A Study of Trastuzumab Emtansine Versus Capecitabine + Lapatinib in Participants With HER2-positive Locally Advanced or Metastatic Breast Cancer: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study of Trastuzumab Emtansine Versus Capecitabine + Lapatinib in Participants With HER2-positive Locally Advanced or Metastatic Breast Cancer (NCT00829166), a randomized phase 3 study. The primary outcome is Percentage of Participants With PD or Death as Assessed by an Independent Review Committee (IRC).

2. Clinical Question

This is a comparison of assigned treatment strategies in a randomized, open-label, parallel-group trial. The primary endpoint was assessed by an independent review committee, while a secondary progression endpoint was assessed by investigators. Those endpoints use related clinical events but differ in who assessed progression. The clinical question is therefore specifically about the committee-assessed progression-or-death outcome in the intention-to-treat population, rather than every measure of disease control or safety collected during follow-up.

3. Trial Design

The registry identifies NCT00829166 as a completed, randomized, multicenter, phase III, open-label, parallel-group study sponsored by Hoffmann-La Roche. Its title describes a comparison of trastuzumab emtansine against capecitabine plus lapatinib in participants with HER2-positive locally advanced or metastatic breast cancer. The official title specifies prior trastuzumab-based therapy. The recorded enrollment is 991.

Participants were assigned to one of two treatment groups. The trastuzumab emtansine group received 3.6 mg/kg by intravenous infusion over 30–90 minutes on Day 1 of each 21-day cycle. The combination group received lapatinib 1250 mg orally once daily during each cycle and capecitabine 1000 mg/m² orally twice daily on Days 1–14. The registry describes treatment continuation until investigator-assessed progression, unmanageable toxicity, or study termination.

The study was open-label, so participants and treating teams were not masked to assigned treatment. The primary progression assessment, however, was made by an independent review committee. Randomization and intention-to-treat analysis are relevant to interpretation: the primary population included all randomized participants and retained their randomized group assignment for analysis. The reported primary endpoint group sizes are 495 and 496.

The registry lists a start date of 2009-02, primary completion in 2012-07, and completed status. The primary endpoint timeframe extends from randomization to the data cut-off date of 14 Jan 2012, described as up to 2 years, 11 months. After progression was reported, the study summary says participants were to be followed for survival every 3 months until death, loss to follow-up, withdrawal of consent, or study termination. The record does not report the original design’s detailed sample-size assumptions or a full statistical analysis plan.

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: Participants of "Lapatinib + Capecitabine" arm were allowed to cross over to receive trastuzumab emtansine based on statistically significant Overall Survival (OS) benefit in favor of trastuzumab emtansine demonstrated in second interim analysis (cut-off date 31 July 2012). The safety analysis of the arm was then reported.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: ITT population included all randomized participants on the basis of the treatment assigned at 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.

Population: Intent-to-treat (ITT) population included all randomized participants on the basis of the treatment assigned at randomization.

6. Primary Endpoint

The registry lists the primary outcome as Percentage of Participants With PD or Death as Assessed by an Independent Review Committee (IRC), measured over From the date of randomization through the data cut-off date of 14 Jan 2012 (up to 2 years, 11 months).

The primary endpoint was the percentage of participants with disease progression or death from any cause, assessed by an independent review committee. It covered the period from randomization through the data cut-off date of 14 Jan 2012, described as up to 2 years, 11 months. The primary analysis used the intention-to-treat population, including all randomized participants according to their assigned group.

Progression was assessed using modified RECIST. At baseline, measurable lesions were identified as target lesions and their longest diameters were combined into a baseline sum. Progression for target lesions was defined as an increase of at least 20% in that sum, using the smallest sum recorded since treatment began as the reference, or the appearance of one or more new lesions. For non-target lesions, progression was defined as new lesions or unequivocal progression of existing non-target disease. The endpoint also counted death from any cause, whether or not progression had first been recorded.

