A Phase 1b/2 Study of MEDI4736 With Tremelimumab, MEDI4736 or Tremelimumab Monotherapy in Gastric or GEJ Adenocarcinoma: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Phase 1b/2 Study of MEDI4736 With Tremelimumab, MEDI4736 or Tremelimumab Monotherapy in Gastric or GEJ Adenocarcinoma (NCT02340975), a randomized phase 1, phase 2 study. The primary outcome is Percentage of Participants With Objective Response (OR) in Phase 2.

2. Clinical Question

3. Trial Design

NCT02340975 is a completed, randomized, open-label Phase 1b/2 study sponsored by MedImmune LLC. The registry lists a parallel-group design, enrollment of 114, a start date of 2015-03-31, and primary completion on 2019-04-29. Participants and study teams were not masked, according to the registry record.

The study was designed to explore and expand dosing and to evaluate safety, tolerability, antitumor activity, pharmacokinetics, pharmacodynamics, and immunogenicity. It included MEDI4736 plus tremelimumab, MEDI4736 monotherapy, and tremelimumab monotherapy. Phase 1b involved dose exploration; the primary endpoint results are from Phase 2.

Phase 2 included five arms. Arm A was a combination regimen for participants in second-line therapy. Arm B was MEDI4736 monotherapy at 10 mg/kg Q2W, and Arm C was tremelimumab monotherapy at 10 mg/kg Q4W. Arm D was a combination regimen for participants in third-line therapy. Arm E was a combination regimen for participants in second- and third-line therapy with a positive IFN-γ gene expression signature. The registry describes MEDI4736 at 20 mg/kg Q4W and tremelimumab at 1 mg/kg Q4W in the initial combination period; the arm descriptions also refer to subsequent MEDI4736 treatment. The full schedule is not reported in the registry record.

The primary endpoint was objective response in Phase 2. The endpoint used the as-treated population, so participants were analyzed according to the treatment they actually received. The registry record does not report the rationale for the planned sample size or the statistical plan in sufficient detail to establish how the comparisons were prespecified.

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: A total of 114 participants were randomized in the study, out of which 113 participants received study treatment.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: As-treated population included all participants who received any study drug and analyzed according to the treatment they actually received.

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: As-treated population included all participants who received any study drug and analyzed according to the treatment they actually received.

Only the first 6 of 7 groups are shown.

6. Primary Endpoint

The registry lists the primary outcome as Percentage of Participants With Objective Response (OR) in Phase 2, measured over From Day 1 up to End of the Treatment (EOT), 90 days post-EOT, every 3 months (Q3M) after Day 90 post-EOT up to 12 months post-EOT, and every 6 months after month 12 post-EOT (approximately up to 4 years and one month).

The primary endpoint is the percentage of participants with an objective response in Phase 2. A response is defined as a best overall response of confirmed complete response or partial response under RECIST v1.1. Complete response means disappearance of all target and non-target lesions; partial response requires at least 30% decrease in the sum of diameters of target lesions from baseline. The endpoint therefore records whether a participant met a defined tumor-response category during assessment, rather than measuring the amount of time until progression or survival.

Objective response is used as a direct summary of observed tumor shrinkage under a specified response rule. A percentage is readily interpretable as the share of participants meeting that rule, while the confidence interval communicates uncertainty in that share. The registry's assessment window begins on Day 1 and continues through treatment, post-treatment assessments, and later follow-up, approximately up to 4 years and one month. The response definition also specifies how best overall response is selected from recorded assessments and when assessment ends.

The endpoint has limitations as a summary: it does not, by itself, describe response duration, symptom change, survival, or the balance of benefits and harms. Those are separate outcomes. Nor does a response percentage alone establish a causal treatment effect. Interpretation also depends on the trial design, the analysis population, the completeness of assessments, and the uncertainty reflected by the interval estimates.

