Phase II Randomized Trial Evaluating Neoadjuvant Therapy With Neratinib and/or Trastuzumab Followed by Postoperative Trastuzumab in Women With Locally Advanced HER2-positive Breast Cancer: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Phase II Randomized Trial Evaluating Neoadjuvant Therapy With Neratinib and/or Trastuzumab Followed by Postoperative Trastuzumab in Women With Locally Advanced HER2-positive Breast Cancer (NCT01008150), a randomized phase 2 study. The primary outcome is Pathologic Complete Response in Breast and Axillary Lymph Nodes..

2. Clinical Question

The trial asks whether pathologic complete response (pCR) in the breast and axillary lymph nodes differs among neoadjuvant regimens that combine weekly paclitaxel with trastuzumab, neratinib, or both, followed by doxorubicin and cyclophosphamide (AC). The central randomized comparison is each investigational regimen against paclitaxel plus trastuzumab followed by AC.

The trial also records secondary outcomes concerning breast-only pCR, clinical complete response, recurrence-free interval, overall survival, and adverse events. Those outcomes address response at different sites or by different measures, follow-up after treatment, and toxicity. They should not be treated as interchangeable with the primary endpoint: the definitions, assessed populations, and timing differ.

3. Trial Design

This was a completed, randomized, open-label, parallel-group phase II trial sponsored by NSABP Foundation Inc. The registry gives an enrollment of 141 and identifies breast cancer as the condition studied. The recorded start was 2010-10 and primary completion was 2015-09. Masking was not used, so treatment assignment was not concealed from participants or study personnel as described by the registry design fields.

Three treatment groups were randomized. The control group received paclitaxel with trastuzumab, followed by AC. One experimental group received paclitaxel with neratinib, followed by AC. The other randomized experimental group received paclitaxel with trastuzumab and neratinib, followed by AC. The primary outcome counts in these groups were recorded for 42 participants per arm. The registry also reports a non-randomized group receiving the three-drug paclitaxel, trastuzumab, and neratinib combination before AC; its primary outcome count includes 12 participants.

The non-randomized group should be distinguished from the randomized comparison. Its results do not have the same allocation basis as the three randomized arms, and the smaller recorded group also makes its estimates less precise. The randomized comparisons are against the paclitaxel-plus-trastuzumab control regimen, not comparisons of every possible pair of groups.

The registry describes the main aim as determining the rate of pCR in the breast and axillary nodes after neoadjuvant therapy. Secondary aims include breast-only pCR, clinical complete response, two-year recurrence-free interval, two-year overall survival, toxicity, and exploration of molecular and genetic correlates of response. The registry record does not report the original sample-size justification, a detailed analysis plan, interim analyses, or further information about the study’s analysis populations. Those details therefore cannot be used to reconstruct how the trial was planned or how results were analyzed by the investigators.

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:

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 Pathologic Complete Response in Breast and Axillary Lymph Nodes., measured over At time of surgery, approximately 7 months.

The primary endpoint is the number of participants with no histologic evidence of invasive tumor cells in the surgical breast specimen and axillary nodes after neoadjuvant chemotherapy. It is assessed at surgery, approximately 7 months into the study. This is a binary response outcome: each assessed participant either meets the stated definition or does not.

Pathologic assessment uses tissue obtained at surgery, providing a direct measure of whether invasive tumor remains in the specified sites after preoperative treatment. Requiring both the breast specimen and axillary nodes to meet the definition makes this a combined endpoint. It is not the same as a response limited to the breast; the registry separately lists breast-only pCR as a secondary outcome.

In the control arm, 16 of 42 participants met the primary endpoint. In the paclitaxel-plus-neratinib arm, 14 of 42 did so, and in the randomized paclitaxel-plus-trastuzumab-plus-neratinib arm, 21 of 42 did so. The non-randomized group recorded 5 of 12 participants meeting the definition. These are counts of participants, not a time-to-event measure.

pCR is used in this trial as an outcome of neoadjuvant treatment response. The registry’s stated endpoint does not itself establish longer-term outcomes: recurrence-free interval and overall survival are listed separately. The record does not report a broader interpretation linking pCR in this trial to later outcomes, so those outcomes should be considered on their own terms.

