S0800, Nab-Paclitaxel, Doxorubicin, Cyclophosphamide, and Pegfilgrastim With or Without Bevacizumab in Treating Women With Inflammatory or Locally Advanced Breast Cancer: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for S0800, Nab-Paclitaxel, Doxorubicin, Cyclophosphamide, and Pegfilgrastim With or Without Bevacizumab in Treating Women With Inflammatory or Locally Advanced Breast Cancer (NCT00856492), a randomized phase 2 study. The primary outcome is Number of Patients With Pathological Complete Response Rate.

2. Clinical Question

3. Trial Design

The registry describes S0800 as a randomized, parallel, phase II trial and lists masking as none. It records an enrollment of 215 and a completed status. The registry dates the start to 2010-04 and primary completion to 2015-10. These trial-level details describe the planned and recorded study framework; they do not by themselves establish how many participants contributed data to each endpoint.

The treatment comparison was designed around neoadjuvant chemotherapy for women with inflammatory or locally advanced HER-2/NEU negative breast cancer. Arm 1 included nab-paclitaxel with bevacizumab, followed by doxorubicin, cyclophosphamide and pegfilgrastim. Arm 2 used nab-paclitaxel first and then the latter chemotherapy regimen with pegfilgrastim. Arm 3 reversed that sequence, beginning with doxorubicin and cyclophosphamide plus pegfilgrastim before nab-paclitaxel. The registry’s primary analysis combined Arms 2 and 3, so it does not provide separate primary estimates for those sequences.

The primary outcome was pathological complete response assessed using pre-study pathology and post-chemotherapy surgery pathology. Secondary outcomes listed in the registry include overall survival, event-free survival and adverse events possibly, probably or definitely related to study drug. An event-free survival analysis is posted. Numerical results for the other listed secondary outcomes are not reported in the registry record. The registry record does not report the trial’s original sample-size justification, detailed randomization procedures, or further protocol-level information needed to reconstruct the full study conduct.

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: Pre-specified analysis of Arm 1 vs. Arm 2/3 to compare effect of bevacizumab

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 Number of Patients With Pathological Complete Response Rate, measured over pre-study pathology vs. post-chemo surgery pathology (approx. 39-42 weeks post-randomization).

The primary endpoint counts participants with pathological complete response, or pCR. The registry describes pCR as the absence of residual invasive cancer in both the breast and axillary lymph nodes after chemotherapy, assessed at surgery. The endpoint compares pathology before study treatment with pathology after chemotherapy, at approximately 39-42 weeks after randomization. Its basic statistical form is binary: each participant contributes a response or no response to the count.

pCR is used in this setting because the registry states that achieving pCR is associated with a favorable long-term prognosis across breast cancer subtypes, and describes it as a surrogate endpoint for disease-free and overall survival. A surrogate endpoint is an intermediate outcome used to study treatment response when the outcome of interest may occur later. This can make pathology findings relevant to an early treatment comparison, but an association with later prognosis does not make pCR equivalent to survival. A difference in pCR alone does not establish that long-term outcomes differ by the same amount or in the same direction.

The primary analysis was prespecified as Arm 1 versus Arm 2/3 to compare the effect of bevacizumab. Since the comparator is a combined group, the endpoint estimate addresses that grouping, rather than an isolated comparison of bevacizumab against either chemotherapy sequence separately. The registry record does not report detailed rules for handling unevaluable or missing surgery pathology.

7. Results Posted in the Registry

• Arm 2/3 (Nab-Paclitaxel/AC+PEG-G)): 21.2% (24/113) versus 35.7% (35/98) on control; risk difference -14.5 percentage points (95% CI -26.6 to -2.4), risk ratio 0.59 (0.38–0.93); unadjusted two-proportion z-test p = 0.019.

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.

Event-free Survival

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Arm 2/3 (Nab-Paclitaxel/AC+PEG-G)): 21.2% (24/113) versus 20.4% (20/98) on control; risk difference 0.8 percentage points (95% CI -10.1 to 11.8), risk ratio 1.04 (0.61–1.76); unadjusted two-proportion z-test p = 0.882.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Arm 2/3 (Nab-Paclitaxel/AC+PEG-G)) and the control. The interval gives the range of effects that remain compatible with the data.

The registry lists overall survival, event-free survival and adverse events possibly, probably or definitely related to study drug as secondary outcomes. Event-free survival is the secondary endpoint with a posted between-group analysis. Numerical overall survival results and numerical results for the listed related-adverse-event outcome are not reported in the registry record. Consequently, the available secondary information does not support a complete assessment across all listed outcomes.

