A Study to Investigate Efficacy and Safety With Oral AZD9833 Compared With Intramuscular Fulvestrant in Post-menopausal Women at Least 18 Years of Age With Advanced ER-positive HER2 Negative Breast Cancer: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study to Investigate Efficacy and Safety With Oral AZD9833 Compared With Intramuscular Fulvestrant in Post-menopausal Women at Least 18 Years of Age With Advanced ER-positive HER2 Negative Breast Cancer (NCT04214288), a randomized phase 2 study. The primary outcome is Progression-free Survival (PFS).

2. Clinical Question

3. Trial Design

SERENA-2 was a randomised, open-label, parallel-group, multicentre Phase 2 study sponsored by AstraZeneca. The registry record lists an enrollment of 240. The study began on 2020-04-22, and primary completion is recorded as 2022-08-30. The registry status is active, not recruiting.

The study compared oral AZD9833 with intramuscular fulvestrant. The posted outcome groups were AZD9833 75mg, AZD9833 150mg, AZD9833 300mg, and fulvestrant 500 mg. The primary outcome results list 74 patients in the AZD9833 75mg group, 73 in the AZD9833 150mg group, 20 in the AZD9833 300mg group, and 73 in the fulvestrant group. The registry record says recruitment to the 300mg arm was stopped early at 20 patients randomised and cautions that data from this arm are highly variable.

Open-label describes the absence of masking: treatment assignment was not blinded. Randomisation assigns participants to treatment groups by chance, while a parallel-group design means participants remain in their assigned group rather than crossing between treatments. These are general descriptions of the design terms; the record does not report further operational details about the randomisation procedure or masking arrangements beyond the open-label designation.

The primary outcome was progression-free survival assessed by the investigator. Secondary outcomes listed in the record include objective response rate, duration of response, tumour-size change at 16 weeks, overall survival, clinical benefit rate at 24 weeks, plasma concentrations of AZD9833, biomarker changes, and health-related quality of life. Details of the protocol’s original sample-size justification, planned interim analyses, and full statistical analysis plan are not reported in the registry record.

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.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: FAS consisted of all randomised patients, with treatment groups assigned in accordance with the randomisation, regardless of the actual treatment received.

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.

Population: Full Analysis Set (FAS) consisted of all randomised patients, with treatment groups assigned in accordance with the randomisation, regardless of the actual treatment received.

6. Primary Endpoint

The registry lists the primary outcome as Progression-free Survival (PFS), measured over From date of randomisation to date of objective disease progression (or last evaluable assessment in the absence of progression) or death (up to data cut-off of 29 months).

Progression-free survival (PFS) measures the time from randomisation until objective disease progression or death. In this study, the registry definition includes death by any cause in the absence of progression, regardless of whether a patient withdrew from randomised therapy or received another anti-cancer therapy before progression. The outcome was assessed by the investigator using RECIST version 1.1. The stated timeframe runs from randomisation to progression, or to the last evaluable assessment when progression is absent, with follow-up up to the data cut-off of 29 months.

A time-to-event endpoint uses both whether an event occurred and how long it took to occur. Participants without a recorded progression or death by their last evaluable assessment can contribute information up to that point; this is generally called censoring. The registry definition identifies the last evaluable assessment as the endpoint for participants without progression, but further details of censoring rules are not reported in the registry record.

The median PFS is the point in the estimated event-time distribution at which half the group is estimated to have experienced progression or death. It describes the middle of the observed time-to-event experience, rather than the experience of every participant. PFS is used in trials because it captures time until disease progression or death, rather than only recording whether progression occurred by a fixed visit. The registry record reports median PFS and 90% confidence intervals for the treatment groups.

7. Results Posted in the Registry

• AZD9833 75mg vs Fulvestrant 500 mg: hazard ratio 0.59 (95% CI 0.42–0.82), reported p = 0.017 (as reported in the registry).
• AZD9833 150mg vs Fulvestrant 500 mg: hazard ratio 0.64 (95% CI 0.46–0.89), reported p = 0.009 (as reported in the registry).

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is a time-to-event measure. For each participant, the relevant time begins at randomisation and ends at objective disease progression or death, as defined in the registry record. Participants without a recorded event are followed to their last evaluable assessment for this endpoint. General statistical practice for time-to-event outcomes allows such participants to contribute information until follow-up ends, rather than treating them as though an event occurred at that time. The exact censoring implementation used for this study is not reported in the registry record.

