T-DM1 and Palbociclib for Metastatic HER2 Breast Cancer: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for T-DM1 and Palbociclib for Metastatic HER2 Breast Cancer (NCT03530696), a randomized phase 2 study. The primary outcome is Estimate Progression-free Survival.

2. Clinical Question

3. Trial Design

This is a Phase II study of T-DM1 with palbociclib in metastatic HER2-positive breast cancer. The official title calls it a single-arm study, and the brief summary says all patients will be treated with the combination. At the same time, the registry record labels allocation as randomized, identifies the model as single group, and lists both T-DM1 with palbociclib and single-agent T-DM1 as treatment groups. The record does not explain how these descriptions fit together, whether the single-agent group was part of the planned design, or how group assignment occurred. The reported group results should therefore be read as registry-listed summaries rather than as proof of a randomized comparison.

The registry reports 55 participants enrolled and a completed study. The study start date is 2018-12-06, and the primary completion date is 2022-12-22. Masking is listed as none. The record does not report further details about recruitment, treatment allocation, or the timing and conduct of assessments beyond the outcome timeframe of up to 4 years.

For the combination group, T-DM1 was administered intravenously every 21 days on day 1 of each cycle, and palbociclib was administered orally on days 5–18. In the single-agent group, T-DM1 was administered intravenously every 21 days on day 1. These are the schedules recorded for the listed groups; further treatment-modification or adherence information is not reported in the registry record.

The primary outcome is median progression-free survival. The record also lists number of participants with response and estimated overall survival as secondary outcomes. The posted information does not include results for those secondary outcomes. No protocol-level sample-size justification or detailed statistical analysis plan is 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: Number of patients included for PFS analysis: T-DM1 with Palbociclib Arm: n=36 T-DM1 Arm: n=11 5 patients lack data on response because they discontinued study participation (or withdrew consent) before the first scan. Thus, the final sample size for responses is n = 47.

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 Estimate Progression-free Survival, measured over Up to 4 years.

Progression-free survival measures the time from the date of first treatment until investigator-determined objective disease progression or death from any cause. Progression is defined using RECIST 1.1. For target lesions, complete response means disappearance of all target lesions; partial response means at least a 30% decrease in the sum of their diameters relative to baseline; progressive disease means at least a 20% increase; and stable disease is neither enough shrinkage for partial response nor enough increase for progressive disease.

The registry reports the median progression-free survival in months, with a 95% confidence interval, over a timeframe of up to 4 years. A median is the time by which half of the analyzed group has had the event or has been censored. It is useful for describing the timing of progression or death in a time-to-event outcome, especially when not all participants have an observed event during follow-up. It does not describe every participant’s outcome or, by itself, quantify a relative treatment effect.

Participants without documented progression or death are censored at their last radiographic tumor assessment, if available, or at randomization if no post-baseline radiographic assessment is available. Censoring indicates that the event time is not observed within the available follow-up; it is not evidence that progression or death never occurred.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

Progression-free survival is a time-to-event outcome. Each participant contributes information from first treatment until progression, death, or the point at which follow-up no longer establishes whether either event has occurred. Unlike a simple count of events by a fixed date, a time-to-event analysis can retain partial follow-up from participants who have not experienced the event by their last assessment. Such participants are censored rather than treated as if they had remained event-free indefinitely.

What does the median represent?

The median progression-free survival is the estimated time at which the progression-free survival curve reaches the midpoint of its probability scale. It gives a central summary of the observed event-time pattern. It is not the average time to progression or death, and it does not say how long any particular participant remained progression-free. A median can be estimated even when some participants are censored, provided the observed follow-up supports estimation around the midpoint. The registry does not report enough detail to assess the amount or timing of censoring in each group.

How should confidence intervals be read?

A confidence interval communicates the precision of an estimated quantity under the analysis method. A wider interval indicates more uncertainty than a narrower one, though interpretation also depends on the estimator and data structure. For single-agent T-DM1, the registry reports a median of 8.3 months with a 95% confidence interval of 2.7 to 8.8 months. For T-DM1 with palbociclib, it reports a median of 16.9 months and a lower confidence bound of 8.3 months, but no upper bound. That incomplete interval does not give the full range of uncertainty. The reason the upper bound is absent is not reported in the registry record.

Why does censoring matter?

Censoring allows participants with incomplete event follow-up to contribute the time for which their status is known. The primary outcome description says participants who have neither progressed nor died are censored at their last radiographic tumor assessment, if available, or at randomization if no post-initiation radiographic assessment is available. Standard time-to-event interpretation relies on assumptions about censoring, including that the censoring process does not systematically conceal a different event risk than the analysis accounts for. The record does not report a diagnostic assessment of these assumptions or describe the number censored in each group.

