A Study of BMS-986253 in Combination With Nivolumab or Nivolumab Plus Ipilimumab in Advanced Cancers: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study of BMS-986253 in Combination With Nivolumab or Nivolumab Plus Ipilimumab in Advanced Cancers (NCT03400332), a randomized phase 1, phase 2 study. The primary outcome is Number of Participants Experiencing Adverse Events (AEs) - Part 1.

2. Clinical Question

The study’s stated purpose was to investigate BMS-986253 in combination with nivolumab or with nivolumab plus ipilimumab in participants with advanced cancers. The primary result addresses the frequency of participants with any adverse event in Part 1. It does not, by itself, determine whether an event was caused by a study drug: the endpoint definition explicitly says that an adverse event does not necessarily have a causal relationship with treatment.

The posted results also include secondary outcomes, including objective response rate in Part 1 and progression-free survival in Part 2. These address different questions and involve different populations and follow-up periods. Their results should not be read as though they were alternate measurements of the primary safety endpoint.

3. Trial Design

The registry record describes a Phase 1/2 study with randomized allocation, quadruple masking and a parallel design. Enrollment is listed as 281, and the status is completed. The study began on 2018-02-12, with primary completion on 2024-04-26. These record-level details describe the trial as a whole; they do not establish that every outcome was assessed in every participant or that all treatment groups contributed to each analysis.

Part 1 contained six posted groups: five BMS-986253 dose-and-schedule groups with nivolumab, and a group receiving BMS-986253 with nivolumab plus ipilimumab. The primary adverse-event endpoint explicitly applies only to treated participants in Part 1. Its displayed group counts range from 12 to 63 participants. The registry record does not report a primary-endpoint analysis plan or a sponsor-reported statistical comparison for these groups.

Part 2 included BMS-986253 plus nivolumab and ipilimumab and a placebo plus nivolumab and ipilimumab group. The posted progression-free survival result compares these groups; the mortality endpoint is also posted for Part 2. Part 1 and Part 2 results therefore answer different questions within the study and should not be combined as if they came from one uniform comparison.

The trial title and summary identify advanced cancers as the setting, and the listed conditions include cancer and melanoma. The registry record does not report the specific tumor types represented in each arm, the detailed eligibility criteria, the allocation ratio, or the procedures used to maintain masking. It also does not report the rationale for the dose schedules or the sample-size justification. Those details should not be inferred from the trial’s overall design labels.

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: 281 participants were randomized and 278 received treatment. Part 1: 159 participants were directly treated and not blinded (open label). Part 2: 122 participants were double-blinded and randomized. 122 participants were randomized in the pre-treatment phase while 119 received treatment. 3 participants were randomized but did not receive treatment.

Only the first 6 of 9 groups are shown.

Pre-Treatment

Reasons for not completing this period:

Treatment

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: All treated participants in Part 1

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

Population: All treated participants for Part 1 and all randomized participants for Part 2. Participants in Part 1 were not randomized.

Only the first 6 of 10 groups are shown.

6. Primary Endpoint

The registry lists the primary outcome as Number of Participants Experiencing Adverse Events (AEs) - Part 1, measured over From first dose up to 100 days after last dose (up to 65 months).

The primary endpoint is the number of treated participants in Part 1 who experienced an adverse event. The time window runs from the first dose through 100 days after the last dose, with the registry record describing this as up to 65 months. An adverse event is defined broadly as a new untoward medical occurrence or worsening of a pre-existing condition after administration of study treatment. The definition does not require the event to be caused by treatment.

This count endpoint is used to describe how many participants had at least one event during the specified observation window. It differs from a count of all individual events: a participant with multiple events still contributes as a participant who experienced an event. The endpoint also differs from a severity-specific measure. Although the registry description explains grades from mild events through fatal events, the primary endpoint itself counts participants with adverse events and does not separate the result by grade.

For safety interpretation, the broad definition matters. A high participant count can reflect events of differing seriousness and causes, while the count alone does not tell readers which events were treatment-related, how many events each participant experienced, or how long participants were observed. Separate serious-event information is posted, but it is not the same endpoint as the primary any-adverse-event count.

The primary endpoint was prespecified for Part 1 only. It does not directly assess response, disease progression or survival, and it cannot establish comparative benefit. Those questions require their own outcome definitions and analyses.

