Capecitabine and Cisplatin (XP)+Sorafenib in Advanced Gastric Cancer (AGC): Sorafenib+XP: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Capecitabine and Cisplatin (XP)+Sorafenib in Advanced Gastric Cancer (AGC): Sorafenib+XP (NCT00565370), a randomized phase 1, phase 2 study. The primary outcome is Number of Participants Who Experienced Dose Limiting Toxicities (DLTs).

2. Clinical Question

The clinical question centers on the safety and tolerability of adding sorafenib to XP, as reflected in the primary endpoint of participants experiencing DLTs. The registry also lists response rate, overall survival and toxicity profile as secondary outcomes. Those outcomes could describe tumor response, time until death, and adverse effects, respectively, but their results are not reported in the registry record.

The record’s design descriptions do not resolve whether the primary question was intended to be evaluated within a single group or through a comparison with XP alone. Randomized allocation and an XP comparator are listed, but the model is identified as single-group, and the primary outcome does not have a reported two-arm analysis. Accordingly, the registry record does not support a comparative estimate of the effect of adding sorafenib. It identifies the intervention, population and toxicity question, while leaving the outcome findings and analysis structure uncertain.

3. Trial Design

This is a phase I-II study of sorafenib with capecitabine and cisplatin in patients with advanced gastric cancer. The registry record lists Asan Medical Center as the sponsor, a start date of 2007-11, primary completion of 2009-12, and a completed status. Enrollment is recorded as 22.

The listed allocation is randomized, masking is none, and the model is single-group. The arm descriptions name XP+sorafenib and XP, but the registry record’s single-group model does not align clearly with those arm descriptions. The record’s skip information also indicates that the primary outcome has fewer than two arms for its analysis. It is therefore important not to treat the listed comparator as evidence that a comparative primary-outcome estimate is available.

The primary outcome is the number of participants who experienced DLTs, expressed as a participant count, with a stated timeframe of 28weeks. The registry lists response rate, overall survival and toxicity profile among secondary outcomes. It identifies five outcome measures overall, but lists no analyses. The registry record does not report the protocol details needed to clarify the phase-specific design, the method of allocation, or the planned handling of the two arm descriptions.

The record identifies a single drug intervention entry naming capecitabine, cisplatin and sorafenib. It does not report enough information to reconstruct treatment schedules, eligibility criteria, follow-up procedures, or the operational definition of a DLT. Those details should not be inferred from the intervention names or the stated endpoint 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.

Overall Study

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: All enrolled patients were analyzed for analysis.

6. Primary Endpoint

The registry lists the primary outcome as Number of Participants Who Experienced Dose Limiting Toxicities (DLTs), measured over 28weeks.

The primary endpoint counts participants who experienced dose-limiting toxicities (DLTs). It is a participant-count outcome rather than a measure of the number or severity of all adverse-event episodes. The registry gives a timeframe of 28weeks, but does not report the detailed DLT definition, the assessment schedule, or the criteria for deciding that an event was dose limiting.

In general, DLT endpoints are used in early-phase research to describe treatment-limiting safety events and inform decisions about whether a regimen can be studied further. A participant-level count can be readily interpreted as how many participants met the endpoint, but it does not by itself describe the timing, duration, recurrence or clinical seriousness of each event. It also does not establish whether an event was caused by a particular drug.

The registry record names this endpoint but does not report the observed count, an estimate of the proportion affected, or a statistical analysis. It also does not provide a primary-outcome result for comparison between XP+sorafenib and XP. Consequently, the endpoint’s intended meaning can be described, but the registry record does not show how many participants experienced DLTs or how that finding was evaluated.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

Statistical methods help translate an endpoint into an estimate and describe how uncertain that estimate is. For a participant-count endpoint, the basic result would ordinarily state how many participants met the endpoint among those assessed. In general, an uncertainty interval around a count or proportion indicates the range of values compatible with the data under the chosen statistical method. It is not a guarantee that the true value lies within that range. The registry record does not report a primary DLT result or an interval, so that interpretation cannot be applied to this trial’s primary endpoint.

What does a participant count measure?

A participant count records people meeting a prespecified event definition, rather than tallying every episode. This distinction matters when a participant may experience more than one event. The primary endpoint is stated as the number of participants who experienced DLTs, but the definition of DLT and the rules for attributing events are not reported in the registry record. Without those details, a reader cannot determine which adverse events qualified, how timing was handled, or whether treatment interruption or dose modification formed part of the endpoint.

Counts also depend on the population assessed. Participants who were enrolled, participants who started treatment, and participants with adequate endpoint follow-up are not automatically the same set. The registry record lists enrollment as 22, while the serious-event safety entry lists 21 at risk. It does not explain the relationship between these populations or specify the denominator used for the primary endpoint.

What would a comparison require?