7. Results Posted in the Registry

• Phase 2 Arm B-M10 mg/kg (Q2W): 0.0% (0/24) versus 11.1% (3/27) on control; risk difference -11.1 percentage points (95% CI -23.0 to 0.7); unadjusted two-proportion z-test p = 0.092.
• Phase 2 Arm C-T10 mg/kg (Q4W): 8.3% (1/12) versus 11.1% (3/27) on control; risk difference -2.8 percentage points (95% CI -22.4 to 16.8), risk ratio 0.75 (0.09–6.49); unadjusted two-proportion z-test p = 0.792.
• Phase 2 Arm D-M20 mg/kg (Q4W) + T1 mg/kg Fw M10 mg/kg (Q2W): 4.0% (1/25) versus 11.1% (3/27) on control; risk difference -7.1 percentage points (95% CI -21.2 to 7.0), risk ratio 0.36 (0.04–3.24); unadjusted two-proportion z-test p = 0.336.
• Phase 2 Arm E-M20 mg/kg (Q4W) + T1 mg/kg Fw M10 mg/kg (Q2W): 15.8% (3/19) versus 11.1% (3/27) on control; risk difference 4.7 percentage points (95% CI -15.6 to 24.9), risk ratio 1.42 (0.32–6.30); unadjusted two-proportion z-test p = 0.643.

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.

Percentage of Participants With Disease Control at 16 Weeks in Phase 2

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Phase 2 Arm B-M10 mg/kg (Q2W): 12.5% (3/24) versus 18.5% (5/27) on control; risk difference -6.0 percentage points (95% CI -25.8 to 13.7), risk ratio 0.68 (0.18–2.53); unadjusted two-proportion z-test p = 0.555.
  • Phase 2 Arm C-T10 mg/kg (Q4W): 16.7% (2/12) versus 18.5% (5/27) on control; risk difference -1.9 percentage points (95% CI -27.5 to 23.8), risk ratio 0.90 (0.20–4.00); unadjusted two-proportion z-test p = 0.889.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Phase 2 Arm B-M10 mg/kg (Q2W) and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Phase 2 Arm C-T10 mg/kg (Q4W) and the control. The interval gives the range of effects that remain compatible with the data.

Percentage of Participants With Disease Control at 24 Weeks in Phase 2

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Phase 2 Arm B-M10 mg/kg (Q2W): 0.0% (0/24) versus 11.1% (3/27) on control; risk difference -11.1 percentage points (95% CI -23.0 to 0.7); unadjusted two-proportion z-test p = 0.092.
  • Phase 2 Arm C-T10 mg/kg (Q4W): 16.7% (2/12) versus 11.1% (3/27) on control; risk difference 5.6 percentage points (95% CI -18.6 to 29.7), risk ratio 1.50 (0.29–7.85); unadjusted two-proportion z-test p = 0.632.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Phase 2 Arm B-M10 mg/kg (Q2W) and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Phase 2 Arm C-T10 mg/kg (Q4W) and the control. The interval gives the range of effects that remain compatible with the data.

Progression Free Survival at 9 Month (PFS-9) in Phase 2

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Phase 2 Arm D-M20 mg/kg (Q4W) + T1 mg/kg Fw M10 mg/kg (Q2W): 4.0% (1/25) versus 11.1% (3/27) on control; risk difference -7.1 percentage points (95% CI -21.2 to 7.0), risk ratio 0.36 (0.04–3.24); unadjusted two-proportion z-test p = 0.336.
  • Phase 2 Arm E-M20 mg/kg (Q4W) + T1 mg/kg Fw M10 mg/kg (Q2W): 5.3% (1/19) versus 11.1% (3/27) on control; risk difference -5.8 percentage points (95% CI -21.4 to 9.7), risk ratio 0.47 (0.05–4.21); unadjusted two-proportion z-test p = 0.488.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Phase 2 Arm D-M20 mg/kg (Q4W) + T1 mg/kg Fw M10 mg/kg (Q2W) and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Phase 2 Arm E-M20 mg/kg (Q4W) + T1 mg/kg Fw M10 mg/kg (Q2W) and the control. The interval gives the range of effects that remain compatible with the data.

