A Study of Trastuzumab Emtansine Versus Taxane in Participants With Human Epidermal Growth Factor Receptor 2 (HER2)-Positive Advanced Gastric Cancer: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study of Trastuzumab Emtansine Versus Taxane in Participants With Human Epidermal Growth Factor Receptor 2 (HER2)-Positive Advanced Gastric Cancer (NCT01641939), a randomized phase 2, phase 3 study. The primary outcome is Overall Survival (OS)- Phase 3.

2. Clinical Question

The trial was randomized, multicenter, parallel-group, and unmasked. It began as an adaptive phase 2/3 study with three treatment arms: two trastuzumab emtansine schedules and standard taxane therapy. After a regimen-selection analysis, the 2.4 mg/kg schedule was selected for further evaluation. The phase 3 results combine participants from the standard taxane and selected trastuzumab emtansine groups across phase 2 and phase 3 (Stage 2), as described in the registry record.

The key statistical question is about comparative survival in the randomized groups, not whether either regimen changed outcomes relative to no treatment. The registry reports a hazard ratio of 1.15, with a 95% confidence interval from 0.87 to 1.60 and p = 0.859. These statistics describe the estimated contrast and its uncertainty; they do not by themselves establish that the regimens are equivalent or that any observed difference is clinically important. The record does not report a prespecified equivalence or non-inferiority conclusion.

3. Trial Design

This was a randomized, multicenter, adaptive phase 2/3 study with parallel treatment groups and no masking. The registry lists enrollment of 415 participants and a terminated status. The study began on 2012-09-03, and its primary completion date was 2015-06-30. The registry does not report why the study was terminated.

At the beginning of the trial, participants were assigned among three arms in a 2:2:1 ratio: trastuzumab emtansine 3.6 mg/kg intravenously every 3 weeks, trastuzumab emtansine 2.4 mg/kg intravenously every week, or investigator-selected standard taxane therapy. The taxane choices were docetaxel or paclitaxel. The regimen-selection analysis was planned after approximately 100 participants across the three arms had received treatment for at least 12 weeks. An Independent Data Monitoring Committee was to select the trastuzumab emtansine schedule for the second stage.

The primary outcome was overall survival. The phase 3 results for the selected 2.4 mg/kg regimen and taxane group combine phase 2 and phase 3 (Stage 2) data. The primary analysis description says the confirmatory comparison was restricted to the taxane arm and the selected trastuzumab emtansine arm. The primary population included all randomized participants, grouped by randomized assignment.

The registry record does not report the original target sample size, the rationale for the enrollment target, detailed randomization procedures, allocation concealment, or the full adaptive decision rules. It also does not report the statistical analysis plan in enough detail to establish how all design decisions were prespecified. These details should not be inferred from the adaptive label alone.

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: A total of 415 participants were randomized, of these 117 participants in taxane arm, 228 participants in 2.4 milligram per kilogram (mg/kg) trastuzumab emtansine arm, and 70 participants in 3.6 mg/kg trastuzumab emtansine arm received at least one dose of the treatment.

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; participants grouped according to the therapy they were randomized to receive.

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) analysis population included all randomized participants; participants grouped according to the therapy they were randomized to receive.

6. Primary Endpoint

The registry lists the primary outcome as Overall Survival (OS)- Phase 3, measured over Date of randomization until death (up to 2 years 3 months).

Overall survival was defined as the time from randomization to death from any cause. The registry reports median overall survival in months, estimated using the Kaplan-Meier method, and a hazard ratio from a log-rank analysis. Participants without a reported death before the analysis cutoff of 30 June 2015 were censored at the latest date before the cutoff when they were known to be alive.

Overall survival is a time-to-event endpoint: it uses both the timing of events and the fact that some participants had not yet experienced the event at the analysis cutoff. The median is the estimated time by which half of the relevant survival distribution has experienced the event. A hazard ratio summarizes the relative event rate over follow-up under the model and analysis used; it is not a difference in median survival and does not directly describe an individual participant’s outcome.

