A Study To Evaluate the Efficacy and Safety Of Atezolizumab or Placebo in Combination With Neoadjuvant Doxorubicin + Cyclophosphamide Followed By Paclitaxel + Trastuzumab + Pertuzumab In Early Her2-Positive Breast Cancer: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study To Evaluate the Efficacy and Safety Of Atezolizumab or Placebo in Combination With Neoadjuvant Doxorubicin + Cyclophosphamide Followed By Paclitaxel + Trastuzumab + Pertuzumab In Early Her2-Positive Breast Cancer (NCT03726879), a randomized phase 3 study. The primary outcome is Percentage of Participants With Pathological Complete Response (pCR) in the PD-L1-Positive Population (IC 1/2/3).

2. Clinical Question

3. Trial Design

NCT03726879 was a Phase III, randomized, double-blind, placebo-controlled, parallel-group trial sponsored by Hoffmann-La Roche. The registry record gives the study status as completed. The recorded start date is 2019-01-11, and the primary completion date is 2021-02-05. Enrollment is listed as 454.

The trial compared atezolizumab with placebo when each was combined with neoadjuvant dose-dense doxorubicin plus cyclophosphamide, followed by paclitaxel, trastuzumab, and pertuzumab. The record identifies the regimen as ddAC-PacHP. Participants were allocated to the atezolizumab or placebo group under double masking; the shared background regimen makes the randomized contrast focused on the atezolizumab component.

The primary endpoint was the percentage of participants with pCR in the PD-L1-positive population. The record defines this population as participants in the Intent-to-Treat population whose PD-L1 status was IC1/2/3 at randomization. The pCR assessment was made after completion of neoadjuvant systemic therapy, on examination of the resected breast specimen and sampled regional lymph nodes.

The registry also lists secondary outcomes covering pCR in other populations, event-free and disease-free survival, overall survival, patient-reported function and global health status, and adverse events. The record includes additional outcome measures, but the posted primary result is the key basis for the principal treatment comparison. Details of the trial’s original sample-size justification and other design assumptions are not reported in the registry record.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Before assignment: A total of 669 participants were screened, of which a total of 454 participants were enrolled.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: The PD-L1-positive population is defined as participants in the Intent-to-Treat (ITT) population whose PD-L1 status is IC1/2/3 at the time of randomization.

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 Percentage of Participants With Pathological Complete Response (pCR) in the PD-L1-Positive Population (IC 1/2/3), measured over From randomization to approximately 6 months.

The primary endpoint was the percentage of participants with pathological complete response, or pCR, in the PD-L1-positive population. The registry defines pCR as the absence of residual invasive cancer on hematoxylin and eosin evaluation of the complete resected breast specimen and all sampled regional lymph nodes after neoadjuvant systemic therapy. The stated staging definition is ypT0/is ypN0 under the current American Joint Committee on Cancer staging system, 8th edition.

This endpoint records whether invasive disease remains detectable in the specified surgical specimens after preoperative treatment. It therefore summarizes tumor response at the time of surgery, rather than measuring the duration of disease control or survival. pCR is useful in this setting because it provides a direct pathological assessment of response following neoadjuvant treatment and allows a comparison between randomized groups.

The endpoint was assessed from randomization to approximately 6 months. The target population was not all participants without qualification: it was the PD-L1-positive subset, defined by status IC1/2/3 at randomization within the Intent-to-Treat population. The primary result should consequently be interpreted as applying to that defined population and endpoint. It does not, by itself, answer whether treatment affects the separately listed survival outcomes or quality-of-life measures.

7. Results Posted in the Registry

• Atezolizumab +ddAC-PacHP: 64.2% (70/109) versus 72.5% (79/109) on control; risk difference -8.3 percentage points (95% CI -20.6 to 4.0), risk ratio 0.89 (0.74–1.06); unadjusted two-proportion z-test p = 0.190.

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.

Percentage of Participants With pCR in the PD-L1-Negative Population

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Atezolizumab +ddAC-PacHP: 60.7% (71/117) versus 53.8% (64/119) on control; risk difference 6.9 percentage points (95% CI -5.7 to 19.5), risk ratio 1.13 (0.90–1.41); unadjusted two-proportion z-test p = 0.284.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Atezolizumab +ddAC-PacHP and the control. The interval gives the range of effects that remain compatible with the data.

Cmin of Trastuzumab Emtansine in Serum

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • Atezolizumab +ddAC-PacHP: mean 2.18 versus 1.22 on control; difference 0.96 (95% CI -0.47 to 2.39), standardized effect d = 0.25; z-test p = 0.189.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Atezolizumab +ddAC-PacHP and the control. The interval gives the range of effects that remain compatible with the data.