The secondary outcomes broaden the picture beyond whether a participant achieved an objective response. They include disease control at 16 weeks and 24 weeks, progression-free survival at 9 months in Phase 2, and other measures such as duration of stable disease, best percentage change in target-lesion measurements, and safety outcomes. The registry also lists Phase 1b outcomes, including objective response and disease-control measures. The posted secondary results show varied estimates across arms, with wide intervals where estimates are available; they do not form a consistent, precise pattern that resolves the uncertainty in the primary endpoint.

Some secondary estimates are not reported for every arm. For progression-free survival at 9 months, values are listed for Arms A, D, and E, while the results for Arms B and C are not reported in the registry record. The registry record lists 24 secondary outcomes as posted, and the posted endpoint descriptions do not give results for every named secondary measure. Missing or unreported results should not be interpreted as zero events or as evidence of no effect.

Secondary outcomes can help describe disease control over time and provide context for the primary response result, but they do not independently confirm it. Their interpretation depends on how the outcomes were defined, assessed, and analyzed. When multiple secondary outcomes and arms are examined, unadjusted results carry less weight because repeated testing can produce apparently notable findings even in the absence of a true effect. A multiplicity strategy and full analysis plan are not reported in the registry record. These outcomes are therefore best read as exploratory or descriptive evidence alongside the primary endpoint, rather than as a set of independent confirmatory claims.

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

• Percentage of Participants With Objective Response in Phase 1b
• Duration of Stable Disease (DSD) in Phase 1b
• Median Best Percentage Change From Baseline of the Sum of Longest Diameters (SLD) of Target Lesions in Phase 1b
• Percentage of Participants With Disease Control at 16 Weeks in Phase 1b
• Percentage of Participants With Disease Control at 24 Weeks in Phase 1b
• Progression Free Survival at 6 Month in Phase 1b
• Number of Participants With Treatment-emergent Adverse Events (TEAEs) and Treatment Emergent Serious Adverse Events (TESAEs) in Phase 2
• Number of Participants With Clinical Laboratory Abnormalities Reported as TEAEs in Phase 2

9. Statistical Interpretation

10. Statistical Methods Explained

The primary outcome is binary at the participant level: each participant either met the confirmed complete or partial response definition or did not. A percentage summarizes the proportion meeting that definition within an arm. The registry reports a 95% confidence interval for each arm-specific percentage. This interval represents the uncertainty associated with estimating an underlying response proportion from the observed participants, under the method used to construct the interval. It is not a range containing 95% of individual participants or their outcomes.

What does the recalculated risk difference represent?

The simple recalculations compare each listed arm with Arm A by subtracting the reference-arm proportion from the comparison-arm proportion. The direction depends on that convention: a negative value indicates a lower observed proportion in the comparison arm, and a positive value indicates a higher one. The recalculated values use posted arm summaries and are unadjusted. They do not account for covariates, the trial's randomization procedures, or any analysis choices that may have been used by the sponsor. The registry flags a discrepancy between these recalculations and the posted sponsor analyses, so the sponsor's own analysis takes precedence.

How should the p-values be read?

An unadjusted p-value summarizes how compatible the observed data are with a specified null model of no difference, under the assumptions of the test. It is not the probability that the null hypothesis is true, nor does it measure the probability that a result is due to chance. It also does not express the size or importance of an effect. With small arm sizes and few observed responses, the test may have limited ability to distinguish differences. A result that does not meet a conventional threshold is not proof of no difference. The registry record does not report a multiplicity adjustment for the recalculated comparisons.

Why do the confidence intervals matter?

The intervals help show the range of effect sizes that remain compatible with the observed summaries and the interval method. Wide intervals indicate weak precision: the data allow both relatively small and considerably larger underlying response proportions. Intervals for arm-specific proportions are not themselves direct tests of the difference between two arms; the recalculated risk-difference intervals address that contrast. Because the trial was open-label, assessment procedures and missing or unevaluable disease assessments may also matter, but details of how these issues were handled are not reported in the registry record.

What are the planning illustrations?