General: Overall survival is often used in clinical trials because death is an objective event and can be defined without relying on subjective symptom or imaging judgments. It can nevertheless require extended follow-up, and censoring means that not every participant’s event time is observed during the study. The interpretation therefore depends on the follow-up, censoring, and analysis procedures. The registry describes the censoring rule and Kaplan-Meier estimation, but additional details of the survival analysis plan are not reported in the registry record.

7. Results Posted in the Registry

• Standard Taxane Therapy vs Trastuzumab Emtansine 2.4 mg: hazard ratio 1.15 (95% CI 0.87–1.60), reported p = 0.859 (as reported in the registry).

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 Disease Progression or Death According to Modified Response Evaluation Criteria in Solid Tumors (mRECIST v1.1) - Phase 3

• 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:
  • Trastuzumab Emtansine 2.4 mg: 93.0% (212/228) versus 88.9% (104/117) on control; risk difference 4.1 percentage points (95% CI -2.5 to 10.7), risk ratio 1.05 (0.97–1.13); unadjusted two-proportion z-test p = 0.195.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Trastuzumab Emtansine 2.4 mg and the control. The interval gives the range of effects that remain compatible with the data.

Progression Free Survival (PFS) According to Modified Response Evaluation Criteria in Solid Tumors (mRECIST v1.1) - Phase 3

• 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:
  • Standard Taxane Therapy vs Trastuzumab Emtansine 2.4 mg: hazard ratio 1.13 (95% CI 0.89–1.43), reported p = 0.308 (as reported in the registry).
• Interpretation:
  • The 95% confidence interval includes no difference (a ratio of 1), so these data do not show a statistically significant difference between Standard Taxane Therapy vs Trastuzumab Emtansine 2.4 mg and the control. The interval gives the range of effects that remain compatible with the data.

Percentage of Participants With Objective Response According to mRECIST v1.1 - Phase 3

• 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:
  • Trastuzumab Emtansine 2.4 mg: 20.6% (42/204) versus 19.6% (20/102) on control; risk difference 1.0 percentage points (95% CI -8.5 to 10.5), risk ratio 1.05 (0.65–1.69); unadjusted two-proportion z-test p = 0.841.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Trastuzumab Emtansine 2.4 mg and the control. The interval gives the range of effects that remain compatible with the data.

Across the posted secondary outcomes, the results cover several different dimensions: progression or death, progression-free survival, objective response, duration of response, quality of life, and gastric-cancer symptoms. These outcomes are not interchangeable. A response proportion describes tumor response under the stated criteria, while a time-to-event outcome incorporates both timing and censoring. Quality-of-life and symptom outcomes address participant-reported experience rather than survival or tumor response.

The overall pattern does not form a clear, uniform separation across all outcome types. The time-to-event comparison for progression-free survival has a sponsor-reported confidence interval that includes 1, while the posted descriptive results for other outcomes use different scales and analysis populations. The registry does not report sponsor inferential analyses for the posted binary secondary outcomes. Those results can describe what was observed, but they do not establish confirmatory differences on their own.

Secondary outcomes can help characterize a trial beyond its primary endpoint, but their interpretation depends on the prespecified hierarchy, multiplicity control, missing-data rules, and analysis populations. The registry record does not report a complete multiplicity strategy or full methods for all secondary endpoints. When many outcomes are considered without a reported adjustment plan, a nominally notable finding may occur by chance. Unadjusted secondary results therefore carry less weight than a prespecified primary analysis and should be viewed as supportive or exploratory unless the record establishes otherwise.

The secondary results do not replace the overall-survival comparison. They provide additional outcome-specific information, but differences in definitions, timing, and available participant counts limit direct comparison across outcomes. The reasons for differing analysis counts and the handling of missing secondary outcome data are not reported in the registry record.

