Neoadjuvant Her2-targeted Therapy and Immunotherapy With Pembrolizumab [IIT2018-04-MCARTHUR-NEOHP]: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Neoadjuvant Her2-targeted Therapy and Immunotherapy With Pembrolizumab [IIT2018-04-MCARTHUR-NEOHP] (NCT03747120), a randomized phase 2 study. The primary outcome is Pathological Complete Response (pCR).

2. Clinical Question

The primary response definition relies on pathologist assessment of the resected breast specimen and sampled regional lymph nodes after neoadjuvant systemic therapy. Thus, the question concerns a pathological response at surgery, not a direct comparison of long-term survival or recurrence. The registry also lists outcomes such as event-free survival, invasive disease-free survival, overall survival and cardiac events, but numerical results for those outcomes are not reported in the registry record.

3. Trial Design

This is a phase 2, open-label, randomized, parallel-group, multicenter trial sponsored by the University of Texas Southwestern Medical Center. The registry lists enrollment of 138 and a trial status of active, not recruiting. The listed start date is 2019-01-25, and the primary completion date is 2024-11-08. The registry record does not report the original sample-size justification or the planned allocation proportions.

Three treatment groups were studied. Arm A received paclitaxel weekly for 12 weeks with trastuzumab and pertuzumab. Arm B received the same regimen with pembrolizumab added. Arm C received paclitaxel weekly for 12 weeks with trastuzumab and pembrolizumab, without pertuzumab. After surgery, participants could receive standard-of-care systemic therapy at the treating physician’s discretion.

Randomization helps create groups for comparison, but the trial was not masked: treatment assignment was open-label. The reported primary outcome is pathological complete response in the breast and axilla, assessed at definitive surgery and framed as occurring 16 weeks from randomization. The primary population is intention-to-treat, including randomized participants regardless of whether they completed treatment. Participants who did not receive study medication were counted as non-responders for this outcome.

The efficacy summaries report 58 participants in Arm A, 58 in Arm B and 22 in Arm C. The smaller Arm C group means its comparison with Arm A is based on fewer participants than the Arm B comparison. The registry record does not report further details on the randomization procedure, concealment, site-level allocation, stratification, or the trial’s planned statistical analysis.

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: Data shown are for the intention-to-treat population, which is defined as any subject who was randomized regardless of treatment completion. Participants were classified as non-responders if they did not receive study medication.

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 Pathological Complete Response (pCR), measured over 16 weeks from randomization.

The primary endpoint is pathological complete response (pCR), defined as no residual invasive cancer in the breast and axilla after neoadjuvant systemic therapy. The registry describes assessment at definitive surgery using hematoxylin and eosin evaluation of the complete resected breast specimen and all sampled regional lymph nodes by a pathologist. The stated timeframe is 16 weeks from randomization.

This endpoint captures whether invasive disease remains in the examined breast and regional nodes at surgery. It is a binary outcome for each participant: classified as a responder or non-responder under the registry’s definition. For the primary analysis population, randomized participants are included regardless of treatment completion; anyone who did not receive study medication is classified as a non-responder. This rule makes treatment non-initiation count against response rather than excluding those participants from the outcome denominator.

Pathological response is used as an early measure of tumor response to neoadjuvant treatment. It differs from outcomes that require longer follow-up, such as event-free survival, invasive disease-free survival and overall survival. A pCR result therefore describes pathology at surgery and does not, by itself, establish what happens to participants over longer follow-up or provide a complete account of treatment harms.

The registry separately lists pCR in the breast only and invasive plus in situ response in the breast and axilla. Those are related but distinct endpoint definitions. The breast-and-axilla primary endpoint should not be treated as interchangeable with a breast-only endpoint, even where numerical entries happen to match in the posted results.

7. Results Posted in the Registry

• Arm B: THP-Pembrolizumab: 67.2% (39/58) versus 48.3% (28/58) on control; risk difference 19.0 percentage points (95% CI 1.3 to 36.6), risk ratio 1.39 (1.01–1.92); unadjusted two-proportion z-test p = 0.039.
• Arm C: TH-Pembrolizumab: 22.7% (5/22) versus 48.3% (28/58) on control; risk difference -25.5 percentage points (95% CI -47.3 to -3.8), risk ratio 0.47 (0.21–1.06); unadjusted two-proportion z-test p = 0.038.

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.

