A Study to Evaluate Rivoceranib Plus Best Supportive Care Compared to Placebo Plus Best Supportive Care in Participants With Gastric Cancer: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study to Evaluate Rivoceranib Plus Best Supportive Care Compared to Placebo Plus Best Supportive Care in Participants With Gastric Cancer (NCT03042611), a randomized phase 3 study. The primary outcome is Overall Survival (OS).

2. Clinical Question

The trial asked whether rivoceranib added to best supportive care (BSC) was associated with different clinical outcomes from placebo added to BSC in participants with advanced or metastatic gastric cancer. The primary outcome was overall survival (OS), measured from randomization until death. The registry also lists progression-free survival, response and disease-control outcomes, quality-of-life measures, and safety outcomes.

The comparison is between two randomized groups, not between treatment and no care: both groups were assigned BSC, which the record describes as palliative, non-cancer therapy. One group received rivoceranib; the other received matching placebo. The registry reports median OS and confidence intervals for each group, but does not report a hazard ratio or a comparative OS analysis. The medians therefore describe survival in each randomized group, while the registry record does not provide a single sponsor-reported estimate of the relative OS effect.

The central interpretive question is what the posted outcomes say about the contrast between the assigned regimens, and how much precision and context the registry results offer. Randomization supports a group comparison, but it does not by itself resolve questions about the statistical analysis, missing outcomes, or the clinical importance of observed differences. Those details are considered only where they are reported.

3. Trial Design

The registry describes a prospective, randomized, double-blinded, placebo-controlled, multinational, multicenter, parallel-group Phase III study. It lists 460 participants as the enrollment and identifies the study as completed. The listed study start was 2017-03-14 and primary completion was 2019-02-15. The title uses rivoceranib, while the official title refers to apatinib; the record describes the intervention as rivoceranib.

Participants were assigned to rivoceranib plus BSC or placebo plus BSC. The arm descriptions specify rivoceranib 700 milligrams orally once per day in each 28-day cycle. The placebo group received matching placebo on the same schedule. BSC is defined in the registry as palliative, non-cancer therapy. These descriptions establish the comparison as the addition of the assigned study drug or matching placebo to BSC.

The masking field is recorded as triple, while the official title describes the study as double-blinded. The record does not identify which parties were included in the masking designation or explain the difference in wording, so the masking procedures beyond those descriptions are not reported in the registry record. The randomization method, allocation concealment procedures, stratification factors, and planned allocation ratio are also not reported in the registry record.

The posted primary outcome is OS, with a timeframe from Day 1 of randomization up to approximately 36 months. The record defines the OS analysis population as all randomized participants, analyzed in the group to which they were assigned. This includes participants still in OS follow-up at the time of primary analysis. The registry reports 308 participants in the rivoceranib group and 152 in the placebo group for the OS outcome. The reasons those outcome counts differ from the enrollment figure are not reported in the registry record.

Progression-free survival, objective response rate, disease control, and several quality-of-life measures are listed as secondary outcomes. Numeric results are posted for some, but not all, of these outcomes. The registry record does not report a full statistical analysis plan or explain the planned approach to multiplicity.

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 were randomized to receive either rivoceranib +best supportive care (BSC) or placebo +BSC for approximately 24 months. Participants who benefited from rivoceranib were permitted to continue rivoceranib during an extension period of the study, up to approximately 36 months. Participants in the extension period were followed for survival status only. Only participants whose reason for study discontinuation was death are included. See the AE module for All-Cause Mortality Deaths.

Core Phase

Reasons for not completing this period:

Extension Phase

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: The ITT set for the OS final analysis consisted of data from all participants who were randomized, including participants who were still in OS follow-up at the time of primary analysis. In the ITT set, participants were included in the group to which they were randomized.

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: The Intent to Treat (ITT) set consists of data from all participants who were randomized. In the ITT set, participants were included in the group to which they were randomized.

6. Primary Endpoint

The registry lists the primary outcome as Overall Survival (OS), measured over Day 1 (randomization) up to approximately 36 months.

The primary endpoint is overall survival: the time from randomization to death. Participants who were alive or lost to follow-up at the end of study were censored, according to the registry definition. Censoring means that a participant’s exact time to death is not observed within the follow-up available for the analysis; the participant contributes follow-up information up to the point specified by the study record.

OS is used in clinical trials because it measures time to a clearly defined event and does not depend on an investigator’s classification of tumor response. It can be affected by follow-up duration and by whether deaths are observed, which is why the registry’s censoring definition and analysis population matter when interpreting the result.