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

• Percentage of Participants With Disease Progression or Death According to Modified Response Evaluation Criteria in Solid Tumors (mRECIST v1.1) - Phase 3
• Progression Free Survival (PFS) According to Modified Response Evaluation Criteria in Solid Tumors (mRECIST v1.1) - Phase 3
• Percentage of Participants With Objective Response According to mRECIST v1.1 - Phase 3
• Duration of Objective Response (DOR) - Phase 3
• Percentage of Participants With Clinically Significant Improvement in European Organisation for Research and Treatment of Cancer Quality of Life Core Module 30 (EORTC QLQ-C30) Score - Phase 3
• Percentage of Participants With Clinically Significant Improvement in Quality of Life Questionnaire Stomach Cancer Module 22 (QLQ-STO22) Score - Phase 3
• Percentage of Participants With Advanced Gastric Cancer (AGC) Symptom Progression - Phase 3
• Time to Advanced Gastric Cancer (AGC) Symptom Progression - Phase 3

9. Statistical Interpretation

10. Statistical Methods Explained

General: Overall survival is a time-to-event outcome because the outcome includes both whether death occurred and how long it took to occur. Participants who were alive at the analysis cutoff did not have an observed death time within the analysis period. Rather than treating those participants as if they had died at the cutoff, the described analysis censored them at the latest date they were known to be alive before the cutoff.

What does censoring mean?

General: Censoring records that a participant’s event time is not fully observed, while retaining the information that the participant was event-free up to a known time. Kaplan-Meier estimation uses this follow-up information to estimate the survival distribution. Its interpretation depends on assumptions about censoring and follow-up. The registry gives the censoring rule for the primary endpoint, but does not report further diagnostic checks or sensitivity analyses for censoring.

How should the median be read?

General: The median survival is the estimated point at which the survival curve reaches half of its starting level. It is a summary of each group’s estimated survival distribution, not an estimate of the difference between groups. The primary results report medians of 8.6 months for standard taxane therapy and 7.9 months for trastuzumab emtansine 2.4 mg. Comparing those medians alone does not provide the same information as a formal time-to-event comparison, because the full pattern of event timing and censoring contributes to the analysis.

What does the hazard ratio describe?

General: A hazard ratio compares event rates over time according to the contrast and method used. A value of 1 represents equal hazards under that comparison; values on either side represent different relative hazards. The registry reports the primary hazard ratio as 1.15 for the standard taxane versus trastuzumab emtansine comparison. Its 95% confidence interval is 0.87 to 1.60. Because the interval includes 1, the result is compatible with no difference in the relative hazard as well as with values on either side. The ratio is not a ratio of medians and should not be read as a percentage change in median survival.

The registry identifies the log-rank method for the primary comparison and reports p = 0.859. The log-rank test compares the observed and expected event patterns between groups over follow-up. It is a test of evidence against the null hypothesis used for that comparison; a large p-value does not demonstrate that the null is true. The registry does not report the detailed model specification, checks of proportional hazards, or alternative analyses of the primary endpoint. The hazard ratio is reported by the sponsor, and the time-to-event interpretation uses that reported value rather than a newly calculated estimate.

What is the role of the intention-to-treat population?

The registry defines the primary intention-to-treat population as all randomized participants, grouped according to the therapy to which they were randomized. This preserves the randomized group assignment for the primary comparison, regardless of whether treatment was completed. The registry does not provide further detail on exclusions, protocol deviations, or the amount of outcome data available for each analysis. For some secondary outcomes, the participant counts differ from the randomized group counts; the reasons for those differences are not reported in the registry record.

The primary analysis combines the standard taxane group and selected trastuzumab emtansine 2.4 mg group across phase 2 and phase 3 (Stage 2), according to the registry description. The trial began with two trastuzumab emtansine schedules and an adaptive selection process. Such a design can make the selection procedure and its relationship to confirmatory testing important to interpretation. The registry says that the confirmatory analyses were restricted to the taxane arm and the selected trastuzumab emtansine arm, but does not report whether or how the analysis accounted statistically for the adaptive regimen selection.