7. Results Posted in the Registry

• Part 1: 1200 Q4W: 93.3% (14/15) versus 100.0% (16/16) on control; risk difference -6.7 percentage points (95% CI -19.3 to 6.0), risk ratio 0.93 (0.82–1.07); unadjusted two-proportion z-test p = 0.294.
• Part 1: 2400 Q4W: 100.0% (18/18) versus 100.0% (16/16) on control; risk difference 0.0 percentage points (95% CI 0.0 to 0.0), risk ratio 1.00 (1.00–1.00); unadjusted two-proportion z-test p = n/a.
• Part 1: 1200 Q2W: 100.0% (12/12) versus 100.0% (16/16) on control; risk difference 0.0 percentage points (95% CI 0.0 to 0.0), risk ratio 1.00 (1.00–1.00); unadjusted two-proportion z-test p = n/a.
• Part 1: 2400 Q2W: 95.2% (60/63) versus 100.0% (16/16) on control; risk difference -4.8 percentage points (95% CI -10.0 to 0.5), risk ratio 0.95 (0.90–1.01); unadjusted two-proportion z-test p = 0.374.

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.

Objective Response Rate (ORR) - Part 1

• 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:
  • Part 1: 1200 Q4W: 6.7% (1/15) versus 6.2% (1/16) on control; risk difference 0.4 percentage points (95% CI -16.9 to 17.7), risk ratio 1.07 (0.07–15.57); unadjusted two-proportion z-test p = 0.962.
  • Part 1: 2400 Q4W: 16.7% (3/18) versus 6.2% (1/16) on control; risk difference 10.4 percentage points (95% CI -10.5 to 31.3), risk ratio 2.67 (0.31–23.14); unadjusted two-proportion z-test p = 0.347.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Part 1: 1200 Q4W 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 Part 1: 2400 Q4W and the control. The interval gives the range of effects that remain compatible with the data.

Progression Free Survival (PFS) - Part 2

• 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:
  • Part 2 Arm A: BMS+Nivo+Ipi vs Part 2 Arm B: Placebo+Nivo+Ipi: hazard ratio 0.94 (95% CI 0.61–1.45), reported p = 0.742 (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 Part 2 Arm A: BMS+Nivo+Ipi vs Part 2 Arm B: Placebo+Nivo+Ipi and the control. The interval gives the range of effects that remain compatible with the data.

Number of Participants Who Died - Part 2

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Part 2 Arm A: BMS+Nivo+Ipi: 48.4% (30/62) versus 29.8% (17/57) on control; risk difference 18.6 percentage points (95% CI 1.4 to 35.8), risk ratio 1.62 (1.01–2.61); unadjusted two-proportion z-test p = 0.039.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically higher for Part 2 Arm A: BMS+Nivo+Ipi than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

The secondary outcomes cover different aspects of the study and should not be collapsed into one overall treatment conclusion. Part 1 includes objective response rate and duration of response, along with pharmacokinetic measures such as concentration, exposure, clearance and concentration at the end of a dosing interval. Part 2 includes progression-free survival. These outcomes differ in definition, time window and population; some describe tumor outcomes, while others describe drug disposition.

The posted objective response estimates vary across Part 1 groups, but the confidence intervals are broad. This pattern means that the point estimates alone do not establish a stable ordering among dose and schedule groups. The Part 2 progression-free survival estimate is also uncertain: the reported hazard-ratio interval includes the null value. This secondary evidence does not resolve the limitations of the primary safety endpoint, which counts any adverse event in Part 1 rather than measuring response or progression.

The posted mortality endpoint is another distinct secondary measure. It counts participants who died during its specified observation window, not deaths attributed to treatment. Its unadjusted comparison has a p-value of 0.039, but a nominal p-value does not remove concerns about multiplicity, incomplete reporting of the analysis plan or the interpretation of the endpoint. The registry record does not report a multiplicity adjustment across the outcomes.

Unadjusted secondary results carry less weight because many endpoints and comparisons can create opportunities for chance findings, and because the registry record does not explain which analyses were confirmatory. The different analysis populations and observation periods also make it inappropriate to use one secondary outcome as direct confirmation of another. In assessing the overall pattern, readers should distinguish descriptive estimates from prespecified confirmatory evidence and give attention to confidence intervals, endpoint definitions and the absent details about missing data and multiplicity.

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

• Objective Response Rate (ORR) - Part 1
• Duration of Response (DOR) - Part 1
• Maximum Concentration (Cmax) - Part 1
• AUC(0-T)-Area Under Curve From Time Zero up to Last Quantifiable Concentration - Part 1
• AUC(TAU)-Area Under Curve in 1 Dosing Interval - Part 1
• Observed Serum Concentration at the End of a Dosing Interval (Ctau) - Part 1
• Total Body Clearance (CLT) - Part 1
• Average Serum Concentration Over a Dosing Interval (Css-avg) - Part 1

9. Statistical Interpretation

10. Statistical Methods Explained

What does a participant count measure?

