A Study to Compare Zolbetuximab (IMAB362) and Chemotherapy With Placebo and Chemotherapy in Adults With Gastric Cancer.: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study to Compare Zolbetuximab (IMAB362) and Chemotherapy With Placebo and Chemotherapy in Adults With Gastric Cancer. (NCT03504397), a randomized phase 3 study. The primary outcome is Progression Free Survival (PFS).

2. Clinical Question

The registry describes a phase 3, global, multicenter, randomized, double-blind, parallel-group study. Its main clinical question is framed as whether the combination can prevent or delay worsening, with progression-free survival (PFS) as the primary endpoint. The posted results report a median PFS for each treatment group and a sponsor-reported hazard ratio from a log-rank analysis.

The population is restricted to tumors meeting the CLDN18.2-positive and HER2-negative criteria and to locally advanced unresectable or metastatic disease. The results therefore address the registered population and treatment comparison; they do not directly answer questions about other biomarker groups, earlier disease settings, or different treatment combinations. The registry record does not report further detail about how biomarker status was established or the eligibility thresholds used.

3. Trial Design

This was a randomized, double-blind, parallel-group phase 3 study. The registry reports enrollment of 565. Participants were assigned to mFOLFOX6 plus zolbetuximab or mFOLFOX6 plus matched placebo. Because the chemotherapy backbone was shared, the randomized comparison estimates the difference between the two assigned regimens, rather than the effect of chemotherapy versus no chemotherapy.

The registry lists the study as global and multicenter. It records a study start date of 2018-06-21 and a primary completion date of 2022-09-09. Its current status is active, not recruiting. The primary outcome timeframe is described as extending from randomization until progression or death, with follow-up reported up to 62 months and 18 days. The registry record does not report the number of sites or countries, the allocation ratio, or details of any stratification factors.

The intervention arm received zolbetuximab with mFOLFOX6; the comparator arm received placebo with mFOLFOX6. The intervention description includes intravenous administration and dosing every 3 weeks after the initial dose. Participants in the placebo group also received mFOLFOX6. Further complete detail about chemotherapy delivery and treatment discontinuation rules is not reported in the registry record.

Double masking is intended to reduce the influence of treatment knowledge on participant care and outcome assessment. Randomization supports a comparison between groups, while the parallel-group structure means participants were assigned to one regimen rather than crossing between regimens as part of the stated design. The registry record does not report the randomization procedure, concealment details, or whether masking applied to every role involved in the trial.

The clinical trial record lists 15 outcome measures and 7 analyses. PFS is the primary outcome. Overall survival and several quality-of-life, response, and safety measures are among the secondary outcomes. The registry record does not report the original sample-size justification or the complete statistical analysis plan.

4. Participant Flow and Accounting

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

Before assignment: Participants who met all inclusion criteria and none of the exclusion criteria were enrolled in the study. Randomization was stratified by region (Asia vs Non-Asia), number of organs with metastatic sites (0 to 2 vs ≥ 3) and prior gastrectomy (Yes or No).

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: FAS

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: Full analysis set(FAS) included all participants who were randomized to 1 of the treatment arms.

6. Primary Endpoint

The registry lists the primary outcome as Progression Free Survival (PFS), measured over From date of randomization until the date of first documented radiological progression or date of death from any cause, whichever occurred first (up to 62 months and 18 days).

PFS is a time-to-event endpoint. In this trial, it is the time from randomization until the earliest of radiological progressive disease or death from any cause. Progression was defined using RECIST 1.1 and assessed by an independent review committee. The registry describes the analysis as using Kaplan–Meier estimates.

This endpoint combines progression and death because either event can mark worsening in the trial’s specified outcome. Measuring time to the first event preserves information about when that event occurred, rather than reducing follow-up to a simple count of participants with an event. A median summarizes the estimated time by which half of the event-time distribution has occurred, under the Kaplan–Meier estimate; it does not describe every participant’s experience or the duration of benefit for an individual.

Time-to-event methods can use information from participants whose event has not been observed by the end of their available follow-up, subject to assumptions about censoring. That is a general property of these methods; the registry record does not report the censoring rules or the amount of follow-up available for each participant. The endpoint is suited to comparing how event times are distributed between randomized groups, but it does not by itself explain which component—progression or death—accounts for a difference in the composite outcome.

