Study of BGB-A317 in Participants With Previously Treated Unresectable HCC: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Study of BGB-A317 in Participants With Previously Treated Unresectable HCC (NCT03419897), a na phase 2 study. The primary outcome is Objective Response Rate (ORR) Assessed by Independent Review Committee (IRC).

2. Clinical Question

In practical statistical terms, this is a question about the proportion of participants meeting a defined response criterion in a treated group, not a question about whether tislelizumab caused better outcomes than an alternative. The registry lists other outcomes, including investigator-assessed ORR, duration of response, progression-free survival, and overall survival. Those outcomes may describe additional aspects of the study, but their names alone do not provide their results. The record describes the study as investigating efficacy, safety, and pharmacokinetics; the primary endpoint information concerns response as assessed by the IRC.

3. Trial Design

This was a phase 2, open-label, multicenter, single-group study sponsored by BeiGene. It investigated tislelizumab in participants with previously treated unresectable HCC. The recorded regimen was 200 milligrams once every 3 weeks. The study enrolled 249 participants and is marked completed in the registry record. The listed start date is 2018-04-09, and the primary completion date is 2021-06-30.

Because the study had a single group, participants were not allocated between a study treatment and a concurrent control. Allocation is recorded as NA; masking is recorded as none. The study’s design can describe outcomes observed among participants receiving the study drug, but it cannot by itself establish how those outcomes compare with outcomes under another treatment. Comparisons with external studies would involve other populations and study conditions and are not reported as a controlled comparison in this record.

Open-label means that the study was not masked. The primary endpoint was assessed by an IRC, which is a distinct assessment source from the investigators named in several secondary outcomes. The record identifies the primary response assessment as IRC-assessed but does not report details of reviewer blinding or procedures for resolving disagreements.

The primary endpoint timeframe runs from the date of first dose to the primary analysis data cut-off date of 30-June-2021, described as up to approximately 3 years and 3 months. That is an outcome observation window, not a statement that every participant was followed for the same duration. Participant-level follow-up times and the handling of incomplete follow-up are not reported in the registry record. Nor is a detailed protocol or statistical analysis plan 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.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Safety analysis set included all participants who received ≥ 1 dose of 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.

6. Primary Endpoint

The registry lists the primary outcome as Objective Response Rate (ORR) Assessed by Independent Review Committee (IRC), measured over From date of first dose to primary analysis data cut-off date of 30-June-2021 (up to approximately 3 years and 3 months).

The primary endpoint is objective response rate assessed by an independent review committee. ORR is the percentage of participants whose best overall response was complete response or partial response, determined using RECIST version 1.1. The record defines complete response as disappearance of all target lesions and partial response as at least a 30% decrease in the sum of target-lesion diameters.

ORR is a binary-outcome summary: each participant is classified according to whether the response definition was met, and the endpoint reports the percentage meeting it. It is used to describe the frequency of a prespecified degree of tumor response during the stated assessment period. An IRC assessment provides an assessment source distinct from investigator assessment; investigator-assessed ORR is separately listed as a secondary outcome. The registry record does not explain the IRC’s operating procedures or whether its assessors were masked.

The stated primary analysis population is the safety analysis set, comprising participants who received at least one dose of study drug. The timeframe begins at first dose and ends at the primary analysis data cut-off date of 30-June-2021. Although the endpoint definition and timeframe are described, the observed ORR estimate and its 95% confidence interval are not reported in the registry record. Consequently, the endpoint’s numerical result and the precision of that result cannot be interpreted from the record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry record identifies a binary percentage endpoint, an IRC assessment, a stated analysis population, and a 95% confidence interval as the dispersion measure. The numerical estimate and interval are not reported in the registry record. The concepts below explain how such methods are generally interpreted, without assigning an unreported result to this study.

What does a percentage endpoint describe?

In general, a binary endpoint assigns each participant to one of two categories: meeting or not meeting a defined criterion. For ORR, the criterion is a best overall response of CR or PR according to RECIST version 1.1. A percentage summarizes the fraction of the analysis population classified as responders. Its interpretation depends on the response rules, the population included, the assessment schedule, and the approach to participants whose outcome cannot be classified. The registry record supplies the response definition and names the safety analysis set, but does not report the observed percentage, a responder count, or the treatment of missing response assessments.