Progression is investigator-determined according to RECIST 1.1. The endpoint therefore combines radiographic assessment of disease progression with death from any cause. It is clinically interpretable as time without documented progression or death, but the timing and frequency of scans can influence when progression is detected. The registry does not report scan schedules, independent review, or additional endpoint-adjudication procedures.

What can a group comparison establish?

A comparative time-to-event analysis often uses a hazard ratio to describe the relative event rate over follow-up, alongside a confidence interval and a statistical test. A hazard ratio has assumptions and does not mean the same thing as subtracting one group’s median from another. The registry information lists no analyses, and no sponsor-reported hazard ratio or formal comparison result is included. The posted medians are therefore not substitutes for an inferential comparison. No recalculated comparison is presented, and the record’s reported analysis would take precedence if a recalculated result conflicted with it.

The design description also constrains method interpretation. A randomized comparison can support stronger causal interpretation when assignment, conduct, and analysis preserve the randomized groups. Here, the record labels allocation as randomized but also calls the model single group and the study single arm; it lists a single-agent group without explaining assignment. The statistical implications cannot be resolved without design details that are not reported. The group-level outcome summaries can be described, but the record does not establish the basis for treating them as a controlled comparison.

How should planning illustrations be understood?

Any sample-size planning illustration based on an observed effect is retrospective and illustrative, not the study’s original design justification. The stated assumptions for such calculations are a two-sided significance level of 0.05, equal allocation, and a normal approximation, with the target effect expressed as a fraction of the observed effect. For time-to-event outcomes, the event-count calculation is based on the total number of events. These assumptions describe a planning exercise; they do not establish how the trial was designed or what effect it was powered to detect. The registry record does not report the original sample-size justification, and a planning illustration must not be interpreted as observed power or as evidence that a result is conclusive.

Over-dispersion is a general statistical concern when variability exceeds what a model assumes; it can make standard errors and tests too optimistic if ignored. The posted endpoint is a time-to-event median, and the registry does not report an analysis that addresses over-dispersion. More generally, statistical methods should be matched to the endpoint and its data structure. No unreported model, adjustment, interim analysis, or testing procedure can be inferred from these results.

10. Multiplicity, Missing Data and Analysis Populations

The registry lists one primary outcome, median progression-free survival, and two secondary outcomes: number of participants with response and estimated overall survival. The posted information does not include results for those secondary outcomes. It is therefore not possible to assess whether the outcome pattern is consistent across endpoints or to determine how any secondary findings relate to the primary outcome.

When a study examines multiple outcomes or performs multiple statistical tests, the chance of at least one apparently notable result can increase. A prespecified testing strategy can address that concern, but a multiplicity plan is not reported in the registry record. The record also does not report interim analyses, stopping rules, or a hierarchy for testing outcomes. Those omissions do not establish that such procedures were absent; they mean they cannot be evaluated from this record.

The registry states that 5 patients lacked response data because they discontinued study participation or withdrew consent before the first scan. It reports a final response sample size of 47. This statement concerns response data and should not be assumed to describe missing progression-free survival data. For progression-free survival, the record gives analysis counts of 36 for T-DM1 with palbociclib and 11 for single-agent T-DM1, but does not explain all differences between those analysis populations and total enrollment of 55.

The endpoint description specifies how participants without documented progression or death are censored. It does not provide counts by group, reasons for missing radiographic assessments, or sensitivity analyses for missing data and censoring. The registry record also does not report whether the analysis followed an intention-to-treat principle or another prespecified analysis population. These gaps limit assessment of how missingness, group membership, and outcome selection may have affected the reported summaries.

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 5% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: non-cardiac chest pain (7.1%), abdominal infection (7.1%), aspartate aminotransferase increased (7.1%), platelet count decreased (7.1%), hydrocephalus (7.1%).

The registry reports serious adverse events for 8 of 38 participants at risk in the T-DM1 with palbociclib group and 3 of 14 at risk in the single-agent T-DM1 group. It also reports other adverse events affecting 38 of 38 and 14 of 14 participants at risk, respectively. These are counts within the registry’s stated groups; they do not by themselves describe event duration, recurrence, severity beyond the serious-event classification, or the timing of events.

The listed serious adverse event terms include non-cardiac chest pain, abdominal infection, aspartate aminotransferase increased, platelet count decreased, and hydrocephalus, each reported as 7.1%, and anaemia reported as 2.6%. The record does not provide enough context in this information to determine how these term-specific percentages were calculated or whether participants could contribute more than one event.

Deaths are reported as 5 in the combination group and 4 in the single-agent group. These counts do not establish that deaths were caused by treatment. The registry information does not provide cause, timing, or an attribution assessment for these deaths. Group sizes and follow-up may differ, so raw counts alone are not a complete basis for comparing safety. The posted adverse-event summaries should be read as descriptive registry counts, not as estimates of comparative risk.

