   0.5603
       Final                   Z   1.9930
      N: 256         p (1-sided)   0.0231
 Events: 229        ~HR at bound   0.7681
   Month: 60    P(Cross) if HR=1   0.0250
             P(Cross) if HR=0.65   0.9000



Cross-check, upper boundaries: gsDesign 3.71, 2.511, 1.993 | rpact 3.71, 2.511, 1.993

15. Analyses Reported by the Sponsor

The registry reports the sponsor's primary analysis as a log-rank comparison of PFS in the PIK3CA mutant cohort. The reported hazard ratio is 0.65, with a 95% confidence interval from 0.50 to 0.85. The sponsor's analysis is the principal inferential result for the primary endpoint; it is not replaced by a calculation from the displayed medians.

The posted secondary ORR and CBR comparisons are simple unadjusted recalculations from group summary statistics, and the registry does not list sponsor-reported analyses for those comparisons. Such recalculations are useful for examining the summaries and their uncertainty, but they may not reproduce the sponsor's analysis choices, including any prespecified methods or adjustments. They should be read as supplementary rather than as official trial analyses.

No discrepancy is recorded for the primary analysis. The registry record does not provide additional details about the complete sponsor analysis plan, such as model diagnostics, sensitivity analyses, or multiplicity procedures. The reported primary estimate and its interval should therefore be presented as the sponsor's recorded result, while recognizing that the record's level of statistical detail is limited.

16. Multiplicity, Missing Data and Analysis Populations

Multiplicity arises when a study tests several outcomes, populations, or comparisons. In general, making many statistical comparisons increases the chance that at least one result appears notable by chance, unless the analysis plan addresses that issue. This trial registry lists multiple secondary outcomes and reports outcomes for separate PIK3CA mutation cohorts. The registry record does not report whether a multiplicity adjustment or a hierarchical testing strategy was used. Consequently, the secondary recalculated comparisons should be treated cautiously and should not be read as though each were an isolated confirmatory test.

The primary endpoint description identifies the mutant cohort by central testing of hotspot mutations in tumor tissue and includes randomized participants meeting that definition. The record also specifies the PFS event definition, Kaplan-Meier estimation, and censoring at the last adequate tumor assessment for participants without an event. It does not report a fuller description of analysis populations for all outcomes, how protocol deviations were handled, or whether analyses were conducted according to intention-to-treat principles.

Missing-data reporting is incomplete for interpretation beyond the stated PFS censoring rule. The registry record does not report the extent or pattern of missing assessments, reasons for missingness, or sensitivity analyses addressing missing observations. Those details can affect how confidently results are generalized beyond participants with observed assessments. The record also does not report how missing data were handled for response, clinical benefit, survival, or quality-of-life outcomes.

These are limits of what can be learned from the registry record, not evidence that a particular procedure was or was not performed. The original sample-size justification, interim analyses, and complete statistical analysis plan are not reported in the registry record. No conclusions about those design or analysis features should be inferred from the posted estimates alone.

17. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 5% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: hyperglycaemia (9.9%), diarrhoea (3.2%), osteonecrosis of jaw (3.2%), anaemia (2.1%), abdominal pain (2.1%).

The registry posts serious adverse event figures for the treatment groups: 110 participants affected among 284 at risk in the alpelisib plus fulvestrant group, and 54 among 287 at risk in the placebo plus fulvestrant group. It also lists serious-event terms, with hyperglycaemia at 9.9%, diarrhoea and osteonecrosis of jaw at 3.2% each, and anaemia, abdominal pain, and acute kidney injury at 2.1% each. These figures describe the registry's reported safety data; they do not by themselves establish that a treatment caused an event.

Deaths are reported separately in treatment and post-treatment survival follow-up group entries. The registry lists 9 and 12 deaths in the treatment groups, and 151 and 156 in the respective post-treatment follow-up entries. These entries have different at-risk counts, including zero in the post-treatment entries, so they should not be combined or interpreted as comparable treatment-period risks. The record does not provide enough detail to infer causes, timing, exposure duration, or causal attribution from the counts alone.

Safety comparisons require attention to observation period, definitions, exposure, and which participants were included in each reporting group. Those details are not fully described in the posted figures. The serious adverse event summaries are therefore descriptive registry data, not a complete assessment of relative safety.

18. What the Result Does, and Does Not, Mean

The posted primary result describes a difference in investigator-assessed progression-free survival between randomized groups within the PIK3CA mutant cohort. The median PFS estimates and the sponsor-reported hazard ratio point in the same direction: the estimated progression-or-death experience during observed follow-up favored the alpelisib plus fulvestrant group relative to placebo plus fulvestrant in that cohort. The confidence interval provides a measure of statistical precision around the hazard ratio, but does not eliminate uncertainty or address every source of bias.