The registry reports PFS medians of 11.04 months for mFOLFOX6 plus zolbetuximab and 8.94 months for mFOLFOX6 plus placebo. It also reports a hazard ratio, which describes the relative event rate over follow-up under the proportional-hazards interpretation. A median and a hazard ratio summarize different aspects of the time-to-event data and should be read together rather than treated as interchangeable measures.

7. Results Posted in the Registry

• mFOLFOX6 + Zolbetuximab vs mFOLFOX6+ Placebo: hazard ratio 0.73 (95% CI 0.59–0.91), reported p = 0.002 (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.

Overall Survival (OS)

• 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:
  • mFOLFOX6 + Zolbetuximab vs mFOLFOX6+ Placebo: hazard ratio 0.78 (95% CI 0.64–0.95), reported p = 0.007 (as reported in the registry).
• Interpretation:
  • The 95% confidence interval excludes no difference (a ratio of 1), so the hazard of the event is statistically lower for mFOLFOX6 + Zolbetuximab vs mFOLFOX6+ Placebo than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Time to Confirmed Deterioration (TTCD) Using Physical Functioning (PF) as Measured by European Organization for Research and Treatment of Cancer Quality of Life Questionnaire - Core Questionnaire (EORTC QLQ-C30)

• 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:
  • mFOLFOX6 + Zolbetuximab vs mFOLFOX6+ Placebo: hazard ratio 1.29 (95% CI 0.99–1.69), reported p = 0.028 (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 mFOLFOX6 + Zolbetuximab vs mFOLFOX6+ Placebo and the control. The interval gives the range of effects that remain compatible with the data.

Time to Confirmed Deterioration (TTCD) Using Oesophago-gastric Questionnaire (OG25) on Abdominal Pain and Discomfort as Measured by EORTC QLQ-OG25

• 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:
  • mFOLFOX6 + Zolbetuximab vs mFOLFOX6+ Placebo: hazard ratio 0.71 (95% CI 0.47–1.07), reported p = 0.051 (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 mFOLFOX6 + Zolbetuximab vs mFOLFOX6+ Placebo and the control. The interval gives the range of effects that remain compatible with the data.

The available secondary results add context to the primary PFS comparison but do not form a uniformly consistent pattern. The overall-survival estimate points in the same broad direction as the primary hazard ratio. The quality-of-life time-to-deterioration results are less uniform: the physical-functioning estimate is on the opposite side of 1, while the abdominal-pain estimate is below 1 but has an interval that includes 1. These outcomes address different aspects of the disease experience and should not be collapsed into one overall conclusion.

Secondary outcomes generally carry less confirmatory weight than the primary endpoint unless the statistical plan specifies how they are tested and how multiplicity is controlled. When many outcomes are assessed, the chance of at least one seemingly notable result arising by chance can increase. This is a general statistical concern, not a claim about which particular result is false. The registry record does not report a multiplicity adjustment, a testing hierarchy, or which secondary hypotheses were designated as confirmatory.

Several other secondary outcomes are listed, including response, duration of response, and safety measures, but their numerical results are not reported in the registry record in the registry record for this analysis. The listed outcomes also include quality-of-life measures and ECOG performance status. Without complete results, analysis-population details, and missing-data methods, it is not possible to assess their precision or their relationship to PFS in detail. Secondary outcomes can help describe broader patterns, but their interpretation depends on prespecification, outcome completeness, and appropriate attention to multiple testing.

Accordingly, the primary endpoint remains the most direct statistical answer to the trial’s central question. The secondary findings may be informative alongside it, but they do not independently establish the primary result’s meaning or resolve the limitations of the primary analysis.

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

• Overall Survival (OS)
• Time to Confirmed Deterioration (TTCD) Using Physical Functioning (PF) as Measured by European Organization for Research and Treatment of Cancer Quality of Life Questionnaire - Core Questionnaire (EORTC QLQ-C30)
• Time to Confirmed Deterioration (TTCD) Using Oesophago-gastric Questionnaire (OG25) on Abdominal Pain and Discomfort as Measured by EORTC QLQ-OG25
• Time to Confirmed Deterioration (TTCD) Using Global Health Status as Measured by EORTC QLQ-C30
• Objective Response Rate (ORR)
• Duration Of Response (DOR)
• Number of Participants With Treatment Emergent Adverse Events (TEAEs)
• Number of Participants With Eastern Cooperative Oncology Group (ECOG) Performance Status

9. Statistical Interpretation

10. Statistical Methods Explained

Progression-free survival is analyzed as a time-to-event outcome because both whether an event occurred and when it occurred matter. The registry defines the event as radiological progression or death from any cause, whichever came first. It states that progression was evaluated under RECIST 1.1 by an independent review committee. This independent assessment is part of the endpoint definition; the registry record does not report further detail about review procedures or adjudication.