Some recalculated comparisons include planning illustrations for participants per arm for 80% power and for 90% power. They use the observed effect, or a stated fraction of it, with a two-sided significance level of 0.05, equal allocation, and a normal approximation. These are hypothetical planning calculations, not observed power and not the trial's sample-size justification. The registry record does not report the trial's own sample-size rationale. The illustrations can become very large when the observed contrast is small, because a smaller target difference is harder to distinguish from random variation. They should not be interpreted as evidence that the trial was designed to detect those particular effects.

What can be concluded from a multi-arm comparison?

Each comparison is relative to Arm A, not a comparison among every possible pair of arms. There are several comparisons, and testing several hypotheses raises the issue of multiplicity: even when there is no true difference, repeated testing can increase the chance of an apparently notable result. A prespecified testing hierarchy or adjustment can address this, but the applicable approach is not reported in the registry record. The simple recalculations therefore serve as descriptive statistical context rather than confirmatory evidence.

11. Confidence Intervals and Precision

The primary response estimates are accompanied by 95% confidence intervals, and those intervals are wide across the arms. For Arm A, the estimate is 11.1% with an interval from 2.4% to 29.2%. Arm B is estimated at 0.0%, with an interval from 0.0% to 14.2%. Arm C is estimated at 8.3%, with an interval from 0.2% to 38.5%; Arm D at 4.0%, with an interval from 0.1% to 20.4%; and Arm E at 15.8%, with an interval from 3.4% to 39.6%.

These intervals communicate substantial uncertainty about the underlying response percentages. Their breadth is consistent with the limited information available within each arm, particularly where the arm size is smaller. The zero estimate for Arm B does not establish that the true response probability is zero: the interval extends to 14.2%. Likewise, the higher point estimate in Arm E does not precisely identify its underlying response proportion, as its interval extends from 3.4% to 39.6%.

Precision of an arm-specific proportion and precision of a between-arm difference are related but not identical. The recalculated risk-difference intervals also include no difference for the comparisons available, so they do not provide a clear estimate of a directional separation from Arm A. A confidence interval is not a guarantee that the true value lies within its limits; it is a procedure for expressing uncertainty under its statistical assumptions.

The registry record does not report interval-construction details beyond identifying 95% confidence intervals for the outcome. In addition, the record flags a discrepancy between sponsor-reported analyses and simple recalculations. That discrepancy is another reason not to treat recalculated precision measures as the sponsor's definitive inferential result.

12. Analyses Reported by the Sponsor

The registry lists sponsor analyses for Arm A compared with Arm B and Arm C. For Arm A versus Arm B, the reported method is Fisher Exact, the parameter is labeled Mean Difference (Net), the estimate is 11.1, the 95% confidence interval is -0.7 to 23.0, and the reported p-value is 0.238. For Arm A versus Arm C, the method is Fisher Exact, the parameter is labeled Median Difference (Net), the estimate is 2.8, the 95% confidence interval is -16.8 to 22.4, and the reported p-value is 1.000.

The simple recalculations from the posted arm summaries use unadjusted risk differences and yield different p-values: 0.092 for Arm B, 0.792 for Arm C, 0.336 for Arm D, and 0.643 for Arm E. The registry record does not explain why the sponsor's reported statistics differ in label or result from the simple recalculations, and it contains sponsor-reported analyses for only the Arm A comparisons with Arms B and C. In view of the flagged discrepancy, the sponsor's own analysis takes precedence. The recalculations should be treated as supplementary checks based on aggregate summaries, not as a correction or substitute for the sponsor's analysis.

Even when reading the sponsor results, the estimates and intervals should be considered together. The posted intervals span values on both sides of no difference, and the results do not give a precise account of the size of any contrast. The registry record does not provide enough detail to reconcile the statistical methods or determine the intended estimand.