The registry lists secondary outcomes that extend beyond the primary pCR comparison. They include pCR in other populations, event-free survival, disease-free survival, overall survival, patient-reported function and global health status, and adverse events. These outcomes address different aspects of response, longer-term disease outcomes, experience of health, and safety; they should not be collapsed into a single overall treatment effect.

The secondary pCR result in the PD-L1-negative population asks a related but distinct question from the primary analysis in the PD-L1-positive population. Differences between subgroup estimates do not, by themselves, establish that treatment effects differ by PD-L1 status. That requires an appropriate comparison of effects across the populations, and such an interaction result is not reported in the registry record.

Other secondary outcomes use different measurements and follow-up. Survival endpoints concern events over time, whereas patient-reported measures concern changes in health or function, and pharmacokinetic measures describe drug concentration. Their results cannot be interpreted as direct replications of the primary pCR comparison. The available registry summaries do not support a single coherent conclusion across all of these distinct outcomes.

Secondary findings generally carry less inferential weight when many outcomes are examined and the analysis plan does not report how multiplicity is handled. Testing many outcomes creates more opportunities for apparently noteworthy results by chance. The registry record does not report a multiplicity adjustment or a hierarchical testing strategy. Secondary estimates can inform further study and provide context, but they should not override the prespecified primary comparison or be treated as independently confirmatory without details of the testing plan.

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

• Percentage of Participants With pCR Based on Hormone Receptor Status
• Percentage of Participants With pCR in the PD-L1-Negative Population
• Event-Free Survival (EFS)
• Disease-Free Survival (DFS)
• Overall Survival (OS)
• Mean Changes From Baseline in Function (Role, Physical)
• Mean Changes From Baseline in Global Health Status
• Percentage of Participants With Adverse Events

9. Statistical Interpretation

10. Statistical Methods Explained

The primary outcome is binary: each participant either met the registry’s definition of pCR or did not. A group percentage expresses the proportion meeting that definition. Comparing percentages is intuitive, but the numerical result depends on the scale used. A risk difference describes the subtraction of the two proportions; an odds ratio compares the odds of response between groups. These measures answer related but not identical questions, and their numerical values should not be read as interchangeable.

The sponsor reports an odds ratio from a Cochran-Mantel-Haenszel analysis. This method combines information across strata rather than treating every participant as if the stratification variables were irrelevant. In this trial the record states that treatment comparisons were stratified by disease stage, T2 versus T3-4, and hormone receptor status, defined using estrogen receptor and/or progesterone receptor positivity. Stratification can make the comparison account for those factors in the analysis, while the randomized allocation remains the basis for the treatment contrast.

An odds ratio of 1 represents equal odds in the two groups. Values below or above 1 indicate a contrast in opposite directions, according to which group is the numerator or reference group. The reported estimate of 0.67 is below 1, while its 95% confidence interval, 0.37 to 1.21, crosses 1. That interval is therefore compatible with the no-difference value. The estimate alone is not enough: its uncertainty is essential to understanding how precisely the trial estimates the contrast.

The posted group percentages can also be used for a simple, unadjusted risk-difference calculation. This recalculation uses the displayed summaries and does not reproduce the sponsor’s stratified method. It is useful as a descriptive comparison on an absolute scale, but it does not account for the stated strata in the way the sponsor analysis does. The recalculated p-value is 0.19; the sponsor-reported p-value is 0.185. Differences in displayed effect measure, adjustment, and rounding mean these quantities should not be treated as duplicate outputs of a single analysis.

A confidence interval describes a range of effect estimates compatible with the data under the method’s assumptions. It is not a statement that the true effect has a specified probability of lying inside the observed interval. A wide interval signals limited precision; an interval spanning the no-difference value indicates that the data do not distinguish that value from a range of alternatives at the stated confidence level. The interval’s inclusion of effects in both directions matters as much as the point estimate’s direction.

A p-value measures how unusual data at least as incompatible with the null hypothesis would be if that null and the analysis assumptions held. It is not the probability that the null is true, and it does not measure the size or clinical importance of an effect. A result with p = 0.185 is not evidence that the groups are identical. Conversely, the direction of a point estimate does not establish that a treatment difference is real. The estimate, its interval, the design, and the outcome definition need to be considered together.

