Lapatinib With Carboplatin and Paclitaxel in Esophagus and Gastroesophageal Junction (GEJ): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Lapatinib With Carboplatin and Paclitaxel in Esophagus and Gastroesophageal Junction (GEJ) (NCT01395537), a na phase 1, phase 2 study. The primary outcome is PHASE I: Number of Patients That Experience a Grade 3-4 Dose Limiting Toxicity.

2. Clinical Question

The study rationale describes combining lapatinib with carboplatin and paclitaxel in recurrent or metastatic disease. The registry record lists overall survival as a secondary outcome, but does not report an analysis or result for it. Consequently, the recorded question is primarily one of toxicity occurrence in a treated group, not whether the combination performs better than another approach. Neither a treatment effect nor a comparative benefit can be estimated from a single group without additional assumptions or external data.

3. Trial Design

The registry describes a Phase I/II pilot study of lapatinib combined with carboplatin and paclitaxel for recurrent or metastatic adenocarcinoma of the esophagus and gastroesophageal junction. It is listed as single-group, with no masking and allocation recorded as not applicable. The enrollment is 13. These design features indicate that the posted record does not describe a randomized comparison or a blinded assessment.

The primary outcome is a count of participants experiencing a grade 3-4 dose-limiting toxicity. The outcome population is described as all participants who received treatment, and the listed timeframe is after 9 weeks (3 cycles) of treatment. The registry also supplies a toxicity definition referring to events during the first 3 week course. Because these descriptions do not give the same apparent observation window, the exact intended assessment period is not clear from the record.

The record lists Phase I and Phase II, but does not report how the phases were organized, whether enrollment or treatment differed between them, or what criteria governed progression between phases. A dose-escalation plan, treatment schedule, dose levels, and protocol-specific decision rules are not reported in the registry record. The status is terminated; the reason for termination is not reported in the registry record.

The start is recorded as 2011-08 and primary completion as 2014-12. These dates describe the registry record and do not establish how long each participant was followed. The record lists overall survival as a secondary outcome but does not supply its analysis details or results. No comparator arm is described, so interpretation must remain descriptive of the treated group rather than comparative.

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.

Phase I

Phase II

Analysis population for the primary outcome, as defined in the registry: All participants that received treatment

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

6. Primary Endpoint

The registry lists the primary outcome as PHASE I: Number of Patients That Experience a Grade 3-4 Dose Limiting Toxicity, measured over after 9 weeks (3 cycles) of treatment.

The primary endpoint counts participants who experienced a grade 3-4 dose-limiting toxicity. Its stated population is all participants who received treatment. The listed timeframe is after 9 weeks (3 cycles) of treatment. The registry definition describes a DLT as any grade 3-4 non-hematologic toxicity, an increase in bilirubin >= 2 mg/dL (or >2X baseline in participants with Gilbert's syndrome), or AST/ALT elevation > 3.0 X ULN during the first 3 week course of therapy.

A count endpoint gives the number of people meeting the event definition, rather than the number of separate events. It can describe how often a defined toxicity occurred in the treated group, provided the population, event criteria, and observation window are applied consistently. In early-phase research, a toxicity endpoint can inform assessment of tolerability and treatment administration. That general purpose does not make the endpoint a measure of efficacy, nor does the endpoint definition alone establish the safety of a regimen.

There is an apparent timing inconsistency: the endpoint timeframe says after 9 weeks (3 cycles), whereas the definition says during the first 3 week course. The registry record does not clarify whether the definition applies to the full timeframe or only the first course. It also does not report the observed number meeting the endpoint, so the primary endpoint result cannot be described from this record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does a single-group design allow?

In a single-group study, observations come from participants receiving the study regimen without a separate concurrent control group. The result can describe outcomes in that group, but it cannot isolate the regimen's effect from other influences. General statistical interpretation of a treatment effect usually relies on a comparison, and no randomized comparison is described in this record. The design is listed as single-group, allocation as not applicable, and masking as none.

