MEDI4736 and Tremelimumab in Treating Patients With Metastatic HER2 Negative Breast Cancer: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for MEDI4736 and Tremelimumab in Treating Patients With Metastatic HER2 Negative Breast Cancer (NCT02536794), a na phase 2 study. The primary outcome is Overall Response Rate (ORR) in Patients With Metastatic HER2 Negative Breast Cancer Treated With Durvalumab in Combination With Tremelimumab.

2. Clinical Question

3. Trial Design

This was a phase II, single-group study sponsored by Northwestern University. It is listed as completed. The registry gives a planned enrollment of 30 and lists the study start as 2016-01-14 and primary completion as 2020-02-06. These are study record details, not evidence that all enrolled patients were evaluable for the primary endpoint. The number actually analyzed for response is not reported in the registry record.

The registered treatment group received MEDI4736 and tremelimumab. The arm description specifies intravenous administration on day 1 and treatment repeated every 28 days for 4 courses, unless disease progression or unacceptable toxicity occurred. It then begins to describe treatment after the last combination dose but ends mid-description. The subsequent treatment details are not reported in the registry record. The interventions also list laboratory biomarker analysis and a pharmacological study, but their procedures and relationship to the primary endpoint are not reported in the registry record.

The design is not randomized and has no separate control group. Allocation is listed as not applicable, and masking is listed as none. Thus, this study can describe outcomes observed in the treated group but cannot estimate a randomized between-group treatment contrast. A single-group response result, if available, would remain subject to uncertainty about how outcomes might have differed without the study treatment.

The record lists 14 outcome measures and zero analysis entries. The primary outcome is titled overall response rate, although the outcome titles repeatedly describe treatment with durvalumab in combination with tremelimumab, while the study title and intervention description use MEDI4736. This naming inconsistency is present in the registry record; its explanation is not reported in the registry record. No separate comparator, statistical analysis plan, sample-size justification, or interim analysis is described 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.

Reached 1st Response Assessment

Reasons for not completing this period:

Continued Treatment After 1st Response

Reasons for not completing this period:

Follow up for 3 Years or Death

Analysis population for the primary outcome, as defined in the registry: number of HER2 negative breast cancer patients that met criteria to be evaluable for this endpoint

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 Overall Response Rate (ORR) in Patients With Metastatic HER2 Negative Breast Cancer Treated With Durvalumab in Combination With Tremelimumab, measured over Up to a maximum of 49 cycles where 1 cycle = 4 weeks for the first 4 cycles and than 1 cycle =2 weeks for 45 cycles.

The primary endpoint is overall response rate, defined as the number of evaluable patients whose best qualifying response is complete response or partial response under RECIST version 1.1. A complete response is defined in the record as disappearance of all lesions. A partial response is defined as at least a 30% decrease in the sum of diameters of target lesions, using the baseline sum of diameters as the reference. The endpoint timeframe is described as up to a maximum of 49 cycles; the record states that the first 4 cycles are 4 weeks each and the following 45 cycles are 2 weeks each.

In general, response endpoints are used in oncology studies to classify whether measurable disease has decreased according to prespecified criteria. Combining complete and partial responses into an overall response measure provides a count of patients meeting either response definition. The record frames this outcome as a number of participants, rather than reporting a rate with a denominator. A rate would require both the response count and a defined analysis population.

The number of complete and partial responses, the evaluable population size, and an overall response estimate are not reported in the registry record. Nor is the method for handling patients without an evaluable response described there. Consequently, the endpoint definition is available, but the posted results do not establish the observed response frequency or its precision.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does the primary endpoint count?

The primary endpoint is a count of patients with a complete response or partial response, among patients considered evaluable for the endpoint. Complete response means disappearance of all lesions. Partial response means at least a 30% decrease in the sum of target-lesion diameters relative to the baseline sum. These definitions follow the record's description of RECIST version 1.1. The record does not report the response count or the size of the evaluable population.