Secondary endpoints address questions different from the primary pathology endpoint. Event-free survival and overall survival concern outcomes over follow-up, while pCR concerns the pathology result after neoadjuvant treatment. A secondary endpoint can add context, but it does not automatically confirm or overturn a primary finding. The registry does not report a set of complete, harmonized estimates for all secondary outcomes from which to judge whether their directions and magnitudes form a consistent pattern.

Unadjusted secondary results generally carry less weight when many outcomes or comparisons are examined. If each is tested without accounting for the number of tests, the chance of at least one apparently unusual result can increase. This is a general statistical concern, not evidence that a particular secondary result is false. The registry record does not report a multiplicity adjustment or a hierarchy for interpreting the secondary outcomes. Secondary findings should therefore be considered supportive or exploratory unless a more complete analysis plan establishes otherwise, and should not be treated as independent confirmation of the primary result.

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

• Overall Survival
• Event-free Survival
• Number of Adverse Events That Are Possibly, Probably or Definitely Related to Study Drug

9. Statistical Interpretation

10. Statistical Methods Explained

What does the primary endpoint count?

pCR is recorded as a count of participants meeting the pathology definition, alongside the number of participants in each analysis group. For a binary endpoint, each person falls into a response category or a non-response category. The reported counts therefore describe how often the endpoint occurred in each group, while the group denominators provide the context needed to compare those counts. A larger count alone does not establish a larger response proportion when group sizes differ.

The registry’s primary analysis compares Arm 1 with the combined Arm 2/3 group. The unadjusted recalculation uses the posted group summaries and expresses the contrast as a risk difference. A risk difference subtracts the event proportion in one group from the event proportion in the other, with the sign determined by the stated order of comparison. A value on the no-difference point corresponds to equal observed proportions. The direction of the estimate must always be read alongside the specified group order.

What does the chi-squared test assess?

The registry reports a chi-squared analysis for the primary comparison. In general, a chi-squared test assesses whether the observed distribution of a categorical outcome across groups is compatible with no association between group and outcome. It does not estimate the size of the difference in the same way as a risk difference, and its p-value does not express the probability that a treatment is effective. Here, the test result is reported as p = 0.019. The registry does not report further detail about the test implementation or any adjustment to that analysis.

The recalculated comparison is an unadjusted analysis of the posted summary statistics. It does not incorporate participant-level covariates or any adjustment not contained in those summaries. The primary comparison has a prespecified grouping stated in the registry, but additional details about the analysis population and handling of missing endpoint data are not reported in the registry record.

How should a p-value be read?

A p-value describes how incompatible the observed data are with a specified null model, assuming that model and the test assumptions apply. Smaller values indicate greater incompatibility with that null model; they do not tell how large or important an effect is. A p-value is also not the probability that the null hypothesis is true. Interpretation should consider the effect estimate, its confidence interval, the endpoint definition and the analysis plan, not the p-value alone.

The posted primary p-value and the recalculated unadjusted p-value are both 0.019. Their agreement is a useful check that the summary-statistic calculation is consistent with the reported result, but it does not prove that every analytic choice or protocol detail is known. No discrepancy is flagged in the registry record.

Why are secondary methods different?

The event-free survival outcome is analyzed with a log-rank test in the registry. In general, a log-rank test compares the timing of events between groups over follow-up, rather than comparing only a single binary status count. It is commonly used when participants can have different follow-up times. The registry describes disease assessment over time, but further details about censoring rules, survival estimates and the analysis population are not reported in the registry record.

The unadjusted recalculation for event-free survival uses the posted summary counts and reports a risk difference. That summary does not have the same structure as a time-to-event comparison, and its p-value need not match the log-rank p-value. The registry’s log-rank result is the sponsor-reported analysis for that endpoint. The two approaches answer related but not identical questions: a summary proportion compares event counts over the specified observation, whereas a time-to-event method incorporates event timing and follow-up information when available.

General statistical practice distinguishes estimation from testing. An estimate gives the direction and magnitude of a contrast, while an interval communicates uncertainty around that estimate. A test addresses compatibility with a null hypothesis under its assumptions. For a binary endpoint, risk differences and relative measures such as risk ratios or odds ratios can express the contrast on different scales. Each scale has its own interpretation; an odds ratio is not the same quantity as a risk difference. The registry supplies several such recalculated measures, but the central interpretation of the primary result can be made from the risk difference, its interval and the reported chi-squared result.

All inferential results depend on how the outcome and analysis population are defined. Randomization supports comparisons between assigned groups, but missing outcomes, exclusions after randomization, departures from assigned treatment and covariate adjustment can affect the interpretation of a particular analysis. Details needed to evaluate those issues are not reported in the registry record. The stated results should therefore be read as the posted group-level analysis, without assuming unreported analysis procedures.