How the Kaplan–Meier estimate works

In general, the Kaplan–Meier method estimates the probability of remaining free of the event over time. It updates that estimate when events occur and can retain information from participants who leave observation or reach the end of their available follow-up without a recorded event, using censoring. Censoring means the event time is not known beyond the participant’s last observed time; it does not mean that an event was observed at that time. Valid interpretation relies on assumptions about how censoring relates to the event process. The registry identifies Kaplan–Meier estimation but does not report the censoring rules or describe how follow-up losses were handled.

The median time-to-event is the time at which the estimated event-free proportion reaches the midpoint of the distribution. If the estimate does not reach that point during observed follow-up, a median may not be estimable. That general issue is relevant when a registry entry gives a missing median or a confidence limit without a complete interval; missing entries should not be filled in by inference. The PFS medians are reported for both groups, while some secondary endpoint median entries are incomplete.

How to read the hazard ratio

A hazard is an event rate at a particular point in follow-up among participants who have not yet experienced the event. A hazard ratio compares the hazards in two groups. Under the usual proportional-hazards interpretation, a ratio below 1 favors a lower event hazard in the first-listed group, and a ratio above 1 favors a higher event hazard in that group. This interpretation assumes the hazard ratio is a useful summary across time; the registry record does not report a test of the proportional-hazards assumption or plots of time-varying effects.

The reported PFS hazard ratio is 0.734. Its 95% confidence interval is 0.591 to 0.910. The interval describes uncertainty in the estimated ratio under the analysis model and procedure; it is not a range containing a fixed proportion of individual participant outcomes. The log-rank test compares event-time patterns between groups and is the method named for the sponsor-reported PFS analysis. A p-value is a measure of compatibility between the observed data and a null hypothesis, conditional on the model and test; it is not the probability that the null hypothesis is true.

What the median adds

The median PFS estimates are 11.04 months and 8.94 months. These summarize the timing of events in each estimated distribution and can be easier to describe than a relative hazard. They do not show the full shape of the event-time curves, nor do they capture differences at every follow-up time. The median and hazard ratio answer related but distinct questions: the medians mark a point on each group’s estimated distribution, while the hazard ratio compares event hazards over follow-up.

Confidence intervals for the medians are also reported: 9.69 to 12.52 months for mFOLFOX6 plus zolbetuximab and 8.21 to 10.41 months for mFOLFOX6 plus placebo. Such intervals communicate uncertainty in each median estimate. Comparing whether separate confidence intervals overlap is not a substitute for the reported between-group test, because the test is calculated for the contrast and has its own uncertainty.

Analysis population and limits of the record

The primary endpoint identifies the FAS as its analysis population. The registry record does not report what inclusion rules define that population, how participants were treated in the analysis if they discontinued assigned treatment, or how protocol deviations were handled. Randomization is a design feature that supports comparability at assignment, but the exact analysis population and follow-up rules are important for understanding the estimate.

The trial is described as randomized, double-blind, and parallel-group. In general, random assignment helps balance prognostic factors between groups on average, while masking can limit some behavioral or assessment biases. These design features do not eliminate uncertainty or guarantee that all sources of bias are absent. The registry record does not report the allocation-generation procedure, concealment process, details of masking success, or the full statistical analysis plan.

The listed sponsor analysis reports a log-rank hazard ratio and its confidence interval and p-value. There is no recorded discrepancy flag. Any simple re-analysis based on summary data would have less information than an analysis of participant-level time-to-event data and would not replace the sponsor’s stated analysis. The registry does not report the full underlying event-time data, so independent reconstruction of the Kaplan–Meier estimates or log-rank test is not possible from these summary details alone.

11. Confidence Intervals and Precision

Precision concerns how much uncertainty remains around an estimate. For the primary hazard ratio, the 95% confidence interval is 0.591 to 0.910 around the estimate of 0.734. The interval stays below 1, so the reported range is consistent with a lower event hazard in the zolbetuximab combination group under the usual hazard-ratio interpretation. It still spans a range of possible relative effects, and the estimate should not be treated as exact.