7. Results Posted in the Registry

• Arm 2: Paclitaxel + Neratinib Then A C: 33.3% (14/42) versus 38.1% (16/42) on control; risk difference -4.8 percentage points (95% CI -25.2 to 15.7), risk ratio 0.88 (0.49–1.56); unadjusted two-proportion z-test p = 0.649.
• Arm 3: Paclitaxel + Trastuzumab + Neratinib Then A C: 50.0% (21/42) versus 38.1% (16/42) on control; risk difference 11.9 percentage points (95% CI -9.2 to 33.0), risk ratio 1.31 (0.80–2.14); unadjusted two-proportion z-test p = 0.272.
• Arm 3 NR: Paclitaxel + Trastuzumab + Neratininb Then AC: 41.7% (5/12) versus 38.1% (16/42) on control; risk difference 3.6 percentage points (95% CI -28.0 to 35.1), risk ratio 1.09 (0.51–2.37); unadjusted two-proportion z-test p = 0.823.

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.

Pathologic Complete Response in Breast.

• 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: Paclitaxel + Neratinib Then A C: 38.1% (16/42) versus 50.0% (21/42) on control; risk difference -11.9 percentage points (95% CI -33.0 to 9.2), risk ratio 0.76 (0.47–1.24); unadjusted two-proportion z-test p = 0.272.
  • Arm 3: Paclitaxel + Trastuzumab + Neratinib Then A C: 52.4% (22/42) versus 50.0% (21/42) on control; risk difference 2.4 percentage points (95% CI -19.0 to 23.8), risk ratio 1.05 (0.69–1.59); unadjusted two-proportion z-test p = 0.827.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Arm 2: Paclitaxel + Neratinib Then A C and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Arm 3: Paclitaxel + Trastuzumab + Neratinib Then A C and the control. The interval gives the range of effects that remain compatible with the data.

Clinical Complete Response, as Measured by Physical Exam

• 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: Paclitaxel + Neratinib Then A C: 80.6% (25/31) versus 69.4% (25/36) on control; risk difference 11.2 percentage points (95% CI -9.3 to 31.7), risk ratio 1.16 (0.88–1.53); unadjusted two-proportion z-test p = 0.293.
  • Arm 3: Paclitaxel + Trastuzumab + Neratinib Then A C: 82.4% (28/34) versus 69.4% (25/36) on control; risk difference 12.9 percentage points (95% CI -6.9 to 32.7), risk ratio 1.19 (0.91–1.55); unadjusted two-proportion z-test p = 0.208.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Arm 2: Paclitaxel + Neratinib Then A C and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Arm 3: Paclitaxel + Trastuzumab + Neratinib Then A C and the control. The interval gives the range of effects that remain compatible with the data.

Recurrence-free Interval (RFI)

• 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: Paclitaxel + Neratinib Then A C: 92.9% (39/42) versus 100.0% (42/42) on control; risk difference -7.1 percentage points (95% CI -14.9 to 0.6), risk ratio 0.93 (0.85–1.01); unadjusted two-proportion z-test p = 0.078.
  • Arm 3: Paclitaxel + Trastuzumab + Neratinib Then A C: 95.2% (40/42) versus 100.0% (42/42) on control; risk difference -4.8 percentage points (95% CI -11.2 to 1.7), risk ratio 0.95 (0.89–1.02); unadjusted two-proportion z-test p = 0.152.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Arm 2: Paclitaxel + Neratinib Then A C and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Arm 3: Paclitaxel + Trastuzumab + Neratinib Then A C and the control. The interval gives the range of effects that remain compatible with the data.

The secondary outcomes broaden the picture beyond the combined breast-and-node pCR endpoint. The registry records a separate breast-only pCR measure, clinical complete response by physical examination before surgery, and recurrence-free interval at 2 years. The secondary aims also include overall survival and toxicity, but the registry record does not report numerical overall-survival results among the secondary results.

Across the posted numerical secondary results, the point estimates do not all move in the same direction. Breast-only pCR and clinical complete response use different definitions from the primary endpoint, while recurrence-free interval concerns a later outcome. These measures should therefore be read as distinct outcomes, rather than as repeated measurements of one underlying result. The primary endpoint remains the combined pathologic assessment in breast tissue and axillary nodes at surgery.