A binary participant-level endpoint records whether each person did or did not meet a defined condition during an observation window. For this primary outcome, the condition is experiencing at least one adverse event. The reported count is therefore a count of participants with an event, not a count of every event episode. If one participant experiences several adverse events, this endpoint still records that participant once.

When a binary outcome is compared between groups, a risk difference expresses the difference between the observed proportions with the event. A value of zero represents equal observed proportions; a value below zero indicates a lower observed proportion in the comparison group than in the reference group. The meaning depends on which group is defined as the reference. These simple comparisons do not automatically adjust for baseline characteristics, follow-up differences or the study’s full randomization structure.

Why are some primary intervals zero-width?

Several Part 1 groups have the event recorded for every participant, including the reference group. When both groups have an observed proportion at the upper boundary, the reported simple comparison can have a risk difference of zero and an interval with identical lower and upper limits. This is a feature of the observed boundary counts and the calculation used for the displayed comparison. It does not show that the true risks are known without uncertainty. In particular, an interval produced by a simple calculation at a boundary should be read alongside the limited group sizes and the event definition, rather than as proof of equivalence.

How should confidence intervals be read?

A confidence interval describes the uncertainty around an estimate under the assumptions of the method used to calculate it. Wider intervals indicate less precision. For a risk difference, an interval that includes zero is compatible with no difference as well as with differences in either direction at the confidence level used. For a hazard ratio, the corresponding null value is one: an interval spanning one does not isolate a difference in either direction. These are general statistical interpretations, not guarantees about the probability that a particular hypothesis is true.

For the posted Part 1 objective response rates, the confidence intervals are especially broad in several groups. That is consistent with the small group counts and limited numbers of observed responses. A point estimate alone can make a result appear more definite than the interval supports. The interval should be considered when judging how much the observed response rate might vary under repeated sampling and the assumptions of the calculation.

What does the Part 2 hazard ratio represent?

The Part 2 progression-free survival analysis reports a hazard ratio from a log-rank analysis. A hazard ratio compares event rates over follow-up under the model and analysis assumptions; it is not the ratio of the two posted median times. A value below one points in one direction for the reported comparison, while a value above one points in the other. The reported interval spans one, so the estimate remains compatible with no hazard difference as well as with effects in either direction. The registry record does not report details such as assessment of the proportional-hazards assumption or handling of censoring beyond the outcome definition.

The median progression-free survival values are summaries of the time-to-event distributions, while the hazard ratio incorporates information across follow-up according to its analysis method. Those summaries need not point to the same simple numerical contrast. Neither a median nor a hazard ratio alone describes every feature of the survival curves.

What are the planning illustrations?

The sample-size illustrations are based on the observed effect and use a two-sided significance level of 0.05, equal allocation, and a normal approximation. Their stated effect assumptions are fractions of the observed effect. For time-to-event planning, the listed quantity is total events; for the other planning illustrations, it is participants per arm. These are planning scenarios, not observed power and not the sample-size justification used to design this trial. The registry record does not report the trial’s original sample-size rationale or whether any planned assumptions were revised.

Using an observed effect as the basis for a planning illustration can be unstable, especially when the underlying estimate comes from small groups or a wide interval. A small observed effect can imply very large sample requirements, while an exaggerated effect in a limited study can imply unrealistically small requirements. The illustrations are therefore useful for understanding how effect size and precision relate to planning, but they do not establish what a future study should enroll.

What is not established by these calculations?

The displayed comparisons are unadjusted and should not be confused with a complete analysis plan. The registry record does not report a prespecified model for the primary endpoint, covariate adjustment, missing-data methods, interim analyses, or the procedure for controlling multiplicity across the many outcomes. It also does not state how participants with incomplete follow-up were handled for each endpoint. Without those details, the comparisons are best treated as descriptive statistical summaries of the posted data.

11. Confidence Intervals and Precision

The primary endpoint is observed in nearly all participants in the posted Part 1 groups. Such frequent events leave little room for the endpoint to separate groups, particularly when group sizes are limited. The reference group has 16 participants, while the other group sizes range from 12 to 63. A difference of a small number of participants can materially affect a proportion in a group of this size, so the raw counts are important context for the comparison.

For the 1200 Q4W comparison, the reported risk-difference interval extends from a negative value to a positive value. The same is true for the 2400 Q2W comparison, whose upper bound is 0.005. Both intervals therefore include zero. The estimates do not precisely distinguish a difference from no difference. The comparison intervals for groups with all participants experiencing an event are zero-width, but that appearance of precision is a consequence of boundary observations and should not be interpreted as certainty about population risks.