The median PFS estimates also have 95% confidence intervals: 9.69 to 12.52 months for mFOLFOX6 plus zolbetuximab and 8.21 to 10.41 months for mFOLFOX6 plus placebo. Each interval expresses uncertainty about its group’s estimated median. Separate intervals do not directly quantify uncertainty in the difference between medians; the hazard ratio and its interval address a different between-group summary.

A 95% confidence interval is a procedure-based interval: under repeated use of the same method and assumptions, intervals constructed in that way would cover the target parameter at the stated rate. It is not a probability statement that the particular parameter has a 95% chance of being inside the already calculated interval. The interval’s width reflects sampling uncertainty and analysis assumptions, and it does not account automatically for every possible bias, measurement limitation, or mismatch between the trial population and another setting.

Confidence intervals are also useful for secondary outcomes, where estimates may be less precise or compatible with more than one direction. For example, the physical-functioning and abdominal-pain hazard-ratio intervals are comparatively compatible with a range that includes 1. This means the estimates do not pin down the relative hazard as clearly as an interval entirely on one side of 1. The registry record does not report confidence intervals for every secondary measure or the missing endpoints needed to complete some median summaries.

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 sample-size figures are planning illustrations based on the observed effect, not the original design assumptions for this trial. For the time-to-event outcomes, the entries are total events required, not participants per arm. The planning assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The target effect is set to the stated fraction of the observed effect. For any other outcomes represented in the planning material, the figures are participants per arm; the time-to-event entries specifically use total events.

The 80% power column describes the number of events needed, under those planning assumptions, for an 80% chance of detecting the specified effect if that effect is the true one. The 90% power column gives the corresponding event target for a 90% chance. Power is a property of a planned design under assumed conditions; it is not the probability that a result is true, and it does not measure the strength of the observed result.

The rows that plan for 75% or 50% of the observed effect ask what would be needed to detect a smaller effect than the one estimated in this trial. Smaller effects need more participants or events to distinguish them from random variation at the same significance and power targets. For time-to-event analyses, event count is central because information accrues as events occur; the number of people enrolled alone does not determine how much event information is available.

These figures are not observed power. They are forward-looking planning illustrations calculated from the observed effect and stated assumptions. The trial’s own original sample-size justification and design assumptions are not reported in the registry record, so these illustrations should not be read as a reconstruction of how the trial was planned.

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 = 0.734, so ln(HR) = -0.3092.
2. d = 4 × (1.96 + 0.842)² / [ln(0.734)]² = 328.3, rounded up to 329 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 <- 0.734                              # 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(3504397)
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 = 329 (80% power), 440 (90% power)
 75% of observed effect: events = 584 (80% power), 782 (90% power)
 50% of observed effect: events = 1,314 (80% power), 1,758 (90% power)
Simulated power with 329 events: 0.791 (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. Group-Sequential Design Analysis

Many time-to-event trials, especially in oncology, stop or report early when interim data are overwhelming, so a single fixed-sample number understates the real design problem. The calculation below is an illustrative group-sequential design built from this trial's posted control median and hazard ratio. The registry record does not state the planned number of looks or the spending function, so the design is a planning illustration, not a reconstruction of the protocol. Assumed: one-sided α = 0.025, 90% power, three equally spaced looks on information, Lan-DeMets O'Brien-Fleming spending, 24 months of accrual, 60 months in total and a 1% monthly dropout hazard.

Under these assumptions the design needs about 528 participants in total, with the boundaries below.

Fixed design versus group-sequential alternatives

Group-sequential designs cost a small increase in maximum sample size and events compared with the fixed design, in exchange for the chance to stop early for overwhelming efficacy. O'Brien-Fleming spending keeps early boundaries very conservative and the final boundary close to the fixed-design value, whereas Pocock spending spends more alpha early.

Sensitivity to a smaller true effect

Planning for exactly the observed effect is optimistic, because observed effects in published trials tend to be somewhat larger than the true effect. The table repeats the main design (three looks, O'Brien-Fleming spending, 90% power) for smaller true hazard ratios.

For context, this trial enrolled 565 participants.

Cross-check, upper boundaries: gsDesign 3.71, 2.511, 1.993 | rpact 3.71, 2.511, 1.993. Agreement between the two packages confirms the boundary calculation.