This matters for both efficacy and safety interpretation. A count of toxicities can describe what was recorded among treated participants, but without a comparator it cannot establish whether that count is higher or lower than it would have been under another treatment. Similarly, a secondary outcome such as overall survival would need its analysis and context to be interpreted; the registry record does not report such an analysis.

How should a count endpoint be understood?

The primary outcome is a count of participants with a defined event. It is not the total number of toxicity episodes, and a participant who has multiple qualifying events may still contribute as one participant to a participant count. The registry describes the event as grade 3-4 dose-limiting toxicity, with specified non-hematologic toxicity and laboratory criteria. The observed count and any analysis are not reported in the registry record.

For a binary event outcome, analysts often describe the number of participants with the event relative to the number assessed. A confidence interval, when reported, expresses uncertainty around an estimate under the chosen statistical model; it is not a guarantee that the true value lies in a particular range. The registry record does not report a primary endpoint estimate or confidence interval, so no such quantity can be interpreted for this outcome.

Why do endpoint definitions and timing matter?

A statistical result depends on a clearly defined event and assessment period. Here, the listed timeframe is after 9 weeks (3 cycles), but the definition refers to toxicity during the first 3 week course. These descriptions may refer to different periods, and the registry record does not clarify their relationship. The issue is not merely wording: counting events over a longer or shorter observation period can change what the count represents. The record also does not explain how missing assessments or participants who did not complete the stated period would be handled.

The definition includes thresholds expressed as grade 3-4, bilirubin >= 2 mg/dL, bilirubin >2X baseline for participants with Gilbert's syndrome, and AST/ALT > 3.0 X ULN. These criteria identify what the registry calls a DLT; they do not provide the observed number meeting the criteria. The record does not explain the adjudication process or how overlapping criteria were treated.

What does the analysis record tell us?

The registry lists 3 outcome measures and 0 analyses. That means the record does not present analysis results for the listed outcomes. It does not report a primary endpoint estimate, a p-value, a confidence interval, or an analysis population count. Absence of an analysis entry is not evidence that no analysis was conducted outside the registry; it means that results are not reported in the registry record.

No statistical analysis plan, model specification, or sample-size justification is reported in the registry record. The presence of Phase I and Phase II labels does not establish which statistical methods were planned or applied. General methods used in early-phase studies can include descriptive event counts and dose-finding rules, but the record does not say whether any such rules were used in this study. Those methods should not be attributed to this trial without documentation.

How do planning illustrations differ from study results?

Planning calculations are prospective illustrations of how much information might be needed under stated assumptions; they are not observations from enrolled participants. The planning assumptions described for illustrations use a two-sided significance level of 0.05, equal allocation between two arms, a normal approximation, and a target effect expressed as a stated fraction of the observed effect. The record itself describes this as planning based on the observed effect, not the trial's own design. Those assumptions do not match the listed single-group study design, so any such illustration should not be read as a reconstruction of the study's actual protocol or as observed power.

General statistical power is a property of a planned test under specified assumptions. It is not a measure of whether a study's result is true, and it cannot replace a reported estimate with uncertainty. A sample-size illustration based on an observed effect is also not independent confirmation of that effect. Since the registry record does not report endpoint analyses, no study-specific power or effect estimate can be inferred from the analysis record.

Finally, over-dispersion is a general concern when observed variation exceeds what a statistical model assumes. It matters because model-based uncertainty can be understated if the model does not accommodate the actual variability. No over-dispersion assessment or count model is reported for this study, and the primary endpoint is described as a participant count rather than as a reported event-rate analysis. These general concepts should not be mistaken for methods documented in the registry record.

