Bortezomib With or Without Irinotecan in Treating Patients With Cancer of the Gastroesophageal Junction or Stomach: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Bortezomib With or Without Irinotecan in Treating Patients With Cancer of the Gastroesophageal Junction or Stomach (NCT00061932), a non-randomized phase 2 study. The primary outcome is True Response Rate Evaluated for the Combination of Irinotecan and PS341 by Response Evaluation Criteria in Solid Tumors (RECIST).

2. Clinical Question

3. Trial Design

This was a phase II, nonrandomized, parallel trial sponsored by the National Cancer Institute (NCI). The registry records enrollment of 41 and an overall status of completed. The recorded start was 2003-04 and primary completion was 2009-12. No masking was used. The registry describes two strata based on prior treatment rather than randomized allocation.

In Stratum 1, described as previously untreated, participants received bortezomib intravenously over 3-5 seconds on days 1, 4, 8, and 11, together with irinotecan intravenously over 90 minutes on days 1 and 8. In Stratum 2, described as previously treated, participants received bortezomib as in Stratum 1; this stratum was closed to accrual as of 9/19/2006. The description does not list irinotecan for Stratum 2.

Because treatment allocation was nonrandomized and the strata differed in prior treatment status and regimen, comparisons between them cannot be interpreted as if participants had been assigned to otherwise comparable treatment groups. The registry explicitly cautions that the arms are not a like-for-like control comparison. The primary endpoint concerns response to the combination, not a randomized estimate of the combination's effect relative to bortezomib alone.

The registry lists change in patterns of gene expression before and after treatment, performed by GeneChip analysis, as a secondary outcome. Detailed eligibility criteria, the rationale for the enrollment target, and a fuller account of the analysis plan are not reported in the registry record.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Overall Study

Reasons for not completing this period:

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 True Response Rate Evaluated for the Combination of Irinotecan and PS341 by Response Evaluation Criteria in Solid Tumors (RECIST), measured over Up to 6 years.

The primary endpoint is titled as the true response rate evaluated for the combination of irinotecan and PS-341 using Response Evaluation Criteria in Solid Tumors (RECIST). It is a binary clinical outcome: a participant is classified according to whether the measured disease meets the response criteria. The registry describes CT or MRI scans of measurable lesions at baseline and every 8 weeks thereafter, with responses classified using RECIST version 1.0. The stated timeframe is up to 6 years.

A response rate is used to describe how often a defined response is observed in the evaluated participants. It can summarize tumor response in a phase II study without claiming that the regimen caused the response; causal interpretation depends on the design and supporting analysis. In this trial, nonrandomized allocation and the differing strata limit comparisons between regimens.

The registry identifies the endpoint's dispersion as a 95% confidence interval, but the response count and interval are not reported in the registry record. Without those results, the rate itself and its precision cannot be described. The registry record also does not report the endpoint's analysis population or explain how participants without an evaluable follow-up scan were handled.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is a binary response classification. In general, a binary endpoint records whether each participant meets a prespecified outcome definition, rather than measuring a continuously varying quantity. The registry says that measurable lesions were assessed by CT or MRI and classified according to RECIST version 1.0. It does not provide the response count or the numerator and denominator used for the reported rate; those results are not reported in the registry record.

What does a confidence interval communicate?

In general, a confidence interval describes statistical uncertainty around an estimate under the method's assumptions. For a response rate, the interval is used to show how much uncertainty remains about the underlying rate, given the observed data and the calculation method. It is not a guarantee that the true rate falls within a particular interval, and it does not correct for limitations in study design, outcome measurement, or missing data. The registry specifies a 95% confidence interval for the primary endpoint, but the actual interval is not reported in the registry record.

Precision depends in part on how much outcome information is available. A response estimate based on fewer evaluable participants would generally be less precise than one based on more, other things being equal. However, the number enrolled is not necessarily the number analyzed for this endpoint. Since the registry does not report the endpoint's analysis population or the response count, its precision cannot be assessed from enrollment alone.

Why does allocation matter?

Random allocation is a design method intended to make treatment groups comparable, on average, with respect to factors that could affect the outcome. This trial is identified as nonrandomized. Participants were placed in strata defined by prior-treatment status, and the listed regimens differ between the strata. Therefore, an apparent outcome difference between strata would not by itself estimate the causal effect of adding irinotecan. The registry specifically cautions against treating the arms as a like-for-like control comparison. Adjustment methods, if any, are not reported in the registry record.

What do the planning assumptions mean?