General statistical context: a binary endpoint such as this records whether each participant experienced the specified event during the assessment period. Reporting the percentage with an event gives an interpretable group-level summary, but it does not show when events occurred or account directly for differences in follow-up time. Time-to-event outcomes, such as progression-free survival, address timing separately. The committee assessment is useful for applying a common review process in an open-label trial, although the registry record does not describe further details of committee procedures.

7. Results Posted in the Registry

• Lapatinib + Capecitabine: 61.3% (304/496) versus 53.5% (265/495) on control; risk difference 7.8 percentage points (95% CI 1.6 to 13.9), risk ratio 1.14 (1.03–1.28); unadjusted two-proportion z-test p = 0.014.

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 PD or Death as Assessed by the Investigator

• 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:
  • Lapatinib + Capecitabine: 67.5% (335/496) versus 58.0% (287/495) on control; risk difference 9.6 percentage points (95% CI 3.6 to 15.6), risk ratio 1.16 (1.06–1.28); unadjusted two-proportion z-test p = 0.002.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically higher for Lapatinib + Capecitabine than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

PFS as Assessed by the Investigator

• Endpoint type: Time to event: compared as a hazard ratio.
• Test: Hazard ratio as reported by the sponsor (Log Rank); it was not recalculated.
• Result:
  • Trastuzumab Emtansine vs Lapatinib + Capecitabine: hazard ratio 0.66 (95% CI 0.56–0.77), reported p < 0.001 (as reported in the registry).
• Interpretation:
  • The 95% confidence interval excludes no difference (a ratio of 1), so the hazard of the event is statistically lower for Trastuzumab Emtansine vs Lapatinib + Capecitabine than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Percentage of Participants With Objective Response (OR) as Assessed by an IRC

• 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:
  • Lapatinib + Capecitabine: 30.8% (120/389) versus 43.6% (173/397) on control; risk difference -12.7 percentage points (95% CI -19.4 to -6.0), risk ratio 0.71 (0.59–0.85); unadjusted two-proportion z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically lower for Lapatinib + Capecitabine than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

The posted secondary results broadly follow the direction of the primary result, but they examine different outcome definitions. Investigator-assessed progression or death is another event-percentage measure. Progression-free survival uses event timing, and objective response asks whether a specified tumor response occurred according to an independent review committee. The overall pattern is therefore one of related findings across progression and response measures, rather than repeated estimates of the same endpoint.

This consistency can make the primary result easier to place in context, but it does not make the secondary outcomes interchangeable or independently confirm the primary endpoint. The investigator assessment is susceptible to a different assessment process from committee review. Progression-free survival depends on time-to-event methods and follow-up. Objective response uses a different denominator from the primary analysis. Each result answers a distinct question and has its own uncertainty and assumptions.

The registry lists additional outcomes, including duration of objective response, clinical benefit, treatment failure, time to treatment failure, and symptom progression. Numerical results for these outcomes are not reported in the registry record. They should not be presumed to have matched the direction of the posted endpoints.

General statistical context: when many endpoints are tested, the chance of obtaining at least one apparently notable result can rise unless the testing strategy accounts for multiplicity. Secondary analyses are often less definitive than a prespecified primary comparison, particularly when their testing hierarchy or adjustment procedure is not known. The registry does not report a multiplicity plan for these outcomes. The unadjusted secondary results therefore carry less weight as confirmatory evidence than a result from a clearly specified primary testing strategy, and their p-values should be interpreted as endpoint-specific rather than as a single overall test of the study.

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

• Percentage of Participants With PD or Death as Assessed by the Investigator
• PFS as Assessed by the Investigator
• Percentage of Participants With Objective Response (OR) as Assessed by an IRC
• Duration of Objective Response (DOR) as Assessed by an IRC
• Percentage of Participants With Clinical Benefit as Assessed by an IRC
• Percentage of Participants With Treatment Failure
• Time to Treatment Failure
• Percentage of Participants With Symptom Progression

9. Statistical Interpretation

10. Statistical Methods Explained

General statistical context: randomized trials compare outcomes across groups assigned by chance. Randomization supports a fair comparison on average, while the intention-to-treat approach analyzes participants according to assigned group, regardless of later treatment changes. In this trial, the primary analysis population is explicitly defined as all randomized participants on the basis of treatment assigned at randomization. The record does not provide additional details about deviations from assigned treatment or how these were handled.