Randomization supports a fair comparison by assigning participants to treatment groups without choosing treatment according to measured or unmeasured prognosis. Double masking and placebo control can reduce the influence of treatment knowledge on care and outcome assessment. Those design features do not remove uncertainty due to the number of observed responses, missing outcome information, or limitations in what the endpoint measures. Details on blinding procedures, allocation concealment, and handling of missing primary endpoint data are not reported in the registry record.

The registry defines the primary PD-L1-positive population within the Intent-to-Treat population, using IC1/2/3 status at randomization. This makes the biomarker definition part of the question being answered. The record describes the pCR comparison as stratified by the stated clinical factors, but further detail on model implementation and any additional covariate adjustments is not reported in the registry record. The registry record also does not report whether interim analyses were planned or performed, or how the original sample size was justified.

These methods concern a response endpoint observed after neoadjuvant treatment and surgery. They do not by themselves estimate a treatment effect on time-to-event outcomes. Event-free survival, disease-free survival, and overall survival are separately listed outcomes and require methods appropriate to their follow-up and event data. Results for those outcomes should not be inferred from the pCR analysis.

11. Confidence Intervals and Precision

The sponsor’s primary estimate is an odds ratio of 0.67 with a 95% confidence interval from 0.37 to 1.21. The interval includes 1, the value representing no contrast in odds. It also covers values below and above 1, so the data are compatible with a range of possible directions and magnitudes. The point estimate suggests lower odds of pCR in the atezolizumab group under the reported comparison, but the interval shows that this direction is not estimated precisely enough to rule out no difference or an effect in the opposite direction.

The group percentages also have uncertainty intervals: 54.47 to 73.17 for atezolizumab and 63.1 to 80.6 for placebo. The intervals overlap, but overlap between separate group intervals is not itself the formal test of the treatment contrast. The sponsor’s confidence interval for the odds ratio is the direct inferential interval for the reported comparison.

The unadjusted recalculated risk-difference interval also extends across zero. Zero is the no-difference value on that scale. This aligns with the broad conclusion that the posted summaries do not yield a precise distinction between treatment groups, while the recalculation remains a different, unadjusted analysis from the sponsor’s stratified result.

Confidence intervals do not capture every source of uncertainty. They reflect sampling uncertainty under the analysis method; they do not resolve questions about missing data, outcome assessment, applicability to other populations, or multiplicity across outcomes. Those issues require information beyond the interval itself, and several relevant analysis details are not reported in the registry record.

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 illustrations based on the observed effect, not the trial’s original design calculation. For this binary pCR outcome, the entries are participants per arm. The general convention in the planning material is that time-to-event outcomes use total events, while other outcomes use participants per arm; the present endpoint is not a time-to-event outcome.

The planning assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The scenarios consider an effect equal to 100% of the observed effect, 75% of the observed effect, and 50% of the observed effect. The estimates are planning illustrations for detecting those assumed effects under those assumptions; they are not the actual enrollment or outcome counts in this completed trial.

The 80% power column represents the planned per-arm sample size under which the calculation would have an 80% chance of detecting the assumed effect at the stated significance level, if that effect and the calculation assumptions were true. The 90% column uses a higher target chance of detecting that assumed effect and therefore generally calls for more participants. A smaller assumed effect is harder to distinguish from random variation, so the planning requirement rises as the target effect is reduced.

These figures are not observed power. Observed power is not a useful substitute for reading the estimated effect and its confidence interval after a trial has finished. The original trial’s own sample-size assumptions and justification are not reported in the registry record, so the illustrative calculations should not be taken as evidence about how the study was designed.

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 = 0.725; treatment proportion p₁ = 0.642; average p̄ = 0.683.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.6578 = 1.2892.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.6552 = 0.5514.
4. n = (1.2892 + 0.5514)² / (-0.0826)² = 496.9, rounded up to 497 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 <- 0.724770642201835                          # control proportion
p1 <- 0.642201834862385                          # 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(3726879)
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 = 497 (80% power), 665 (90% power)
 75% of observed effect: n per arm = 869 (80% power), 1,163 (90% power)
 50% of observed effect: n per arm = 1,918 (80% power), 2,567 (90% power)
power.prop.test: 496.9 
Simulated power at n = 497 per arm: 0.805 (20000 replicates)

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

14. Analyses Reported by the Sponsor

The sponsor’s reported primary analysis is the Cochran-Mantel-Haenszel odds ratio: 0.67, with a 95% confidence interval from 0.37 to 1.21 and p = 0.185. The method accounts for the disease-stage and hormone-receptor strata stated in the registry record. This is the analysis to use when describing the sponsor’s inferential result.