The registry record reports no investigator analyses for the secondary outcomes. The comparisons available are unadjusted recalculations, and the secondary outcomes are additional to the primary question. When multiple outcomes and comparisons are examined, some apparent differences can occur by chance even if there is no corresponding underlying effect. Unless the analysis plan specifies how multiplicity is handled, unadjusted secondary p-values carry less weight as confirmatory evidence.

The registry does not report a multiplicity strategy or a hierarchy for interpreting secondary outcomes. It is therefore appropriate to treat these results as descriptive and exploratory rather than as a collection of independent definitive tests. The secondary outcomes can complement the primary result by showing how other response definitions and follow-up measures were recorded, but they cannot resolve the primary comparison’s uncertainty by themselves.

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

• Pathologic Complete Response in Breast.
• Clinical Complete Response, as Measured by Physical Exam
• Recurrence-free Interval (RFI)
• Overall Survival
• Adverse Events Experienced by Participants as a Measure of Toxicity

9. Statistical Interpretation

10. Statistical Methods Explained

The primary outcome is binary and is reported as a participant count: participants either met the histologic definition of pCR in both the breast and axillary nodes or did not. For an unadjusted comparison of two arms, the risk difference expresses the difference between the observed proportions meeting the endpoint. A negative value means the proportion is lower in the named experimental arm than in the control arm; a positive value means it is higher. This is a general interpretation of the statistic, not a statement about the cause of any observed difference.

What does the confidence interval describe?

A confidence interval gives a range of values compatible with the observed data under the method used to calculate it. General statistical interpretation: a wider interval indicates less precision, while a narrower interval indicates greater precision. An interval that includes zero does not establish that the true difference is exactly zero. It indicates that the data, under the calculation method, do not distinguish zero from the range of differences included in that interval at the corresponding confidence level.

For the comparison of paclitaxel plus neratinib with control, the risk difference was -0.048 and its interval extended from -0.252 to 0.157. For the randomized three-drug regimen versus control, the risk difference was 0.119 and the interval extended from -0.092 to 0.330. Both intervals include zero. The reported unadjusted p-values were 0.649 and 0.272, respectively. A p-value describes how unusual data at least as discrepant as those observed would be under a specified null model and assumptions. It is not the probability that the null hypothesis is true, nor a measure of clinical importance.

Why use unadjusted comparisons?

The registry record reports no analysis results for the primary endpoint, and the comparisons described are recalculated, unadjusted summaries from the posted counts. An unadjusted contrast compares the observed arm summaries without accounting for additional covariates. This makes the arithmetic and interpretation straightforward, but it does not reproduce an adjusted model or necessarily match an investigator’s prespecified analysis. The registry record does not report whether an adjusted primary analysis was planned or performed.

Risk ratios and odds ratios are alternative ways to express a binary-outcome contrast. A risk ratio compares the proportions with the endpoint, whereas an odds ratio compares the corresponding odds. Their numerical scale and interpretation differ from those of a risk difference; they should not be interchanged. The posted primary comparisons include risk-ratio and odds-ratio summaries, but the risk difference is used for the main interpretation because it states an absolute contrast in the observed endpoint proportions. No adjustment or causal interpretation follows from choosing one effect scale.

How should the non-randomized group be treated?

Random allocation supports a comparison of the randomized groups by making treatment assignment independent of participant characteristics on average, subject to the trial’s conduct and analysis. The extra group is explicitly identified as non-randomized. Its observed result can be described, but it should not be given the same comparative interpretation as the randomized treatment assignments. The record does not report further details about how participants entered that group or how it was analyzed.

General statistical explanation: in a binary outcome, a small number of participants can make the estimated proportion sensitive to each individual outcome. The non-randomized group contains 12 participants, compared with 42 in each randomized arm, and its interval is broad. This limits the precision of its contrast and reinforces the importance of distinguishing it from the randomized comparisons.

What do the planning illustrations mean?