Pathological Complete Invasive and in Situ Response Rate (Breast and Axilla)

• 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 B: THP-Pembrolizumab: 51.7% (30/58) versus 43.1% (25/58) on control; risk difference 8.6 percentage points (95% CI -9.5 to 26.7), risk ratio 1.20 (0.82–1.77); unadjusted two-proportion z-test p = 0.353.
  • Arm C: TH-Pembrolizumab: 18.2% (4/22) versus 43.1% (25/58) on control; risk difference -24.9 percentage points (95% CI -45.5 to -4.4), risk ratio 0.42 (0.17–1.07); unadjusted two-proportion z-test p = 0.038.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Arm B: THP-Pembrolizumab and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically lower for Arm C: TH-Pembrolizumab than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Pathological Complete Response Rate (pCR) in Breast Only

• 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 B: THP-Pembrolizumab: 67.2% (39/58) versus 48.3% (28/58) on control; risk difference 19.0 percentage points (95% CI 1.3 to 36.6), risk ratio 1.39 (1.01–1.92); unadjusted two-proportion z-test p = 0.039.
  • Arm C: TH-Pembrolizumab: 22.7% (5/22) versus 48.3% (28/58) on control; risk difference -25.5 percentage points (95% CI -47.3 to -3.8), risk ratio 0.47 (0.21–1.06); unadjusted two-proportion z-test p = 0.038.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically higher for Arm B: THP-Pembrolizumab than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically lower for Arm C: TH-Pembrolizumab than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

The registry identifies several secondary outcomes intended to describe response beyond the primary pCR definition and to cover surgical, survival and cardiac outcomes. The posted numerical summaries include invasive plus in situ response in the breast and axilla and pCR in the breast only. The first of these has a different endpoint definition from the primary outcome. For the breast-only endpoint, the arm counts and comparison estimates match the posted primary outcome entries; that numerical match should not be taken to mean that breast-only and breast-and-axilla definitions are identical.

Across the posted response summaries, the direction of the Arm B comparison with Arm A is positive, while the Arm C comparison with Arm A is negative. The invasive-plus-in-situ comparison for Arm B has a risk-difference interval that includes zero; the corresponding Arm C interval is entirely below zero. These related results provide context for the primary response pattern, but they do not replace the primary endpoint or resolve its limitations.

Secondary results generally carry less weight because they may be among multiple outcomes examined, and the registry record does not report a multiplicity plan or a hierarchy for testing them. A nominal p-value from a secondary comparison can be more vulnerable to chance findings when many comparisons are considered. The record also does not report a separate analysis plan or named analyses for these outcomes.

Although the registry lists residual cancer burden, breast-conserving surgery, event-free survival, invasive disease-free survival, overall survival and cardiac events among secondary outcomes, their numerical results are not reported in the registry record. Consequently, the available response summaries cannot establish whether the observed pathology pattern is accompanied by differences in longer-term outcomes or other listed measures. Each secondary outcome also has its own definition, timeframe and analysis requirements, which should be considered rather than treating them as one collective result.

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

• Pathological Complete Invasive and in Situ Response Rate (Breast and Axilla)
• Pathological Complete Response Rate (pCR) in Breast Only
• Residual Cancer Burden (RCB)
• Breast Conserving Surgery Rate
• Event Free Survival
• Invasive Disease-free Survival (IDFS)
• Overall Survival
• Symptomatic Cardiac Events and Asymptomatic LVEF Events

9. Statistical Interpretation

10. Statistical Methods Explained

For a binary outcome such as pCR, the basic observation is whether each participant meets the response definition. The registry presents the number of responders and the number of participants in each arm. A response proportion describes the share of the arm classified as responding; the reported arm comparisons then express the difference between the comparator and the relevant treatment group on the risk-difference scale. Positive values in these comparisons indicate a higher response proportion in the treatment group than in Arm A, while negative values indicate a lower response proportion.

What does a risk difference express?

A risk difference is an absolute contrast between two response proportions. It keeps the comparison on the same scale as the outcome itself, so the estimate describes how far apart the groups are in response probability rather than expressing a relative multiple. In this trial, the estimates are 0.190 for Arm B versus Arm A and -0.256 for Arm C versus Arm A. Their signs indicate the direction of the observed comparisons; their confidence intervals describe the range of values compatible with the data under the interval method used.

How should the p-values be read?

