T-DM1 Alone Versus T-DM1 and Metronomic Temozolomide in Secondary Prevention of HER2-Positive Breast Cancer Brain Metastases Following Stereotactic Radiosurgery: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for T-DM1 Alone Versus T-DM1 and Metronomic Temozolomide in Secondary Prevention of HER2-Positive Breast Cancer Brain Metastases Following Stereotactic Radiosurgery (NCT03190967), a randomized phase 1, phase 2 study. The primary outcome is Phase II: Median Amount of Time Participant Survives Without New Brain Lesions After Starting Treatment.

2. Clinical Question

The study addressed secondary prevention: participants had already received stereotactic radiation or surgery for brain metastases, and the research question concerned subsequent progression. The registry describes evaluation of brain lesions using RANO-BM criteria and systemic disease using RECIST. Its progression definition included a 25% increase in the sum of products of measurable lesions over baseline if there was no decrease, the appearance of new lesions, or failure to return for evaluation because of death or deteriorating condition.

The clinical question therefore has two important boundaries. First, it concerns a specific population with HER2-positive breast cancer and previously treated brain metastases, not breast cancer generally. Second, the planned randomized phase II comparison did not yield the primary outcome data. The record describes the question the study intended to address, not an estimate of whether either regimen changed progression time.

3. Trial Design

NCT03190967 was a phase I/II study with randomized allocation recorded, no masking, and a sequential model. It enrolled 12 participants. The registry lists a start date of 2018-04-18, a primary completion date of 2021-06-28, and a terminated status. The sponsor was the National Cancer Institute (NCI).

The phase I portion evaluated T-DM1 with temozolomide through dose escalation. T-DM1 was listed at 3.6 mg/kg, while the temozolomide dose levels were 30 mg/m^2, 40 mg/m^2, and 50 mg/m^2. The phase I aims included recording grade 3 and/or grade 4 adverse events and dose-limiting toxicities at each dose level. The planned phase II portion included T-DM1 alone and T-DM1 with temozolomide at the recommended phase 2 dose.

In the study description, treatment was organized in 3-week cycles, with T-DM1 on Day 1 of each cycle. Participants in the combination group also took temozolomide capsules by mouth every day. The description also mentions medication diaries, repeated assessments, and questions about general well-being and functioning. These describe the intended study procedures; the registry record does not establish which procedures each participant completed.

Although the record identifies randomized allocation and planned phase II arms, it also states that the trial was terminated early because drug support for the planned randomized phase II cohorts was withdrawn. The primary outcome was not done, and data were never collected. Thus, the phase I dose-escalation experience and the planned phase II comparison should not be treated as if both produced a completed randomized efficacy analysis. Details of the randomization procedure and the number assigned to each intended phase II arm 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: The phase II portion was never started as we could no longer get the drug from the manufacturer.

Level 1 Dose Escalation

Level 2

Level 3

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 Phase II: Median Amount of Time Participant Survives Without New Brain Lesions After Starting Treatment, measured over From the start of treatment until new brain lesions.

The primary outcome was median time to progression, described in the registry title as the median amount of time participants survived without new brain lesions after starting treatment. Time-to-event outcomes measure how long it takes for a defined event to occur. A median time-to-event value is the time by which half of the relevant population has experienced the event, provided the data and analysis permit that estimate. This is a general description of the statistic, not a result from this trial.

The endpoint was intended to capture disease progression after treatment began. The registry specified RANO-BM criteria for evaluating brain lesions and RECIST for systemic evaluation. Its progression definition was broader than the appearance of a new brain lesion: it also included a specified increase in measurable lesions, and failure to return for evaluation because of death or deteriorating condition. Consequently, the endpoint’s operational definition included events beyond new brain lesions alone.

Time-to-event outcomes can be useful when the timing of progression matters and participants may have different lengths of follow-up. In general, their analysis must account for participants whose event time is not observed during follow-up. The registry record does not report how such observations would have been handled in this study, nor does it report an analysis method for the endpoint.