General: When multiple outcomes or comparisons are tested, the chance of at least one apparently positive result can increase. A multiplicity plan may define which outcomes are confirmatory, how testing is ordered, and whether p-values or confidence intervals are adjusted. The registry record does not report a complete multiplicity strategy for the outcomes in the record. Accordingly, secondary results without sponsor-reported inferential analyses should be treated as descriptive or exploratory rather than as independent confirmatory evidence.

The registry reports Kaplan-Meier estimates and a log-rank comparison for overall survival. It reports the same general time-to-event method for progression-free survival. The posted binary secondary results include percentages and confidence intervals for objective response, while other outcomes are listed without a sponsor-reported comparison in the results data described. The registry does not report the full statistical methods for those secondary outcomes, including the handling of missing outcome data, covariate adjustment, or any multiplicity correction.

11. Confidence Intervals and Precision

The primary median estimates have 95% confidence intervals: 7.1 to 11.2 months for standard taxane therapy and 6.7 to 9.5 months for trastuzumab emtansine 2.4 mg. These intervals convey uncertainty about the estimated medians. They do not mean that 95% of participants survived within those ranges, nor do they directly give a confidence interval for the difference between the medians.

The reported hazard ratio is 1.15, with a 95% confidence interval from 0.87 to 1.60. The interval includes 1, so the data are compatible with no difference in the relative hazard under the reported comparison. It is also broad enough to include values on either side of 1. This limits precision: the point estimate alone gives an incomplete picture, and the interval shows that the estimated comparative effect remains uncertain.

General: A confidence interval describes the range of values reasonably compatible with the data under the statistical method and its assumptions. It is not a guarantee that the true value lies within the interval, and it is not a measure of the probability that one treatment is better. Wider intervals generally indicate less precision than narrower intervals, but interpretation also depends on what values are clinically important. The registry record does not report a threshold for clinical importance or a prespecified margin for equivalence or non-inferiority.

The reported p-value, 0.859, is a measure of how unusual the observed test statistic would be under the null hypothesis and the test assumptions. It does not quantify the probability that the null hypothesis is true. Taken with the interval, it indicates that the posted results do not provide strong statistical evidence of a difference, while leaving uncertainty about the size and direction of any difference. Precision should therefore be judged from the confidence interval, not from the p-value alone.

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 describe sample sizes under stated assumptions; they are not the trial’s original design calculations. For the time-to-event outcomes, the entries are total numbers of events, not participants per arm. For the other outcomes, entries are participants per arm. The assumptions used for these illustrations are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The target effect is a stated fraction of the observed effect. The registry record does not report the original design assumptions or sample-size justification.

The 80% power column gives the planned number of events or participants under the illustration’s assumptions that would yield an 80% chance of detecting the chosen effect if that effect were the true one. The 90% power column gives the corresponding requirement for a 90% chance. These are planning concepts, not guarantees that a study of that size will detect a result. The displayed values are based on the observed effect, so they are retrospective illustrations rather than evidence about the trial’s actual probability of success.

Rows describing 100%, 75%, and 50% of the observed effect ask what sample size would be needed to detect effects of those relative sizes under the stated assumptions. Smaller effects require more participants or, for time-to-event analyses, more events. The figures are not observed power, and they should not be interpreted as evidence that the study was or was not adequately powered under its original plan.

For the primary survival comparison, the displayed planning values are total events. For binary outcomes, values are participants per arm. These distinctions matter: a total-event target is not a per-group enrollment target, and the per-arm figures should not be mistaken for the combined number across both groups. The illustrations do not account for every possible feature of the trial, including its adaptive design, and do not replace the trial’s original sample-size rationale, which is not reported in the registry record.

