Neoadjuvant Sunitinib With Paclitaxel/Carboplatin in Patients With Triple-Negative Breast Cancer: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Neoadjuvant Sunitinib With Paclitaxel/Carboplatin in Patients With Triple-Negative Breast Cancer (NCT00887575), a non-randomized phase 1, phase 2 study. The primary outcome is Phase II: The Number of Subjects Exhibiting Pathologic Complete Response to Neoadjuvant Treatment With Sunitinib/Paclitaxel/Carboplatin.

2. Clinical Question

3. Trial Design

This is a completed Phase I/II clinical trial registered as NCT00887575. Its stated purpose was to examine the combination of sunitinib with paclitaxel and carboplatin as neoadjuvant treatment for locally advanced breast cancer. The registry describes the study as non-randomized, unmasked and single group. The enrolled population is listed as 54, but the record does not provide a participant-flow account that connects enrollment to each dose level, surgery or endpoint assessment.

The dose descriptions distinguish Dose Level I, Dose Level II and Dose Level III. Dose Level I lists paclitaxel IV at 70 mg/m^2 on days 1, 8 and 15 of each cycle, carboplatin IV at AUC = 5 on day 1 of every cycle, and sunitinib PO at 25mg daily. It also describes maintenance sunitinib PO at 25mg daily. Dose Level II lists paclitaxel IV at 80 mg/m^2 on days 1, 8 and 15 of each cycle, carboplatin IV at AUC = 5 on day 1 of every cycle, and sunitinib PO at 25mg daily. Dose Level III lists the same paclitaxel and sunitinib schedules as Dose Level II, with carboplatin IV at AUC = 6 on day 1 of every cycle.

The primary endpoint is explicitly labeled Phase II and concerns the number of subjects with pathologic complete response. Its defined population includes Dose Level I participants from Phase I and Phase II who underwent surgery. The registry record does not explain the dose-level assignment procedure, the rationale for each dose, or how participants moved between study phases. Because assignment was not randomized, dose-level comparisons cannot be treated as randomized evidence. The record also does not report blinding procedures beyond the unmasked classification, or a control treatment.

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.

Phase I

Phase II - Neoadjuvant Therapy

Phase II - Maintenance Therapy

Analysis population for the primary outcome, as defined in the registry: Patients treated at Dose Level I (Phase I and Phase II) who underwent surgery

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: Between June 2009 and July 2012, 54 patients were enrolled (15 in phase I and 39 additional patients in phase II). Three patients were later found to be ineligible and were excluded from the analysis. An additional patient did not receive any study treatment and is also excluded from the analysis.

6. Primary Endpoint

The registry lists the primary outcome as Phase II: The Number of Subjects Exhibiting Pathologic Complete Response to Neoadjuvant Treatment With Sunitinib/Paclitaxel/Carboplatin, measured over at weeks 26-30.

The primary endpoint is a count of participants with pathologic complete response, assessed at weeks 26-30. The registry defines pathologic complete response as no residual invasive breast cancer in final breast or axillary lymph node samples. This definition makes the outcome a binary participant-level classification: each assessed participant either meets the stated definition or does not. The registered endpoint is the number meeting it, rather than a continuous measurement or time-to-event measure.

Pathologic response is used in neoadjuvant studies to describe what remains in tissue samples after treatment and surgery. In general, a clearly defined binary endpoint can provide an interpretable count for a specified treatment group, provided the eligible population, assessment timing and handling of unevaluable cases are also clear. The endpoint’s meaning depends on the pathology definition and the population in which it is assessed; it does not by itself establish a causal treatment effect.

Here, the registered primary population is patients treated at Dose Level I in Phase I and Phase II who underwent surgery. That restriction matters: the endpoint does not automatically describe every enrolled participant, every dose level, or people without a surgical assessment. The registry record does not report the observed response count, the number assessed for this outcome, or a statistical estimate of response. It also does not report whether an independent pathology review was used or how missing or indeterminate pathology assessments were handled.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does a binary endpoint count?

The primary outcome is a participant count: the number who meet the pathologic complete response definition. In general, a binary endpoint assigns each eligible participant to one of two categories according to a prespecified definition. A count is interpretable only alongside its denominator, because the same count can represent different levels of response in populations of different sizes. A proportion can summarize the count relative to the number assessed, while a confidence interval can express uncertainty in an estimated proportion. The registry record gives neither a primary count nor its denominator, so neither a response estimate nor its precision can be described.

