Study of Ociperlimab (BGB-A1217) in Combination With Tislelizumab in Advanced Solid Tumors: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Study of Ociperlimab (BGB-A1217) in Combination With Tislelizumab in Advanced Solid Tumors (NCT04047862), a non-randomized phase 1 study. The primary outcome is Phase 1: Number of Participants With Treatment Emergent Adverse Events (TEAEs) and Treatment Emergent Serious Adverse Events (TESAEs).

2. Clinical Question

The study was designed as an early-phase investigation rather than a randomized test of one regimen against another. Its questions therefore span dose exploration, safety characterization, and preliminary tumor-response assessment. The endpoint emphasis differs by phase: safety and dose selection in Phase 1, and cohort-specific response assessment in Phase 1b. Because the cohorts have different diseases, regimens, and doses, their outcomes should be read within their own context rather than treated as if they came from interchangeable arms. The registry record describes enrollment of 446 participants, but that enrollment figure should not be assumed to be the denominator for every endpoint or cohort. The safety analysis set is defined separately as participants who received at least one dose of any study drug.

3. Trial Design

This was a Phase 1 study with Phase 1b components, conducted as a non-randomized, unmasked, parallel-group trial. The registry lists an enrollment of 446 and a completed status. The recorded study start was 2019-08-15, and primary completion was 2024-08-07. These record dates describe the study timeline; they do not establish that all participants had the same follow-up duration.

In Phase 1 dose escalation, participants received ociperlimab at 50 mg, 150 mg, 450 mg, 900 mg, or 1800 mg with tislelizumab 200 mg. The first cycle was described as 28 days, with later cycles every 21 days. Dose verification included an ociperlimab 900 mg monotherapy cohort and a cohort receiving ociperlimab 900 mg with tislelizumab 200 mg. These cohorts addressed dose and safety questions, not randomized treatment comparisons.

Phase 1b dose expansion covered distinct tumor populations and treatment regimens. The registry describes cohorts for squamous and non-squamous non-small cell lung cancer, small-cell lung cancer, esophageal squamous cell carcinoma, esophageal adenocarcinoma, head and neck squamous cell carcinoma, and gastric or gastroesophageal junction adenocarcinoma. Some received chemotherapy alongside ociperlimab and tislelizumab; regimens and eligibility criteria differed among cohorts. The dose-optimization cohorts evaluated ociperlimab 450 mg, 900 mg, or 1800 mg with tislelizumab 200 mg in participants with metastatic non-small cell lung cancer meeting the specified PD-L1 criterion.

Since allocation was not randomized and masking was absent, differences between cohorts can reflect differences in patient populations, disease characteristics, regimens, or follow-up as well as dose. The design does not support interpreting cohort contrasts as unbiased causal estimates of one regimen versus another. Protocol details beyond the cohort descriptions, including a statistical analysis plan, are not reported in the registry record.

4. Participant Flow and Accounting

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

Only the first 6 of 19 groups are shown.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Analysis was performed on the safety analysis set that included all participants who received \>= 1 dose of any study drug.

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: Baseline analysis included participants from safety analysis set, that is, (i.e.) participants who received more than or equal to (\>=) 1 dose of any study drug.

Only the first 6 of 20 groups are shown.

6. Primary Endpoint

The registry lists the primary outcome as Phase 1: Number of Participants With Treatment Emergent Adverse Events (TEAEs) and Treatment Emergent Serious Adverse Events (TESAEs), measured over Up to 30 days after the last dose of study interventions (up to 35.7 months [Dose escalation cohorts] and up to 13.3 months [Dose verification cohorts]).

The Phase 1 primary endpoint counted participants with treatment-emergent adverse events (TEAEs) and treatment-emergent serious adverse events (TESAEs). An adverse event was defined as an unfavorable and unintended sign, symptom, or disease associated with study-drug use, whether or not considered related to the drugs. A TEAE was an event that began, or worsened in severity from baseline, on or after the first dose and within the defined period after discontinuation or initiation of new systemic anticancer therapy.

The endpoint also distinguishes serious events. The registry definition includes events resulting in death, being life-threatening, requiring or prolonging hospitalization, causing disability or a congenital anomaly, or otherwise considered medically significant. Severity was assessed using NCI-CTCAE v.5.0, with grades ranging from mild through death due to an adverse event.