A comparative analysis would need a clearly defined outcome in each relevant group, an analysis population, and an estimate of the between-group contrast. Random allocation can support a fair comparison when implemented and analyzed as planned, but a registry label alone does not establish which participants were randomized or how the analysis was conducted. Here, the record lists randomized allocation and XP and XP+sorafenib arm descriptions, yet it identifies a single-group model and gives no two-arm primary-outcome analysis. The comparison method and estimate are not reported in the registry record.

In general, a confidence interval communicates precision as well as direction: a wide interval indicates greater uncertainty about the range of plausible effects, while a narrow interval indicates more precision under the model. A p-value addresses how compatible the observed data are with a specified null model; it does not measure the size or clinical importance of an effect. Neither a p-value nor an interval for the primary endpoint is reported. Thus, there is no basis for describing the statistical strength or precision of a primary-outcome comparison.

What do the planning assumptions mean?

The registry’s planning assumptions describe sample-size illustrations based on the observed effect, a two-sided significance level of 0.05, equal allocation between arms, and a normal approximation. The assumed effect is expressed as a fraction of the observed effect. In general, such calculations ask how many participants per arm would be needed to detect an assumed effect under stated modeling choices and a chosen power target. They are planning illustrations, not results from the enrolled participants and not observed power.

The registry record does not report numerical planning results or the primary observed effect on which the illustrations depend. Those values therefore cannot be interpreted for this trial. The assumptions also do not establish that the trial was designed using those methods: the record does not report the trial’s own sample-size justification or statistical plan. The record’s stated assumptions should be kept distinct from protocol decisions that are not reported.

For count data, a normal approximation may be less suitable when counts are small or highly uneven; general statistical practice may instead consider methods that better represent the data’s distribution. This is a general point, not a claim about the trial’s analysis. The registry record does not give enough primary-outcome data to assess distributional assumptions or determine whether a particular model was appropriate.

Why do over-dispersion and event definitions matter?

In general, over-dispersion occurs when observed variation is greater than a statistical model expects. If a model underestimates variation, its uncertainty estimates can be too small. This issue is especially relevant to event counts when participants differ substantially in event risk or when events cluster, although the primary endpoint counts affected participants rather than event episodes. The registry record does not report the data distribution or a model, so there is no basis to claim that over-dispersion was present or addressed.

Finally, a safety count and a serious adverse-event listing answer different questions from a DLT endpoint unless the protocol explicitly defines them as equivalent. The registry record presents serious-event figures but does not report such an equivalence. Statistical interpretation should therefore preserve the endpoint labels and avoid combining these entries into a single outcome.

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("XP Plus Sorafenib")
n      <- c(21)
events <- c(3)
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  XP Plus Sorafenib  21  3  14.3  3.0  36.3  5.0  34.6  5.2  34.9

11. Multiplicity, Missing Data and Analysis Populations

The registry lists five outcome measures, including the primary DLT count and secondary outcomes of response rate, overall survival and toxicity profile. It lists no analyses. The statistical plan for selecting among outcomes, controlling multiplicity, or distinguishing confirmatory from exploratory analyses is not reported in the registry record. In general, testing multiple outcomes can increase the chance of findings that appear notable by chance, so the planned interpretation of multiple tests is important. No adjustment or testing hierarchy can be inferred.

Missing-data handling is also not reported in the registry record. The enrollment figure is 22, while the serious-event entry lists 21 participants at risk. The record does not explain whether these counts differ because of eligibility, treatment exposure, follow-up, or another reason. It does not state how participants without endpoint assessments were handled, whether missingness was expected, or whether any sensitivity analyses were planned. No missing primary-outcome count is reported.

The analysis populations are not described. The registry record does not state whether analyses were intended to include all randomized participants, only participants who received treatment, or another defined group. This is particularly important given the randomized allocation label, the single-group model, and the absence of a reported two-arm primary analysis. Without a defined population and an account of exclusions, the denominator for interpreting the primary endpoint is unclear.

These omissions limit reproducibility and make it difficult to assess whether outcome reporting followed a prespecified analysis plan. They do not establish that appropriate procedures were absent; the relevant details are simply not reported in the registry record.

12. 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: deep vein thrombosis (9.5%), diarrhea (4.8%), febrile neutropenia (4.8%), gastric perforation (4.8%).

For the XP Plus Sorafenib group, the registry lists 21 participants at risk for serious adverse events and 4 participants affected. These are participant counts. The record also lists Deep vein thrombosis at 9.5%, Diarrhea at 4.8%, Febrile neutropenia at 4.8%, and Gastric perforation at 4.8% among the top serious terms. The denominator and counting convention for those term percentages are not reported in the registry record, so they should be read as posted rather than converted into participant counts.

A serious adverse event is a reporting category, while the primary endpoint is DLTs. The registry does not state that these categories are equivalent. The serious-event figures therefore cannot be used to fill in the missing DLT result. Nor do the term percentages indicate, by themselves, whether an event was caused by a particular treatment or whether a participant experienced more than one listed event.

The entry also lists 21 participants at risk and 21 affected for other adverse events. Deaths are not reported in the registry record. The safety information is limited to these entries; timing, duration, severity grading beyond the serious-event designation, and treatment attribution are not reported.