The outcome definition refers to residual invasive breast cancer in final breast or axillary lymph node samples. Its timing is listed as weeks 26-30. The population is specifically Dose Level I participants in Phase I and Phase II who underwent surgery. These details define the scope of the endpoint. General interpretation should not extend it to all enrolled participants, all dose levels, or people without surgical samples unless an analysis population is separately reported.

What does a confidence interval tell the reader?

In general, a confidence interval describes the statistical uncertainty around an estimate under the assumptions of the analysis method. A narrow interval indicates greater precision than a wide interval, while neither interval alone resolves possible bias in the design or data collection. For a single-group binary outcome, an interval around a response proportion would describe uncertainty about the response in the analyzed group, not a treatment effect relative to a control. No such interval is reported for the primary endpoint in this registry record.

Why does the design affect interpretation?

Random allocation can help make treatment groups comparable, on average, with respect to both measured and unmeasured baseline characteristics. This trial is listed as non-randomized, unmasked and single group. Its dose levels may describe different treatment schedules, but the record does not say that participants were assigned randomly among them. The registry’s caution that the arms are not a like-for-like control comparison is therefore central: differences across dose levels, if reported, could reflect differences in participants or study conduct as well as differences in dosing. A response count from one group is descriptive and does not by itself estimate what would have happened to the same participants under another treatment.

In general, masking can reduce some risks of biased outcome assessment or treatment decisions, depending on the outcome and procedures. The registry lists no masking. Pathology is a tissue-based endpoint, but the record does not report whether pathologists were blinded to treatment or whether samples underwent independent review. No conclusion about the presence or absence of those procedures can be drawn from the masking classification alone.

What can the registry’s analysis entry establish?

The record lists 0 analyses. This is a statement about analyses listed in the registry record, not proof that researchers performed no analyses. The record does not provide a primary endpoint estimate, a p-value, a confidence interval, or a model specification. It also does not report whether the primary analysis was planned as descriptive or inferential, how any uncertainty was calculated, or whether covariates were considered. Those features are not reported in the registry record and should not be reconstructed from the endpoint title.

General statistical practice distinguishes prespecified analysis methods from analyses chosen after results are observed. Prespecification can help readers assess whether results follow an intended plan. The registry record does not report a statistical analysis plan, hypothesis-testing procedure, adjustment method, or interim analysis approach. No claim about those procedures is warranted. The study is labeled Phase I/II, but that label alone does not establish a particular dose-finding model, stopping rule or sample-size rationale.

How should sample-size illustrations be understood?

General sample-size calculations are planning tools: they estimate the amount of information that might be needed to detect a specified effect under stated assumptions. They do not describe the observed strength of evidence in a completed study and should not be called observed power. The planning assumptions associated with the illustrations state a two-sided significance level of 0.05, equal allocation between two arms, a normal approximation, and an effect set as a stated proportion of an observed effect. For time-to-event outcomes, the assumption describes a total event count; other outcomes are expressed as participants per arm.

These are planning assumptions, not the design description of this trial. The registered design is single group and non-randomized, whereas the illustration assumptions refer to equal allocation between two arms. The record does not report a sample-size justification for the trial itself. It also does not report an observed primary endpoint effect from which a trial-specific planning illustration could be interpreted. Such calculations cannot turn this study into a randomized comparison, establish that a target effect was detected, or replace the missing primary endpoint estimate.

Finally, the registry’s safety entries are counts and listed percentages, but it does not report a comparative safety model or uncertainty intervals. General comparisons of event rates require clearly defined risk periods, event definitions and denominators. Without further detail, the entries are best treated as descriptive registry reporting rather than as adjusted or causal estimates.

10. Exact Estimates for Each Arm

No two-arm comparison is shown for this trial, but each arm's result can still be summarized with exact interval estimates computed from the posted counts. The Clopper-Pearson interval is exact and conservative; the Wilson interval is a common closer-to-nominal alternative; the Bayesian credible interval uses a uniform Beta(1, 1) prior. For small groups the three can differ noticeably.