What does the primary percentage represent?

The primary measure is an event percentage: the proportion of participants in each group who had progression assessed by an independent review committee or died from any cause during the stated period. The registry reports 53.5% for trastuzumab emtansine and 61.3% for lapatinib plus capecitabine. A lower percentage means fewer participants in that group were recorded with the combined endpoint by the cut-off. It does not indicate how soon those events occurred, and it does not separate progression from death.

For a binary outcome, a risk difference expresses the absolute contrast in event probability between groups. The simple recalculation uses the posted counts and group sizes without adjustment for other variables. A risk ratio instead compares the event probabilities as a ratio, while an odds ratio compares event odds. These measures answer related but different questions, and their numerical scales are not interchangeable. The registry’s primary endpoint has no sponsor-reported effect estimate or confidence interval listed, so interpretation should distinguish the reported group percentages from the recalculated comparison.

Why are the secondary outcomes different?

Progression-free survival is a time-to-event outcome: it uses event timing as well as whether an event was observed. The registry reports medians with 95% confidence intervals and a hazard ratio estimated using a log-rank method. General statistical context: the median is the time by which half the relevant event-time distribution has experienced the event, under the analysis definition. A hazard ratio compares event rates over follow-up under the model and assumptions used; it is not a ratio of medians and does not directly give the difference in median time.

The posted progression-free survival hazard ratio is 0.658, with a 95% confidence interval from 0.560 to 0.774 and p < 0.001 by the reported log-rank analysis. The reported median values are 9.4 months for trastuzumab emtansine and 5.8 months for lapatinib plus capecitabine. These are registry-reported secondary results, not the primary binary endpoint. The record does not describe further model diagnostics or the detailed censoring implementation.

Objective response is another binary outcome, assessed by an independent review committee. The posted response percentages are 43.6% and 30.8%, with group-specific 95% confidence intervals. The registry reports a Mantel-Haenszel chi-squared test for a difference in response rates, with p< 0.001. The arm denominators for this outcome are 397 and 389, rather than the primary endpoint denominators. The registry does not explain the difference in denominators, so it should not be attributed to a particular reason.

What do confidence intervals and p-values convey?

General statistical context: a confidence interval describes a range of estimates compatible with the observed data and the interval procedure, under its assumptions. It communicates uncertainty about an estimate; it is not a range containing a fixed proportion of individual participant outcomes. If an interval for a difference includes the no-difference value, the data are compatible with no difference at the corresponding confidence level, though that alone does not establish equivalence. A p-value measures how unusual data at least as extreme as those observed would be under a specified null model. It is not the probability that the null hypothesis is true.

The posted hazard-ratio interval lies below 1, while the interval for each group’s median shows uncertainty around that group’s estimate. The response outcome has its own group-specific intervals and reported test. These summaries should be read with their endpoint definitions and analysis populations, rather than treated as a single uniform analysis. The primary simple recalculation is unadjusted; its p-value is not a substitute for a sponsor-reported analysis.

Why can event percentages and survival analyses tell different things?

A percentage with an event treats the stated follow-up period as a window and summarizes who experienced an event within it. Time-to-event methods retain information about the timing of events and generally account for participants whose event time is not observed through the end of follow-up using censoring methods. General statistical context: the validity of those methods depends on assumptions about censoring and the analysis definition. The registry reports the time-to-event analysis method and estimates for progression-free survival, but detailed censoring rules and diagnostics are not reported in the registry record.

The trial also lists further secondary outcomes, including duration of objective response, clinical benefit, treatment failure, and symptom progression. Their titles are present in the registry, but numerical results for those outcomes are not reported in the registry record. No conclusions about their values or direction should be inferred from the endpoints that do have results.

