Neoadjuvant Nivolumab and Chemotherapy in Patients With Localized Triple-negative Breast Cancer: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Neoadjuvant Nivolumab and Chemotherapy in Patients With Localized Triple-negative Breast Cancer (NCT04331067), a randomized phase 1, phase 2 study. The primary outcome is Percent Change in Tumor Infiltrating Lymphocytes (TILs).

2. Clinical Question

3. Trial Design

This was a Phase Ib/II, parallel-group study listed as randomized and unmasked in the registry. It studied neoadjuvant treatment for localized triple-negative breast cancer. The registry reports enrollment of 15 and a terminated status. The study start date is 2020-11-19, and the primary completion date is 2023-03-20. The reason for termination is not reported in the registry record.

The original treatment structure included Arm A, with neoadjuvant chemotherapy plus nivolumab, and Arm B, which added cabiralizumab to chemotherapy plus nivolumab. The chemotherapy drugs were paclitaxel and carboplatin. The registry describes paclitaxel as intravenous at 80 mg/m^2 weekly for 12 weeks and carboplatin as intravenous at AUC 5 every 3 weeks for 12 weeks. The complete nivolumab schedule is not reported in the registry record.

An amendment approved on 10/13/2022 changed the study: Arm B would no longer enroll, cabiralizumab would no longer be given, and remaining patients would be enrolled in a single-arm study. The posted primary analysis groups Arm A and the unrandomized arm together because the treatment received was the same, while reporting Arm B separately. Thus, the analysis groups reflect both the original multi-arm structure and the later change in enrollment. The registry does not report enough detail to determine how allocation proceeded for every participant or whether the original randomization remained applicable after the amendment.

The primary endpoint was percent change in stromal TILs between baseline and week 5. Other registered outcomes included pathological complete response, recurrence-free survival, adverse events, and comparisons of TAM and TIL change by pCR status. The registry lists 11 outcome measures, but 8 secondary outcomes are posted. It lists no analyses for a two-arm comparison of the continuous primary endpoint; the reason recorded is that values are missing or not estimable for that comparison. The statistical analysis plan, sample-size justification, and reason for termination 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.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Arm A and Unrandomized Arm were analyzed together as the treatment received was the same. There were 2 participants with missing data in Arm A/Unrandomized Arm and 2 participants with missing data in Arm B and these participants were not included in the analysis.

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 Percent Change in Tumor Infiltrating Lymphocytes (TILs), measured over Baseline and week 5.

The primary endpoint is percent change in tumor-infiltrating lymphocytes (TILs) from baseline to week 5. The registry defines the stromal TIL score as the percentage of tumor stroma area occupied by mononuclear inflammatory cells. The outcome therefore concerns a measured feature of the tumor’s immune-cell environment, rather than a direct measure of tumor size or a clinical event.

A change in TILs can be used to describe whether the measured immune-cell presence in tumor stroma differs between the baseline and an on-treatment assessment. The registry’s stated hypothesis was that TILs would increase following neoadjuvant nivolumab with chemotherapy, alongside a reduction in tumor-associated macrophages. The primary outcome is the mean percent change in TILs; it does not itself report the TAM outcome.

Using a continuous measure preserves variation in the measured change rather than reducing the result to categories such as increased or not increased. The mean describes the average change in the analyzed group, while the standard deviation describes how spread out individual changes are around that mean. General statistical interpretation requires attention to both: a mean alone does not indicate whether participants had similar responses or widely varying changes. The registry does not report the measurement procedure in enough detail to assess measurement reliability, nor does it report a clinical threshold defining a meaningful TIL change.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is continuous: each analyzed participant has a percent-change measure, and the posted group summaries use a mean and standard deviation. In general, a mean is the arithmetic average of observed measurements. It gives a compact description of the group’s central tendency, but it can be affected by unusually large or small observations. A mean does not show whether all participants changed in the same direction.

In general, the standard deviation describes the spread of observations around the mean, in the same measurement units as the outcome. A larger standard deviation indicates more variation among observed measurements. Here, the standard deviations are 121.0 for the combined Arm A and unrandomized group and 82.5 for Arm B; the corresponding means are 95.2 and 1.63. Those summaries indicate that individual changes varied considerably. They do not reveal the distribution’s shape or show whether the variation arose from a small number of extreme observations.