10. Exact Estimates for Each Arm

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

R code and output

arm    <- c("Lapatinib With Carboplatin and Paclitaxel")
n      <- c(13)
events <- c(1)
for (i in seq_along(n)) {
  ci <- binom.test(events[i], n[i])$conf.int          # Clopper-Pearson exact interval
  cr <- qbeta(c(0.025, 0.975), 1 + events[i], 1 + n[i] - events[i])   # Bayesian, uniform Beta(1,1) prior
  wi <- (function(x, n) { p <- x / n; z <- qnorm(0.975); d <- 1 + z^2 / n
    c((p + z^2 / (2 * n) - z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d, (p + z^2 / (2 * n) + z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d) })(events[i], n[i])
  cat(sprintf("ARM|%s|%d|%d|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f\n", arm[i], n[i], events[i], 100 * events[i] / n[i],
              100 * ci[1], 100 * ci[2], 100 * wi[1], 100 * wi[2], 100 * cr[1], 100 * cr[2]))
}

ARM  Lapatinib With Carboplatin and Paclitaxel  13  1  7.7  0.2  36.0  1.4  33.3  1.8  33.9

11. Multiplicity, Missing Data and Analysis Populations

The registry lists 3 outcome measures, including the primary Phase I DLT count and the secondary outcome of overall survival. It lists 0 analyses. A multiplicity plan describing how interpretation would account for testing or reporting more than one outcome is not reported in the registry record. Without analysis results, the record also does not show whether any adjustment was performed. The presence of several outcomes does not by itself establish that formal hypothesis tests were conducted.

The primary endpoint population is described as all participants who received treatment, while enrollment is 13. The number who received treatment is not reported separately. The record does not state how participants without a complete assessment, those who stopped treatment, or missing outcome information would be handled. Reasons for missing data, the extent of missingness, and any imputation or sensitivity analysis are not reported in the registry record.

These omissions limit interpretation because a count based only on participants with available assessments may differ from a count using a broader analysis population. The record does not explain whether follow-up was complete for the primary timeframe or how death or withdrawal before assessment would affect endpoint classification. No assumptions about missingness or analysis-population membership can therefore be attributed to the study.

Overall survival is listed as a secondary outcome, but its analysis population, censoring rules, follow-up, and result are not reported. Those details are important for interpreting a time-to-event outcome, so the outcome listing alone cannot establish what was analyzed or what the findings were.

12. Safety Results

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

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 0% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: febrile neutropenia (15.4%), diarrhea (7.7%), infections and infestations - other, specify (7.7%), neutropenic fever (7.7%).

The serious adverse event record lists 4 participants affected among 13 at risk in the lapatinib with carboplatin and paclitaxel group. It lists other adverse events for 11 participants among 13 at risk, and records 13 deaths. The registry does not give further detail in this record about the timing, causes, or treatment relationship of the deaths. These counts should be read as registry-reported safety information, not as proof that the treatment caused each event.

The listed serious event terms include Febrile neutropenia at 15.4%, Diarrhea at 7.7%, Infections and infestations - Other, specify at 7.7%, and Neutropenic Fever at 7.7%. These term-specific percentages do not describe the primary DLT endpoint unless the terms also meet its criteria and timeframe. The record does not explain whether terms overlap within participants, how events were coded, or the duration of safety follow-up.

Serious adverse events and dose-limiting toxicities are distinct reporting concepts. The DLT definition includes grade and laboratory thresholds, while the safety listing groups reported adverse events by seriousness and term. The registry record does not report the primary DLT count, so these safety figures cannot be used to fill that gap or to estimate a comparative risk.

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

The record establishes that a Phase I/II single-group study of lapatinib with carboplatin and paclitaxel was registered for recurrent or metastatic adenocarcinoma of the esophagus and gastroesophageal junction, with enrollment of 13. It defines a primary Phase I toxicity outcome and gives descriptive safety counts. It does not report the primary endpoint result, an analysis of overall survival, or a comparative treatment effect.

The distinction between a registered question and a reported result is important. An endpoint description says what investigators intended to measure; it does not tell readers how many participants experienced the event. In this case, the record does not give the DLT count. Likewise, listing overall survival as a secondary outcome does not provide a survival estimate or establish what follow-up was achieved.