The posted summary for each group is a median survival time with a 95% confidence interval. A median describes the time by which half of the observed survival distribution has reached the event, under the analysis method; it is not an average survival time and does not describe every participant’s outcome. A confidence interval conveys uncertainty around the estimated median. It does not, by itself, establish the size or statistical significance of a difference between groups.

The registry provides group-specific medians but no sponsor-reported hazard ratio, comparative confidence interval, or OS hypothesis-test result. Therefore, these OS summaries do not constitute a complete comparative effect estimate. The sponsor’s reported comparative OS analysis is not reported in the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry’s primary OS result is expressed as a median time with a 95% confidence interval. The endpoint is time-to-event: each participant is followed from randomization until death or censoring. Censoring is necessary when the event time is not observed during the available follow-up. The record specifies that participants alive or lost to follow-up at the end of study were censored, but it does not report the statistical method used to estimate the medians or their confidence intervals.

What does a median survival time describe?

A median is the point in the estimated survival distribution at which the estimated proportion remaining alive reaches one half. Unlike a mean, it does not use the full magnitude of each participant’s survival time as an average. It is a compact summary of the survival curve, but it does not show the curve’s shape, the timing of events across follow-up, or whether the groups differ consistently throughout follow-up.

The two posted medians are group-specific. Comparing them informally is not equivalent to estimating a hazard ratio or conducting a time-to-event test. A hazard ratio is a relative measure of event rates over follow-up under a model and assumptions; neither a hazard ratio nor its confidence interval is reported for OS in this registry record. A comparative OS p-value is also not reported. Accordingly, the median summaries can be described, but they cannot stand in for the absent sponsor-reported comparison.

How should confidence intervals be read?

A confidence interval is a range of estimates compatible with the observed data under the method used to construct it. A 95% confidence interval describes the method’s long-run coverage property when the method’s assumptions hold; it is not a statement that there is a 95% probability that a fixed parameter lies in this particular interval. For a difference between groups, a confidence interval directly addresses uncertainty about that contrast. The OS intervals in the registry are for each group’s median, not for the difference between medians or for a hazard ratio.

The ORR and DCR summaries are proportions with 95% confidence intervals. The displayed risk differences and unadjusted p-values are simple unadjusted recalculations from the posted arm summaries, not reported sponsor analyses. Their interpretation is limited to the stated comparisons and does not account for a multiplicity plan that is not reported. The confidence interval for a risk difference expresses uncertainty in the absolute difference between group proportions. The p-value describes how compatible the observed data are with a specified no-difference null under the calculation’s assumptions; it does not measure the probability that the null is true or the size of a clinical effect.

What do the quality-of-life summaries show statistically?

The global health status/quality-of-life outcome is reported as mean change from baseline with a standard deviation. A standard deviation describes variation among participant-level changes; it is not the standard error of the mean and does not directly describe uncertainty in the group mean. The recalculated mean difference and confidence interval compare the posted group means using an unadjusted approach. That comparison does not account for any adjustment, repeated-measures model, or missing-data method the sponsor may have used, because such methods are not reported in the registry record.

Quality-of-life measurements can be missing when participants do not complete questionnaires. The registry provides the number contributing to this outcome in each group, but does not report the missing-data pattern, reasons for missingness, or the analysis method used to address missing responses. Without those details, it is not possible to assess whether the observed questionnaire results represent all randomized participants equally.

What is the role of the planning illustrations?

Planning illustrations use the observed effect as a reference and show how hypothetical per-arm sample sizes change when the target effect is set to a stated fraction of that observed effect. The registry’s illustration assumptions are a two-sided significance level of 0.05, equal allocation, and a normal approximation. These are planning scenarios, not a reconstruction of the trial’s original design calculation and not observed power. The trial’s own sample-size justification and design assumptions are not reported in the registry record.

Such illustrations are sensitive to the effect used as a starting point. An effect estimated in one trial can be imprecise, and a smaller target effect generally requires more participants under otherwise similar planning assumptions. The numbers are therefore best understood as hypothetical planning aids rather than evidence that the trial was adequately or inadequately powered. No time-to-event planning illustration is posted for the primary endpoint in these results.

Why does over-dispersion matter in general?

In general, over-dispersion means that observed variability exceeds what a statistical model assumes. When it occurs, standard errors can be too small if the model does not account for it, making confidence intervals too narrow and tests too optimistic. This is a general statistical consideration, not a documented feature of this trial. The registry record does not report whether over-dispersion was assessed or relevant to the analyses.