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

The registry record establishes that a study of sorafenib combined with capecitabine and cisplatin in advanced gastric cancer was completed and that DLTs were named as the primary endpoint. It also provides limited serious adverse-event reporting for the XP Plus Sorafenib group. These registry entries describe what the registry lists; they do not constitute a full analysis of safety or treatment effects.

The record does not report the primary DLT count, its denominator, a comparative estimate, or an uncertainty interval. It also does not report results for response rate or overall survival. As a result, the posted information cannot establish the frequency of DLTs, the precision of any estimate, or whether outcomes differed between XP+sorafenib and XP. No conclusion about treatment benefit or lack of benefit follows from the missing results.

The design information also requires caution. Randomized allocation and an XP comparator are listed, but the model is single-group and the primary outcome has no reported two-arm analysis. This leaves the relationship between the stated arms and the primary analysis unresolved. An arm description is not a substitute for an analyzed comparison.

In general, a completed study status means the record identifies the study as completed; it does not mean that all results are available or that the findings are conclusive. Similarly, serious adverse-event term percentages do not, without their denominator and context, fully describe participant-level risk. Readers can use the record to understand the study’s stated question and some safety reporting, but not to infer an effect estimate or make a comparative statistical judgment.

14. Why This Trial Matters Statistically

This trial matters statistically because its registry record illustrates how a named endpoint can be clear in concept while its results and analysis remain unavailable. Counting participants with DLTs is an interpretable safety objective, but meaningful interpretation also requires the observed count, a denominator, an event definition and a description of the analysis population. Those details are not reported for the primary endpoint.

The entry also shows why design labels need to be read together. Randomized allocation, XP and XP+sorafenib arm descriptions, and a single-group model appear in the same record. Without a reported two-arm primary analysis, the information does not establish a comparative estimate. Statistical interpretation must distinguish the existence of a comparator description from evidence generated by a comparison.

Another important issue is the separation of endpoints. The serious adverse-event entries do not automatically answer the DLT question, and the listed secondary outcomes do not supply efficacy findings when results are absent. Keeping endpoints distinct prevents readers from treating related but non-equivalent measures as interchangeable.

Finally, the planning assumptions underscore the difference between design illustrations and observed evidence. Sample-size calculations rely on assumptions about an effect, allocation and statistical method. They do not describe outcomes among enrolled participants and should not be read as observed power. Because the numerical planning results and observed primary effect are not reported in the registry record, those assumptions cannot resolve the missing trial findings.

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 primary limitation is that the registry record does not report the primary DLT result or its denominator. It also does not give an uncertainty interval, a statistical test, or an analysis population. The stated 28weeks timeframe does not replace the missing operational definition or assessment details for DLTs.

The design descriptions are difficult to reconcile: allocation is randomized, arm descriptions include XP+sorafenib and XP, and the model is single-group. The registry indicates fewer than two arms for the primary outcome and lists no analyses. It does not explain how these entries relate to the conduct of the study, so a comparative interpretation cannot be reconstructed.

Enrollment is recorded as 22, whereas the serious-event safety group lists 21 at risk. The registry record does not explain this difference or describe participant flow. Eligibility criteria, treatment exposure, follow-up completeness, missing-data procedures and analysis-population definitions are not reported in the registry record.

Safety reporting is limited. The record lists 4 participants affected by serious adverse events and several serious-event terms with percentages, but the denominator and counting convention for the term percentages are not reported. Event timing, attribution, detailed grading and deaths are not reported in the registry record. The serious-event listing also cannot be assumed to represent the DLT endpoint.

The registry lists secondary outcomes but does not report their results or analyses. It also does not report a multiplicity strategy, a statistical analysis plan, or the trial’s own sample-size justification. These limitations concern what the registry record reports and do not establish what may have been specified or analyzed elsewhere.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Capecitabine and Cisplatin (XP)+Sorafenib in Advanced Gastric Cancer (AGC): Sorafenib+XP — NCT00565370.
• Registry record last updated 2020-01-18; status last verified 2020-01.
• Results first posted 2015-08-31.
• 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 completed phase I-II study of sorafenib with capecitabine and cisplatin in advanced gastric cancer, sponsored by Asan Medical Center. It lists enrollment of 22, randomized allocation, no masking, and a single-group model. The arm descriptions include XP+sorafenib and XP.

The primary endpoint is the number of participants experiencing DLTs during 28weeks. Response rate, overall survival and toxicity profile are listed as secondary outcomes. The registry lists five outcome measures and no analyses. It does not report the primary endpoint result or results for the secondary outcomes.

For XP Plus Sorafenib, the safety entry lists 21 participants at risk and 4 affected by serious adverse events. Serious-event terms and percentages are listed, while deaths are not reported in the registry record. The mismatch among the allocation, model and arm descriptions, together with the absence of a primary analysis, prevents a comparative interpretation from the registry information alone.