13. Multiplicity, Missing Data and Analysis Populations

The study includes several treatment arms and many outcomes. The primary endpoint compares response percentages across Phase 2 arms, while secondary measures cover disease control at different assessment times, progression-free survival, lesion change, stable disease, and safety. In this setting, multiplicity matters: examining many outcomes and contrasts increases the opportunity for chance findings. The registry record does not report a multiplicity adjustment, a testing hierarchy, or enough of the statistical plan to determine which comparisons were intended to be confirmatory. The simple recalculated p-values are explicitly unadjusted and should be interpreted cautiously.

The primary analysis population is described as as-treated: participants who received any study drug were analyzed according to the treatment they actually received. The record does not report a parallel intention-to-treat analysis for this endpoint. An as-treated analysis can differ from an analysis according to randomized assignment if participants do not receive the assigned treatment or if treatment received is associated with other factors. The registry record does not provide enough information to assess whether such differences occurred or how they would affect these results.

The endpoint definition explains the response criteria and assessment window, but missing-data handling is not reported in the registry record. In particular, the record does not explain how participants without an evaluable response, with missed assessments, or with early discontinuation were classified in the primary analysis. The outcome definition mentions not evaluable as a possible best overall response category, but that alone does not establish how such cases entered the response percentage.

The registry record also does not report the trial's sample-size justification, interim analyses, a detailed statistical analysis plan, or the rationale for the selected comparisons. These gaps limit assessment of prespecification and the impact of analysis choices. The planning illustrations attached to recalculations are not evidence of the trial's intended power or design rationale.

14. 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: abdominal pain (16.7%), colitis (16.7%), pyrexia (16.7%), transaminases increased (16.7%), decreased appetite (16.7%).

The safety summaries report participants affected by serious adverse events and by other adverse events, as well as deaths, in the Phase 1b and Phase 2 groups. For example, the serious-adverse-event counts are 2 among 6 at risk in the Phase 1b combination group, 14 among 27 in Arm A, 16 among 24 in Arm B, 10 among 12 in Arm C, 15 among 25 in Arm D, and 12 among 19 in Arm E. These are descriptive counts within groups; they do not establish that the study drug caused an event or that one regimen has a different safety profile from another.

The registry lists deaths separately by group, but the record does not provide sufficient detail in these summaries to attribute deaths to treatment or to distinguish causes. The listed serious adverse-event terms include abdominal pain, colitis, pyrexia, transaminases increased, decreased appetite, and hypoglycaemia, each reported as 16.7%. The registry record does not clarify the denominator for those term percentages in the registry record, so they should not be compared as if their basis were established.

Comparisons of adverse-event counts require attention to exposure time, follow-up, event definitions, and the clinical context. The summaries do not provide those details or adjusted comparisons. They are useful as a record of observed safety events, not as a causal estimate or a complete evaluation of relative safety.

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

The results describe objective response observed in five Phase 2 arms among participants with metastatic or recurrent gastric or gastroesophageal junction adenocarcinoma. The response percentage is a measure of confirmed complete or partial response under RECIST v1.1. The arm estimates range from 0.0% to 15.8%, and their confidence intervals are broad. The data therefore provide limited precision about the underlying response proportions and do not clearly distinguish the arms.

These results do not establish that the combination regimens are better or worse than either monotherapy. Nor do they establish that the treatments are equivalent. Failure to demonstrate a clear difference is not proof that there is no difference, particularly when the estimates are uncertain. The arm point estimates are descriptive, and their apparent ordering should not be treated as a dependable ranking.

The findings also do not answer whether any observed response lasts, whether progression-free survival or overall survival differs, or how harms compare in a causal way. Those questions require their own outcomes and analyses. Secondary results add descriptive context, but they are numerous and unadjusted results carry less inferential weight. The sponsor's reported analyses differ from the simple recalculations; the sponsor analysis takes precedence, and the registry does not report the reason for the discrepancy.

General statistical interpretation: an estimate is most useful when considered together with its uncertainty, the population to which it applies, and the design that produced it. Here, eligibility differs across arms, including treatment line and the biomarker criterion for Arm E. Thus, the results should be understood as findings from this trial's selected groups and analysis population, not as a direct estimate for every person with this cancer. Nothing in these results alone establishes a treatment recommendation.