11. Confidence Intervals and Precision

Precision is best assessed using the uncertainty intervals attached to the estimates, while keeping each interval tied to its endpoint and population. For progression-free survival, the median was 9.4 months for trastuzumab emtansine, with a 95% confidence interval from 7.49 to 10.78, and 5.8 months for lapatinib plus capecitabine, with a 95% confidence interval from 5.59 to 6.93. The registry also reports a hazard ratio of 0.658 with a 95% confidence interval from 0.560 to 0.774.

For objective response, the posted percentages were 43.6% and 30.8%. Their group-specific 95% confidence intervals were 38.6% to 48.6% and 26.3% to 35.7%, respectively. These intervals express uncertainty about the response percentage in each analysis group; they are not individual-level ranges. The response analysis used denominators of 397 and 389, which differ from those for the primary endpoint.

The primary endpoint results are posted as percentages, and the unadjusted recalculation supplies a p-value, but a confidence interval for that primary comparison is not reported in the registry record. A p-value alone does not show the range of plausible effect sizes and should not be treated as a measure of precision. Likewise, the width of an interval is affected by the amount of information available and by the variability of the outcome. General statistical context: wider intervals indicate less precise estimation, while narrower intervals indicate greater precision, subject to the model and sampling assumptions.

Intervals for different endpoints should not be compared as if they estimate the same quantity. The hazard-ratio interval concerns relative event rates over time; median intervals concern group-specific event times; response intervals concern percentages. The record does not report a confidence interval for every secondary outcome, and intervals do not resolve questions about multiplicity, missing data, or clinical importance.

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 illustrations use the observed effect as a reference and show the number of participants per arm for the binary outcomes. For the time-to-event outcome, the figures are total events, not participants per arm. These are planning illustrations based on a two-sided significance level of 0.05, equal allocation between the two arms, a normal approximation, and an effect to detect set at the stated fraction of the observed effect. The trial’s original design assumptions are not reported in the registry record.

The 80% power column gives the planned group size, or event count for time-to-event planning, associated with an 80% chance of detecting the specified effect under those assumptions if that effect is real. The 90% power column uses a higher target chance of detection, so it generally requires more information. These planning values are not observed power; observed power is not a useful substitute for estimating uncertainty from the results already observed.

Rows based on 100% of the observed effect use the full observed contrast as the planning target; rows based on 75% or 50% use smaller target effects. Smaller effects are harder to distinguish from random variation and therefore need more participants or events. For example, the primary binary outcome’s planning figures range from 638 participants per arm for 80% power at 100% of the observed effect to 2,578 participants per arm at 50% of the observed effect. The corresponding 90% figures are 853 and 3,451 participants per arm. These are illustrative calculations, not a reconstruction of the trial’s original sample-size justification.

For progression-free survival, the analogous figures are total events. The observed-effect illustration gives 180 total events for 80% power and 240 for 90% power; the 50% effect illustration gives 717 and 960 total events, respectively. Event targets cannot be read as participant totals because participants may have different follow-up and not all will experience the event during observation.

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 = 0.535; treatment proportion p₁ = 0.613; average p̄ = 0.574.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.6993 = 1.3706.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.6971 = 0.5867.
4. n = (1.3706 + 0.5867)² / (0.0775)² = 637.0, rounded up to 638 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 <- 0.535353535353535                          # control proportion
p1 <- 0.612903225806452                          # 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(829166)
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 = 638 (80% power), 853 (90% power)
 75% of observed effect: n per arm = 1,140 (80% power), 1,526 (90% power)
 50% of observed effect: n per arm = 2,578 (80% power), 3,451 (90% power)
power.prop.test: 637 
Simulated power at n = 638 per arm: 0.801 (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. Multiplicity, Missing Data and Analysis Populations

The registry record does not report a multiplicity adjustment or a testing hierarchy for the secondary outcomes. It lists multiple secondary measures, and numerical results are posted for investigator-assessed progression or death, progression-free survival, and objective response. Without a reported plan describing which comparisons were confirmatory and how repeated testing was handled, the secondary p-values should not be treated as if each were the only hypothesis tested. General statistical context: testing many outcomes can increase the chance of at least one small p-value under a setting where none of the effects is real. Adjustment methods or a prespecified hierarchy can address that issue, but neither is described in the registry record.