Counting participants with events is useful in early-phase studies because safety and tolerability are central to evaluating dose regimens and deciding whether further investigation is warranted. Counts show how many participants in each cohort experienced an event, while the cohort denominator gives context for those counts. A participant can experience more than one event, so participant counts do not describe the total number of event episodes or their duration. The primary endpoint timeframe extended to 30 days after the last dose, with the record describing maximum periods of up to 35.7 months for dose-escalation cohorts and up to 13.3 months for dose-verification cohorts. Those different exposure and follow-up windows matter when comparing raw counts. Numerical summaries for the primary endpoint are not reported in the registry record as a single overall treatment effect.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does a participant count represent?

The primary outcome is a count of participants who experienced TEAEs and TESAEs, rather than a continuous measurement such as a laboratory value. For a cohort, the event count describes the number of participants meeting the event definition during the stated observation period. The denominator describes participants considered at risk for that outcome in the posted cohort result. A count alone is difficult to interpret without its denominator; conversely, a proportion would summarize the count relative to that denominator. No new proportions should be inferred from the registry counts when the posted presentation gives counts.

Participant-based outcomes differ from event-based outcomes. One participant may experience multiple adverse events, but a count of participants with at least one event does not count each episode separately. It also does not convey event duration, recurrence, severity distribution, or whether the event led to a change in treatment. The registry defines severity using NCI-CTCAE v.5.0, but the cohort counts do not by themselves summarize all of those dimensions.

Why are the cohort counts not a treatment comparison?

Random allocation can help make groups comparable, on average, for measured and unmeasured characteristics. This study was non-randomized. Participants were grouped by dose, phase, and tumor cohort, and the regimens were not identical across groups. Some Phase 1b cohorts received chemotherapy in addition to the study drugs, while others did not. The registry also includes ociperlimab monotherapy in dose verification. Accordingly, raw differences in event counts can reflect differences in participants, disease settings, combinations, dose, or time under observation.

Even within dose-optimization cohorts, a comparison across ociperlimab doses is not randomized according to the record. The design therefore does not justify a causal interpretation of differences between those cohorts. Formal adjustment could in principle address measured differences under a stated method, but such an analysis plan and adjusted estimates are not reported in the registry record.

What do precision measures usually add?

As a general statistical concept, an estimate’s standard error describes its expected sampling variability under a model. A confidence interval uses that uncertainty to express a range of values compatible with the data and method. With sparse counts or small cohorts, estimates can be imprecise; a difference between two observed counts may be compatible with a wide range of underlying rates. The registry record does not report confidence intervals for the safety counts or numerical efficacy estimates. Thus, the precision of cohort-level event frequencies cannot be assessed from a confidence interval in the record.

As a general concept, a p-value evaluates how unusual data would be under a specified null hypothesis and analysis assumptions. It is not a measure of effect size or clinical importance. The record lists 0 analyses, and does not report efficacy hypothesis tests, a comparison model, or a statistical analysis plan. No p-value should be inferred from the descriptive counts.

What do the planning illustrations mean?

The planning illustrations are calculations based on an observed effect, not the study’s original design justification and not a measure of observed power. Their stated assumptions are a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation, with the effect to detect set as a stated fraction of the observed effect. For non-time-to-event outcomes, the reported planning quantity is participants per arm for 80% power or another stated power level; time-to-event calculations use a total event count. These are hypothetical planning quantities, not observed participants or trial results.

Those assumptions do not describe the actual allocation in this non-randomized multi-cohort study. Equal allocation between two arms is a planning simplification, not a design feature of the listed cohorts. Moreover, no efficacy effect estimate is reported in the registry record, so the underlying clinical result cannot be evaluated from the record alone. A planning calculation based on an effect does not establish that the study had that power, that the effect is reliable, or that the trial was designed around it. The original sample-size justification is not reported in the registry record.

Over-dispersion is a general concern when outcome variability exceeds what a simple statistical model expects. It can make standard errors too small if the model fails to account for the extra variability, leading to overly narrow confidence intervals or misleading tests. The primary endpoint is a participant count, but the registry does not report a model for analyzing those counts or any over-dispersion assessment. It is therefore not possible to say that a particular count model, variance adjustment, or robust method was used.

For safety outcomes, the observation window also affects interpretation. The TEAE definition includes a post-treatment period, and the record gives different maximum follow-up durations for dose-escalation and dose-verification cohorts. A longer observation period can provide more opportunity to record events. Raw counts across cohorts do not account for differences in exposure time, and the registry record does not report person-time rates or a time-to-event safety analysis.