A p-value measures how inconsistent the observed data are with a specified null hypothesis, under the assumptions of the test. It is not the probability that the null hypothesis is true, nor does it measure the importance or size of an effect. The primary comparisons have unadjusted p-values of 0.039 and 0.038. Because the registry does not report multiplicity handling, these values should be treated as unadjusted rather than as evidence of a confirmed trial-wide error-controlled finding. The registry record does not identify the exact test or adjustment strategy used for the displayed comparisons.

What does intention-to-treat contribute?

An intention-to-treat population retains randomized participants in the group to which they were assigned, regardless of treatment completion. This preserves the comparison created by randomization more closely than excluding participants based on treatment received or completed. For this endpoint, the registry explicitly states that participants who did not receive study medication were classified as non-responders. The handling of other kinds of missing endpoint data is not reported in the registry record.

Why does open-label status matter?

In an open-label study, treatment assignment is not masked. This can affect behaviors, co-interventions, assessments, or decisions made during care. Pathological assessment uses tissue examination, but the registry record does not report whether pathologists were blinded to treatment assignment or how assessment consistency was monitored. Therefore, no conclusion about the extent of possible assessment bias can be drawn from the record alone.

What can the secondary comparisons add?

Secondary endpoints can show whether observed patterns extend to other definitions or types of outcome. For binary secondary response measures, the same general distinction applies between an absolute response difference, its uncertainty, and a nominal p-value. The registry lists additional outcomes, including residual cancer burden, breast-conserving surgery, survival outcomes and cardiac events. Their listed presence does not mean numerical results or analysis details are available for each one; where results are absent, they are not reported in the registry record.

In general, when several endpoints or comparisons are examined, the chance of observing at least one small p-value can increase unless the analysis plan accounts for multiplicity. The registry record does not say whether a hierarchy, correction, or other multiplicity strategy was specified. This is why secondary and unadjusted results warrant more cautious interpretation than a clearly prespecified, multiplicity-controlled primary analysis.

Finally, the displayed effect estimates and intervals summarize statistical uncertainty conditional on the data and analysis. They do not account for every source of uncertainty, including the choice of endpoints, follow-up completeness, potential bias, or applicability to patients outside the trial’s eligibility criteria. The statistical summaries should be interpreted together with the trial design and the amount of information in each arm.

11. Confidence Intervals and Precision

The 95% confidence interval for the primary risk difference comparing Arm B with Arm A runs from 0.013 to 0.366. For Arm C compared with Arm A, it runs from -0.473 to -0.038. In general, a confidence interval gives a range of values compatible with the observed data under the method’s assumptions; it is not a guarantee that the true effect lies within that particular interval, nor does it describe the probability that an individual participant will respond.

The Arm B interval is entirely above zero, while the Arm C interval is entirely below zero. In the risk-difference scale, zero represents no difference between the groups. The interval limits indicate that the estimated direction is consistent across the range in each interval, but the width means the magnitude remains uncertain. In particular, the interval for Arm C spans a wide range of negative effects, reflecting the limited information in a comparison that includes 22 participants in that arm.

Confidence intervals should be considered alongside the arm counts and the endpoint definition. They quantify sampling uncertainty for the estimated comparison, but they do not capture every potential source of bias or uncertainty. Open-label conduct, unreported details of outcome assessment, and unreported handling of missing data beyond the stated non-responder rule are separate considerations.

The registry also gives relative-risk and odds-ratio summaries. These use different scales from a risk difference and can communicate a comparison in relative terms. The risk difference is useful for reading the absolute separation in response proportions, while the other measures express relative contrasts. Estimates on different scales need not have confidence intervals with identical implications. The record does not report adjusted estimates or a multiplicity strategy, so the unadjusted intervals should not be treated as definitive evidence across all outcomes and comparisons.

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 figures are participants per arm for the binary response comparisons, not total events. They are illustrations based on the observed effect, not the trial’s original sample-size design. The stated planning assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The scenarios use the full observed effect and smaller fractions of that observed effect.

The 80% power column describes the number of participants per arm that the planning calculation estimates would be needed to have an 80% chance of detecting the specified effect, under those assumptions, if that effect were the true one. The 90% power column uses a higher target probability of detecting it and therefore requires more participants per arm than the corresponding 80% target. These are planning concepts, not evidence that the completed trial had a particular probability of finding a result.

Smaller effects are harder to distinguish from random variation, so the planning illustrations require more participants when the assumed effect is reduced. The figures are comparisons of hypothetical design scenarios, not observed participant counts or an assessment of the trial’s success. They should not be interpreted as the trial’s original design rationale; the original design assumptions are not reported in the registry record.