R code and output

arm    <- c("Phase II- Sunitinib/Paclitaxel/Carboplatin")
n      <- c(32)
events <- c(12)
for (i in seq_along(n)) {
  ci <- binom.test(events[i], n[i])$conf.int          # Clopper-Pearson exact interval
  cr <- qbeta(c(0.025, 0.975), 1 + events[i], 1 + n[i] - events[i])   # Bayesian, uniform Beta(1,1) prior
  wi <- (function(x, n) { p <- x / n; z <- qnorm(0.975); d <- 1 + z^2 / n
    c((p + z^2 / (2 * n) - z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d, (p + z^2 / (2 * n) + z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d) })(events[i], n[i])
  cat(sprintf("ARM|%s|%d|%d|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f\n", arm[i], n[i], events[i], 100 * events[i] / n[i],
              100 * ci[1], 100 * ci[2], 100 * wi[1], 100 * wi[2], 100 * cr[1], 100 * cr[2]))
}

ARM  Phase II- Sunitinib/Paclitaxel/Carboplatin  32  12  37.5  21.1  56.3  22.9  54.7  22.9  54.9

11. Multiplicity, Missing Data and Analysis Populations

The registry lists a primary endpoint and several secondary outcomes: adverse events as a measure of safety and tolerability, overall response rate, disease-free survival, and overall survival. It does not report a multiplicity strategy, such as a hierarchy for testing outcomes or an adjustment for conducting multiple statistical tests. In general, when several outcomes are tested, the chance of at least one apparently unusual result can increase unless the analysis plan addresses that issue. No conclusion about the control of that risk can be made for this trial because the registry record does not describe such a plan.

The registry lists 0 analyses, and it does not report inferential results for the primary or secondary outcomes. The absence of analyses in the record should not be interpreted as evidence that investigators did not analyze the data. It means that estimates, uncertainty measures and testing procedures cannot be assessed from this registry entry. Interim analyses and any stopping rules are not reported in the registry record.

Missing-data handling is also not reported. The primary outcome population includes participants who underwent surgery, but the registry does not state how many participants were eligible for that assessment, how many had evaluable pathology, or how unevaluable or missing results were treated. A participant who lacks an endpoint assessment cannot automatically be classified as either having or not having a pathologic complete response. In general, missingness can affect both the denominator and interpretation of a binary outcome, especially if the reason for missingness is related to prognosis or treatment experience.

The record does not provide a broader analysis-population definition, such as a rule for including participants according to treatment received or assigned. Nor does it report whether safety analyses used all treated participants or another population. These details are not reported in the registry record, so no specific intention-to-treat, per-protocol or safety population should be assumed.

12. 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: fatigue (4.9%), nausea (4.9%), neutropenia (4.9%), thrombocytopenia (4.9%), vomiting (4.9%).

The registry safety entry lists 41 participants at risk for serious adverse events and 3 participants affected. These figures describe the serious adverse-event summary for Dose Level I. The record also lists 41 at risk and 41 affected for other adverse events. They are registry counts, and the record does not supply a fuller account of how adverse events were collected, categorized or attributed. The 41 at-risk entry should not be assumed to be the denominator for every endpoint or every outcome in the trial.

The listed serious adverse-event terms include fatigue, nausea, neutropenia, thrombocytopenia and vomiting, each with a recorded percentage of 4.9%, and acute renal failure at 2.4%. The registry does not explain whether these term entries refer to participants, events or another reporting unit, whether participants could have more than one term, or how the percentages were calculated. The term list is therefore not a complete description of event severity, duration, timing or clinical context.

In general, adverse-event counts are descriptive unless the record also provides a suitable comparison, consistent follow-up, clear definitions and an analysis addressing uncertainty. This trial is non-randomized and single group, so these entries do not establish comparative risk or causation. The registry record does not report an attribution method, event grades, exposure-adjusted rates, or uncertainty estimates. The figures should be read as the safety information recorded for this group, not as a complete estimate of harms across all participants or as evidence about the contribution of any individual drug.

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

The record establishes that a completed Phase I/II study examined neoadjuvant sunitinib with paclitaxel and carboplatin in patients with locally advanced triple-negative breast cancer. It describes three dose levels, identifies a pathologic complete response endpoint, defines that endpoint using final breast or axillary lymph node samples, and lists selected safety figures. These features clarify the question the study was designed to address and the scope of its primary endpoint.

However, the registry does not report the primary response count or the number of participants assessed for that outcome. It therefore does not show the observed response frequency or how precisely it was estimated. The enrollment entry of 54 cannot fill that gap because the primary endpoint applies to the subset treated at Dose Level I who underwent surgery, and the record does not provide the size or flow of that subset. No conclusion about the primary outcome can be drawn from the endpoint definition alone.

The study design also limits causal interpretation. Non-randomized allocation and a single-group model do not provide a randomized control comparison. Dose-level descriptions are not evidence of an unbiased dose comparison, and differences between groups, if available elsewhere, would not automatically represent effects caused by the dosing differences. The registry’s own caution says the arms are not a like-for-like control comparison.