The recalculation uses the posted group response percentages to obtain an unadjusted risk-difference comparison. It changes both the effect scale and the analysis approach; it is not a recreation of the sponsor’s stratified odds-ratio calculation. Its interval spans zero, and its unadjusted p-value is 0.19. These features are broadly consistent with the sponsor analysis in not separating the comparison clearly from no difference, but the numerical results should not be compared as if they were the same statistic.

The recalculated quantities are simple checks based on the posted summaries, not replacements for the sponsor’s analysis. The registry record does not flag a discrepancy between the posted summaries and the sponsor’s reported analysis. Interpretation should therefore retain the sponsor’s method and estimand as the primary reported result, while treating the unadjusted calculation as an additional descriptive view.

15. Multiplicity, Missing Data and Analysis Populations

The primary endpoint is defined as pCR in the PD-L1-positive population, and its analysis is reported using a stratified Cochran-Mantel-Haenszel test. The registry also lists multiple secondary outcomes, including pCR in another biomarker-defined population, time-to-event outcomes, patient-reported outcomes, and adverse events. Considering several endpoints raises a multiplicity question: repeated testing can increase the chance of at least one apparently positive result when there is no underlying effect.

A method for controlling multiplicity, such as a prespecified testing sequence or adjustment across endpoints, is not reported in the registry record. The record also does not report whether secondary analyses were intended to be confirmatory or exploratory. As a result, secondary p-values, where available, should be interpreted cautiously and should not be treated as if they had the same role as the primary test.

The primary population is described as participants in the Intent-to-Treat population with PD-L1 status IC1/2/3 at randomization. The registry record does not provide enough detail to establish how participants with unavailable or unevaluable PD-L1 status were handled, or whether all randomized participants in the defined population contributed to the primary endpoint analysis. Although arm-specific denominators are posted, the handling of missing pCR assessments and other missing outcome data is not reported in the registry record.

The registry description identifies the stratification variables used in the primary comparison, but further details about the analysis population, missing-data assumptions, sensitivity analyses, and any protocol-defined deviations are not reported in the registry record. Interim analyses and any resulting adjustment to the final inference are also not reported. These omissions limit how fully the analysis can be reconstructed from the posted summaries; they do not establish that such procedures were absent from the trial’s protocol or analysis documents.

16. Safety Results

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

Deaths

Other (non-serious) adverse events

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

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: febrile neutropenia (4.4%), pneumonitis (4.4%), cardiac failure (2.2%), adrenal insufficiency (1.8%), pneumonia (1.8%).

The safety summary reports serious adverse events for 65 of 226 participants in the atezolizumab group and 47 of 225 in the placebo group. These are participant counts with at least one serious adverse event as represented in the registry summary; they are not counts of all serious-event episodes. The record also lists other adverse events for 226 of 226 and 225 of 225 participants, respectively. These categories should be read as reported and not assumed to be mutually exclusive.

The listed serious adverse event terms include febrile neutropenia and pneumonitis at 4.4% each, cardiac failure at 2.2%, adrenal insufficiency and pneumonia at 1.8% each, and pyrexia at 1.3%. The registry summary does not assign these listed term percentages to a treatment arm in the information, so they should not be attributed to one group alone.

Deaths are listed for 11 participants in the atezolizumab group and 13 in the placebo group. The registry record does not report the cause or treatment relationship for these deaths in this summary. The counts describe reported safety outcomes; they do not, by themselves, establish causation or provide a complete comparison of the nature, timing, or severity of adverse events.

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

The primary result describes pCR after neoadjuvant treatment and surgery in the registry-defined PD-L1-positive population. The posted percentages were lower in the atezolizumab group than in the placebo group, while the sponsor’s odds-ratio interval includes the no-difference value. The estimates are therefore uncertain: the data do not establish a clear difference in pCR between the groups under the reported analysis, and they do not establish equivalence.

This is not a conclusion about every participant with early HER2-positive breast cancer. The primary analysis is limited to the PD-L1-positive population defined at randomization, and the study description specifies a particular disease-stage range. Generalizing beyond that population requires evidence about the relevant groups and outcomes.

pCR is a pathological assessment at surgery. It does not directly state how long participants remain free of disease, whether they survive longer, or how treatment affects their function and global health status. Those outcomes are listed separately in the registry, and they require their own results and uncertainty assessments. No treatment recommendation follows from this statistical comparison.

The confidence interval and p-value describe uncertainty under the reported analysis; neither tells the probability that a treatment is beneficial. A point estimate in one direction is not proof of a true effect, and a result that does not cross a conventional significance threshold is not proof of no effect. The appropriate reading is that the reported primary analysis did not precisely distinguish the groups, while leaving a range of possible effects compatible with the data.