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

d is the total number of events needed (Schoenfeld's formula) with 1:1 allocation and proportional hazards.

Worked calculation for this trial

1. Hazard ratio HR = 1.150, so ln(HR) = 0.1398.
2. d = 4 × (1.96 + 0.842)² / [ln(1.150)]² = 1,607.3, rounded up to 1,608 events in total 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
hr <- 1.15                              # observed hazard ratio
events_needed <- function(power, frac = 1) ceiling(round(4 * (z_a + qnorm(power))^2 / log(hr^frac)^2, 6))   # Schoenfeld, 1:1 allocation
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: events = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(events_needed(0.80, f), big.mark = ","), format(events_needed(0.90, f), big.mark = ",")))

# Simulation: given d total events under proportional hazards, events in the treatment arm are Binomial(d, hr / (1 + hr)).
# The score (log-rank type) statistic is (D1 - d / 2) / sqrt(d / 4).
set.seed(1641939)
d    <- events_needed(0.80)
reps <- 20000
D1 <- rbinom(reps, d, hr / (1 + hr))
z  <- (D1 - d / 2) / sqrt(d / 4)
cat(sprintf("Simulated power with %s events: %.3f (%d replicates)\n", format(d, big.mark = ","), mean(abs(z) > z_a), reps))

Output from running the code

100% of observed effect: events = 1,608 (80% power), 2,152 (90% power)
 75% of observed effect: events = 2,858 (80% power), 3,826 (90% power)
 50% of observed effect: events = 6,430 (80% power), 8,607 (90% power)
Simulated power with 1,608 events: 0.795 (20000 replicates)

The simulation uses the fact that, with d events in total and proportional hazards, the number of events in the treatment arm is binomial. It is an approximation that conditions on the total number of events, and it should land close to the target power.

14. Analyses Reported by the Sponsor

The registry reports the sponsor’s primary time-to-event analysis as a log-rank hazard ratio of 1.15, with a 95% confidence interval from 0.87 to 1.60 and p = 0.859. The sponsor-reported comparison is the principal inferential result for overall survival. The reported analysis is not replaced by an unadjusted recalculation from group summaries.

For time-to-event outcomes, a hazard ratio generally cannot be recovered reliably from the two medians and their confidence intervals alone. The timing of events and censoring across follow-up contributes to the analysis. The posted primary interpretation therefore uses the sponsor’s reported hazard ratio and log-rank result, alongside the group median estimates. No independent hazard-ratio recalculation is used as a substitute.

For secondary binary outcomes, simple unadjusted recalculations from posted summary statistics may help describe the observed contrast, but they do not reproduce an adjusted or prespecified sponsor analysis. The registry does not report sponsor analyses for the posted binary comparisons. Such recalculations should therefore be distinguished from the sponsor’s confirmatory analyses and treated cautiously, especially when multiple outcomes are available. The registry record does not report enough detail to establish whether a secondary analysis was prespecified or adjusted for multiplicity.

15. Multiplicity, Missing Data and Analysis Populations

The registry identifies overall survival as the phase 3 primary outcome and lists a range of secondary outcomes. It describes the confirmatory comparison as restricted to standard taxane therapy and the selected trastuzumab emtansine 2.4 mg regimen. A complete plan for controlling multiplicity across the primary and secondary outcomes, or across any interim or adaptive analyses, is not reported in the registry record. The record also does not report whether secondary outcome testing followed a formal hierarchy or whether p-values and confidence intervals were adjusted.

General: Multiplicity matters when a study examines multiple outcomes or treatment contrasts. Without an appropriate plan, the chance of obtaining at least one apparently significant result can rise as the number of tests increases. This does not make secondary outcomes uninformative, but it affects how strongly isolated findings should be interpreted. In this record, the primary overall-survival analysis should remain distinct from unadjusted or descriptive secondary results.