General sample-size calculations estimate the amount of information that might be needed to detect a chosen effect under specified assumptions. The registry's planning description specifies a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. It describes planning illustrations based on a stated fraction of the observed effect; they are not the trial's own design and are not observed power. The record does not report the trial's sample-size justification or identify a prospective effect target used to design enrollment. Planning illustrations cannot replace the missing response result, and they do not show that the trial achieved a particular level of power.

For a binary outcome, the meaning of the planning calculation depends on what is treated as an event and on the assumed response rates. General calculations also rely on the assumed allocation and statistical approximation. When actual allocation is nonrandomized and the groups are not comparable, a simplified two-arm planning illustration should not be read as proof that an observed difference is attributable to treatment. Any numerical planning results should be understood as hypothetical requirements under their stated assumptions, rather than observed properties of this study.

How should the assessment schedule be read?

The registry describes imaging at baseline and every 8 weeks thereafter, with the primary endpoint timeframe stated as up to 6 years. Repeated assessment provides opportunities to classify response over follow-up, but the record does not specify the precise endpoint rule for assigning a participant's overall response, the assessment window used for the primary analysis, or what happened when scans were unavailable. Those details are not reported in the registry record. They matter because different analysis rules can produce different counts of participants classified as responders.

The registry lists no posted analyses for this record. The statistical test, any adjustment for prognostic factors, the confidence-interval calculation method, and any sensitivity analyses are not reported in the registry record. Without such details and the endpoint data, the response-rate estimate cannot be independently checked or interpreted beyond its stated definition.

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("Stratum 1 (Previously Untreated)", "Stratum 2 (Previously Treated)")
n      <- c(29, 12)
events <- c(3, 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  Stratum 1 (Previously Untreated)  29  3  10.3  2.2  27.4  3.6  26.4  3.8  26.5
ARM  Stratum 2 (Previously Treated)  12  1  8.3  0.2  38.5  1.5  35.4  1.9  36.0

11. Multiplicity, Missing Data and Analysis Populations

The registry lists the RECIST response rate as the primary outcome and change in patterns of gene expression before and after treatment by GeneChip analysis as a secondary outcome. It does not report how many hypotheses were formally tested, whether an order of testing was prespecified, or whether any adjustment for multiple comparisons was used. Multiplicity procedures and the statistical analysis plan are not reported in the registry record.

Multiplicity matters in general when many outcomes, subgroups, or repeated tests are examined: the chance of obtaining an apparently notable result by chance can increase across a collection of tests. That general concern does not establish that such testing occurred in this trial. The registry does not describe the number of analyses actually performed or any strategy for handling multiple analyses.

The record does not report the number of participants included in the primary response analysis, how missing imaging assessments were treated, or whether participants who stopped treatment or lacked follow-up scans were included in a particular analysis population. It also does not describe an intention-to-treat or evaluable-participant analysis. These details are important for understanding the denominator and potential bias in a response rate. The enrollment figure cannot substitute for an analysis denominator, and no response count or missing-data procedure is 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: leukocytopenia (27.6%), neutropenia (27.6%), diarrhea (27.6%), dehydration (25.0%), vomiting (20.7%).

The registry lists serious adverse-event counts for each stratum. For Stratum 1, it records 24 participants affected among 29 at risk; for Stratum 2, it records 8 affected among 12 at risk. It also lists other adverse-event counts of 25 among 29 at risk for Stratum 1 and 10 among 12 at risk for Stratum 2. These are counts, not estimates of comparative treatment effects. The registry does not report a statistical comparison, and the strata are not a like-for-like control comparison.

The listed serious adverse-event terms include leukocytopenia, neutropenia, and diarrhea at 27.6% each; dehydration at 25.0%; vomiting at 20.7%; and anemia at 17.2%. These percentages are recorded for the listed terms, but the registry information does not specify the relevant event window or provide term-specific denominators alongside them. The percentages should not be combined or interpreted as mutually exclusive, since a participant may experience more than one event.

Death counts are not reported in the registry record. The serious-event entries describe recorded events and should not be taken as proof that a treatment caused them.

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

The registry describes a phase II investigation of bortezomib with or without irinotecan in patients with gastroesophageal junction or stomach cancer that could not be removed by surgery. Its primary question concerns the response rate for the combination, assessed using RECIST. The registry also records a gene-expression outcome and serious adverse-event information.