In short, the posted numerical material supports descriptive accounting of participants with events by cohort. It does not provide the details needed to reconstruct an inferential analysis, establish precision, or identify a causal dose or regimen effect. General methods such as confidence intervals, hypothesis tests, and adjusted models require explicit estimates, assumptions, and analysis specifications; those are not reported in the registry record.

10. Multiplicity, Missing Data and Analysis Populations

The study has multiple phases, cohorts, doses, tumor populations, and outcome measures. In such a setting, multiplicity is a general statistical issue: making many inferential comparisons can increase the chance of obtaining at least one apparently notable result by chance. A prespecified strategy may define which analyses are primary and whether any adjustment or hierarchical testing is used. The registry identifies primary objectives by phase, but multiplicity handling and a detailed statistical analysis plan are not reported in the registry record. Because numerical efficacy analyses are not posted, no multiplicity-adjusted or unadjusted efficacy conclusion can be evaluated.

Missing-data handling is also not reported in the registry record. The safety analysis population is described as all participants who received at least one dose of any study drug. The registry does not state how participants with incomplete event follow-up, missing assessments, withdrawal, or other unavailable data were handled for each endpoint. The primary safety endpoint’s observation period is defined, but the definition alone does not explain how incomplete observation was addressed analytically.

The safety analysis set should not be confused with the enrolled population. The record reports enrollment of 446, while the safety population includes participants who received at least one dose. Cohort-specific denominators are listed for the adverse-event results; no single denominator should be substituted for every cohort or outcome. The registry does not report the number excluded from the safety set or an endpoint-specific accounting of missing observations.

For the Phase 1b response objective, response is to be assessed by investigators using RECIST v.1.1. The record does not report the response-evaluable population, how missing or unevaluable scans were treated, or the numerical response results. Therefore, neither the handling of missing efficacy assessments nor the impact of missingness on the response estimate can be assessed from the registry record.

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 3% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: atrial flutter (100.0%), carotid sinus syndrome (33.3%), dyspnoea (33.3%), pneumonia (19.5%), dysphagia (16.7%).

The safety results are cohort-level counts. For example, the Phase 1b dose-expansion Cohort 9 lists 32 of 60 participants with serious adverse events and 59 of 60 with other adverse events. The Phase 1b dose-optimization cohort receiving ociperlimab 900 mg with tislelizumab 200 mg lists 8 of 21 with serious adverse events and 17 of 21 with other adverse events. These figures describe recorded participants within each group; they are not a comparison between equivalent groups.

Serious adverse events and other adverse events are distinct categories in the posted results. The registry provides definitions for seriousness and for treatment-emergence, and describes the event window as up to 30 days after the last dose. TEAEs include events whether considered related to study drugs or not. Accordingly, an event count should not be described as a count of drug-caused events.

The record also lists death counts by cohort, including 42 in Cohort 9 and 35 in Cohort 4. These are reported counts, not evidence by themselves that deaths were caused by treatment. The registry lists serious-event terms including atrial flutter at 100.0%, carotid sinus syndrome at 33.3%, dyspnoea at 33.3%, pneumonia at 19.5%, dysphagia at 16.7%, and nausea at 16.7%. The record does not make the denominator for these term percentages clear in the registry record described, so the percentages should not be converted into participant counts.

Because follow-up durations and cohort sizes differ, raw counts do not directly measure comparative risk. The record does not report a comparative safety model, exposure-adjusted rates, or uncertainty intervals for these counts.

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

The registry describes a broad early-phase investigation of ociperlimab, alone or with tislelizumab, across dose-escalation, dose-verification, dose-expansion, and dose-optimization cohorts. It states that Phase 1 sought to characterize safety and tolerability and inform dose selection, while Phase 1b sought investigator-assessed overall response rate in each expansion cohort. The posted safety counts provide descriptive information about participants recorded with serious and other adverse events in the listed groups.

Those counts do not establish that a treatment caused an event, that one dose is safer than another, or that a regimen is more effective than another. The trial was non-randomized, and the cohorts differ in tumor type, eligibility, dose, and accompanying treatment. Comparisons across them are vulnerable to differences in participant characteristics and observation time. A count of deaths does not, on its own, establish treatment-related mortality.

The registry does not report numerical overall response rate, duration of response, or disease control results in the registry record for this analysis. It also does not report confidence intervals, inferential efficacy tests, or a statistical analysis plan. The clinical effect and its precision therefore cannot be quantified from the registry record. The completed status is not a substitute for posted outcome estimates.