The sample-size illustrations use the observed effect as a planning reference and show participant requirements for specified power levels and fractions of that effect. They are not a reconstruction of the trial’s original design. The registry record does not report the original sample-size justification or full statistical plan. The assumptions stated for the illustrations are a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation; the effect to detect is the stated fraction of the observed effect. The figures are planning aids, not evidence that the trial had a particular probability of detecting its actual effect.

General statistical explanation: power is the probability, under a specified alternative and statistical model, of rejecting a null hypothesis at the chosen significance level. Its value depends on the assumed effect, sample size, outcome variability, and analysis method. A planned power level is not an observed property of the results after the study. The sample-size calculations therefore answer hypothetical planning questions, not whether a result is true or whether a treatment has an effect.

These methods help separate the observed counts, estimated contrasts, and uncertainty from broader conclusions. The registry supplies the endpoint definition and arm counts, while important elements of the investigator-reported statistical plan are not reported in the registry record. Interpretation should accordingly stay close to the unadjusted results and acknowledge their precision and design limitations.

11. Confidence Intervals and Precision

The randomized primary comparisons are not estimated with high precision. For paclitaxel plus neratinib versus control, the risk difference was -0.048, and its 95% confidence interval ran from -0.252 to 0.157. For paclitaxel plus trastuzumab and neratinib versus control, the risk difference was 0.119, with a 95% confidence interval from -0.092 to 0.330. Both intervals include zero and span values consistent with differences in either direction.

These intervals are important because the point estimates alone can look more definite than the data warrant. The negative point estimate for the neratinib-only combination does not establish a lower response, and the positive point estimate for the three-drug randomized regimen does not establish a higher response. The interval widths show that the observed arm sizes leave considerable uncertainty about the magnitude of either contrast.

The non-randomized comparison is even less precise: its risk difference was 0.036, with a 95% confidence interval from -0.280 to 0.351. The group had 12 participants, and this estimate should also be separated from randomized evidence because allocation was not random.

In general, confidence intervals quantify sampling uncertainty under a calculation method and its assumptions. They do not incorporate every possible source of uncertainty, such as bias from missing data, deviations from the intended analysis, or limitations in outcome assessment. The registry record does not report enough detail to assess all such sources. The intervals therefore describe precision for the recalculated unadjusted contrasts, not the full uncertainty surrounding every interpretation of the trial.

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 calculations, not a record of the trial’s actual enrollment targets or original design rationale. The figures for binary outcomes are participants per arm for 80% power and participants per arm for 90% power, assuming equal allocation between two arms. They are not totals across both arms. The trial’s own original sample-size justification is not reported in the registry record.

The stated planning assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect to detect is set as the stated fraction of the observed effect. Entries labelled as planning for 100% of the observed effect use the full observed contrast as the target; the other rows consider smaller fractions. Smaller effects need more participants to detect, all else equal, which is why the illustrative requirements rise as the target effect becomes smaller.

The 80% power column describes the modeled sample size at which a study would have an 80% chance of detecting the specified effect under the stated assumptions. The 90% column applies the same idea at a higher probability of detection, generally requiring a larger sample. These are hypothetical design quantities, not findings from the enrolled participants. This is not observed power.

For time-to-event outcomes, the sample-size assumptions specify a different unit: total number of events, using the Schoenfeld method. The calculations shown for the binary endpoint and other non-time-to-event outcomes are participants per arm. A reader should therefore check the outcome type and unit before comparing any planning figure with enrollment. None of these illustrations establishes that the completed trial was designed to achieve a particular power or that a nonsignificant result reflects the absence of a meaningful effect.

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 = 0.381; treatment proportion p₁ = 0.333; average p̄ = 0.357.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.6776 = 1.3281.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.6768 = 0.5696.
4. n = (1.3281 + 0.5696)² / (-0.0476)² = 1,588.2, rounded up to 1,589 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 <- 0.380952380952381                          # control proportion
p1 <- 0.333333333333333                          # 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(1008150)
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 = 1,589 (80% power), 2,126 (90% power)
 75% of observed effect: n per arm = 2,845 (80% power), 3,809 (90% power)
 50% of observed effect: n per arm = 6,447 (80% power), 8,630 (90% power)
power.prop.test: 1588.2 
Simulated power at n = 1,589 per arm: 0.802 (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. Multiplicity, Missing Data and Analysis Populations

The registry record does not report a multiplicity adjustment, a prespecified hierarchy among the primary and secondary outcomes, or a detailed statistical analysis plan. This matters because the trial includes several outcomes and treatment comparisons. General statistical explanation: when many hypotheses are tested, the chance of at least one apparently notable result can increase. A multiplicity method or ordered testing plan can address that issue, but none is described in the registry record.