What does the confidence interval add?

General explanation: a point estimate is a concise description of the sample, while a confidence interval communicates uncertainty about the underlying quantity. Repeated studies could produce different estimates because they enroll different participants and observe different outcomes. A confidence interval uses the observed data and a statistical procedure to express that uncertainty. It is not a guarantee that the underlying value lies within the reported range, and it does not determine whether an effect is important. The record specifies a 95% confidence interval for the primary endpoint but does not report its limits or the method used to calculate it.

Why does the assessment source matter?

The primary ORR is assessed by an IRC, while investigator-assessed ORR is listed separately as a secondary outcome. Different assessment sources can apply the same response criteria but may differ in the information and review process available to them. Using a defined committee assessment can distinguish the primary assessment from investigators’ assessments, but the registry record does not describe reviewer masking, adjudication, or other operating details. It is therefore possible to identify who assessed the primary endpoint, but not to evaluate those procedures from the record alone.

How should the analysis population be read?

The primary endpoint description names the safety analysis set, defined as all participants who received at least one dose of study drug. In general, the population used for an endpoint determines whose outcomes contribute to its estimate. An analysis restricted to participants who began treatment may differ from one based on everyone enrolled if some enrollees never receive a dose. The record lists 249 enrolled and lists 249 at risk in the safety information. It does not provide the primary endpoint’s responder count, an explicit analyzed denominator, or additional population definitions.

What does the single-group design imply?

General explanation: a single-group study estimates outcomes within the treated group, but does not create a randomized estimate of the difference between treatment and control. Randomization is one way to balance measured and unmeasured characteristics between concurrent groups; no such comparison is described. A single-group response result can be useful as a description of outcomes under the study conditions, but causal interpretation requires caution. The study’s allocation is recorded as NA and its model as single-group. The record does not give a comparator arm.

What can planning assumptions tell readers?

General explanation: a sample-size calculation uses assumptions about a target effect, variability, significance level, allocation, and desired power to illustrate how many participants or events might be required for a future design. Such an illustration is not an observed result and is not observed power. The registry record describes a planning illustration based on the observed effect, with a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation; it says the effect to detect is a stated fraction of the observed effect. Time-to-event illustrations use a total event count, while other illustrations use participants per arm. These assumptions do not describe the actual single-group design, and the registry record does not report a trial-specific sample-size justification. No numerical planning outputs are stated in the record.

The record lists one analysis and 17 outcome measures. These counts do not establish which inferential procedures were used, whether analyses were prespecified, or how uncertainty was handled across multiple outcomes. A detailed statistical analysis plan, any interim analyses, and multiplicity procedures are not reported in the registry record. The endpoint’s named confidence interval and outcome definitions should therefore be distinguished from unreported implementation details.

10. Exact Estimates for Each Arm

No two-arm comparison is shown for this trial, but each arm's result can still be summarized with exact interval estimates computed from the posted counts. The Clopper-Pearson interval is exact and conservative; the Wilson interval is a common closer-to-nominal alternative; the Bayesian credible interval uses a uniform Beta(1, 1) prior. For small groups the three can differ noticeably.

R code and output

arm    <- c("Tislelizumab")
n      <- c(249)
events <- c(32)
for (i in seq_along(n)) {
  ci <- binom.test(events[i], n[i])$conf.int          # Clopper-Pearson exact interval
  cr <- qbeta(c(0.025, 0.975), 1 + events[i], 1 + n[i] - events[i])   # Bayesian, uniform Beta(1,1) prior
  wi <- (function(x, n) { p <- x / n; z <- qnorm(0.975); d <- 1 + z^2 / n
    c((p + z^2 / (2 * n) - z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d, (p + z^2 / (2 * n) + z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d) })(events[i], n[i])
  cat(sprintf("ARM|%s|%d|%d|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f\n", arm[i], n[i], events[i], 100 * events[i] / n[i],
              100 * ci[1], 100 * ci[2], 100 * wi[1], 100 * wi[2], 100 * cr[1], 100 * cr[2]))
}