The registry defines the primary analysis population as all randomized participants, analyzed according to assigned treatment. This is the intention-to-treat population. The primary arm sizes are reported, as are different denominators for objective response. The reason for those different denominators and the handling of missing response assessments are not reported in the registry record. The registry also does not describe imputation, sensitivity analyses for missing outcomes, or how participants with incomplete assessment were classified for the primary binary endpoint.

For progression-free survival, the registry reports a log-rank method and a hazard ratio, but detailed censoring rules and missing-data sensitivity analyses are not reported in the registry record. These details matter because time-to-event results depend on how follow-up ends for participants without a recorded event. No assumptions about those procedures should be inferred from the reported estimate alone.

The randomized design, open-label status, independent review of the primary progression endpoint, and intention-to-treat definition are stated in the record. Further information about protocol deviations, analysis-set exclusions, interim analyses, or the original statistical plan is not reported in the registry record. These gaps limit assessment of whether all analyses followed a prespecified plan, but they do not change the numerical summaries that are posted.

15. Safety Results

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

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: diarrhoea (3.5%), vomiting (2.0%), pulmonary embolism (1.8%), pyrexia (1.6%), respiratory tract infection (1.5%).

Serious adverse events were recorded for 92 of 490 participants at risk in the trastuzumab emtansine group and 99 of 488 in the lapatinib plus capecitabine group. The registry also lists a safety group titled lapatinib plus capecitabine/trastuzumab emtansine, with 19 of 136 participants at risk affected. Other adverse events were recorded for 474 of 490, 471 of 488, and 115 of 136 participants, respectively, for the groups as listed. These are counts within the registry’s safety populations; they are not the primary efficacy analysis denominators.

The most frequently listed serious adverse-event terms include diarrhoea (3.5%), vomiting (2.0%), pulmonary embolism (1.8%), pyrexia (1.6%), respiratory tract infection (1.5%), and thrombocytopenia (0.8%). The registry lists deaths as null, so death counts for the safety reporting are not reported in the registry record. The record does not provide enough detail to infer comparative duration, severity distributions beyond the serious-event classification, or event causality. These safety counts should be read as descriptive reporting, not as a standalone treatment comparison.

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

The posted primary results show that the percentage with committee-assessed progression or death by the data cut-off was lower in the trastuzumab emtansine group than in the lapatinib plus capecitabine group. This statement describes the recorded group-level outcome in this randomized trial. It does not, by itself, establish the overall clinical value of either treatment or support an individual treatment decision.

The endpoint combines progression and death and is bounded by the stated follow-up period. It does not distinguish how much of the result came from progression versus death, identify when events occurred, or summarize all dimensions of health and treatment experience. The secondary progression-free survival outcome contributes timing information, while response and safety measures address other aspects. None of these measures alone provides a complete account of benefits and harms.

Randomization and intention-to-treat analysis strengthen the comparison of assigned groups, while open-label treatment and unreported details of missing-data handling and multiplicity constrain interpretation. Independent review of the primary progression endpoint provides a distinct assessment process, but it does not remove every possible source of bias or uncertainty. The primary endpoint’s confidence interval is not reported in the registry record, so the unadjusted p-value should not be mistaken for a complete description of precision.

General statistical context: statistical evidence and clinical importance are separate questions. A small p-value does not quantify how meaningful an effect is to participants, and a difference in an outcome percentage does not show whether the balance of outcomes is favorable in every setting. The trial results pertain to its enrolled population, treatment regimens, follow-up, endpoint definitions, and analysis methods. The registry record does not establish that results transfer unchanged to populations or circumstances beyond those studied.

The right interpretation is therefore bounded: the posted summaries support a between-group contrast for particular measured outcomes during the recorded study period, with varying levels of detail and precision. They do not establish a mechanism, guarantee outcomes for individuals, or replace a full assessment of safety, quality of life, treatment burden, and other relevant evidence. Those additional matters are not all reported in the registry record.