11. Confidence Intervals and Precision

A confidence interval gives a range of values reasonably compatible with the observed data under the method used to calculate it. It communicates precision: a narrower interval indicates less statistical uncertainty about the estimated contrast, while a wider interval indicates more. The interval is not a range that contains every participant’s response, and it is not a guarantee that the true effect lies within its bounds.

For the primary risk difference, the recalculated interval lies entirely below the no-difference value. In the stated comparison order, that means the interval supports a lower pCR proportion in Arm 2/3 than in Arm 1. The interval still spans a range of possible effect sizes, so the result is more precise about direction than about the exact magnitude. The registry’s reported chi-squared p-value is consistent with a difference, but a p-value cannot substitute for the interval’s information about effect size and uncertainty.

The event-free survival risk-difference interval includes the no-difference value. That means its posted summary estimate is compatible with both some difference in either direction and no difference under the interval method. The interval should be interpreted as uncertain evidence, not proof that the groups are identical. The registry’s sponsor-reported log-rank analysis is a separate time-to-event analysis.

These recalculated intervals use posted summary statistics and are unadjusted. The registry record does not report adjusted confidence intervals for the primary comparison. Confidence intervals also do not account for every possible source of uncertainty, such as missing outcome data, analysis choices or multiple comparisons, unless those features are incorporated into the method.

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 planning calculations are illustrations based on the observed effect, not the trial’s original design assumptions. The registry record does not report the original sample-size justification. The stated planning assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The assumed effect is a stated fraction of the observed effect. For the binary primary endpoint, the figures are participants per arm. For the time-to-event endpoint, the figures are total events, using the Schoenfeld method. The planning entries therefore have different units and should not be compared as if they represented the same quantity.

The 80% power column describes the planned number of participants per arm, or total events for a time-to-event calculation, under which the assumed effect would have an 80% chance of producing a statistically significant result using the stated assumptions. The 90% power column uses a higher target probability of detecting that assumed effect. These are planning probabilities under a model, not guarantees about a future study. The effect fractions distinguish planning for the full observed effect from planning for smaller fractions of it. When the effect to detect is smaller, more participants or events are needed to distinguish it from random variation.

The curve represents the relationship between an assumed effect size and the calculated planning requirement. It does not show the trial’s observed power or a forecast of clinical benefit. These calculations are not observed power; they are illustrations of what a future design might require under the listed assumptions. They do not establish what the original trial planned, and they do not account for design features not included in the assumptions.

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

n is the number of participants in each arm, p₀ and p₁ are the control and treatment proportions, and p̄ is their average.

Worked calculation for this trial

1. Control proportion p₀ = 0.357; treatment proportion p₁ = 0.212; average p̄ = 0.285.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.6382 = 1.2509.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.6300 = 0.5302.
4. n = (1.2509 + 0.5302)² / (-0.1448)² = 151.4, rounded up to 152 participants per arm 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
p0 <- 0.357142857142857                          # control proportion
p1 <- 0.212389380530973                          # treatment proportion
n_per_arm <- function(power, frac = 1) {
  q1 <- p0 + (p1 - p0) * frac               # smaller effect = fraction of the observed difference
  pbar <- (p0 + q1) / 2
  z_b <- qnorm(power)
  ceiling(round((z_a * sqrt(2 * pbar * (1 - pbar)) + z_b * sqrt(p0 * (1 - p0) + q1 * (1 - q1)))^2 / (q1 - p0)^2, 6))
}
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: n per arm = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(n_per_arm(0.80, f), big.mark = ","), format(n_per_arm(0.90, f), big.mark = ",")))

# Check against R's built-in function (unrounded n per arm)
cat("power.prop.test:", round(power.prop.test(p1 = p0, p2 = p1, power = 0.80)$n, 1), "\n")

# Simulation: how often does the z-test reject at the 80%-power sample size?
set.seed(856492)
n    <- n_per_arm(0.80)
reps <- 20000
x0 <- rbinom(reps, n, p0); x1 <- rbinom(reps, n, p1)
ph <- (x0 + x1) / (2 * n)
z  <- ((x1 - x0) / n) / sqrt(ph * (1 - ph) * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(z) > z_a, na.rm = TRUE), reps))

Output from running the code

100% of observed effect: n per arm = 152 (80% power), 203 (90% power)
 75% of observed effect: n per arm = 281 (80% power), 375 (90% power)
 50% of observed effect: n per arm = 652 (80% power), 873 (90% power)
power.prop.test: 151.4 
Simulated power at n = 152 per arm: 0.805 (20000 replicates)

The simulation draws many pairs of event counts at the planned sample size and counts how often the two-proportion z-test rejects the null. It should land close to the target power; at very small required sample sizes it can fall below it, because the normal approximation is rough there.