The posted comparisons are unadjusted recalculations from arm-level counts. They should not be mistaken for a complete investigator analysis. The record reports no analyses for the primary endpoint and no reported analyses for the listed secondary results. It is not possible from the registry record to determine whether analyses were conducted elsewhere, what model or covariate adjustments might have been used, or whether an analysis plan was finalized before results were examined.

Available population details are also limited. Arm-level denominators are reported for the outcomes, and the denominator differs for some secondary measures. The record does not report a full account of exclusions, missing outcome data, reasons for non-assessment, or whether the analysis followed an intention-to-treat principle. Consequently, the counts cannot establish how unassessed participants were handled or whether the reported groups correspond to a particular analysis population.

These omissions do not invalidate the posted counts, but they limit how fully the analyses can be audited and interpreted. Missingness can affect a binary outcome if the probability of being assessed is related to treatment or outcome status. General statistical explanation: handling missing data requires assumptions about why values are absent; different assumptions can lead to different estimates. Since the registry record does not describe missing-data methods, no such assumptions should be attributed to the investigators.

Accordingly, the unadjusted comparisons are useful for describing the recorded results, while conclusions about prespecification, multiplicity control, adjusted effects, and analysis populations remain limited by information not reported in the registry record.

15. 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: cellulitis (8.3%), cholecystitis (8.3%), pyrexia (4.8%), diarrhoea (4.8%), embolism (4.8%).

The adverse-event results report serious adverse events in 7 of 42 participants in the control arm, 7 of 42 in the paclitaxel-plus-neratinib arm, 10 of 42 in the randomized three-drug arm, and 2 of 12 in the non-randomized group. The registry also reports other adverse events affecting 41 of 42 participants in the control arm and all participants in each of the other listed groups: 42 of 42, 42 of 42, and 12 of 12, respectively.

These are participant counts within the registry’s adverse-event categories. They do not indicate how many events each person experienced, their duration, severity beyond the serious-event category, or whether an event was judged related to a particular treatment. The record lists cellulitis and cholecystitis at 8.3% each, and pyrexia, diarrhoea, and embolism at 4.8% each, among the most frequent serious terms; infection is listed at 2.4%. Those terms should be read as recorded event listings, not as a complete safety profile.

Deaths are not reported in the registry record. The counts alone do not support a direct claim of comparative safety: the groups differ in size and one group was non-randomized. Adverse-event interpretation also depends on definitions, observation time, severity grading, and attribution details that are not fully reported in the registry record.

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

The posted primary results show different observed pCR counts across the randomized arms, with 16 of 42 participants meeting the endpoint in the paclitaxel-plus-trastuzumab control arm, 14 of 42 in the paclitaxel-plus-neratinib arm, and 21 of 42 in the randomized arm receiving paclitaxel, trastuzumab, and neratinib. The recalculated unadjusted comparisons have confidence intervals that include zero and span a range of differences in either direction. The results therefore leave substantial uncertainty about the size and direction of the contrasts.

That uncertainty is not the same as proof that the regimens are equivalent, and it is not evidence that the observed differences are necessarily real. A nonsignificant unadjusted p-value does not answer whether an effect is absent or clinically important. The interval and study size both matter when interpreting the results.

The trial’s primary endpoint is a measure of pathologic response at surgery. It does not by itself establish effects on recurrence-free interval or overall survival, which are separate outcomes with different follow-up and interpretation. The posted secondary results offer additional outcome-specific information, but the registry record does not report investigator analyses or a multiplicity plan for them. The non-randomized group is descriptive and should not be treated as an additional randomized comparison.