Most importantly, the primary outcome was not collected. The record attributes this to early termination following withdrawal of drug support for the planned randomized phase II cohorts. There is therefore no trial estimate of median progression time to interpret, and no estimate of a treatment contrast for this endpoint.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

Several design and statistical ideas help clarify what the registry record can and cannot establish. They describe how the study was structured and what a completed analysis would generally require; they are not claims that a particular analysis was carried out.

What does a sequential phase I/II design mean?

In general, a phase I dose-escalation study investigates treatment doses and safety before a later phase evaluates a clinical outcome. A sequential phase I/II design links these stages: information from the earlier component can inform the dose used in the later component. In this trial, the phase I combination groups used temozolomide dose levels of 30 mg/m^2, 40 mg/m^2, and 50 mg/m^2, with T-DM1 listed at 3.6 mg/kg. The planned phase II combination arm was described as using the recommended phase 2 dose. The registry record does not report the dose-selection rules or the formal method used to identify that dose.

The staged structure matters because the phase I dose groups answer a different kind of question from a randomized phase II comparison. Dose escalation focuses on safety and dose-limiting toxicity across dose levels. A randomized comparison is designed to compare outcomes between treatment groups. The record lists both intentions, but says the study ended before the planned randomized phase II cohorts and that its primary outcome was never collected. The phase I group listings cannot substitute for a phase II outcome comparison.

What would randomization contribute?

In general, random assignment aims to make treatment groups comparable at the time of assignment, on average, by allocating participants without choosing their group based on their characteristics. This supports a causal comparison when follow-up and analysis are appropriately conducted. Randomization does not guarantee identical groups, especially in small studies, and it cannot make missing outcome data available. The registry record identifies randomized allocation, but does not report the randomization sequence, allocation concealment procedure, or the number assigned to each planned phase II arm. Because the phase II cohorts were not completed, there is no reported randomized primary-outcome contrast to interpret.

How should the primary endpoint be understood?

The intended primary outcome was median time to progression. In general, a time-to-event analysis uses the timing of an event rather than only whether it occurred. This can preserve information about when progression happened. A median summarizes the event-time distribution, while a comparison between treatment groups would require an analysis plan and observed follow-up data. The registry record describes criteria for progression using RANO-BM for brain lesions and RECIST for systemic evaluation. It also includes a 25% increase in the sum of products of measurable lesions over baseline if there was no decrease, appearance of new lesions, or failure to return for evaluation due to death or deteriorating condition.

In general, some participants in a time-to-event study may not have an observed event by the end of follow-up. Analyses commonly need a defined approach to such incomplete event times, often called censoring. The registry record does not report the planned censoring rules, time-to-event estimator, comparison method, or handling of competing events. Since the endpoint was never collected, there are no observed event times on which to apply these methods.

What do confidence intervals and p-values normally add?

In general, a confidence interval expresses the statistical uncertainty around an estimate under an analysis model. A narrower interval indicates greater precision under that model; a wider interval indicates less precision. A p-value measures how compatible data are with a specified null hypothesis under stated assumptions. Neither quantity can replace the estimated effect, and neither establishes clinical importance on its own. No primary-outcome estimate, confidence interval, or p-value is reported for this trial because the outcome data were not collected.

Interpretation of any interval or test would also depend on the statistical model, analysis population, endpoint definition, follow-up, and missing-data approach. These details are not reported in the registry record for a primary-outcome analysis. No inferential conclusion should be reconstructed from the planned design alone.

How do the planning illustrations differ from observed power?

The registry describes planning illustrations based on an observed effect, with a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. For time-to-event outcomes, the stated illustration concerns the total number of events; other illustrations concern participants per arm. These assumptions describe a planning calculation, not the trial’s own design or its observed results. A sample-size illustration is not observed power: it cannot show what the completed trial demonstrated, and it does not establish that a treatment effect exists.