R code and output

suppressWarnings(suppressMessages({ library(gsDesign); library(rpact) }))
med_control <- 8.94          # control-arm median, months (converted from the posted unit)
hr          <- 0.734          # hazard ratio used for planning
lambdaC     <- log(2) / med_control
eta         <- 0.01               # dropout hazard per month (assumption)
acc_months  <- 24                 # accrual duration (assumption)
study_months<- 60                 # total study duration (assumption)

design <- function(k, sfu, beta, h = hr) gsSurv(k = k, test.type = 1, alpha = 0.025, beta = beta, sfu = sfu,
  lambdaC = lambdaC, hr = h, hr0 = 1, eta = eta, gamma = 1, R = acc_months, T = study_months, minfup = study_months - acc_months)

# Main design: 3 looks, Lan-DeMets O'Brien-Fleming spending, one-sided alpha 0.025, 90% power
x <- design(3, sfLDOF, 0.10)
cat(sprintf("N_TOTAL|%s\n", format(ceiling(sum(x$gamma * x$R)), big.mark = ",")))
for (i in 1:3) cat(sprintf("GS|%d|%.1f|%.0f|%.3f|%.5f|%.5f\n", i, x$T[i], x$n.I[i], x$upper$bound[i], pnorm(-x$upper$bound[i]), cumsum(x$upper$spend)[i]))
cat("\n"); print(gsBoundSummary(x))

# Comparison of designs
cmp <- list(list("Fixed design, one analysis", 1, sfLDOF), list("Two looks, O'Brien-Fleming", 2, sfLDOF),
            list("Three looks, O'Brien-Fleming", 3, sfLDOF), list("Three looks, Pocock", 3, sfLDPocock))
cat("\n")
for (b in c(0.20, 0.10)) for (d in cmp) {
  y <- design(d[[2]], d[[3]], b)
  cat(sprintf("DES|%s|%d|%.0f|%s|%.1f\n", d[[1]], round(100 * (1 - b)), tail(y$n.I, 1), format(ceiling(sum(y$gamma * y$R)), big.mark = ","), tail(y$T, 1)))
}

# Sensitivity: what if the true effect is smaller than the observed one? (3 looks, O'Brien-Fleming, 90% power)
cat("\n")
for (fr in c(1, 0.75, 0.5)) {
  y <- design(3, sfLDOF, 0.10, exp(log(hr) * fr))
  cat(sprintf("SENS|%.2f|%.3f|%.0f|%s\n", fr, exp(log(hr) * fr), tail(y$n.I, 1), format(ceiling(sum(y$gamma * y$R)), big.mark = ",")))
}

# Independent cross-check of the boundaries with rpact
r <- getDesignGroupSequential(kMax = 3, alpha = 0.025, beta = 0.10, sided = 1, typeOfDesign = "asOF", informationRates = x$timing)
cat(sprintf("\nCross-check, upper boundaries: gsDesign %s | rpact %s\n", paste(round(x$upper$bound, 3), collapse = ", "), paste(round(r$criticalValues, 3), collapse = ", ")))

Method: LachinFoulkes 
    Analysis               Value Efficacy
   IA 1: 33%                   Z   3.7103
      N: 382         p (1-sided)   0.0001
 Events: 148        ~HR at bound   0.5431
   Month: 17    P(Cross) if HR=1   0.0001
             P(Cross) if HR=0.73   0.0338
   IA 2: 67%                   Z   2.5114
      N: 528         p (1-sided)   0.0060
 Events: 296        ~HR at bound   0.7467
   Month: 27    P(Cross) if HR=1   0.0060
             P(Cross) if HR=0.73   0.5603
       Final                   Z   1.9930
      N: 528         p (1-sided)   0.0231
 Events: 444        ~HR at bound   0.8275
   Month: 60    P(Cross) if HR=1   0.0250
             P(Cross) if HR=0.73   0.9000



Cross-check, upper boundaries: gsDesign 3.71, 2.511, 1.993 | rpact 3.71, 2.511, 1.993

15. Analyses Reported by the Sponsor

The registry reports the sponsor’s primary analysis as a log-rank comparison, with a hazard ratio of 0.734, a 95% confidence interval from 0.591 to 0.910, and p = 0.002. These are the study’s reported inferential results and should be used as the main statistical analysis of the randomized PFS contrast. The reported median estimates are 11.04 months and 8.94 months.

The registry record does not flag a discrepancy between the sponsor analysis and the posted summary. A simple re-calculation from summary statistics, when available, is unadjusted and cannot reproduce all aspects of a participant-level time-to-event analysis. It may not incorporate the original event times, censoring patterns, stratification, or analysis details. For that reason, a summary-based recalculation is not a replacement for the sponsor’s stated log-rank result.