What the posted results can establish from the information recorded is limited. The primary response count and 95% confidence interval are not reported in the registry record. As a result, the response rate and its precision cannot be stated, and the endpoint description cannot be treated as evidence of a particular response outcome. The missing result does not mean that no responses occurred; it means the registry information does not report the result.

The design also limits what can be inferred from comparing the strata. Allocation was nonrandomized, the strata differed by prior-treatment status, and the listed regimens were not the same. The registry explicitly says the arms are not a like-for-like control comparison. A difference between strata would not, on its own, show that one regimen produced a better outcome than another.

The safety entries give counts and event-term percentages, but they do not provide a comparative analysis or establish causation. Nor does the enrollment figure indicate the analysis population for response. The registry does not report the response denominator, missing-data handling, or detailed analysis methods.

In general, trial results are most informative when the outcome estimate, its uncertainty, the analysis population, and the design limitations are considered together. For this record, the endpoint definition and study structure are described, while key response results and analysis details are not reported in the registry record. No treatment recommendation follows from these registry entries.

14. Why This Trial Matters Statistically

This trial is statistically informative as an example of why study design and outcome reporting must be considered together. The registry identifies a clinically interpretable binary endpoint, response under RECIST, and describes how measurable lesions were assessed. It also specifies a 95% confidence interval as the measure of endpoint dispersion. However, without the response count and interval, readers cannot evaluate either the observed response rate or its precision.

The nonrandomized structure is central to interpreting any comparison. The strata represent different prior-treatment histories and different listed regimens, so they do not form a randomized control comparison. This illustrates a general principle: treatment groups can differ for reasons other than the treatment itself, and an unadjusted difference between them cannot automatically be assigned a causal interpretation.

The registry also makes clear that safety counts and response outcomes answer different questions. Serious adverse-event counts can describe the recorded safety experience, while a response endpoint describes tumor response under a defined assessment system. Neither outcome alone supplies a complete account of the study, and the reported safety percentages cannot compensate for the missing response estimate.

Finally, the absence of analysis-population and missing-data details matters for a binary endpoint. A rate requires a defined denominator, and readers need to know how participants without evaluable assessments were handled to understand that denominator. Those details are not reported in the registry record. Taken together, the record is useful for understanding the intended question and design, but the missing primary result prevents a quantitative interpretation of the response outcome.

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.

Important outcome details are absent from the registry record. Although the primary endpoint is defined as a RECIST response rate with a 95% confidence interval, the response count, rate, interval, and analysis denominator are not reported. The method used to calculate the interval is also not reported in the registry record.

The nonrandomized design and the differences between strata limit direct comparisons. The previously untreated stratum received bortezomib with irinotecan, while the previously treated stratum received bortezomib as described. The prior-treatment distinction and regimen difference mean that the strata are not a like-for-like comparison, as the registry notes. The record does not describe adjusted analyses that might address differences between participants.

The registry does not report detailed eligibility criteria, the rationale for enrollment, or the complete statistical analysis plan. It also does not state how missing scans, unevaluable responses, treatment discontinuation, or changes in follow-up were handled. Although imaging is described at baseline and every 8 weeks thereafter, the exact rule for deriving a participant's primary response classification is not reported in the registry record.

The safety information lists counts and selected serious adverse-event percentages, but does not include death counts, term-specific denominators, or a comparative analysis. Details such as the event-reporting window and how repeated events were counted are not reported in the registry record. These limitations restrict both efficacy interpretation and comparisons of safety across strata.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Bortezomib With or Without Irinotecan in Treating Patients With Cancer of the Gastroesophageal Junction or Stomach — NCT00061932.
• Registry record last updated 2021-02-02; status last verified 2021-01.
• Results first posted 2015-06-03.
• 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

NCT00061932 is a completed phase II, nonrandomized, parallel study sponsored by the National Cancer Institute (NCI). The registry records enrollment of 41, a start of 2003-04, and primary completion in 2009-12. The trial studied bortezomib with or without irinotecan in patients with gastroesophageal junction or stomach cancer that could not be removed by surgery.

The primary endpoint was the response rate for the combination of irinotecan and PS-341 under RECIST version 1.0, with a stated 95% confidence interval and a timeframe of up to 6 years. The registry describes imaging at baseline and every 8 weeks thereafter. It does not report the response count, rate, or interval.

The two nonrandomized strata differed by prior-treatment status and listed regimen, and the registry cautions that they are not a like-for-like control comparison. Gene-expression change was listed as a secondary outcome. Serious adverse-event counts and selected term percentages are recorded, while death counts and detailed analysis methods are not reported in the registry record.