14. Analyses Reported by the Sponsor

For pCR, the registry reports the sponsor’s chi-squared analysis with p = 0.019. The unadjusted recalculation from the posted group summaries also gives p = 0.019 for the primary comparison. The registry record does not flag a discrepancy between those results. The recalculation provides an unadjusted estimate and interval from the summary counts; it does not replace the sponsor’s own analysis or add information from participant-level data.

The sponsor’s result should be read in the context of the test it reports and the endpoint and group definition stated in the registry. A matching p-value is not evidence that all analytic details are identical, nor does it resolve questions about missing data, analysis populations or multiplicity. Those details are not reported in the registry record.

For event-free survival, the sponsor reports a log-rank p-value, while the recalculation is based on posted summary statistics and a risk difference. These are different methods, so their p-values need not agree. The sponsor-reported time-to-event result is the relevant registry analysis for that endpoint; the recalculation is a simple unadjusted comparison of the summary counts. Neither result should be treated as an adjusted analysis unless the registry reports such an adjustment.

15. Multiplicity, Missing Data and Analysis Populations

The registry identifies a prespecified primary comparison of Arm 1 with the combined Arm 2/3 group. It also lists secondary outcomes, including overall survival, event-free survival and adverse events related to study drug. Testing several outcomes can raise a multiplicity concern: without an adjustment or a clearly stated hierarchy, the chance of at least one nominally notable result can be greater than it would be for a single planned test. The registry record does not report a multiplicity adjustment, a testing hierarchy, or enough detail to determine how secondary results were intended to affect the primary conclusion.

The registry reports the primary result as a chi-squared test and the event-free survival result as a log-rank test. It does not report detailed analysis-population definitions, whether analyses followed intention-to-treat principles, or the handling of participants with missing endpoint assessments. For pCR, the analysis requires surgery pathology after treatment, but the registry does not report how unavailable, incomplete or unevaluable pathology was handled. For time-to-event analyses, censoring rules and detailed follow-up procedures are not reported in the registry record.

These omissions do not establish that the trial failed to use appropriate procedures; they limit what can be verified from the registry results. In particular, the posted group denominators should not be assumed to represent every randomized participant or to arise from a particular exclusion rule. The registry records enrollment as 215 and gives endpoint analysis denominators of 98 and 113 for the primary comparison; why those records differ is not reported in the registry record. Further detail on allocation concealment, protocol deviations, missing-data sensitivity analyses and adjusted analyses is also not reported.

16. Safety Results

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

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: neutrophil count decreased (5.2%), febrile neutropenia (4.2%), white blood cell decreased (4.2%), vomiting (2.1%), sepsis (2.1%).

The registry’s serious adverse event group records report 22 affected participants among 96 at risk in Arm I, 3 among 60 at risk in Arm II, and 3 among 51 at risk in Arm III. These are counts in the registry’s safety groups; they are not a calculation of comparative risk, and they do not by themselves establish that the study drug caused an event. The registry also lists other adverse events affecting 92 among 96, 58 among 60 and 50 among 51 in the respective groups.

The serious-event terms listed include neutrophil count decreased at 5.2%, febrile neutropenia at 4.2%, white blood cell decreased at 4.2%, vomiting at 2.1%, sepsis at 2.1% and skin infection at 2.1%. These are the registry-listed percentages; the record does not make the denominator and arm attribution for each listed term clear in the information posted. Death counts are not reported in the registry record. The registry’s listed related-adverse-event secondary outcome is distinct from these group-level safety counts, and its numerical result is not reported.

Differences in group size and the recorded safety denominators matter when interpreting counts. The registry record does not report enough information about event collection windows, severity grading details or exposure-adjusted analyses to make a fuller comparative safety assessment.

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

The primary result describes a difference in pathological complete response between the registry’s stated treatment groups. The posted counts and unadjusted statistical results indicate that pCR was more frequent in Arm 1 than in the combined Arm 2/3 group in this trial. The confidence interval communicates uncertainty around the size of that difference. This is evidence about the measured pathology endpoint in the studied population, not a direct measurement of long-term survival.

The trial was randomized, which is a design feature intended to support fair comparisons between assigned groups. However, the registry information alone does not establish every detail of trial conduct, endpoint completeness or analysis-population handling. The primary comparison also combines two different chemotherapy sequences in the comparator group. Its estimate should not be interpreted as a separate result for either sequence, or as proof that the regimens are otherwise identical.