More generally, randomization helps make the assigned groups comparable at baseline in expectation, but it does not specify the analysis model, guarantee complete outcome data, or eliminate uncertainty. The posted registry results do not describe the full statistical methods for OS, the secondary outcomes, or safety. Those missing details constrain how far the summary statistics can be interpreted.

10. Multiplicity, Missing Data and Analysis Populations

The registry identifies one primary endpoint, OS, and multiple secondary outcomes. It does not report whether the study used a formal hierarchy for testing outcomes, adjusted significance levels, or another multiplicity procedure. Without that information, unadjusted secondary p-values should be treated as exploratory summaries rather than as a family of confirmatory tests. This limitation applies even when an individual unadjusted p-value is small, because the number and relationship of analyses considered are not described.

The record reports no sponsor analyses for the posted secondary endpoints. The binary risk differences and p-values, and the quality-of-life mean difference, are simple unadjusted recalculations from the posted arm summaries. The primary OS result is given as arm-specific medians and confidence intervals, without a sponsor-reported comparative hazard ratio or test. A complete statistical analysis plan, including prespecified estimands, model specifications, covariates, censoring rules beyond the basic definition, and analysis timing, is not reported in the registry record.

For OS, the stated population is the ITT set: all randomized participants, including participants still in OS follow-up at primary analysis, analyzed according to randomized group. The record does not provide the number of participants still being followed, details of follow-up completeness, or the number and timing of censoring events. It also does not report sensitivity analyses for censoring or loss to follow-up.

For the quality-of-life outcomes, the registry gives outcome-specific group counts but does not report missing questionnaire data, reasons for non-completion, or the method used to handle missing responses. The analysis population definition for each secondary outcome is not reported in the registry record. For response outcomes, the record posts group counts and proportions but does not explain how participants without an evaluable assessment were handled.

These omissions do not establish that the trial lacked a statistical plan or that missing data were handled improperly. They mean that those features cannot be evaluated from the registry results. The posted summaries should therefore be interpreted with attention to the available definitions and with caution about unreported multiplicity, missingness, and analysis choices.

11. 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: asthenia (5.3%), abdominal pain (4.9%), ileus (2.6%), cholangitis (2.6%), decreased appetite (2.6%).

The safety record lists the number at risk and the number affected by serious adverse events and other adverse events for each group. For serious adverse events, 307 participants were at risk in the rivoceranib plus BSC group and 149 were affected; in the placebo plus BSC group, 151 were at risk and 66 were affected. For other adverse events, the corresponding at-risk and affected counts are 307 and 299 in the rivoceranib group, and 151 and 140 in the placebo group. The record also lists deaths of 283 and 134, respectively. These counts are descriptive; they do not by themselves establish that an event was caused by the assigned treatment.

The registry lists asthenia at 5.3%, abdominal pain at 4.9%, and ileus, cholangitis, decreased appetite, and pneumonia each at 2.6% among the top serious adverse-event terms. The record does not specify the arm attribution for these term-level percentages in the posted summary, so they should not be assigned to one group from this information alone.

Safety counts depend on the population at risk, the event definitions, observation time, and how events were collected. The registry does not provide exposure-adjusted rates, event onset or duration, severity grading details, or a comparative statistical analysis for these safety summaries. Counts from groups of different sizes should not be treated as directly comparable rates without appropriate context. The posted figures describe reported event occurrence; they do not establish causation or a complete safety profile.

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

The registry results describe a randomized comparison of rivoceranib plus BSC with placebo plus BSC in participants with advanced or metastatic gastric cancer. They report group-specific OS medians and confidence intervals, selected secondary outcome summaries, and safety counts. They do not report a sponsor-reported comparative OS estimate or test. Consequently, the primary endpoint cannot be fully characterized as a comparative treatment effect from the posted OS summaries alone.

The arm-specific medians are estimates, not predictions of how long an individual participant will live. A median does not mean that every participant experienced a similar duration, and it does not indicate the survival experience beyond the observed follow-up. The registry’s timeframe is up to approximately 36 months, but the posted summary does not include a survival curve, number at risk over time, or comparative hazard ratio.

The secondary outcomes add information about response, disease control, and quality of life, but they do not resolve the limitations in the primary OS reporting. Some numeric outcomes are missing from the registry record, and the multiplicity plan and analysis details are not reported. The unadjusted recalculations are not sponsor analyses and should be distinguished from prespecified, adjusted analyses, if any.