In general, a binary endpoint assigns each assessed participant to one of two categories, such as meeting or not meeting a response definition. A count can be converted to a proportion only when the relevant denominator is specified. It is important not to assume that planned enrollment and the analyzed population are identical: eligibility, missing assessments, or other analysis rules may affect who contributes. The endpoint population rules and missing-response handling are not reported in the registry record.

What can a confidence interval add?

In general, a confidence interval communicates statistical uncertainty around an estimated quantity. For a response proportion, it would help readers judge how much variation is compatible with the observed count and denominator under the chosen method. It does not measure the chance that a particular treatment effect is true, and its interpretation depends on the statistical procedure and assumptions. No confidence interval or method for calculating one is reported in the registry record.

For a single-group study, an interval around a response estimate would quantify uncertainty in that group's response frequency. It would not create a control group or remove the limitations of a nonrandomized comparison. The registry record has no concurrent comparator and reports no between-group estimate.

What is the role of the analysis population?

The primary endpoint description refers to patients who met criteria to be evaluable. In general, an analysis population should be defined before interpreting a count or rate, since excluding participants or treating missing assessments in different ways can change the denominator and potentially the estimate. A protocol may distinguish participants enrolled, treated, and evaluable, but the registry record does not state those populations or provide their sizes. It also does not report how early discontinuation, progression before assessment, or unavailable response measurements were handled.

What does the design allow?

This study is listed as single group, with allocation not applicable and no masking. In general, random allocation to a concurrent control group can support a more direct comparison by balancing measured and unmeasured prognostic factors on average. A single-group design instead describes outcomes in the group receiving the study regimen. It does not estimate the difference between that regimen and a control within the same trial.

Masking concerns whether participants or study personnel know treatment assignment. The record lists no masking. In general, knowledge of treatment may influence care or assessment, although the potential relevance depends on the endpoint and trial procedures. The registry record does not describe blinded response review, assessment procedures, or safeguards for endpoint evaluation.

How should the planning assumptions be read?

The record characterizes sample-size calculations as planning illustrations based on the observed effect, not as the trial's own design. It lists a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation, with the effect to detect expressed as a stated fraction of the observed effect. It also distinguishes time-to-event planning outputs, which are total event counts, from other outputs, which are participants per arm.

In general, a planning calculation estimates the amount of information a future study might need under specified assumptions. It does not describe the number of participants or events actually observed in a completed study, and it is not observed power. Planning results depend on the assumed effect, variability, allocation, and statistical method. Here, the record does not give a numerical observed effect or numerical planning results, and its two-arm allocation assumptions do not match the registered single-group design. The illustrations should not be read as the protocol's sample-size justification; that justification is not reported in the registry record.

Why can over-dispersion matter?

In general, over-dispersion occurs when observed variability exceeds what a statistical model assumes. It can make model-based uncertainty too small if the model does not account for the extra variability. The primary endpoint is described as a binary response count, and the registry record does not report a model, an over-dispersion assessment, or an adjustment for it. No claim about such an adjustment can be made from this record.

What does the registry say about analyses?

The record lists 14 outcome measures but zero analysis entries. It does not report the statistical test, estimation method, confidence-interval method, covariate adjustment, subgroup analysis, or analysis timing for the primary endpoint. In general, these details are needed to reproduce an analysis and understand how the reported estimate was obtained. Their absence does not establish that no analyses were conducted; it means their methods are not reported in the registry record.