The result does not mean that every participant in one group experienced a longer period without progression than every participant in the other group. A group-level estimate does not predict an individual's outcome. Nor does PFS alone answer every question about survival, symptoms, quality of life, or treatment burden. Those are separate outcomes, and numerical registry results for several of them are not reported in the registry record.

The result is limited to the trial population and endpoint as defined: men and postmenopausal women with hormone receptor-positive, HER2-negative advanced breast cancer after aromatase inhibitor treatment, with the primary analysis focused on centrally identified PIK3CA mutant tumors. It should not be generalized automatically to people outside those eligibility characteristics or to the non-mutant cohort. Secondary results may add context but are less definitive when their analyses are unadjusted and their multiplicity handling is not reported.

As a general statistical point, a confidence interval and a p-value address sampling uncertainty under specified assumptions; neither establishes clinical importance on its own. The registry record does not provide a threshold for judging a clinically meaningful PFS difference, nor does it supply all details needed to assess missingness, assessment processes, or robustness. The findings describe the posted randomized comparison, not a recommendation about care or a guarantee of benefit or harm.

19. Why This Trial Matters Statistically

This trial is statistically informative because it uses randomization and a concurrent placebo-plus-fulvestrant comparison to estimate the contrast associated with adding alpelisib to a common treatment regimen. The primary question is restricted to a biologically defined subgroup identified through central tumor-tissue testing, making the target population explicit rather than treating all enrolled participants as a single undifferentiated group.

Its primary endpoint also illustrates why time-to-event methods are used in clinical research. Participants can have different lengths of follow-up, and some may not have an observed progression or death before their last adequate assessment. Kaplan-Meier estimation and a log-rank comparison use the timing of events while incorporating censoring under the stated rules. The posted median and hazard ratio offer complementary summaries, though neither fully describes every feature of the event-time distributions.

The registry results also illustrate the distinction between a primary analysis and secondary comparisons. Response and clinical benefit summaries can help describe the broader outcome pattern, but the posted secondary recalculations are unadjusted and the registry does not report multiplicity procedures. Other listed outcomes lack numerical results in the registry record. This makes transparent separation of confirmatory and exploratory evidence important.

Finally, the record shows why statistical interpretation depends on reporting quality as well as estimates. Confidence intervals communicate precision, while missing details about the full analysis plan, missing data, and sensitivity analyses limit assessment of robustness. A statistically clear primary estimate can still leave important questions about generalizability and other endpoints unanswered.

20. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

The primary analysis applies to a specific mutation-defined cohort and relies on central testing of tumor tissue for PIK3CA hotspot mutations. The result should not be assumed to apply to the non-mutant cohort or populations outside the eligibility description. The registry record does not report detailed information about the availability of tumor tissue, test failures, or how such cases were handled.

Progression was assessed by local radiology assessment according to RECIST 1.1. The registry record does not report whether an independent central review was conducted, whether assessment schedules were identical between groups, or the extent of missing tumor assessments. Because PFS depends on documented progression and censoring, these details would help readers assess possible measurement and follow-up effects.

The record reports the sponsor's primary hazard ratio and confidence interval but does not report a complete statistical analysis plan, assessment of proportional hazards, or sensitivity analyses. It also does not report multiplicity handling for the many listed outcomes, or full results for several secondary endpoints. The posted ORR and CBR calculations are unadjusted recalculations rather than sponsor-reported analyses.

Safety entries require caution because treatment-period and post-treatment follow-up figures are presented separately, and the latter list zero participants at risk alongside death counts. The record does not provide enough information in these figures to compare causal risk, timing, or exposure-adjusted rates. These reporting limitations do not overturn the primary result, but they constrain broader interpretation and the assessment of its robustness.

21. Related Tutorials

22. Related Statistical Calculators

23. Sources

• ClinicalTrials.gov: Study Assessing the Efficacy and Safety of Alpelisib Plus Fulvestrant in Men and Postmenopausal Women With Advanced Breast Cancer Which Progressed on or After Aromatase Inhibitor Treatment. — NCT02437318.
• Registry record last updated 2025-02-13; status last verified 2025-02.
• Results first posted 2019-08-13.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

24. Record Summary

NCT02437318 is a completed randomized, triple-masked, parallel-group phase III study of alpelisib plus fulvestrant versus placebo plus fulvestrant in men and postmenopausal women with hormone receptor-positive, HER2-negative advanced breast cancer after aromatase inhibitor treatment. The registry lists enrollment of 572 and study dates from 2015-07-23 to primary completion on 2018-06-12.

The primary endpoint was investigator-assessed PFS in the centrally identified PIK3CA mutant cohort. The sponsor reported a hazard ratio of 0.65, with a 95% confidence interval from 0.50 to 0.85; median PFS was 11.0 months versus 5.7 months. Secondary response and clinical benefit summaries are available, but several listed outcomes and important analysis details are not reported in the registry record. The findings are specific to the reported population, endpoint, and analysis.