Uncertainty is also prominent in the Part 1 objective response results. For example, the 2400 Q4W response rate is 16.7%, with a 95% confidence interval from 3.6% to 41.4%. In the 1200 Q2W group, the observed rate is 0.0%, with an interval from 0.0% to 26.5%. These ranges illustrate that a point estimate, including a zero estimate, may coexist with substantial uncertainty about the underlying rate.

For Part 2 progression-free survival, the hazard ratio is 0.94 and its 95% confidence interval is 0.61 to 1.45. The interval includes one and allows for effects in either direction. The reported median estimates also have overlapping confidence intervals. Taken together, these summaries do not give a sharply determined comparative estimate. Precision is distinct from whether a result crosses a conventional significance threshold: a broad interval can leave clinically important possibilities unresolved. The registry record does not report a clinical threshold for interpreting these endpoints.

12. Multiplicity, Missing Data and Analysis Populations

The registry record lists 29 outcome measures and one analysis in its overall record summary, and it also lists 22 posted secondary outcomes. The primary endpoint is restricted to treated participants in Part 1, whereas the posted Part 2 endpoints involve different groups. The record does not report how statistical testing across outcomes, doses, schedules or time points was organized. It does not report a family-wise error strategy, false-discovery procedure, hierarchy of endpoints or a distinction between confirmatory and exploratory tests. Consequently, nominal comparisons should not be treated as if they were protected from the increased chance of false-positive findings across multiple analyses.

For the primary endpoint, the registry record does not list a sponsor-reported inferential comparison. The displayed group comparisons are unadjusted. It is not reported in the registry record whether these comparisons were specified in advance, whether any adjustment for baseline factors was planned, or whether there were interim analyses. No sample-size justification for the primary endpoint is reported in the registry record.

Missing-data handling is not reported in the registry record. The posted denominators show the group sizes used for particular outcomes, but they do not explain how missing observations, early discontinuation, incomplete follow-up or unavailable assessments were treated. For adverse events, the stated window extends through 100 days after the last dose; the record does not describe the completeness of observation through that window for each participant. The posted summaries cannot establish whether follow-up duration was comparable between groups.

Analysis-population details are only partly stated. The primary endpoint population is described as all treated participants in Part 1. The registry record does not provide a fuller analysis-population definition for each secondary endpoint, such as whether the analyses followed intention-to-treat principles or used another population. It also does not explain how participants were assigned to analysis groups if treatment exposure differed from assignment. These omissions limit assessment of selection, missingness and the role of randomization in each comparison. The results can be described as posted, but the registry record does not support reconstruction of a complete statistical analysis plan.

13. 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: malignant neoplasm progression (44.4%), dysphagia (13.3%), pyrexia (13.3%), dyspnoea (12.5%), colitis (12.3%).

The serious adverse-event figures are separate from the primary count of participants with any adverse event. In Part 1, the posted serious-event counts range from 7 of 12 in the 1200 Q2W group to 31 of 63 in the 2400 Q2W group; the other groups also have participants with serious adverse events. Part 2 reports 44 of 62 in the BMS-986253 combination group and 42 of 57 in the placebo combination group. These are participant counts, not numbers of event episodes.

The listed serious-event terms include malignant neoplasm progression, dysphagia, pyrexia, dyspnoea, colitis and pneumonia. The registry record gives percentages for these terms but does not explain in this summary whether terms can overlap within a participant or how attribution was assessed. The figures should therefore be read as reported safety descriptions, not as a causal comparison between treatment groups.

Deaths are also posted as a distinct outcome, with a specified window from first dose through 100 days after last dose. The count alone does not establish cause or treatment relationship. Comparisons of safety counts require attention to group size, follow-up, event definitions and attribution; those details are not all reported in the registry record.

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

The primary result describes the number of treated participants in each Part 1 group who had at least one adverse event during the stated observation window. It shows that such events were recorded for most participants in every posted group. That is a description of observed events under the study’s broad definition, not a finding that BMS-986253 caused those events or that one dose schedule is safer than another.

The endpoint does not distinguish events by severity, causal attribution, duration or clinical importance. The registry’s definition says an adverse event need not be causally related to treatment, and the primary measure counts participants rather than individual episodes. A participant-level count cannot tell whether one participant had one event or several. Separate serious-event summaries add information about seriousness, but they do not establish causation or replace a full safety analysis.