This is not observed power. The completed trial’s estimates and confidence intervals describe the observed data and their precision; a retrospective power calculation based on the observed effect would not add an independent measure of evidence. The planning figures are most useful for showing how the required enrollment changes with the effect size and desired power under the stated calculation 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 = 0.483; treatment proportion p₁ = 0.672; average p̄ = 0.578.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.6985 = 1.3691.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.6855 = 0.5770.
4. n = (1.3691 + 0.5770)² / (0.1897)² = 105.3, rounded up to 106 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 <- 0.482758620689655                          # control proportion
p1 <- 0.672413793103448                          # 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(3747120)
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 = 106 (80% power), 141 (90% power)
 75% of observed effect: n per arm = 191 (80% power), 255 (90% power)
 50% of observed effect: n per arm = 434 (80% power), 581 (90% power)
power.prop.test: 105.3 
Simulated power at n = 106 per arm: 0.793 (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 trial compares two pembrolizumab-containing regimens with a common THP comparator and reports multiple secondary outcomes. This creates more than one opportunity to observe a nominally small p-value. Statistical methods often address this issue through a prespecified hierarchy, adjustment, or other multiplicity strategy. The registry record does not report whether any such plan was used, so the posted p-values are interpreted as unadjusted and should not be treated as controlling the overall chance of false-positive findings across the set of comparisons.

The record identifies the primary outcome population as intention-to-treat and specifies that randomized participants were included regardless of treatment completion. Participants who did not receive study medication were classified as non-responders. This is an explicit rule for those participants, but it does not fully describe how every possible missing endpoint assessment was handled. Further missing-data methods, including rules for participants without evaluable surgery or pathology information, are not reported in the registry record.

The registry reports no separate named analyses for the posted outcomes. It does not describe whether estimates were adjusted for baseline factors, stratification variables, study center, or other covariates. The exact inferential method behind the posted comparisons and any prespecified analysis population beyond the stated intention-to-treat definition are also not reported in the registry record.

The efficacy summaries include 58 participants in Arm A, 58 in Arm B and 22 in Arm C. The safety summaries use 20 participants at risk in Arm C. These denominators differ across outcome summaries, and the registry record does not explain the reason for the difference. They should therefore be kept distinct rather than assumed to represent the same analysis population. The record also does not report a detailed account of exclusions, follow-up completeness, or missingness by treatment group.

15. Safety Results

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

Deaths

Other (non-serious) adverse events

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

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: pneumonitis (5.0%), diarrhea (1.7%), anemia (1.7%), congestive heart failure (1.7%).

The serious adverse event summaries report 2 affected participants among 58 at risk in Arm A, 3 among 58 in Arm B, and 1 among 20 in Arm C. No deaths are reported in any of these groups. These counts describe the number of participants with a serious adverse event, not the total number of serious events or the duration of exposure. The registry record does not provide further detail about event timing, severity grading, relatedness to treatment, or how participants with multiple events were counted.

The registry also lists pneumonitis at 5.0%, and diarrhea, anemia and congestive heart failure at 1.7% each among the top serious adverse event terms. The record does not specify in this summary which treatment group each term belongs to, so these percentages should not be assigned to an individual arm. Other adverse events affected 39 of 58 participants in Arm A, 33 of 58 in Arm B and 11 of 20 in Arm C.

Safety counts are descriptive and should be read with their stated denominators. The Arm C safety denominator is 20, whereas its primary efficacy denominator is 22. The registry record does not explain that difference or report enough information to compare the clinical importance of individual events across arms.

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

The primary outcome data describe pathological response at surgery in this randomized phase 2 trial. The posted counts and unadjusted risk differences indicate a higher observed response proportion for Arm B than Arm A and a lower observed response proportion for Arm C than Arm A. The confidence intervals show that the estimates remain imprecise, especially for the smaller Arm C group, and the registry does not report multiplicity handling for the comparisons.

These results do not establish that one regimen is clinically superior overall. The endpoint is an early pathological measure, not a direct measure of long-term survival, recurrence, quality of life or net benefit. While the registry lists several longer-term and safety outcomes, their numerical results are not reported in the registry record. The response estimates alone therefore cannot answer whether any observed difference persists over time or outweighs potential harms.