ARM  Tislelizumab  249  32  12.9  9.0  17.7  9.3  17.6  9.3  17.6

11. Multiplicity, Missing Data and Analysis Populations

The registry record lists 17 outcome measures and one analysis. It identifies ORR assessed by IRC as the primary endpoint and lists secondary outcomes that include investigator-assessed ORR, duration of response, progression-free survival, and overall survival. When a study evaluates multiple outcomes, the chance of obtaining at least one apparently notable result can increase if each is tested separately. General approaches to multiplicity include prespecifying a primary outcome, adjusting inference across tests, or treating additional analyses as exploratory. The registry record does not report a multiplicity adjustment, testing hierarchy, or detailed statistical analysis plan; the listed counts alone do not establish which approach was used.

Missing outcome data can affect a percentage endpoint if some participants lack assessments needed to determine their best overall response. General methods may classify missing outcomes in specified ways or use other modeling approaches, but the choice and its assumptions matter. The registry record does not report how missing, unevaluable, or discontinued participants were handled for the primary endpoint. It also does not provide a responder count or explicit primary endpoint denominator. The stated safety analysis set comprises participants who received at least one dose, but that definition does not itself explain missing response handling.

The record lists 249 participants enrolled and 249 at risk in the safety information. It does not report a separate endpoint analysis population, exclusions, or sensitivity analyses. Interim analyses and stopping rules are not reported in the registry record. Consequently, readers can identify the named endpoint, analysis population description, and outcome inventory, but cannot determine from this record how multiplicity or missing response assessments influenced the primary estimate.

12. 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: ascites (3.6%), hepatic failure (2.8%), pneumonia (2.0%), aspartate aminotransferase increased (2.0%), upper gastrointestinal haemorrhage (1.6%).

The safety information lists 249 participants at risk in the tislelizumab group. It records 94 participants with serious adverse events and 226 with other adverse events. These are counts, not evidence that the events were caused by the study drug. The record does not provide a comparison group, participant-level event histories, exposure-adjusted rates, or details of attribution in the safety figures. A count also does not show how many events each participant experienced.

The serious adverse event terms listed most frequently include ascites at 3.6%, hepatic failure at 2.8%, pneumonia at 2.0%, and aspartate aminotransferase increased at 2.0%. Upper gastrointestinal haemorrhage and hepatic function abnormal are each listed at 1.6%. These percentages describe the listed terms in the registry safety information; they do not establish that a term was treatment-related. The record does not give definitions or further detail for these terms in the information summarized.

The safety information also lists 180 deaths. Timing, cause, relationship to treatment, and follow-up context for these deaths are not reported in the registry record. Deaths should not be assumed to be caused by treatment on the basis of this count alone. These figures describe what the record lists, not comparative safety or individual risk.

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

The registry record documents a phase 2 study of tislelizumab in previously treated participants with unresectable HCC, identifies an IRC-assessed response endpoint, and gives the response criteria used. It also provides selected safety counts and lists additional outcomes. These details help establish what the study set out to assess and the broad structure of the recorded results.

The primary response result itself is not reported in the registry record. As a result, no observed ORR percentage, responder count, or 95% confidence interval can be interpreted from the record. The endpoint definition cannot show whether the response criterion was met by any particular proportion of participants, and the enrollment count is not a substitute for the missing endpoint estimate.

Even if a response estimate were available, a single-group study would describe outcomes within the treated group rather than provide a randomized comparison with another treatment. The study is listed as open-label, with no masking and no comparator. Thus, the design does not isolate a treatment contrast or establish that an observed outcome was caused by the study drug. This is a general distinction between descriptive outcomes in one group and comparative evidence.

The listed serious adverse events and deaths also require cautious interpretation. They are recorded counts without a concurrent comparator or enough information in the registry record to assess timing, attribution, or exposure-adjusted occurrence. They do not independently establish either treatment causation or comparative risk.