Safety counts provide additional descriptive information, but counts alone do not show whether an event was caused by study treatment or explain its clinical context. The summary also does not provide enough detail to assess the timing and course of each event. The efficacy and safety findings should therefore be understood as distinct parts of the registry record, not as a complete account of every potential benefit or harm.

18. Why This Trial Matters Statistically

This trial raises several important statistical issues common to randomized studies of combination treatment. Because the background chemotherapy and HER2-directed regimen was shared between groups, the randomized comparison focuses on the addition of atezolizumab versus placebo within that treatment context. Correctly describing the contrast avoids attributing the result to a difference in the entire regimen.

The primary endpoint is restricted to a biomarker-defined population. That makes the definition and timing of PD-L1 status central to interpretation: the registry defines positivity as IC1/2/3 at randomization within the Intent-to-Treat population. A result in this group should not automatically be extended to the PD-L1-negative group or to all patients with the disease. Subgroup-specific conclusions require attention to how subgroup effects are compared, not only whether separate estimates look different.

The analysis also illustrates why effect scale and adjustment matter. The sponsor reports a stratified odds ratio, while a simple recalculation from the group percentages gives an unadjusted risk-difference perspective. Each can clarify a different feature of the comparison, but they should not be treated as interchangeable. The confidence interval is essential because the primary point estimate alone would give an incomplete picture of precision.

Finally, pCR is a response endpoint, while the registry lists longer-term survival and patient-reported outcomes separately. A sound statistical interpretation keeps those questions distinct and accounts for the multiple outcomes when considering secondary findings. Missing-data and multiplicity details are not reported in the registry record, which limits the extent to which the full inferential plan can be assessed from the posted summaries.

19. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.
• Participant counts were reconstructed from posted percentages and arm sizes, so small rounding differences are possible.

The primary conclusion is confined to pCR in the PD-L1-positive population, not to all participants or all people with early HER2-positive breast cancer. The registry’s population definition depends on PD-L1 status IC1/2/3 at randomization, and the study description identifies the stage range T2-4, N1-3, M0. Applicability beyond that population cannot be established from this primary result alone.

pCR is assessed in surgical specimens after neoadjuvant treatment. It is not itself a measure of event-free survival, disease-free survival, or overall survival. Although those outcomes are listed in the registry, their results are not part of the primary comparison. The primary finding therefore cannot answer whether the groups differ on longer-term outcomes.

The reported primary analysis and the unadjusted recalculation use different effect measures and methods. The recalculation is based on summary proportions and does not account for the stated stratification; it is not an independent participant-level analysis. Further details about analysis populations, missing outcome handling, sensitivity analyses, and the original sample-size rationale are not reported in the registry record.

The registry lists numerous secondary outcomes, but a multiplicity strategy is not reported. This limits the strength of any isolated secondary finding as confirmatory evidence. The safety summary gives participant counts and selected serious-event terms, but does not provide full event-level timing, attribution, or arm-specific detail for the listed terms in this summary. These reporting limits constrain interpretation without establishing that the omitted details were unavailable to the trial investigators.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: A Study To Evaluate the Efficacy and Safety Of Atezolizumab or Placebo in Combination With Neoadjuvant Doxorubicin + Cyclophosphamide Followed By Paclitaxel + Trastuzumab + Pertuzumab In Early Her2-Positive Breast Cancer — NCT03726879.
• Registry record last updated 2024-11-05; status last verified 2024-10.
• Results first posted 2022-03-25.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

23. Record Summary

NCT03726879 was a completed Phase III randomized, double-blind, placebo-controlled trial sponsored by Hoffmann-La Roche. It compared atezolizumab with placebo alongside neoadjuvant ddAC-PacHP in early HER2-positive breast cancer. The registry lists enrollment of 454, a start date of 2019-01-11, and a primary completion date of 2021-02-05.

The primary endpoint was pCR in the PD-L1-positive population, defined by IC1/2/3 status at randomization within the Intent-to-Treat population. The posted pCR percentage was 64.2 with atezolizumab and 72.5 with placebo. The sponsor’s stratified odds ratio was 0.67, with a 95% confidence interval from 0.37 to 1.21 and p = 0.185. The interval includes the no-difference value, leaving uncertainty about the treatment contrast.

The registry lists secondary efficacy, patient-reported, and safety outcomes. Important details on multiplicity, missing data handling, and the trial’s original sample-size justification are not reported in the registry record.