Planning calculations depend on assumptions about the effect to detect and the analysis. If those assumptions differ from the eventual setting, the resulting planning figures may not describe the study’s actual information. In this case, the primary endpoint was never collected, so an illustration based on an observed effect cannot be interpreted as evidence from a completed randomized phase II analysis. No sample-size justification or original protocol calculation is reported in the registry record.

Finally, the phase I safety counts are descriptive. In general, estimates from small groups can be imprecise, and counts alone do not establish differences between doses or treatment regimens. The registry lists adverse events and deaths by phase I dose level, but does not report a formal inferential comparison for those outcomes. That distinction—descriptive event recording versus a prespecified comparative analysis—is essential when reading the available results.

10. Multiplicity, Missing Data and Analysis Populations

The registry record does not report a statistical plan for handling multiplicity, meaning the increased chance of a false-positive finding when many outcomes, comparisons, or analyses are considered. The listed outcomes include safety and time-to-event measures, but the record does not specify a testing hierarchy, adjustment method, or which secondary outcome tests, if any, would have been confirmatory. With no primary outcome analysis reported, there is no reported multiplicity adjustment to assess.

The primary outcome data are not simply described as incomplete: the registry states that the outcome was not done and that data were never collected, following early termination and withdrawal of drug support for the planned randomized phase II cohorts. This prevents estimation of the primary endpoint and is different from a reported analysis in which some participants have missing follow-up values.

The analysis populations are not reported in the registry record. It does not specify whether intended analyses would use all randomized participants, participants who started treatment, or another defined group. The record also does not describe a primary-outcome missing-data strategy, outcome adjudication process, or rules for excluding observations. These details cannot be inferred from the design label or from the enrollment count.

For the adverse-event listings, the registry gives counts by dose level, but does not report an analysis population definition or a statistical comparison. The available counts should therefore be treated as descriptive listings, not as results from a documented multiplicity-controlled testing plan.

11. 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: dysphasia (33.3%), muscle weakness right-sided (33.3%), neoplasms benign, malignant and unspecified (incl cysts and polyps) - other, malignant neoplasm -cml (16.7%).

The registry lists serious and other adverse-event counts for three phase I dose levels. For level 1, 3 participants were at risk for serious events and 1 was affected; 3 were at risk for other events and 3 were affected. For level 2, the corresponding counts were 3 and 0 for serious events, and 3 and 3 for other events. For level 3, the corresponding counts were 6 and 1 for serious events, and 6 and 6 for other events. These are participant counts within dose-level groups, not treatment-effect estimates.

The listed serious-event terms include dysphasia at 33.3%, muscle weakness right-sided at 33.3%, and neoplasms benign, malignant and unspecified (incl cysts and polyps) - Other, Malignant Neoplasm -CML at 16.7%. The registry does not explain whether participants could appear under more than one term, so the terms should not be assumed to represent mutually exclusive groups.

The record also lists deaths by phase I dose level: 2 at level 1, 1 at level 2, and 1 at level 3. It does not report that these deaths were caused by treatment, nor does it provide a causal assessment in these figures. Serious adverse events, other adverse events, and deaths are distinct reported categories and should not be treated as interchangeable measures. The counts describe the recorded experience; they do not establish comparative safety between regimens.

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

The record shows that NCT03190967 was intended to examine T-DM1 alone and T-DM1 with temozolomide after local treatment of HER2-positive breast cancer brain metastases. It also records phase I dose-escalation groups and adverse-event counts in those groups. These are relevant descriptions of the study’s conduct and reported safety information.

It does not show whether adding temozolomide changed the time to progression. The primary outcome was not done because the trial ended early after withdrawal of drug support for the planned randomized phase II cohorts, and the data were never collected. No efficacy estimate or comparison between the intended phase II arms is reported. Accordingly, the record cannot establish that the regimens produced the same outcome, nor can it establish that one produced a better outcome.

The listed safety figures also have a limited interpretation. They are counts within phase I dose levels, not a randomized comparison of T-DM1 alone against T-DM1 plus temozolomide. The record does not provide causal attribution for the listed deaths. These data can describe what was recorded in the phase I groups, but they do not by themselves determine whether a treatment caused an event or support a comparative safety conclusion.