17. Why This Trial Matters Statistically

This trial illustrates why endpoint definition matters in biostatistics. The primary result is a percentage with a combined progression-or-death event, while progression-free survival uses event timing and objective response measures whether a response was recorded. The results may align directionally, but they estimate different quantities and rely on different analysis populations and methods. Keeping these distinctions clear prevents a binary percentage from being mistaken for a time-to-event estimate or a response rate.

It also shows the value and limits of independent outcome assessment in an open-label study. Treatment assignment was not masked, but progression for the primary endpoint was assessed by an independent review committee using modified RECIST. That design feature is relevant to potential assessment bias, yet it does not answer every question about conduct, missing observations, or analysis choices. Those details are not all reported in the registry record.

The registry contains several layers of statistical evidence: arm-level percentages, an unadjusted recalculation for the primary binary outcome, reported survival analysis, and a reported test for objective response. A reader should distinguish calculations based on posted summaries from sponsor-reported analyses. It is also important that multiple secondary endpoints exist without a reported multiplicity plan, so secondary p-values do not automatically carry the same confirmatory status as a prespecified primary test.

Finally, the planning illustrations demonstrate how sample-size needs depend on the effect targeted and the desired power, and why event counts differ from participant counts in time-to-event planning. They are not the original trial design justification. Taken together, the record is a useful example of interpreting endpoint-specific evidence while keeping precision, multiplicity, analysis population, and unreported design details in view.

18. 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.

Several trial-specific details needed for a fuller assessment are not reported in the registry record. The original sample-size justification, detailed statistical analysis plan, interim-analysis procedures, and multiplicity strategy are not reported. The planning illustrations are calculated from the observed effects and cannot substitute for the original design rationale.

The study was open-label. Although the primary progression outcome was assessed by an independent review committee, the record does not describe committee procedures in detail or explain how assessment disagreements were handled. The secondary progression endpoint was investigator-assessed, so its assessment process differs from the primary endpoint. The registry does not report whether or how that distinction affected treatment decisions or outcome evaluation.

Missing-data procedures are not reported for the primary event percentage or for the secondary response outcome. The objective-response denominators differ from the primary endpoint denominators, but the reason is not reported. The record does not specify imputation, sensitivity analyses, or the handling of participants without an evaluable assessment. For progression-free survival, the detailed censoring rules and related sensitivity analyses are also not reported.

The results cover the data cut-off of 14 Jan 2012, described as up to 2 years, 11 months. The primary endpoint is a composite event percentage and does not show event timing or separate progression from death. The registry lists other secondary outcomes, but numerical results for several of them are not reported in the registry record. Safety reporting includes a combination-treatment group as well as the two main treatment groups, and the registry does not give a death count for the safety results. These limits argue for a focused interpretation of the outcomes actually posted rather than assumptions about unreported analyses.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: A Study of Trastuzumab Emtansine Versus Capecitabine + Lapatinib in Participants With HER2-positive Locally Advanced or Metastatic Breast Cancer — NCT00829166.
• Registry record last updated 2016-10-31; status last verified 2016-09.
• Results first posted 2013-04-23.
• 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.

22. Record Summary

NCT00829166 is a completed, randomized, multicenter, phase III, open-label, parallel-group trial sponsored by Hoffmann-La Roche. It enrolled participants with HER2-positive locally advanced or metastatic breast cancer who had received prior trastuzumab-based therapy. The comparison was trastuzumab emtansine versus lapatinib plus capecitabine.

The primary endpoint was the percentage with disease progression assessed by an independent review committee or death from any cause, analyzed in the intention-to-treat population through the data cut-off of 14 Jan 2012. The posted percentages were 53.5% and 61.3%, respectively. A simple unadjusted recalculation gives p=0.014; a sponsor-reported primary analysis is not listed in the registry record.

Secondary results include investigator-assessed progression or death, progression-free survival, and objective response. Multiple secondary outcomes are listed, but numerical results for several are not reported in the registry record. The original design assumptions, multiplicity plan, and detailed missing-data procedures are also not reported.