What does the mean percent change represent?

Percent change expresses a change relative to a baseline measurement, whereas the stromal TIL score itself is defined in the registry as the percentage of tumor stroma area occupied by mononuclear inflammatory cells. These are related but distinct concepts: the endpoint is a percent change in the score, not simply the week 5 TIL score. The record does not provide the individual baseline and week 5 values or the precise calculation details for each participant’s percent change. It is therefore not possible to reconstruct the endpoint or assess how baseline values influenced the reported averages.

Why is a between-group test not reported?

The registry lists no analysis for a two-arm comparison of the continuous endpoint, noting that values are missing or not estimable for that comparison. Accordingly, no test statistic, p-value, or confidence interval for the difference between groups is reported. A numerical contrast between group means is not a substitute for an inferential comparison: it does not quantify uncertainty or account for variation within groups. No recalculated comparison is available in the posted results, and there is no discrepancy between a recalculation and a sponsor-reported comparison to resolve.

In general, a confidence interval describes a range of values compatible with the data under a specified statistical procedure. It can communicate the precision of an estimate, but its interpretation depends on the method and assumptions. No confidence interval for the primary group means or a between-group contrast is reported in the registry record. A p-value, if reported, would address compatibility with a null model under specified assumptions; it would not by itself describe the magnitude, clinical relevance, or precision of an effect. No primary-outcome p-value is posted.

How should variability and group structure be considered?

In general, when measurements are highly variable, the mean alone can be a limited description, particularly in small groups. The standard deviation helps show the extent of dispersion, but it is not a measure of uncertainty in the mean. A standard error or confidence interval would be needed to describe precision using a particular inferential approach; neither is reported for these means. The record does not report a distributional assessment or the use of robust methods.

The analysis combines Arm A with the unrandomized arm because the treatment was the same. This combines participants from distinct enrollment designations for the primary summary, while Arm B remains separate. The registry does not report whether the groups differed in baseline characteristics or how the amendment affected allocation. Consequently, the posted summaries do not support an assumption that the groups are exchangeable for a causal comparison.

What can planning calculations establish?

In general, sample-size planning calculations describe how many participants a future study might need under specified assumptions about the effect, variability, allocation, significance level, and power. They are planning illustrations, not evidence that a completed study had a particular probability of finding an effect. The registry record does not report the trial’s own sample-size justification, and no estimable two-arm primary comparison is posted from which to ground a reliable planning illustration. Any planning calculation based on an observed effect would be especially uncertain when the underlying groups are small and the primary comparison is not estimable.

The registry does not report whether interim analyses were planned or conducted, which statistical tests were prespecified, whether adjustment for covariates was planned, or how outcome distributions were evaluated. Such choices can affect inference and should not be assumed. The interpretation is therefore limited to the posted descriptive means, standard deviations, analyzed counts, and the registry’s explanation for the unavailable two-arm comparison.

10. Confidence Intervals for Each Arm

No two-arm comparison is shown for this trial, but each arm's mean can be given with a t-based 95% confidence interval computed from the posted mean, spread and number of participants.

R code and output

arm <- c("Arm A and Unrandomized Arm: Neoadjuvant Chemo + Ni", "Arm B: Neoadjuvant Chemo + Nivolumab + Cabiralizum")
n   <- c(7, 4)
m   <- c(95.2, 1.63)
sp  <- c(121, 82.5)
sd  <- sp      # posted standard deviations
for (i in seq_along(n)) {
  se <- sd[i] / sqrt(n[i]); h <- qt(0.975, n[i] - 1) * se
  cat(sprintf("CARM|%s|%d|%.3f|%.3f|%.3f|%.3f\n", arm[i], n[i], m[i], sd[i], m[i] - h, m[i] + h))
}

CARM  Arm A and Unrandomized Arm: Neoadjuvant Chemo + Ni  7  95.200  121.000  -16.706  207.106
CARM  Arm B: Neoadjuvant Chemo + Nivolumab + Cabiralizum  4  1.630  82.500  -129.646  132.906

11. Multiplicity, Missing Data and Analysis Populations

The registry lists 11 outcome measures and posts 8 secondary outcomes, in addition to the primary TIL endpoint. These include pCR, recurrence-free survival, adverse events, and comparisons of TAM and TIL percent change by pCR status. When a study examines multiple outcomes or comparisons, the chance of at least one apparently notable result can increase. In general, multiplicity procedures can control this issue, but the registry record does not report whether such procedures were planned or used. No adjusted set of primary or secondary outcome findings is presented in the posted record.