In general, the strength of a clinical conclusion depends on the outcome data, how participants were followed, the analysis population, and the statistical methods. Here, the central efficacy outcome is absent, and several analysis details are not reported in the registry record. A planned design is not equivalent to completed evidence. The most accurate conclusion is that the clinical question remained unanswered in the posted primary outcome record.

13. Why This Trial Matters Statistically

This trial illustrates the difference between a study’s planned question and the evidence it ultimately records. A phase I/II design can be intended to move from dose evaluation to randomized outcome comparison, but those stages answer different questions. The phase I dose-level listings document a limited safety experience; they do not supply the planned phase II estimate of time to progression.

Early termination has a direct statistical consequence when the primary outcome is never collected: there is no estimate to evaluate, no uncertainty interval to interpret, and no treatment contrast to test. This is not a negative result or a finding of no difference. It is an absence of the data required to answer the main comparative question.

The record also makes endpoint definition important. The outcome description emphasizes time without new brain lesions, while the operational definition includes systemic evaluation, measurable-lesion change, new lesions, and specified failure to return for evaluation. Clear distinctions between an endpoint’s label and its operational criteria matter for interpreting what a completed analysis would represent.

Finally, the study demonstrates why planning assumptions, safety descriptions, and efficacy results must remain distinct. Planning calculations do not become trial findings, and event counts in small dose groups do not establish comparative treatment effects. Reading these categories separately helps prevent overinterpretation when a clinical trial stops before its intended randomized analysis.

14. 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 most consequential limitation is explicit in the registry record: the trial was terminated early because drug support was withdrawn for the planned randomized phase II cohorts. The primary outcome was not done, and the data were never collected. Therefore, no primary efficacy estimate or randomized comparison is available to assess.

The total enrollment was 12, but the registry does not report how many participants entered each planned phase II arm. The phase I adverse-event groups are listed by dose level, and should not be interpreted as a completed randomized comparison of T-DM1 alone with the combination. The registry record also does not describe the detailed randomization procedure or allocation concealment.

The primary endpoint’s description combines a focus on new brain lesions with progression criteria that also include measurable-lesion change and systemic evaluation. This would require careful interpretation in any completed analysis. The registry does not report the planned primary analysis method, analysis population, missing-data approach, or multiplicity strategy. Those details cannot be reconstructed from the outcome definition alone.

The safety listings are counts by phase I dose level. The registry does not report a formal comparative safety analysis or causal attribution for the listed deaths. It also does not state whether serious-event terms could overlap across participants. The counts should be understood as descriptive and should not be generalized into a comparative claim.

These limitations are specific to the status and reporting of this study. They constrain what can be concluded from the registry record; they do not resolve the clinical question that the study was designed to investigate.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: T-DM1 Alone Versus T-DM1 and Metronomic Temozolomide in Secondary Prevention of HER2-Positive Breast Cancer Brain Metastases Following Stereotactic Radiosurgery — NCT03190967.
• Registry record last updated 2023-10-11; status last verified 2023-09.
• Results first posted 2023-10-11.
• 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.

18. Record Summary

NCT03190967 was a phase I/II study of T-DM1 alone versus T-DM1 with metronomic temozolomide for secondary prevention after local treatment of HER2-positive breast cancer brain metastases. The record lists randomized allocation, no masking, a sequential model, enrollment of 12, and National Cancer Institute (NCI) sponsorship. The study began on 2018-04-18, had a primary completion date of 2021-06-28, and is listed as terminated.

Phase I evaluated dose levels of the combination. A randomized phase II comparison was planned, but drug support for the planned cohorts was withdrawn and the study ended early. The primary outcome, median time to progression, was not done and data were never collected. The record includes adverse-event counts by phase I dose level, but no primary efficacy analysis. The registry record does not report an analysis plan for the intended primary comparison.