The primary analysis population is described as all randomized participants, grouped according to randomized treatment. For overall survival, the registry specifies censoring at the latest date before the cutoff when a participant was known to be alive. Further details about missing data, loss to follow-up, and sensitivity analyses are not reported in the registry record. The registry does not state how missing data were handled for the secondary outcomes, including response, quality of life, and symptom measures.

Some secondary outcomes report analysis counts that differ from the randomized group counts. The reasons for those differences are not reported in the registry record, so they should not be attributed to a particular exclusion, missing-data rule, or assessment process. The registry also does not provide a full account of protocol deviations or analysis-population definitions for every endpoint.

The adaptive design included selection of a trastuzumab emtansine regimen after approximately 100 participants across all three arms had received treatment for at least 12 weeks. The record does not report whether the confirmatory testing strategy adjusted for this selection or for any interim decision process. Those design and analysis details are relevant to inferential interpretation, but should not be inferred from the trial’s adaptive designation.

16. 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: gastric haemorrhage (5.8%), pneumonia (4.5%), pyrexia (4.3%), febrile neutropenia (3.6%), anaemia (3.6%).

The registry reports safety counts by treatment group. For serious adverse events, the listed at-risk counts and affected counts are 111 and 31 for standard taxane therapy, 224 and 65 for trastuzumab emtansine 2.4 mg, and 69 and 23 for trastuzumab emtansine 3.6 mg. The listed counts for other adverse events are 111 and 105, 224 and 211, and 69 and 62, respectively. These figures describe the recorded safety groups and should be read using the at-risk denominators shown in the registry.

The most frequently listed serious adverse-event terms include gastric haemorrhage at 5.8%, pneumonia at 4.5%, pyrexia at 4.3%, febrile neutropenia at 3.6%, anaemia at 3.6%, and upper gastrointestinal haemorrhage at 3.6%. The registry does not provide enough information in these figures to determine comparative risk or to establish causation. Death counts for the safety groups are not reported in the registry record.

Safety figures are not efficacy results. Differences in group size, exposure, event definitions, and follow-up can affect how adverse-event counts are interpreted. The registry record does not report all details needed to assess those factors or to make a complete comparative safety analysis. These figures should therefore be treated as a description of the posted adverse-event record rather than a stand-alone measure of treatment benefit or harm.

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

The posted primary result is an estimated overall-survival comparison between participants randomized to standard taxane therapy and those randomized to the selected trastuzumab emtansine 2.4 mg regimen. The medians were 8.6 months and 7.9 months, respectively. The reported hazard ratio was 1.15, with a 95% confidence interval from 0.87 to 1.60 and p = 0.859. The interval includes 1, and the p-value does not provide strong evidence of a difference under the reported test.

That finding does not prove the two regimens produced identical outcomes. The confidence interval includes a range of possible comparative effects, and the registry does not report an equivalence or non-inferiority analysis. A failure to find a statistically significant difference is not evidence that the groups are equivalent. The result also does not establish the probability that either regimen is better; a p-value is not such a probability.

The reported estimates apply to the registry’s described phase 3 comparison and analysis population. They should not automatically be generalized to other populations, treatment settings, regimens, or outcomes. The study enrolled participants with previously treated locally advanced or metastatic HER2-positive gastric cancer, and the selected regimen was trastuzumab emtansine 2.4 mg. The record describes cumulative phase 2 and phase 3 (Stage 2) results for the selected trastuzumab emtansine and standard taxane arms.

Secondary outcomes add descriptive context, but the registry does not report a complete multiplicity plan or sponsor inferential analyses for all of them. They therefore do not provide a basis for overriding the primary result. The posted safety figures describe adverse events but do not, by themselves, establish a comparative benefit-harm balance. The statistical results are summaries of the recorded data under the reported methods, not medical advice or a recommendation about treatment.