What does the median PFS describe?

The median is a summary of the estimated distribution of event times. It marks the point by which half of the group is estimated to have experienced progression or death. It is not the average time for all participants, and it does not describe the timing of every event. The posted medians are 7.2 months for AZD9833 75mg, 7.7 months for AZD9833 150mg, 6.3 months for AZD9833 300mg, and 3.7 months for fulvestrant 500 mg. They are reported alongside 90% confidence intervals, except that the 300mg group has no upper confidence limit listed.

What does a hazard ratio mean?

A hazard is the instantaneous event rate among participants who have not yet experienced the event, considered over the course of follow-up. A hazard ratio compares the estimated hazards between treatment groups. A ratio below 1 indicates a lower estimated hazard in the first group named in the comparison than in the comparator group, in the reported analysis. The sponsor reported hazard ratios of 0.59 for AZD9833 75mg versus fulvestrant 500 mg and 0.64 for AZD9833 150mg versus fulvestrant 500 mg.

A hazard ratio is not a ratio of median PFS values, and it is not a direct statement about how many months an individual may gain or lose. Interpreting a single hazard ratio as a stable relative comparison over follow-up generally relies on assumptions about how the hazards relate over time. Whether those assumptions were assessed, and what diagnostics were used, are not reported in the registry record.

How was the comparison tested?

The sponsor-reported comparisons used the Log Rank method. In general, a log-rank test compares the timing of events across groups over follow-up, accounting for censored observations under the method’s assumptions. The registry record lists a hazard ratio, a 90% confidence interval, and a p-value for each of the two sponsor-reported comparisons. It does not report further details of the test implementation or the method used to estimate the hazard-ratio confidence intervals.

The reported p-values are 0.017 for the comparison of AZD9833 75mg with fulvestrant 500 mg and 0.009 for the comparison of AZD9833 150mg with fulvestrant 500 mg. A p-value expresses how compatible the observed data are with a specified null hypothesis under the test assumptions. It is not the probability that the null hypothesis is true, nor does it measure the size or clinical importance of a difference.

What do the confidence intervals add?

A confidence interval communicates uncertainty around an estimate. The 90% intervals for the reported hazard ratios are 0.42 to 0.82 for AZD9833 75mg versus fulvestrant and 0.46 to 0.89 for AZD9833 150mg versus fulvestrant. For the median PFS estimates, the intervals are 3.7 to 10.9 months for AZD9833 75mg, 5.5 to 12.9 months for AZD9833 150mg, and 2.0 to 6.0 months for fulvestrant 500 mg. The interval for AZD9833 300mg has a lower limit of 1.9 months, while an upper limit is not listed. Confidence intervals reflect uncertainty in the group estimates; they are not ranges of individual outcomes.

The full analysis set consisted of all randomised patients, with treatment groups assigned according to randomisation regardless of treatment actually received. This preserves the randomised group assignment for the defined analysis population. The registry record does not specify how missing outcome assessments were handled, whether additional analysis populations were used, or whether sensitivity analyses were conducted.

General statistical considerations also include multiplicity, which arises when multiple treatment comparisons or endpoints are assessed. The registry record lists multiple treatment groups and secondary outcomes, but the approach to multiplicity adjustment is not reported. The original sample-size justification, planned significance level for the primary trial analysis, interim analyses, and complete statistical analysis plan are also not reported in the registry record. Planning calculations associated with the observed hazard ratios are separate illustrations and should not be mistaken for the trial’s original design plan.

10. Confidence Intervals and Precision

The median PFS estimates are accompanied by 90% confidence intervals that convey uncertainty in the estimated group medians. The interval for AZD9833 75mg extends from 3.7 to 10.9 months around a median of 7.2 months. For AZD9833 150mg, the interval extends from 5.5 to 12.9 months around a median of 7.7 months. For fulvestrant 500 mg, it extends from 2.0 to 6.0 months around a median of 3.7 months. The 300mg group has a median of 6.3 months and a lower confidence limit of 1.9 months; an upper limit is not reported.