For the primary endpoint, 2 participants in Arm A and the unrandomized arm had missing data, and 2 participants in Arm B had missing data; these participants were not included in the analysis. The reported primary analysis includes 7 participants in the combined group and 4 in Arm B. The record does not explain the reasons for missingness, state whether missingness was related to treatment or TIL change, or describe imputation or sensitivity analyses. Excluding participants with missing outcomes can affect a result if the observations available for analysis differ systematically from those not included.

The record says Arm A and the unrandomized arm were analyzed together because the treatment received was the same. This is the stated rationale for combining those groups, but the registry does not report additional analysis-population details or baseline comparability. The analysis population should therefore be understood as the participants with available primary outcome data in the specified groupings, rather than all enrolled participants. The registry does not report a prespecified analysis plan, handling of protocol deviations, or methods for any other missing outcome data.

12. 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 0% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: myositis (33.3%), anemia (16.7%), vomiting (16.7%), sepsis (16.7%), neutrophil count decreased (16.7%).

The serious adverse event summaries use group-specific counts of participants at risk and affected. Arm A reports 1 affected participant among 6 at risk; Arm B reports 3 among 6; the unrandomized arm reports 0 among 3. The record also reports 1 death in Arm A, 1 in Arm B, and 0 in the unrandomized arm. These figures are descriptive. The registry does not establish from these counts whether an event or death was caused by treatment.

The listed top serious-event terms are myositis at 33.3%, and anemia, vomiting, sepsis, neutrophil count decreased, and platelet count decreased at 16.7% each. The registry does not provide enough information alongside these terms to interpret them as comparative risks across treatment groups. Event counts and participant counts are not interchangeable: one participant may experience more than one event, and the record does not explain the overlap for these listings.

All participants recorded as at risk for other adverse events were affected in each group: 6 of 6 in Arm A, 6 of 6 in Arm B, and 3 of 3 in the unrandomized arm. These broad counts do not describe severity or causation. The registry lists several specific grade 3 or higher adverse-event outcomes but does not post their detailed results in the registry record for interpretation. Group sizes are small, so the figures should be read as counts from this study rather than stable estimates of event frequency.

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

The posted result describes measured stromal TIL change between baseline and week 5 among participants with available data. The mean percent change was 95.2 in the combined Arm A and unrandomized group and 1.63 in Arm B. The associated standard deviations, 121.0 and 82.5, show substantial variability around those means. The result is an observation about these analyzed groups; it is not a demonstrated difference between treatments.

The registry’s hypothesis after amendment concerned increased TILs and reduced TAMs following neoadjuvant nivolumab with chemotherapy. The posted primary outcome addresses TIL change, but the registry does not report an inferential test for the primary comparison or a clinical threshold for interpreting the size of a TIL change. The reported means therefore do not establish that treatment caused the observed changes, whether the changes are clinically meaningful, or how the result relates to longer-term outcomes.

The original study included a cabiralizumab arm, but an amendment ended further enrollment to that arm and stopped cabiralizumab administration. The analysis also combines Arm A and an unrandomized arm that received the same treatment. These features, along with missing primary data and the absence of a reported two-arm comparison, limit what can be inferred from the separate group summaries. The record does not report enough information to assess whether differences in participant characteristics or enrollment timing could explain the observed pattern.

The result does not establish whether nivolumab plus chemotherapy is effective or ineffective, whether adding cabiralizumab changes outcomes, or whether participants experienced a clinical benefit. It is not a substitute for results on pCR, recurrence-free survival, or a complete safety analysis. Those outcomes are listed in the registry, but their detailed findings are not included in the primary result. No medical conclusion should be drawn from the TIL summary alone.