10. Exact Estimates for Each Arm

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

R code and output

arm    <- c("Treatment (MEDI4736, Tremelimumab)")
n      <- c(27)
events <- c(4)
for (i in seq_along(n)) {
  ci <- binom.test(events[i], n[i])$conf.int          # Clopper-Pearson exact interval
  cr <- qbeta(c(0.025, 0.975), 1 + events[i], 1 + n[i] - events[i])   # Bayesian, uniform Beta(1,1) prior
  wi <- (function(x, n) { p <- x / n; z <- qnorm(0.975); d <- 1 + z^2 / n
    c((p + z^2 / (2 * n) - z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d, (p + z^2 / (2 * n) + z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d) })(events[i], n[i])
  cat(sprintf("ARM|%s|%d|%d|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f\n", arm[i], n[i], events[i], 100 * events[i] / n[i],
              100 * ci[1], 100 * ci[2], 100 * wi[1], 100 * wi[2], 100 * cr[1], 100 * cr[2]))
}

ARM  Treatment (MEDI4736, Tremelimumab)  27  4  14.8  4.2  33.7  5.9  32.5  6.1  32.7

11. Multiplicity, Missing Data and Analysis Populations

The registry record lists 14 outcome measures, including overall response rate as the primary endpoint and toxicity, overall survival, progression-free survival, and clinical benefit rate among the secondary outcomes. It does not report a multiplicity strategy, such as a hierarchy of tests or an adjustment for evaluating multiple outcomes. In general, examining multiple endpoints can increase the chance of a nominally unusual result if each is tested separately. Without a stated plan and reported analyses, the effect of multiplicity on any inferential claims cannot be assessed.

Missing data procedures are not reported in the registry record. This is relevant because the primary outcome population is described as patients evaluable for the endpoint, but the criteria for evaluability and the number evaluable are not given. The record does not say how missing response assessments, early withdrawal, or incomplete follow-up were handled. In general, different missing-data rules can change the denominator and interpretation of an endpoint, so the response count alone would not be sufficient without the analysis population and rules.

Analysis populations are also not described beyond the endpoint's reference to evaluable patients. The registry does not report whether analyses used all enrolled participants, participants who received treatment, or another defined group. No sensitivity analyses for missing data are reported. These omissions restrict reproducibility and make it difficult to assess how robust an endpoint result would be to alternative analysis choices.

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: confusion (6.7%), hepatitis (6.7%), disease progression (6.7%), fever and chills (6.7%), pleural effusion (3.3%).

The safety record lists 30 participants at risk for serious adverse events and 18 affected, as well as 30 at risk for other adverse events and 30 affected. It also lists 22 deaths. These are the registry's reported figures; the record does not provide the timing, duration of observation, event-level overlap, or detailed attribution needed to interpret them as treatment-caused outcomes. The death count should not be read as a count of deaths caused by the study drugs.

The listed serious adverse event terms include confusion, hepatitis, disease progression, and fever and chills, each reported as 6.7%, and pleural effusion and vomiting, each reported as 3.3%. The record does not explain whether a participant could contribute to more than one term, how terms were coded, or the follow-up period used for these percentages. They should therefore be read as registry safety summaries, not as comparative risk estimates.

In general, adverse-event frequency depends on the population monitored, observation time, event definitions, and reporting rules. A single-group study cannot establish whether an event was more or less frequent than it would have been under another treatment. No comparator safety data are reported in the registry record.

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

The record establishes that a phase II, single-group study of MEDI4736 with tremelimumab was registered as completed, with a planned enrollment of 30. It also establishes the intended primary endpoint and its response definitions. These registry entries describe the study's question and design, but they do not provide the primary result. A response count, analyzed denominator, overall response estimate, and uncertainty interval are not reported in the registry record.

Accordingly, the posted registry information does not allow a quantitative assessment of how often patients met the response definition or how precisely that frequency was estimated. It does not support a conclusion about the magnitude of antitumor activity from the primary endpoint. The absence of a reported estimate is not evidence that no responses occurred; it means the result is not reported in the registry record.

Even if a response estimate were reported, the design would require careful interpretation. In general, a single-group study can characterize outcomes in the treated group but cannot provide a randomized comparison with an alternative. It cannot, by itself, distinguish treatment effects from differences in patient mix, disease course, assessment, or other factors. No control group is listed.