General statistical interpretation distinguishes evidence of an observed outcome from evidence of a causal treatment effect. Causal comparisons usually require design or analysis features that address how treatment groups differ. Here, non-random allocation and the varied cohort structure limit what can be concluded from cross-cohort contrasts. No treatment recommendation follows from the descriptive registry data. The appropriate reading is narrower: the record gives study objectives, cohort descriptions, endpoint definitions, and safety counts, while key efficacy estimates and analytic details are not reported in the registry record.

13. Why This Trial Matters Statistically

Statistically, this study illustrates why early-phase, multi-cohort trials require careful separation of descriptive safety evidence from comparative treatment evidence. Dose escalation, dose verification, disease-specific expansion, and dose optimization serve different purposes. Pooling their participants into one overall figure can obscure differences in regimen, disease setting, dose, and follow-up. Reading each cohort in context is essential to avoid treating unlike groups as interchangeable.

The endpoint is also a reminder that participant counts answer a limited but important question: how many people in a defined group experienced an event under a stated event definition and observation window? Counts are not the same as event totals, exposure-adjusted incidence, severity distributions, or causal attribution. The TEAE definition includes events whether considered related or not, and the registry separately defines serious events. Keeping these concepts distinct is necessary for accurate interpretation.

The design makes the distinction between dose exploration and randomized comparison especially important. Multiple doses and cohorts can generate useful descriptive observations, but non-random allocation means differences may arise for reasons other than treatment. The registry’s stated Phase 1b response objective is clinically relevant to the study’s aim, but the numerical results and inferential precision are not reported in the registry record. That reporting gap limits the statistical conclusions that can be drawn.

Finally, planning calculations and observed results answer different questions. A sample-size illustration based on an effect and stated assumptions does not show that a study had a particular observed power, nor does it replace a reported effect estimate and uncertainty interval. Keeping those distinctions clear supports transparent interpretation of early-phase evidence without overstating what the posted record can establish.

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 central limitation is the non-randomized design. There is no like-for-like randomized control group, so outcomes across doses or cohorts cannot be interpreted as causal treatment contrasts. The unmasked design may also matter for outcomes involving investigator assessment, although the registry does not report how assessment procedures addressed that issue.

Clinical and treatment heterogeneity is substantial in the cohort descriptions. The populations include multiple solid-tumor settings, and some Phase 1b cohorts received chemotherapy with the study drugs while others did not. The Phase 1b primary response objective is reported by dose-expansion cohort, not as one common population-wide response estimate. These differences limit pooling and cross-cohort comparisons.

The registry lists enrollment of 446, but safety denominators are cohort-specific and the safety analysis set includes participants who received at least one dose. The record does not report a complete accounting of the enrolled, treated, and endpoint-evaluable populations. Follow-up windows also differ: the stated maximum is up to 35.7 months for dose-escalation cohorts and up to 13.3 months for dose-verification cohorts. Raw event counts do not adjust for such differences in observation time.

Although safety counts and selected serious-event terms are reported, the numerical efficacy results are not reported in the registry record. The record also does not report confidence intervals, a detailed statistical analysis plan, methods for missing data, multiplicity handling, or comparative safety models. Without these details, precision, robustness, and the consequences of incomplete follow-up cannot be assessed. These limitations constrain interpretation of the posted record; they do not establish that an unreported analysis was or was not performed outside the registry.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Study of Ociperlimab (BGB-A1217) in Combination With Tislelizumab in Advanced Solid Tumors — NCT04047862.
• Registry record last updated 2025-08-22; status last verified 2025-08.
• Results first posted 2025-08-22.
• 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

NCT04047862 is a completed Phase 1/1b, non-randomized, unmasked parallel study sponsored by BeiGene. The registry reports enrollment of 446 and describes ociperlimab dose escalation and verification, followed by disease-specific Phase 1b expansion and dose optimization. The principal regimen combined ociperlimab with tislelizumab, with chemotherapy included in some expansion cohorts.

Phase 1 primary objectives concerned safety, tolerability, and dose selection. Phase 1b focused on investigator-assessed overall response rate by RECIST v.1.1 in each expansion cohort. The primary Phase 1 endpoint counted participants with TEAEs and TESAEs. Cohort-level safety counts and selected serious-event terms are reported, but numerical efficacy estimates and inferential analyses are not reported in the registry record. The non-randomized, heterogeneous cohort structure limits direct comparisons across doses and populations.