The posted results do not establish that bevacizumab caused a particular long-term outcome, that any regimen is appropriate for an individual, or that the pCR difference translates into a corresponding survival difference. The registry identifies pCR as a surrogate associated with prognosis, but an association is not the same as demonstrating that a treatment-induced change in pCR produces a specified change in survival. Overall survival results are not reported in the registry record, and the secondary information is incomplete.

Nor does a result with a statistically notable p-value describe the size of benefit in clinical terms. That requires considering the endpoint’s meaning, uncertainty, harms and the broader evidence, which cannot be reconstructed from these registry results alone. The appropriate reading is narrower: the posted analysis reports a group difference in pCR, with a confidence interval and an accompanying sponsor analysis, while important questions about other outcomes and some analysis details remain unanswered in the registry record.

18. Why This Trial Matters Statistically

This trial is statistically interesting because its primary question is a randomized comparison of a bevacizumab-containing regimen with a combined group of regimens without bevacizumab. The endpoint is a clinically relevant pathology response measured after neoadjuvant treatment, while the registry also names longer-term outcomes. The design therefore illustrates the distinction between an early response endpoint and outcomes that require follow-up over time.

The comparison also shows why the definition of the analysis groups matters. The primary result pools Arms 2 and 3 even though their chemotherapy sequences differ. That choice addresses the registry’s stated comparison of Arm 1 versus Arm 2/3, but it does not answer whether the two non-bevacizumab sequences differ from one another. Readers should match any estimate to the precise contrast it represents.

Another statistical lesson is that an effect estimate, its confidence interval and a p-value contribute different information. The p-value describes compatibility with a null model under the test assumptions; the estimate gives a direction and scale; the interval conveys uncertainty about the magnitude. Looking at all of these together is more informative than treating a threshold as the whole conclusion.

Finally, the record illustrates the limits of registry outcome reporting. A listed secondary outcome is not the same as a fully reported numerical result, and a planning calculation based on the observed effect is not the original study’s design justification. Keeping those distinctions clear helps prevent a pathology result, an incomplete secondary record or a planning illustration from being given more weight than it can support.

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.

The primary comparison combines Arms 2 and 3, although their treatment sequences differ. The primary result therefore does not provide independent estimates for the two sequences. The registry’s arm description for Arm 1 lists later treatment in weeks 14, 16 and 18, but does not report the remainder of that schedule in the posted description. The registry record also does not report detailed eligibility criteria beyond the stated population, or the original sample-size justification.

The primary endpoint depends on post-treatment surgery pathology, but the rules for unavailable or unevaluable pathology are not reported in the registry record. The registry reports enrollment of 215 and primary analysis denominators of 98 and 113; the reason for the difference is not reported. Analysis-population definitions, missing-data methods, protocol deviations, covariate adjustment and sensitivity analyses are not reported. These limits make it difficult to assess how robust the posted estimates are to alternative analysis choices.

Outcome reporting is incomplete for the listed secondary outcomes: numerical overall survival and related-adverse-event results are not reported in the registry record. Further event-free survival details, including censoring and analysis-population information, are also not reported. The available serious adverse event counts do not include death counts, and the registry record does not report sufficient context about the collection window or term-level denominators to support a detailed comparative safety analysis.

Finally, the recalculated results are unadjusted comparisons of posted summary statistics. They cannot reproduce analyses that might rely on participant-level data or unreported design features. These limitations restrict interpretation of the record; they do not establish that unreported procedures were absent.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: S0800, Nab-Paclitaxel, Doxorubicin, Cyclophosphamide, and Pegfilgrastim With or Without Bevacizumab in Treating Women With Inflammatory or Locally Advanced Breast Cancer — NCT00856492.
• Registry record last updated 2017-06-27; status last verified 2017-05.
• Results first posted 2017-06-27.
• 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

The registry describes a completed, randomized, parallel phase II trial of neoadjuvant regimens in women with inflammatory or locally advanced HER-2/NEU negative breast cancer. Enrollment is recorded as 215. The primary endpoint is the number of participants with pathological complete response, comparing Arm 1, which included bevacizumab, with a combined Arm 2/3 group without bevacizumab.

The posted primary counts are 35 among 98 in Arm 1 and 24 among 113 in Arm 2/3. The recalculated unadjusted primary result and the registry-reported chi-squared result both have p = 0.019. The registry lists overall survival, event-free survival and related adverse events as secondary outcomes; numerical results are not reported for overall survival or the related-adverse-event outcome. The record does not report the original sample-size rationale, detailed missing-data methods, or a multiplicity adjustment.