Several secondary endpoints are named, including duration of response, progression-free survival, and overall survival. The names indicate further outcomes of interest, but do not provide their estimates. Without numerical results, analysis details, and information about missing data, the record does not support conclusions about the size or precision of those outcomes. The appropriate reading is therefore limited to the study design, stated endpoint definitions, and safety figures that the registry record actually reports.

14. Why This Trial Matters Statistically

Statistically, this study is a clear example of why an endpoint definition and an endpoint result are different kinds of information. The registry record defines ORR using CR or PR under RECIST version 1.1 and identifies an IRC assessment. That tells readers what qualifies as a response and which assessment source was named. But without the response estimate and its 95% confidence interval, readers cannot assess either the observed frequency or its precision.

The design also matters for the kind of evidence the study can produce. A single-group study can describe outcomes in participants receiving the recorded regimen, but it has no concurrent randomized comparator. Its estimates therefore answer a different question from a randomized treatment comparison. Recognizing that distinction prevents an observed within-group outcome from being interpreted as a causal contrast.

The record also illustrates why the analysis population and missing-data rules matter. It names a safety analysis set for the primary endpoint, but does not report the analyzed denominator, responder count, or handling of participants without evaluable response assessments. Those details are essential for understanding how a percentage is formed and who it represents.

Finally, the registry lists multiple outcomes and selected safety counts. The presence of additional measures does not reveal how they were prioritized or whether inference was adjusted for multiple testing. Nor do event counts alone establish causation or comparative risk. These distinctions are broadly important when reading clinical trial records: definitions, design, analysis choices, precision, and reported estimates each contribute different information.

15. 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 immediate limitation for interpreting the primary outcome is that the registry record does not report its numerical result. Although it specifies ORR, the IRC assessment, RECIST version 1.1, a 95% confidence interval, and the analysis-population definition, it provides no response estimate, interval limits, or responder count. The primary endpoint’s precision and magnitude therefore cannot be evaluated from the record.

The single-group, open-label design has no concurrent comparator and no masking. It can describe outcomes in the study group, but cannot provide a randomized between-group comparison. The record does not report details of the IRC’s procedures, including reviewer masking or adjudication, so the assessment process cannot be fully evaluated.

The primary endpoint description names participants who received at least one dose as the safety analysis set. However, the registry record does not state the primary endpoint’s analyzed denominator, how missing or unevaluable assessments were handled, or whether sensitivity analyses were conducted. These omissions limit assessment of how the analysis population and incomplete outcomes may have affected the estimate.

The record lists 17 outcome measures and one analysis but does not report a multiplicity strategy, interim analyses, or a detailed statistical analysis plan. A trial-specific sample-size justification is also not reported in the registry record. Although planning assumptions are described, they are illustrations and do not establish the study’s actual design rationale. Finally, safety counts are not accompanied by a concurrent control or fuller attribution and timing details, limiting comparative interpretation.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Study of BGB-A317 in Participants With Previously Treated Unresectable HCC — NCT03419897.
• Registry record last updated 2024-10-26; status last verified 2024-10.
• Results first posted 2023-07-27.
• 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.

19. Record Summary

This completed phase 2 study investigated tislelizumab, also called BGB-A317, in participants with previously treated unresectable HCC. It was open-label, multicenter, and single-group, with no concurrent comparator. The recorded regimen was 200 milligrams once every 3 weeks. The registry lists 249 participants enrolled, a start date of 2018-04-09, and a primary completion date of 2021-06-30.

The primary endpoint was IRC-assessed ORR, defined using CR and PR under RECIST version 1.1. The stated analysis population was participants who received at least one dose. The numerical primary endpoint result and its confidence interval are not reported in the registry record. Safety information lists 94 participants with serious adverse events, 226 with other adverse events, and 180 deaths among 249 at risk. These counts do not establish treatment causation or comparative risk.

The record names additional response, duration, progression-free survival, and overall survival outcomes, but their numerical results are not reported in the registry record. A detailed statistical plan, missing-data handling, and multiplicity procedures are also not reported.