The posted comparisons do not establish equivalence between the Part 1 groups. Several groups have boundary counts, and the intervals for the comparisons that vary include zero. These features mean that the observed data do not precisely determine how underlying event risks differ. Nor does a statistical comparison alone answer whether any difference would be clinically important; the registry record does not provide a threshold for that judgment.

The secondary outcomes do not change the scope of the primary result. Response and progression-free survival address tumor outcomes, while pharmacokinetic measures describe drug concentrations or disposition. The Part 2 progression-free survival estimate remains uncertain, and the mortality result should be interpreted cautiously as an unadjusted secondary comparison. None of these outcomes establishes that the primary adverse-event count reflects a treatment effect.

In short, the posted data support a description of observed safety events and exploratory reading of the distinct secondary outcomes. They do not, on their own, prove benefit, harm, equivalence or causal attribution. More complete interpretation would require the analysis plan, details of follow-up and missing data, and full reporting of outcome-specific methods.

15. Why This Trial Matters Statistically

This trial illustrates how the same registry record can contain outcomes that answer very different statistical questions. The primary endpoint is a binary safety count in Part 1; objective response is a tumor outcome in Part 1; and progression-free survival is a time-to-event outcome in Part 2. Treating these results as interchangeable would obscure differences in population, measurement and analysis.

The primary endpoint also demonstrates a limitation of common binary comparisons when events are nearly universal. If every observed participant in both groups has an event, the calculated difference can be zero with an interval that offers no realistic account of uncertainty at the population level. The group counts and the calculation method must be considered together. A zero-width interval in this setting is not substantive evidence of exact equality.

The Part 2 progression-free survival result provides a useful contrast: a hazard ratio is reported with an interval that spans the null value, making uncertainty visible. The broad response-rate intervals likewise show why small groups and few events can limit precision even when point estimates differ. Statistical interpretation depends not only on the estimated value but also on the interval and the information available.

Finally, the record lists many outcomes but does not report a multiplicity strategy or several important analysis details. This underscores why registry results should be read with attention to confirmatory status, missing-data procedures, analysis populations and the prespecified plan. The study is statistically informative as a case in endpoint interpretation and uncertainty, while the posted summaries leave important questions unresolved.

16. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

The primary analysis is limited to treated participants in Part 1, while the registry’s overall enrollment and Part 2 results refer to a broader study. The posted record does not provide tumor-type distributions by group, detailed eligibility criteria, baseline characteristics or the allocation ratio. Without those details, the comparability and representativeness of particular outcome groups cannot be fully evaluated from the registry summary alone.

The primary endpoint is broad and does not separate adverse events by grade or attribution in the displayed participant counts. The registry description defines grades, but the posted primary result is not a grade-specific analysis. The serious-event summaries provide additional descriptive counts, yet the record does not fully explain attribution, overlap among event terms or completeness of event follow-up. The primary endpoint’s long possible observation period also means that time at risk could matter; group-specific exposure and follow-up summaries are not reported in this record.

Group sizes are modest for many Part 1 comparisons, and the high frequency of any adverse event leaves limited variation for a binary contrast. Boundary counts lead to zero-width intervals in some simple comparisons, a limitation of the displayed calculation rather than evidence of certainty. Secondary response intervals are wide, and the progression-free survival interval spans the null value.

The registry record does not report the original sample-size justification, a complete statistical analysis plan, missing-data methods, interim-analysis details or multiplicity control. It also does not establish which secondary analyses were confirmatory. These reporting gaps limit causal and confirmatory interpretations. The posted statistics remain useful descriptive evidence, but they cannot substitute for complete methods and results reporting.

17. Related Tutorials

18. Related Statistical Calculators

19. Sources

• ClinicalTrials.gov: A Study of BMS-986253 in Combination With Nivolumab or Nivolumab Plus Ipilimumab in Advanced Cancers — NCT03400332.
• Registry record last updated 2026-02-24; status last verified 2026-01.
• Results first posted 2025-06-19.
• 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.

20. Record Summary

NCT03400332 is a completed randomized, quadruple-masked, parallel Phase 1/2 study of BMS-986253 with nivolumab or nivolumab plus ipilimumab in advanced cancers. The primary endpoint counts treated Part 1 participants with at least one adverse event from first dose through 100 days after last dose. Events were recorded in most participants across the posted Part 1 groups, but the broad endpoint does not establish treatment causation.

Posted secondary results include Part 1 objective response rates and Part 2 progression-free survival and mortality outcomes. Their populations and definitions differ, and the registry record does not report multiplicity control, missing-data methods or a complete analysis plan. The results are best read as descriptive and exploratory, with uncertainty and reporting limitations kept in view.