These results also do not establish a treatment recommendation. They describe the registered trial outcomes and the precision of simple unadjusted contrasts. A complete assessment would require information not reported in the registry record, including the original sample-size justification, the full prespecified analysis plan, analysis populations, missing-data handling, and details of any adjusted or investigator-reported analyses. The conclusions that can be drawn should remain bounded by the endpoint definitions, recorded counts, and uncertainty in the available comparisons.

17. Why This Trial Matters Statistically

This trial is statistically informative because it places three randomized neoadjuvant regimens alongside a common control regimen and records a clearly defined pathologic endpoint. The randomized structure gives the main comparisons a stronger basis than a non-randomized contrast, while the open-label design and the limited registry analysis details remain relevant to interpretation.

The trial also illustrates why an observed difference and a precise estimate are not the same thing. The primary endpoint counts vary across arms, but the unadjusted confidence intervals for the randomized contrasts include zero and extend across differences in both directions. The uncertainty is central to the result; reading only the point estimates would overstate what these data can resolve.

It is equally important to distinguish the combined primary endpoint from related secondary outcomes. Breast-only pCR, clinical response, recurrence-free interval, and overall survival address different questions and time points. Their interpretation depends on definitions and analysis choices, and unadjusted secondary findings should not be treated as independent confirmation when the multiplicity strategy is not reported.

Finally, the registry’s sample-size illustrations help explain how planning requirements depend on the assumed effect size and desired power. They are not the trial’s original justification and are not observed power. Together, these features make the record a useful example of how design, endpoint definition, precision, multiplicity, and reporting completeness all shape the meaning of clinical trial results.

18. 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 registry record gives arm counts and unadjusted comparison summaries but does not report the investigators’ detailed statistical analysis plan or analyses. The primary and secondary comparisons described are recalculated from posted counts. They do not account for covariates, stratification, or other model features that may or may not have been part of an investigator analysis; such details are not reported in the registry record.

The analysis population and missing-data handling are not described in sufficient detail to assess the effect of participants whose outcomes may not have been available. Denominators vary for some secondary outcomes, but the registry record does not explain the differences. The primary endpoint counts use 42 participants in each randomized arm, while the clinical complete response denominators are 36, 31, 34, and 9. The reasons for these denominators are not reported in the registry record.

The additional combination group is non-randomized and includes 12 participants. Its comparison with control is not equivalent to the randomized comparisons, and its estimate is imprecise. The registry also does not report a multiplicity procedure despite multiple endpoints and comparisons. This limits the confirmatory interpretation of unadjusted secondary results.

The open-label design means masking was not used. The primary endpoint is pathologic, but the registry record does not provide further information about assessment procedures or whether outcome assessors were masked. The registry’s detailed treatment descriptions are incomplete in the recorded text, so regimen particulars beyond the broad group descriptions should not be inferred. Long-term overall survival results are not reported in the numerical secondary results. These limitations make cautious, endpoint-specific interpretation preferable to broad conclusions about the regimens.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Phase II Randomized Trial Evaluating Neoadjuvant Therapy With Neratinib and/or Trastuzumab Followed by Postoperative Trastuzumab in Women With Locally Advanced HER2-positive Breast Cancer — NCT01008150.
• Registry record last updated 2021-10-25; status last verified 2021-10.
• Results first posted 2018-12-06.
• 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.

22. Record Summary

NCT01008150 is a completed randomized phase II, open-label, parallel-group trial of neoadjuvant regimens for women with locally advanced HER2-positive breast cancer. It recorded an enrollment of 141 participants. The primary outcome was pathologic complete response in the breast and axillary lymph nodes at surgery, approximately 7 months after treatment began.

For the primary endpoint, the randomized arms recorded 16 of 42 participants in the control group, 14 of 42 in the paclitaxel-plus-neratinib group, and 21 of 42 in the paclitaxel-plus-trastuzumab-plus-neratinib group. Unadjusted recalculated comparisons have confidence intervals that include zero. An additional non-randomized group recorded 5 of 12 participants meeting the endpoint and should be interpreted separately.

The registry lists secondary outcomes and adverse-event counts, but does not report a detailed analysis plan, multiplicity strategy, original sample-size justification, or missing-data methods. The available results describe observed counts and uncertain unadjusted contrasts rather than a definitive comparative conclusion.