The safety information is similarly limited to descriptive entries. It records serious adverse events and selected event terms, but does not provide the definitions, follow-up details or comparative analyses needed to interpret relative safety. A listed event count cannot, on its own, show that a treatment caused an event or establish how its risk compares with another treatment.

In general, a complete assessment of a clinical trial result considers the endpoint estimate, its uncertainty, the analysis population, missing data, study design and prespecified analysis methods. Several of these elements are not reported in the registry record. The posted information can be used to understand the registered design and selected recorded outcomes, but it cannot establish treatment benefit, lack of benefit, or comparative safety.

14. Why This Trial Matters Statistically

This trial is statistically notable because its primary endpoint is narrowly defined, while its analyzed population is also specifically restricted. The endpoint concerns a pathologic finding at surgery, and the registry identifies participants treated at Dose Level I in Phase I and Phase II who underwent surgery. This makes the distinction between total enrollment and the endpoint population especially important. Without participant flow and an endpoint denominator, readers cannot tell how much information supports the primary outcome.

The non-randomized, single-group design is another central feature. In general, an outcome count from a treatment group can describe what was observed among assessed participants, but it cannot by itself separate the treatment’s contribution from other explanations. The existence of multiple dose descriptions does not remove that limitation. The registry explicitly cautions that the arms are not like-for-like control groups, which prevents an unjustified interpretation of dose-level differences as randomized effects.

The record also illustrates why outcome definitions and statistical results need to be considered together. A clear pathologic response definition identifies what is being counted, but without the count, denominator and uncertainty, the magnitude and precision of the result remain unknown. Similarly, selected adverse-event figures can signal what was recorded without providing a full comparative safety analysis.

Finally, the registry entry lists no analyses. This limits what can be learned about the statistical methods actually used, including missing-data handling, multiplicity, uncertainty estimation and analysis populations. The entry remains useful for understanding the planned clinical question and design, but its statistical value for evaluating outcomes is constrained by the absence of reported endpoint estimates and analysis details.

15. 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 main limitation of the registry record is that it does not report the primary endpoint result. Although the response definition and assessment timeframe are stated, no response count, endpoint denominator, response proportion or confidence interval is given. As a result, neither the observed frequency nor the precision of pathologic complete response can be assessed.

The registered primary population is limited to Dose Level I participants from Phase I and Phase II who underwent surgery, while total enrollment is listed as 54. The record does not report how many participants were in each phase or dose level, how many underwent surgery, or how many had an evaluable pathology assessment. It also does not give reasons for missing or unavailable endpoint data. The relationship between enrollment and the primary analysis population is therefore unclear.

The single-group, non-randomized design limits causal and comparative conclusions. No randomized control treatment is described, and the registry cautions that the dose-level arms are not like-for-like comparisons. The record does not explain how dose levels were assigned, whether participants could change dose level, or how treatment exposure was handled in analysis.

Important statistical details are not reported in the registry record, including the statistical analysis plan, sample-size rationale for the trial, inferential method, multiplicity approach, interim analyses, stopping rules, and analysis populations. The record also does not describe pathology review procedures or whether assessment was blinded. These absences prevent assessment of potential analytic and measurement biases.

Safety reporting is limited to the recorded counts and selected serious adverse-event terms. Definitions, grading, attribution, event timing and uncertainty are not reported. The entries cannot be treated as a complete safety profile or a comparative estimate. These limitations describe gaps in the registry record; they do not establish that the trial or its investigators lacked procedures or analyses that were not entered into the registry.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Neoadjuvant Sunitinib With Paclitaxel/Carboplatin in Patients With Triple-Negative Breast Cancer — NCT00887575.
• Registry record last updated 2016-09-07; status last verified 2016-09.
• Results first posted 2014-12-17.
• 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.

19. Record Summary

NCT00887575 is a completed Phase I/II, non-randomized, unmasked, single-group trial of neoadjuvant sunitinib with paclitaxel and carboplatin for locally advanced triple-negative breast cancer. The registry lists enrollment of 54 and describes three dose levels. The primary endpoint is the number of Dose Level I participants in Phase I and Phase II who underwent surgery and had no residual invasive breast cancer in final breast or axillary lymph node samples at weeks 26-30.

The registry record does not report the primary endpoint count, its denominator, or an analysis result. It lists 0 analyses. Safety entries include 3 participants with serious adverse events among 41 at risk and selected serious-event terms with listed percentages. Because the design is non-randomized and lacks a like-for-like control comparison, the record supports a description of the study framework and recorded safety entries, but not a causal or comparative estimate of treatment effect.