The hazard-ratio estimates also have 90% confidence intervals. For AZD9833 75mg versus fulvestrant 500 mg, the estimate is 0.59 and the interval is 0.42 to 0.82. For AZD9833 150mg versus fulvestrant 500 mg, the estimate is 0.64 and the interval is 0.46 to 0.89. The intervals indicate that the estimated relative hazard remains uncertain even though the reported intervals do not include 1. They do not imply that the true effect is equally likely to be any value within the interval.

Precision depends in part on the amount of information available, including the number and timing of events and the amount of follow-up. The registry record lists 74 patients for AZD9833 75mg, 73 for AZD9833 150mg, 20 for AZD9833 300mg, and 73 for fulvestrant 500 mg. The record specifically warns that the 300mg data are highly variable because recruitment to that arm stopped early. Its limited group size and absent upper confidence limit make the estimate especially difficult to interpret precisely.

Confidence intervals describe statistical uncertainty under the analysis method; they do not account for every possible limitation, such as unreported analysis choices or missing-data handling. The registry record does not provide enough information to evaluate those aspects in detail.

11. 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 planning figures are event counts, not participants per arm. The time-to-event calculations use the Schoenfeld approach and report the total number of events needed across the two groups for the stated power targets. The figures are planning illustrations based on the observed effect, not the original trial design assumptions.

The illustration assumes a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. It considers effects equal to 100%, 75%, and 50% of the observed effect. For the AZD9833 75mg comparison, the table gives 113 total events for 80% power and 151 for 90% power when planning for 100% of the observed effect; for 75% of the observed effect, it gives 201 and 269 total events; for 50% of the observed effect, it gives 452 and 604 total events. For the AZD9833 150mg comparison, the corresponding totals are 158 and 212, 281 and 376, and 631 and 845.

The 80% power column describes the event total associated with an 80% chance of detecting the specified effect under the illustration’s assumptions, if that effect were present. The 90% power column describes the event total for a higher chance of detecting that specified effect under those assumptions. These are design-planning concepts, not probabilities that the reported result is correct. Smaller effects require more events to detect with the same power, as reflected by the larger totals for planning for 75% or 50% of the observed effect.

This is not observed power. The registry record does not report the trial’s original sample-size justification or design assumptions, so these figures should not be read as a reconstruction of the original planning process.

12. 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 = 0.590, so ln(HR) = -0.5276.
2. d = 4 × (1.96 + 0.842)² / [ln(0.590)]² = 112.8, rounded up to 113 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 <- 0.59                              # 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(4214288)
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 = 113 (80% power), 151 (90% power)
 75% of observed effect: events = 201 (80% power), 269 (90% power)
 50% of observed effect: events = 452 (80% power), 604 (90% power)
Simulated power with 113 events: 0.817 (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.

13. 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 164 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 240 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 <- 3.7          # control-arm median, months (converted from the posted unit)
hr          <- 0.59          # 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: 94         p (1-sided)   0.0001
  Events: 51        ~HR at bound   0.3529
   Month: 14    P(Cross) if HR=1   0.0001
             P(Cross) if HR=0.59   0.0338
   IA 2: 67%                   Z   2.5114
      N: 154         p (1-sided)   0.0060
 Events: 102        ~HR at bound   0.6074
   Month: 22    P(Cross) if HR=1   0.0060
             P(Cross) if HR=0.59   0.5603
       Final                   Z   1.9930
      N: 164         p (1-sided)   0.0231
 Events: 153        ~HR at bound   0.7240
   Month: 60    P(Cross) if HR=1   0.0250
             P(Cross) if HR=0.59   0.9000



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

14. Analyses Reported by the Sponsor

The posted hazard ratios and their confidence intervals and p-values are sponsor-reported analyses using the Log Rank method. For AZD9833 75mg versus fulvestrant 500 mg, the reported hazard ratio is 0.59, with a 90% confidence interval of 0.42 to 0.82 and p = 0.017. For AZD9833 150mg versus fulvestrant 500 mg, the reported hazard ratio is 0.64, with a 90% confidence interval of 0.46 to 0.89 and p = 0.009.

The planning event counts are a separate calculation based on the observed effects and stated planning assumptions. They do not replace the sponsor’s outcome analysis, and they do not independently verify or re-estimate its hazard ratios. No discrepancy is flagged between the posted sponsor analyses and the values used for the planning illustrations. The sponsor-reported analysis remains the relevant source for the reported clinical-trial comparison; the planning figures address a different question about event totals under hypothetical design assumptions.