The reported confidence interval and p-value should be interpreted in the context of the trial’s analysis plan. The registry record does not report the complete plan, including whether the analysis incorporated stratification or how the primary analysis was positioned within any testing hierarchy. The sponsor’s result is therefore the direct reported comparison, while the registry’s incomplete detail limits assessment of every modeling and testing choice behind it.

16. Multiplicity, Missing Data and Analysis Populations

Multiplicity arises when a study examines several endpoints, treatment comparisons, subgroups, or repeated analyses. If each is tested separately without an adjustment or prespecified hierarchy, the overall chance of a false-positive finding can be greater than the nominal level for one test. This general issue is relevant because the registry lists multiple secondary outcomes as well as the primary PFS endpoint.

The registry record does not report whether secondary hypotheses were tested with a formal multiplicity adjustment, whether a fixed testing sequence was used, or which outcomes were intended to support confirmatory claims. The reported p-values should therefore be read as outcome-specific results rather than assumed to have family-wise error control. The record also does not report interim analyses or any stopping rule, and those features should not be inferred.

Missing data can affect time-to-event and quality-of-life analyses in different ways. For a time-to-event outcome, participants may have incomplete event-time information by the analysis cutoff, and methods commonly account for this through censoring. The registry record does not report detailed censoring rules, follow-up completeness, or the handling of participants who withdrew or were lost to follow-up. For questionnaire-based outcomes, unanswered or unavailable assessments can affect who contributes information over time. The registry record does not report the missing-questionnaire strategy or any sensitivity analyses.

The primary population is labeled FAS, but the registry record does not define that population. It does not report the numbers included in each endpoint analysis, how participants were assigned for analysis after treatment discontinuation, or the handling of protocol deviations. These details matter for assessing whether the analysis preserves the intended randomized comparison and how representative the estimates are of all randomized participants.

These omissions limit independent evaluation of the full inferential framework. They do not alter the values reported in the registry, but they constrain conclusions about multiplicity control, missing-data robustness, and the precise analysis population.

17. 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 5% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: vomiting (8.6%), nausea (6.8%), malignant neoplasm progression (4.3%), abdominal pain (3.2%), pneumonia (2.9%).

The registry reports serious adverse event counts by intervention exposure. For zolbetuximab, 133 participants with a serious adverse event are listed among 279 at risk; for placebo, 129 are listed among 278 at risk. The chemotherapy component entries report 262 affected among 557 at risk for each listed component. These are registry counts by intervention category, and the categories can overlap because participants may receive more than one intervention. They should not be added to estimate a number of people with an event or compared as if each intervention category were a separate randomized group.

The registry lists vomiting at 8.6%, nausea at 6.8%, malignant neoplasm progression at 4.3%, abdominal pain at 3.2%, pneumonia at 2.9%, and febrile neutropenia at 2.9% among the top serious-event terms. These terms describe reported event categories, not the total safety profile. The record also lists deaths by intervention category, but those figures do not by themselves establish that a treatment caused a death.

Safety counts require context about exposure, observation time, event definitions, and attribution. The registry record does not report the full adverse-event tables or detailed causality assessments in the registry record. These data describe reported serious events; they are not sufficient on their own for a complete comparison of harms.

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

The primary result describes a time-to-event comparison in adults with CLDN18.2-positive, HER2-negative, locally advanced unresectable or metastatic gastric or gastroesophageal junction adenocarcinoma who were assigned to zolbetuximab plus mFOLFOX6 or placebo plus mFOLFOX6. The reported PFS medians and hazard ratio indicate that the estimated event-time distribution differed between the randomized groups, with the reported hazard ratio below 1. The confidence interval and p-value quantify statistical uncertainty under the stated analysis.

This result does not mean that every participant in one group experienced a later event than every participant in the other group. It does not predict an individual’s outcome, identify the timing of progression for any particular person, or distinguish the contributions of progression and death to the composite endpoint. It also does not by itself establish the balance of benefits and harms; safety and other outcomes require separate consideration.

The finding applies directly to the trial’s registered population, treatment regimens, and outcome definitions. It should not be assumed to apply to people whose tumors do not meet the listed biomarker criteria, to other disease stages, or to different treatment combinations. The registry record does not report detailed subgroup results that would show whether the result varied across participant characteristics.