Because the study is single-group, any reported outcome would describe the treated participants rather than provide a controlled estimate of what the regimen changes relative to another option. A comparison with an external group would require assumptions about comparability that are not addressed in the record. The adverse event counts are also descriptive and do not establish causation.

The record's terminated status is a study-status entry, not an explanation of why the trial ended or a result about the regimen. The reason for termination is not reported in the registry record. The timing descriptions for the primary endpoint also differ, and the intended assessment window is not clarified. These uncertainties further limit interpretation.

Accordingly, the posted information supports a description of design, enrollment, outcome definitions, and reported adverse event summaries. It does not support a conclusion about the frequency of the primary DLT endpoint, overall survival, or relative effects. Those missing results should not be inferred from the study rationale or from the presence of safety terms.

14. Why This Trial Matters Statistically

This study illustrates why design and reporting details matter as much as an endpoint label. A participant-count outcome is interpretable only when readers know who was assessed, what counted as an event, and over what period events were counted. The record gives a DLT definition and a population description, but its timeframe and event-definition wording do not align clearly: the timeframe is after 9 weeks (3 cycles), while the definition refers to the first 3 week course. That uncertainty affects the meaning of any count.

The single-group design also sets a clear boundary on inference. It can describe outcomes among participants receiving the combination, but it does not provide a concurrent control for estimating comparative effects. That distinction is essential when considering the study rationale, which describes hopes for the combination but is not itself evidence of an outcome.

The record also demonstrates the difference between registering outcomes and reporting results. Overall survival is listed as a secondary outcome, and the DLT count is identified as primary, yet analyses are not reported. Without estimates and their uncertainty, the reader cannot assess the observed primary outcome or interpret the secondary outcome statistically. Safety event summaries add useful descriptive information but do not substitute for those analyses.

Finally, the record lists the study as terminated without a reason. A status alone cannot explain the conduct or outcome of a trial. Clear reporting of analysis populations, missing data, endpoint timing, and reasons for study status changes helps readers distinguish what was planned, what was observed, and what remains uncertain.

15. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

The principal limitation is the absence of reported analyses. The registry lists 0 analyses, so the primary DLT count and any estimate of uncertainty are not available in this record. Overall survival is named as a secondary outcome, but its results and analysis details are also not reported.

The study is listed as single-group, with no masking and no separate comparator. This prevents a direct controlled comparison within the study. The enrollment is 13, and the number who received treatment is not separately stated, although the primary endpoint population is described as treated participants. The record does not state how many participants had an assessable endpoint.

The primary endpoint timeframe is ambiguous: it is listed as after 9 weeks (3 cycles), while the DLT definition refers to toxicity during the first 3 week course. The registry record does not clarify this difference. It also does not describe missing-data handling, endpoint adjudication, the analysis plan, or the sample-size justification.

The study status is terminated, but the reason is not reported in the registry record. The safety information does not give the timing, causes, or treatment relationship of the listed deaths, nor does it explain overlap among adverse event terms or the duration of safety observation. These limitations restrict the record to descriptive interpretation and prevent a fuller assessment of the intended outcomes.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Lapatinib With Carboplatin and Paclitaxel in Esophagus and Gastroesophageal Junction (GEJ) — NCT01395537.
• Registry record last updated 2019-03-19; status last verified 2019-03.
• Results first posted 2019-02-26.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

19. Record Summary

The registry record describes a Phase I/II pilot study of lapatinib with carboplatin and paclitaxel for recurrent or metastatic adenocarcinoma of the esophagus or gastroesophageal junction. It is listed as a single-group study with no masking, enrollment of 13, and terminated status. The start is recorded as 2011-08 and primary completion as 2014-12.

The primary endpoint is the number of treated participants with grade 3-4 dose-limiting toxicity. The registry lists overall survival as a secondary outcome. It does not report analyses for these outcomes, and the primary endpoint's timeframe is not clearly reconciled with its toxicity definition. Descriptive adverse event counts are listed, but they do not provide a comparative assessment or replace the missing endpoint analysis.