The record does not give full details of the statistical analysis plan, including the original design justification, missing-data procedures, or multiplicity approach. Those omissions limit how fully the sponsor-reported results can be independently assessed from the registry record alone.

15. Multiplicity, Missing Data and Analysis Populations

The registry record includes several treatment groups, two sponsor-reported AZD9833-versus-fulvestrant hazard-ratio comparisons, and multiple secondary outcomes. When several hypotheses are tested, the chance of obtaining at least one apparently notable result can increase. Statistical methods that address this issue are commonly described as multiplicity adjustments. Whether the trial used such an adjustment, and how it was applied, are not reported in the registry record. The posted p-values should therefore be understood as belonging to their reported comparisons, without assuming that they control error across all outcomes or treatment comparisons.

The primary analysis population is described as the full analysis set: all randomised patients, assigned to groups according to randomisation regardless of the treatment actually received. The record does not report whether other analysis populations were used or how patients with missing endpoint assessments were handled. It also does not describe sensitivity analyses addressing missing data or departures from the planned analysis.

The endpoint definition says that PFS runs from randomisation to objective disease progression or death, or to the last evaluable assessment without progression, up to the data cut-off of 29 months. This describes the endpoint and timeframe but does not provide all operational details needed to evaluate censoring or missing-assessment rules.

The original statistical analysis plan, including any interim analyses, is not reported in the registry record. The original sample-size justification and planned primary-analysis assumptions are also not reported. Consequently, the available record does not establish whether the posted analyses followed a prespecified multiplicity strategy, how missing outcomes were addressed, or whether interim monitoring affected the analysis.

16. 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: breast cellulitis (14.3%), covid-19 pneumonia (14.3%), covid-19 (5.0%), international normalised ratio increased (5.0%), hyponatraemia (1.4%).

The registry record lists serious adverse events and other adverse events by treatment group. Serious adverse events were reported for 6 of 74 patients in the AZD9833 75mg group, 7 of 73 in the AZD9833 150mg group, 2 of 20 in the AZD9833 300mg group, and 4 of 73 in the fulvestrant 500 mg group. The corresponding counts reported for other adverse events were 44 of 74, 61 of 73, 17 of 20, and 31 of 73. Death counts listed by group were 23, 14, 7, and 21, respectively.

The record also lists serious adverse-event terms and percentages, including Breast cellulitis at 14.3%, COVID-19 pneumonia at 14.3%, COVID-19 at 5.0%, International normalised ratio increased at 5.0%, Hyponatraemia at 1.4%, and Cerebrovascular accident at 1.4%. These terms are reported without a treatment-group breakdown in the listed terms, so they should not be assigned to a particular arm on this basis.

These counts describe the posted safety data, not whether an event was caused by a study drug. The record does not provide the additional detail needed to interpret the timing, duration, clinical context, or attribution of each event. Death counts should likewise not be assumed to represent deaths caused by treatment.

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

The posted PFS results show median estimates of 7.2 months for AZD9833 75mg, 7.7 months for AZD9833 150mg, 6.3 months for AZD9833 300mg, and 3.7 months for fulvestrant 500 mg. The sponsor-reported hazard ratios for AZD9833 75mg and 150mg versus fulvestrant are both below 1, with reported 90% confidence intervals that do not include 1. These statistics describe the trial’s group-level time-to-event comparisons; they do not establish how an individual participant would fare.

The results do not, by themselves, establish the clinical importance of the differences or resolve uncertainty about the size of the effect. The confidence intervals show that the estimates have uncertainty, and the 300mg arm is particularly limited: recruitment stopped early at 20 patients randomised, the registry cautions that its data are highly variable, and an upper confidence limit is not listed. The record does not provide a sponsor-reported hazard-ratio comparison for that arm.

The p-values describe compatibility of the observed data with a null hypothesis under the reported test and its assumptions. They are not the probability that a treatment works, nor do they account automatically for multiple outcomes or comparisons. The registry record does not report whether multiplicity was addressed, the full original statistical plan, missing-data procedures, or interim analyses. Those omissions matter when placing individual comparisons in the broader trial analysis.

The study is open-label and the primary endpoint was assessed by the investigator using RECIST version 1.1. The record does not give details about possible safeguards against assessment bias or the implementation of the assessment process. The outcome data should be read in that context, alongside the available uncertainty intervals and the registry’s specific caution about the 300mg group. These results are statistical evidence from the posted comparisons, not medical advice or a recommendation about treatment.