18. Why This Trial Matters Statistically

This trial illustrates how an adaptive phase 2/3 design can create a link between early regimen selection and later confirmatory comparison. Participants were initially assigned among two trastuzumab emtansine schedules and standard taxane therapy; one trastuzumab emtansine regimen was then selected for further evaluation. Understanding how that selection relates to confirmatory inference is important, but the registry record does not report whether statistical adjustments were made for the selection process.

The primary endpoint also demonstrates why time-to-event results require more than a single summary. The group medians describe estimated survival distributions, while the hazard ratio and log-rank test compare the event patterns over follow-up. Confidence intervals are essential because point estimates alone can give a false impression of certainty. In this case, the interval around the reported hazard ratio includes 1 and remains broad enough to leave uncertainty about the comparative effect.

The trial also emphasizes the difference between a primary result and a collection of secondary findings. Secondary outcomes span survival-related events, response, quality of life, and symptoms. Because they address different questions and may involve different analysis populations, their results should not be combined into a single impression of success or failure. Multiplicity and missing-data methods affect how strongly each secondary result can be interpreted, and those methods are not fully reported in the registry record.

Finally, retrospective sample-size illustrations are not the same as a trial’s original design justification. They can show how assumed effect size relates to planning requirements under explicit assumptions, but they cannot establish what the trial was designed to detect or its observed power. The original sample-size rationale is not reported in the registry record.

19. 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.

Several trial-specific limitations affect interpretation. The study was adaptive, with selection of a trastuzumab emtansine regimen after an early treatment period. The registry describes the selection process and the later confirmatory comparison, but does not report whether the primary statistical analysis adjusted for the selection or for any interim decision-making. This leaves uncertainty about the full inferential framework.

The registry reports a terminated status but does not report the reason for termination. It lists enrollment of 415, while the primary phase 3 comparison includes 117 participants in the standard taxane group and 228 in the selected trastuzumab emtansine group. These comparisons are described as cumulative phase 2 and phase 3 (Stage 2) data. The record does not report enough detail to reconcile every enrollment and analysis population count or explain all differences in endpoint-specific denominators.

The primary confidence interval for the hazard ratio is broad and includes 1. Thus, the comparative estimate is uncertain, and the result does not establish equivalence. The record does not report a clinical-importance threshold, equivalence margin, or non-inferiority framework. The original sample-size justification and design assumptions are not reported in the registry record.

Details are also limited for several analysis choices. The registry does not report a complete multiplicity strategy, detailed missing-data methods for secondary outcomes, full model specifications, proportional-hazards checks, or sensitivity analyses. It does not report sponsor inferential analyses for the posted binary secondary outcomes. The safety figures provide counts and selected serious adverse-event terms, but not enough information to assess all factors needed for a comprehensive comparative safety interpretation.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: A Study of Trastuzumab Emtansine Versus Taxane in Participants With Human Epidermal Growth Factor Receptor 2 (HER2)-Positive Advanced Gastric Cancer — NCT01641939.
• Registry record last updated 2017-05-12; status last verified 2017-05.
• Results first posted 2016-08-11.
• 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.

23. Record Summary

The registry record describes a randomized, multicenter, unmasked, adaptive phase 2/3 trial of trastuzumab emtansine versus standard taxane therapy in previously treated HER2-positive advanced gastric cancer. The primary endpoint was overall survival, analyzed using Kaplan-Meier estimates and a log-rank comparison. The confirmatory comparison was between standard taxane therapy and the selected trastuzumab emtansine 2.4 mg regimen, with cumulative phase 2 and phase 3 (Stage 2) results.

Median overall survival was 8.6 months for standard taxane therapy and 7.9 months for trastuzumab emtansine 2.4 mg. The reported hazard ratio was 1.15, with a 95% confidence interval from 0.87 to 1.60 and p = 0.859. The interval includes 1, leaving uncertainty about the comparative effect. The registry lists secondary efficacy, quality-of-life, symptom, and safety outcomes, but does not report a complete multiplicity or missing-data strategy for all outcomes.