The findings also do not provide a direct randomized comparison of Arm B against Arm C. Each posted comparison uses Arm A as the reference. Differences between the two pembrolizumab-containing arms cannot be inferred by comparing their separate estimates as if that were a prespecified head-to-head test.

Interpretation applies to the trial’s stated population: chemo-naive patients with invasive HER2-positive breast cancer meeting the tumor-size and/or clinical lymph-node criteria. The open-label design, limited size of Arm C, and unreported details of the analysis plan and missing-data handling all constrain interpretation. The registry status is active, not recruiting; the posted results should be considered within the scope and completeness of the record rather than as a full account of every endpoint or follow-up outcome.

In general, a statistical result is one component of evidence. The effect estimate gives the observed direction and scale, the confidence interval conveys precision, and the design and analysis details shape how much confidence to place in the comparison. None of these elements alone supports medical advice for an individual patient.

17. Why This Trial Matters Statistically

This trial illustrates the distinction between a randomized treatment comparison and a complete account of treatment effects. Randomization supports comparisons among assigned groups, while an open-label design and unreported analysis details leave questions about assessment and inference. Reading the estimates together with the confidence intervals prevents a small unadjusted p-value from being treated as a complete description of the evidence.

It also shows why a shared comparator does not automatically answer every treatment question. Arm B and Arm C were each compared with Arm A in the posted summaries, but the record does not report a direct comparison between the pembrolizumab-containing arms. Their separate contrasts should not be used as a substitute for that analysis.

The difference in group size is statistically relevant to precision. Arm C has fewer participants than Arms A and B, and its interval is broad. This illustrates why the direction of an estimate and the certainty about its magnitude are separate issues. The planning illustrations also show, in general, that smaller effects require more participants to detect under a given set of assumptions; those illustrations are not the original trial design rationale.

Finally, the primary outcome is a pathological response measure, while the registry lists additional outcomes spanning surgery, survival and safety. Statistical interpretation depends on the endpoint actually measured, the population analyzed, and the completeness of reported results. The record therefore supports a focused reading of pCR comparisons, not broad conclusions about all benefits and harms of the regimens.

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 open-label design means treatment assignment was known, and the registry record does not report whether pathology reviewers were blinded. It also does not describe the randomization method, allocation concealment, stratification, or site-level procedures. These omissions limit the ability to assess potential sources of bias from the registry information alone.

Arm C included 22 participants in the primary efficacy summary, compared with 58 in each of the other arms. This smaller group yields less information for its comparison with Arm A. The safety summary uses 20 participants at risk in Arm C, and the reason for the different denominator is not reported in the registry record. The record does not provide baseline characteristics or enough detail to assess how comparable the groups were in prognostic factors.

The primary endpoint is a short-term pathological outcome. Numerical results for the listed survival outcomes and several other secondary measures are not reported in the registry record. Follow-up duration and completeness for those outcomes are also not reported. The registry does not describe multiplicity control, detailed missing-data procedures, covariate adjustment, or the original sample-size justification.

The posted comparisons are unadjusted. Their intervals and p-values do not address possible bias, endpoint selection, incomplete reporting, or limitations in generalizability. The eligibility criteria also define a particular clinical population, so the results should not be assumed to apply to people outside those criteria. These are limitations of what can be concluded from the registry record, not proof that any particular bias occurred.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Neoadjuvant Her2-targeted Therapy and Immunotherapy With Pembrolizumab [IIT2018-04-MCARTHUR-NEOHP] — NCT03747120.
• Registry record last updated 2026-07-08; status last verified 2026-06.
• Results first posted 2025-05-08.
• 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

NCT03747120 is a randomized, open-label, parallel-group, multicenter phase 2 trial sponsored by the University of Texas Southwestern Medical Center. It compares neoadjuvant THP with THP-pembrolizumab and TH-pembrolizumab in chemo-naive patients with invasive HER2-positive breast cancer meeting the stated tumor or lymph-node criteria. Enrollment is listed as 138; the record lists the trial as active, not recruiting.

The primary outcome is pathological complete response in the breast and axilla, assessed at definitive surgery at 16 weeks from randomization in an intention-to-treat population. The posted counts are 28 of 58 in Arm A, 39 of 58 in Arm B and 5 of 22 in Arm C. Unadjusted comparisons with Arm A have risk-difference estimates of 0.190 for Arm B and -0.256 for Arm C. Interpretation is limited by the open-label design, uncertainty, multiple comparisons without a reported adjustment plan, and incomplete numerical reporting for several listed outcomes.