14. Why This Trial Matters Statistically

This study illustrates a common challenge in early-phase clinical research: a biologically oriented measure can be informative about changes in the tumor environment, while remaining distinct from a clinical outcome. Stromal TIL score quantifies the proportion of tumor stroma occupied by mononuclear inflammatory cells. Tracking its change during neoadjuvant treatment can describe an immune-related response pattern, but the measure does not automatically establish a patient-important benefit or a causal treatment effect.

It also demonstrates why the analysis population and trial history matter. The study was initially randomized and included a cabiralizumab combination arm, then changed enrollment after an amendment and included an unrandomized arm. The primary summary combines two treatment designations because they received the same regimen. Clear reporting of these changes is important for understanding what the posted means represent and why a direct comparison may not be available.

Statistically, the combination of small analyzed groups, substantial variability, missing data, and no reported inferential comparison makes precision and group comparability central concerns. A difference in descriptive means can attract attention, but without an estimate of uncertainty and a suitable comparison, it cannot distinguish a treatment-related pattern from variation among participants. The record’s explicit statement that a two-arm continuous comparison was not estimable is therefore essential context, not a minor technical detail.

The registry also shows the value of separating a planned biological hypothesis from the result actually posted. The primary endpoint is TIL change; other outcomes include pCR, recurrence-free survival, and adverse events. Each addresses a different question, and findings for one cannot stand in for the others. Interpreting each outcome with its own analysis population, missing-data information, and statistical uncertainty is necessary for a balanced account of the evidence.

15. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

The primary outcome analysis excludes participants with missing TIL data: the registry reports 2 missing participants in the combined Arm A and unrandomized group and 2 in Arm B. It does not give reasons for missing data, evaluate whether missingness was associated with observed outcomes, or report sensitivity analyses. The posted means may therefore describe a subset whose measurements were available rather than every participant enrolled.

The analyzed groups are small, and the reported standard deviations indicate substantial variability. The registry does not post participant-level measurements, distributions, confidence intervals, or a test of the primary between-group comparison. It states that values are missing or not estimable for a two-arm comparison. This prevents assessment of statistical precision and makes it inappropriate to treat the difference in group means as an established effect.

Study design also limits interpretation. The registry describes a randomized study, but an amendment ended enrollment to Arm B, discontinued cabiralizumab, and directed remaining enrollment to a single-arm study. The primary analysis combines Arm A and the unrandomized arm because treatment was the same. The registry does not report enough detail to assess group comparability, allocation after amendment, or the influence of enrollment timing on the posted summaries.

The record does not report the trial’s sample-size justification, prespecified statistical analysis plan, interim analyses, or reason for termination. It does not give a clinical threshold for meaningful TIL change or enough detail to evaluate measurement reliability. Although pCR, recurrence-free survival, TAM measures, and adverse events are listed among outcomes, detailed results for those outcomes are not included in the primary result details. These limitations constrain both causal interpretation and assessment of clinical relevance.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Neoadjuvant Nivolumab and Chemotherapy in Patients With Localized Triple-negative Breast Cancer — NCT04331067.
• Registry record last updated 2026-07-09; status last verified 2026-06.
• Results first posted 2024-05-16.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

19. Record Summary

NCT04331067 is a Phase Ib/II study of neoadjuvant nivolumab and chemotherapy in localized triple-negative breast cancer. The registry lists randomized allocation, no masking, parallel groups, enrollment of 15, and a terminated status. The primary endpoint is mean percent change in stromal TILs from baseline to week 5.

The posted primary summaries report a mean of 95.2 and standard deviation of 121.0 for Arm A combined with the unrandomized arm, with 7 participants analyzed, and a mean of 1.63 and standard deviation of 82.5 for Arm B, with 4 analyzed. Two participants in each analysis grouping had missing data and were excluded. The registry says a two-arm continuous comparison was not estimable. The treatment-arm amendment, missing observations, small analyzed groups, and absence of inferential results limit conclusions about comparative effects.