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

The posted results show that the registry recorded median progression-free survival of 16.9 months for T-DM1 with palbociclib and 8.3 months for single-agent T-DM1. The corresponding analysis groups contain 36 and 11 participants. The combination-group confidence interval is incomplete in the record, while the single-agent estimate has a reported 95% confidence interval of 2.7 to 8.8 months. These details indicate that the medians have different amounts of reported uncertainty and that the combination estimate cannot be fully characterized from the interval as posted.

These summaries do not establish that the combination caused a longer progression-free survival time. The record does not include a hazard ratio, a formal statistical test, or a reported between-group confidence interval. In addition, the study is described as single-arm in its title and summary, but the record lists two treatment groups and labels allocation as randomized while specifying a single-group model. Without clarification of design and assignment, a causal comparison is not supported by the registry information alone.

The endpoint is based on investigator-determined progression or death, with censoring for participants whose event status is not observed through follow-up. The registry does not report enough information about scans, censoring, or analysis assumptions to assess how those features affected the estimates. The reported medians are group-level summaries, not predictions for an individual participant.

The record also lists response and overall survival as secondary outcomes, but their results are not included in the posted information. Safety counts are descriptive and do not establish treatment attribution or comparative risk. A complete interpretation would require the design and analysis details, complete uncertainty estimates, and outcome results that are not reported in the registry record.

13. Why This Trial Matters Statistically

This trial is statistically notable because its registry descriptions do not align cleanly. The official title and brief summary describe a single-arm study of T-DM1 with palbociclib, yet the record includes a single-agent T-DM1 group, labels allocation as randomized, and identifies the model as single group. That discrepancy matters: the meaning of group-level medians depends on whether the groups were planned as a valid randomized comparison or arose in another way. The record does not explain the apparent mismatch.

The endpoint also illustrates why a median alone is not a complete treatment comparison. Progression-free survival incorporates censored follow-up, and its median is a central time-to-event summary rather than a measure of relative event rate. A complete comparative interpretation would usually need an appropriate treatment-effect estimate and its uncertainty. Neither a sponsor-reported hazard ratio nor a formal comparison result is included in the posted analyses.

Unequal analysis-group sizes and incomplete uncertainty reporting add to the interpretive limits. The registry reports 36 participants in the combination analysis and 11 in the single-agent analysis; the combination median has no upper confidence bound reported. These features make it especially important not to interpret the apparent separation of medians as proof of a reliable difference.

The record is also a reminder that enrollment, endpoint-specific analysis populations, and response populations can differ. The record reports 55 enrolled, gives separate progression-free survival group counts, and notes that 5 patients lacked response data before the first scan. Those figures answer different questions and should not be conflated. For researchers, the case emphasizes the value of transparent design descriptions, complete uncertainty estimates, explicit analysis populations, and accessible outcome-specific results.

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 principal limitation is ambiguity in the registry’s account of the design. The study title and brief summary say single arm, the model is listed as single group, allocation is labeled randomized, and two treatment groups are reported. The registry record does not explain whether the single-agent group was planned, how participants were allocated, or what role that group played in the analysis. This prevents a confident interpretation of the medians as results from a randomized comparison.

The outcome reporting is incomplete for interpretation. The combination-group median is reported with a lower confidence bound but no upper bound. No hazard ratio, formal comparison test, or other sponsor analysis is reported. The record lists response and overall survival as secondary outcomes but does not include their results in the posted information. The original sample-size justification and detailed statistical plan, including any interim or multiplicity procedures, are not reported in the registry record.

Analysis-population details are limited. Total enrollment is 55, while the progression-free survival groups include 36 and 11 participants. The record separately reports 5 patients without response data before the first scan and a final response sample size of 47. It does not fully reconcile these populations or report all reasons for non-inclusion in each analysis.

Important time-to-event details are also absent, including group-specific numbers censored, follow-up distributions, scan schedules, and sensitivity analyses. These details would help readers evaluate the precision and robustness of the progression-free survival estimates. The safety summaries lack event timing and attribution information, so raw counts cannot establish comparative or causal safety effects.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: T-DM1 and Palbociclib for Metastatic HER2 Breast Cancer — NCT03530696.
• Registry record last updated 2024-08-13; status last verified 2024-07.
• Results first posted 2024-07-24.
• 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

The registry describes a completed Phase II study of T-DM1 with palbociclib in metastatic HER2-positive breast cancer, with 55 participants enrolled. It lists median progression-free survival as the primary outcome, defined from first treatment to investigator-determined progression under RECIST 1.1 or death from any cause. The reported median is 16.9 months for the combination group and 8.3 months for the single-agent T-DM1 group.

The design descriptions are inconsistent: the title and summary describe a single-arm study, while the record lists two treatment groups and labels allocation as randomized alongside a single-group model. The combination estimate has only a lower confidence bound reported; no formal between-group analysis is posted. Response and overall survival are listed as secondary outcomes, but their results are not included in the reported information. These limitations mean the medians are best read as descriptive registry summaries rather than evidence of a confirmed comparative effect.