Statistical evidence is not the same as certainty. The confidence interval leaves a range of plausible hazard ratios, and design or analysis limitations can affect how well the estimate represents the target population. The registry record does not report the complete statistical analysis plan, multiplicity approach, missing-data handling, or definition of the primary analysis population. Those omissions limit a full assessment of robustness.

The result is therefore best read as the sponsor-reported primary comparison from this randomized trial, rather than as a guarantee about an individual outcome or a conclusion about every relevant clinical question. The registry record supplies an estimate and uncertainty measures for PFS; interpreting it alongside safety, secondary outcomes, and the missing methodological detail gives a more complete statistical perspective.

19. Why This Trial Matters Statistically

Statistically, this trial matters because it uses a randomized, double-blind comparison to examine the addition of a biomarker-directed agent to a shared chemotherapy backbone. That structure makes the treatment contrast relatively clear: the two groups differ in whether zolbetuximab or matched placebo is added to mFOLFOX6. The primary endpoint is also a time-to-event measure with a defined progression assessment method and death included as an event.

The reported PFS hazard ratio is accompanied by a confidence interval and a log-rank p-value, allowing readers to consider both the estimated direction and its precision. The medians provide a complementary description of the event-time distributions. Reading these measures together is more informative than relying on a single summary statistic.

The trial also illustrates why secondary outcomes need careful interpretation. Overall survival has a reported estimate in a similar broad direction, whereas quality-of-life estimates are not uniform. Multiple endpoints, incomplete results, and unreported details about multiplicity and missing data mean that secondary findings should be treated as contextual rather than automatically confirmatory.

Finally, the registry record distinguishes a reported treatment comparison from the limits of the available information. It gives key design and outcome details but not the original sample-size justification or complete analysis plan. Recognizing those boundaries is part of sound statistical interpretation: a clear estimate can be informative while still leaving questions about assumptions, robustness, and generalizability unanswered.

20. 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 registry record does not report the original sample-size justification, the full statistical analysis plan, or the allocation procedure. It also does not define the FAS used for the primary analysis. These gaps make it difficult to assess how the planned analysis population, treatment discontinuation, and protocol deviations were handled.

The record identifies a log-rank analysis and Kaplan–Meier estimation but does not report censoring rules, proportional-hazards assessment, or detailed follow-up completeness. The hazard ratio is a useful summary when its assumptions are appropriate; the registry record does not provide enough information to evaluate how well a constant relative hazard describes the data over time.

Several secondary outcomes are listed, but complete numerical results are not reported in the registry record. Some posted secondary time-to-deterioration median entries are incomplete, and the registry does not report the missing-data procedures for questionnaire outcomes. The multiplicity approach is also not reported, which limits the strength of confirmatory interpretation for secondary p-values.

Safety information includes selected serious-event counts and top event terms, rather than a complete account of all adverse events, exposure duration, or attribution. Counts are listed across intervention categories that may overlap, so they should not be treated as independent randomized groups or summed into participant totals.

The trial population is limited to the registered biomarker-defined, advanced disease setting and the specified regimens. The registry record does not report detailed subgroup analyses or results that would establish whether estimates differ across other populations. These limitations concern what can be assessed from the registry record and do not replace review of the full trial report.

21. Related Tutorials

22. Related Statistical Calculators

23. Sources

• ClinicalTrials.gov: A Study to Compare Zolbetuximab (IMAB362) and Chemotherapy With Placebo and Chemotherapy in Adults With Gastric Cancer. — NCT03504397.
• Registry record last updated 2026-10-20; status last verified 2026-09.
• Results first posted 2025-02-26.
• 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.

24. Record Summary

The registry describes a randomized, double-blind, parallel-group phase 3 study of zolbetuximab plus mFOLFOX6 compared with placebo plus mFOLFOX6 in adults with CLDN18.2-positive, HER2-negative, locally advanced unresectable or metastatic gastric or gastroesophageal junction adenocarcinoma. Enrollment is reported as 565, and the primary outcome is PFS assessed from randomization to radiological progression or death.

The sponsor-reported primary analysis gives median PFS estimates of 11.04 months and 8.94 months and a hazard ratio of 0.734, with a 95% confidence interval from 0.591 to 0.910 and p = 0.002. The registry also reports overall-survival and selected quality-of-life results, alongside serious adverse-event counts. The record does not report the complete analysis plan, original sample-size justification, multiplicity procedure, or several details needed to assess missing data and analysis populations.