16. Why This Trial Matters Statistically

This trial illustrates several recurring statistical challenges in early-phase oncology studies. It combines dose exploration and dose expansion, evaluates monotherapy and combination regimens, and includes multiple Phase 2 arms with differing treatment-line and biomarker eligibility. Those design features make it important to distinguish descriptive arm estimates from a prespecified confirmatory comparison. The registry record does not report enough of the statistical plan to determine how those roles were assigned.

The outcome is a participant-level response percentage, an intuitive summary, but the uncertainty is substantial when arm sizes are limited and responses are uncommon. Wide confidence intervals show why point estimates alone can be misleading: the largest estimate need not indicate a reliably superior regimen, and an observed zero does not establish a zero underlying response proportion. Interpretation requires attention to precision and not just ranking.

The record also makes a useful distinction between sponsor-reported analyses and simple recalculations from posted summaries. Those calculations can help readers understand the data, but they may use a different estimand or method. With a flagged discrepancy, the sponsor's own analysis takes precedence; unresolved differences limit independent interpretation.

Finally, the trial demonstrates why multiple outcomes, missing-data handling, and analysis populations matter. Response, disease control at specified times, progression-free survival, and adverse events describe different aspects of a study. They should not be combined into a single claim, and unadjusted secondary findings should not be treated as independent confirmations. The registry record leaves several of these methodological details unreported, which appropriately narrows the conclusions that can be drawn.

17. 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.
• Participant counts were reconstructed from posted percentages and arm sizes, so small rounding differences are possible.

The Phase 2 arms are small, and their response estimates have broad confidence intervals. That limits precision and makes arm rankings unstable. The study is open-label, and the registry does not describe the extent to which response assessments were independently reviewed or otherwise protected from assessment bias. The response endpoint is based on best overall response, so interpretation also depends on the timing and completeness of disease assessments.

The arms are not identical in eligibility. Arms A, B, and C are described for second-line therapy; Arm D is for third-line therapy; Arm E includes second- and third-line participants with a positive IFN-γ gene expression signature. Differences in response across arms may therefore reflect differences in the populations as well as differences in treatment. The available results do not provide a basis for separating those influences.

Some secondary outcome estimates are not available for every arm, and handling of missing or unevaluable assessments is not reported in the registry record. The record also does not provide detailed exposure-adjusted safety analyses or enough information to interpret deaths causally. These limitations are specific to the posted summaries and do not imply that the study did not collect additional information.

There is a flagged discrepancy between sponsor analyses and simple recalculations from the posted summaries. The reason is not reported in the registry record, and sponsor analyses take precedence. The statistical plan, multiplicity approach, sample-size justification, and interim-analysis procedures are also not reported in the registry record. Together, these omissions constrain independent evaluation of prespecification and confirmatory intent.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: A Phase 1b/2 Study of MEDI4736 With Tremelimumab, MEDI4736 or Tremelimumab Monotherapy in Gastric or GEJ Adenocarcinoma — NCT02340975.
• Registry record last updated 2020-06-09; status last verified 2020-05.
• Results first posted 2020-05-11.
• 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.

21. Record Summary

NCT02340975 is a completed, randomized, open-label, parallel Phase 1b/2 study of MEDI4736, tremelimumab, and their combination in metastatic or recurrent gastric or gastroesophageal junction adenocarcinoma. The primary Phase 2 outcome was the percentage of participants with confirmed complete or partial response by RECIST v1.1, analyzed in the as-treated population.

Observed response estimates were 11.1% in Arm A, 0.0% in Arm B, 8.3% in Arm C, 4.0% in Arm D, and 15.8% in Arm E. The 95% confidence intervals are broad. The simple unadjusted recalculations do not provide a precise separation from Arm A, and the registry flags a discrepancy with sponsor-reported analyses; those sponsor analyses take precedence. Differences in eligibility across arms, incomplete secondary reporting, and unreported details on multiplicity and missing-data handling further limit interpretation. The results are descriptive and do not establish superiority, equivalence, or causal safety differences among regimens.