The safety figures describe events recorded in the registry, but they do not establish causality or comparative safety. The record does not provide event timing, exposure-adjusted rates, or a control group. The listed deaths should not be treated as deaths attributed to treatment in the absence of such information.

The record also contains a naming difference: study and intervention details use MEDI4736, while endpoint titles refer to durvalumab with tremelimumab. The registry record does not explain this difference. It should remain an unresolved documentation issue rather than being used to infer an unreported detail about the intervention or results.

14. Why This Trial Matters Statistically

Statistically, this trial illustrates the distinction between a well-defined endpoint and a reported result. The registry gives an operational definition of overall response under RECIST version 1.1, including the criteria for complete and partial response. Yet without the response count and evaluable denominator, the endpoint cannot be quantified from the registry record. A definition makes an outcome interpretable; it does not substitute for the observed data.

The design also highlights why the comparator structure matters. A single-group study can describe the experience of participants receiving a regimen, but it does not offer a concurrent estimate of what would have happened under a control condition. The trial's allocation and masking fields reinforce that the record does not describe a randomized comparison. Any interpretation should distinguish within-group description from comparative evidence.

The endpoint population is another important statistical detail. The primary outcome is defined for patients considered evaluable, while the record does not state the evaluability criteria or analyzed number. Without those details, readers cannot determine how participants were included in the response analysis or how missing assessments affected interpretation. Transparent reporting of denominators and missing-data rules is essential for reproducibility.

Finally, the outcome titles' use of durvalumab alongside the study's MEDI4736 terminology, and the absence of analysis entries, make documentation quality relevant. These features do not show that an analysis was not conducted, but they limit what can be verified from the registry record. Clear alignment between intervention names, endpoint definitions, analysis populations, and results is important for interpreting clinical trial 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.

A central limitation of the available registry information is that the primary endpoint result is not reported in the registry record. The response count, analysis denominator, estimate, and uncertainty interval are absent, so the primary outcome cannot be quantitatively interpreted from the posted results. The number enrolled is a planned enrollment figure and does not resolve the unknown size of the evaluable population.

The single-group design has no concurrent control, so it cannot produce a randomized treatment contrast. Allocation is listed as not applicable and masking as none. The registry record does not describe independent or blinded response assessment, nor does it report an analysis plan that could clarify methods for estimation and inference.

There is no reported method for handling missing response data, no evaluability criteria, and no description of analysis populations. The record lists 14 outcome measures but zero analysis entries; multiplicity procedures and the statistical methods for the primary and secondary outcomes are not reported in the registry record. Numerical results for the listed secondary outcomes are also not reported there.

The safety summaries lack detail on observation periods, event timing, attribution, and overlap among event categories. They do not provide a comparator group or sufficient information for comparative safety interpretation. In addition, outcome titles refer to durvalumab while the trial title and intervention details name MEDI4736. The registry record does not explain this naming inconsistency.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: MEDI4736 and Tremelimumab in Treating Patients With Metastatic HER2 Negative Breast Cancer — NCT02536794.
• Registry record last updated 2022-05-17; status last verified 2022-04.
• Results first posted 2021-04-23.
• 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

The registry record describes a completed phase II, single-group study sponsored by Northwestern University, with planned enrollment of 30. The study evaluated MEDI4736 with tremelimumab in patients with metastatic HER2-negative breast cancer. Its primary endpoint is overall response rate, defined by complete or partial response under RECIST version 1.1 among patients evaluable for the endpoint.

The primary response count, analyzed denominator, response estimate, confidence interval, and statistical analysis are not reported in the registry record. Secondary outcomes are listed, but their numerical results are not reported there. Safety summaries list serious and other adverse events and deaths, without enough detail to establish attribution or make a comparison with another group. The registry identifies no comparator arm and does not report missing-data, multiplicity, or analysis-population procedures.