The findings also cannot establish that observed events were caused by study treatment. The safety counts and serious adverse-event terms are descriptive, and the registry does not provide the exposure-adjusted comparisons and event-level context needed for a fuller assessment.

These results are not a recommendation for any participant or a conclusion about what care should be chosen. They describe the evidence posted for this trial and its limits. Interpretation beyond those posted estimates would require additional information, particularly the sponsor’s comparative OS analysis, analysis plan, missing-data methods, and complete reporting of the listed secondary outcomes.

13. Why This Trial Matters Statistically

This trial is statistically notable because it provides a randomized, placebo-controlled comparison with OS identified as the primary endpoint, yet the registry results do not include the comparative OS statistic most useful for interpreting the difference between the randomized groups. Group-specific medians are meaningful descriptive summaries, but they answer a narrower question than a hazard ratio or a formal between-group test. This distinction illustrates why an endpoint’s label and its posted analysis need to be considered together.

The record also distinguishes outcomes with different statistical structures. OS is a time-to-event outcome with censoring; ORR and DCR are binary outcomes; and quality of life is reported as mean change with standard deviations. Each requires interpretation suited to its data type. A simple comparison of group proportions or means cannot substitute for a time-to-event analysis, and a confidence interval for an arm-specific median does not directly quantify uncertainty in a comparative effect.

The secondary summaries show why multiplicity and prespecification matter. Several outcomes are listed, but the registry does not describe a testing hierarchy or adjustment method. Unadjusted p-values can be informative as descriptive measures for individual comparisons, but they should not be treated as confirmatory evidence when the broader analysis plan is unavailable.

Finally, differences in outcome-specific participant counts and the absence of reported missing-data methods limit the assessment of who contributed to each result. These issues do not invalidate the posted estimates, but they affect how confidently the results can be generalized across the randomized population. A complete statistical interpretation depends not only on numeric estimates, but also on transparent reporting of analysis populations, follow-up, censoring, and missingness.

14. 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 most important trial-specific limitation in the posted results is the absence of a comparative OS estimate. The registry reports median OS and confidence intervals by group, but no hazard ratio, comparative confidence interval, or OS p-value. This prevents a direct description of the primary endpoint’s relative effect based on the sponsor’s reported analysis. The statistical method used to estimate the medians and intervals is not reported in the registry record.

Numeric results are not reported for several listed secondary outcomes, including progression-free survival, stomach-specific quality of life, and EQ-5D-5L measures. The number of outcome measures listed is eight, while numeric results are available for only some outcomes. The registry record does not explain whether the remaining results were not collected, not analyzed, or not posted.

The analysis counts differ by outcome, and the record does not explain the reasons. For OS, the population is defined as all randomized participants, but outcome-specific populations and missing-data handling for the secondary outcomes are not reported. The number and timing of censoring events, follow-up completeness, and sensitivity analyses for loss to follow-up are also not reported.

Interpretation of secondary findings is further limited by the absence of a reported multiplicity plan and sponsor analyses for those comparisons. Recalculated comparisons are unadjusted and rely only on the posted summary statistics. They do not reveal whether a prespecified model, adjustment, or alternative analysis was used by the sponsor.

The safety record provides event counts and a list of top serious adverse-event terms, but does not give arm attribution for the term-level percentages in the posted summary. It also lacks exposure-adjusted comparisons and further event-level context. These limitations mean that the safety figures are descriptive rather than a complete comparative assessment.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: A Study to Evaluate Rivoceranib Plus Best Supportive Care Compared to Placebo Plus Best Supportive Care in Participants With Gastric Cancer — NCT03042611.
• Registry record last updated 2022-07-08; status last verified 2022-06.
• Results first posted 2022-07-08.
• 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.

18. Record Summary

NCT03042611 is a completed randomized Phase III study comparing rivoceranib plus BSC with placebo plus BSC in participants with advanced or metastatic gastric cancer. The registry lists enrollment of 460 participants, a study start of 2017-03-14, and primary completion on 2019-02-15. The intervention description specifies rivoceranib 700 milligrams orally once per day during each 28-day cycle; both groups received BSC.

OS was the primary endpoint, defined as time from randomization to death, with participants alive or lost to follow-up at end of study censored. Group-specific medians and 95% confidence intervals are posted, but a comparative OS analysis is not reported in the registry record. Selected secondary outcome results and safety counts are also posted. The registry does not report a complete statistical analysis plan, multiplicity approach, or missing-data methods.