18. Why This Trial Matters Statistically

This trial illustrates several important features of interpreting clinical time-to-event results. A median PFS gives a useful summary of the event-time distribution, while a hazard ratio compares event hazards across follow-up. The two statistics answer related but distinct questions: the median describes a central point in each group’s estimated event-time distribution, whereas the hazard ratio summarizes a relative comparison of event hazards. Neither should be treated as a complete description of every participant’s experience.

The registry record reports both estimated effects and uncertainty intervals. This makes it possible to distinguish the point estimate from its precision: the hazard ratios are 0.59 and 0.64, while their 90% confidence intervals span 0.42 to 0.82 and 0.46 to 0.89. The intervals are important context for the estimates, not decorative additions to them. The 300mg arm also demonstrates why group size and recruitment history matter when reading an estimate; the record explicitly warns that its data are highly variable after recruitment stopped early at 20 patients randomised.

The example also highlights the difference between trial results and planning calculations. Event totals calculated from observed effects can illustrate what might be required under stated assumptions, but they do not reveal the original trial design or establish observed power. Here, the planning assumptions are stated, while the original design justification is not reported in the registry record.

Finally, the record lists multiple treatment groups and outcomes but does not report the multiplicity strategy, missing-data procedures, or full analysis plan. Statistical interpretation depends not only on a point estimate and p-value, but also on how analyses were selected and implemented. Keeping those distinctions visible helps readers avoid treating a reported comparison as a complete account of the trial.

19. 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.

Several limitations are specific to the available trial record. The AZD9833 300mg arm stopped recruitment early at 20 patients randomised, and the registry explicitly describes its data as highly variable. The median PFS is reported as 6.3 months, with a lower confidence limit of 1.9 months, but no upper confidence limit is listed. The posted sponsor analyses do not include a hazard-ratio comparison for this arm.

The study was open-label, and the primary endpoint was investigator-assessed. The record describes RECIST version 1.1 as the assessment standard, but does not report additional details about assessment safeguards or whether independent review was used. The record also does not report the full statistical analysis plan, original sample-size justification, planned interim analyses, multiplicity procedures, or missing-data methods. These gaps restrict assessment of prespecification and analytical robustness.

The primary outcome timeframe extends up to a data cut-off of 29 months, but the registry excerpt does not provide a detailed account of follow-up completeness or the number and timing of PFS events by group. The confidence intervals communicate uncertainty in the posted estimates, but cannot address unreported design or analysis details.

The safety record lists adverse-event counts and selected serious-event terms, but does not provide, in this record, enough context to determine attribution, timing, duration, or severity for individual events. The listed deaths should not be interpreted as treatment-caused deaths. The sponsor-reported statistical comparisons and the available safety summaries should therefore be read within the limits of the registry information.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: A Study to Investigate Efficacy and Safety With Oral AZD9833 Compared With Intramuscular Fulvestrant in Post-menopausal Women at Least 18 Years of Age With Advanced ER-positive HER2 Negative Breast Cancer — NCT04214288.
• Registry record last updated 2026-08-26; status last verified 2026-08.
• Results first posted 2023-12-12.
• 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.

23. Record Summary

SERENA-2 is a randomised, open-label, parallel-group, multicentre Phase 2 study of oral AZD9833 versus intramuscular fulvestrant in post-menopausal women at least 18 years of age with advanced ER-positive HER2-negative breast cancer. The registry record lists enrollment of 240 and a primary outcome of investigator-assessed PFS defined using RECIST version 1.1.

The posted median PFS estimates are 7.2 months for AZD9833 75mg, 7.7 months for AZD9833 150mg, 6.3 months for AZD9833 300mg, and 3.7 months for fulvestrant 500 mg. Sponsor-reported hazard ratios versus fulvestrant are 0.59 for AZD9833 75mg and 0.64 for AZD9833 150mg, each with a 90% confidence interval and reported p-value. The record cautions that the 300mg arm stopped recruitment early at 20 patients randomised and its data are highly variable.

The registry record does not report the original sample-size justification, full statistical analysis plan, multiplicity approach, or missing-data procedures. Planning event totals are illustrations based on observed effects and stated assumptions, not the original trial design.