A Study of Atezolizumab and Bevacizumab in Hepatocellular Carcinoma: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study of Atezolizumab and Bevacizumab in Hepatocellular Carcinoma (NCT04829383), a na phase 2 study. The primary outcome is Grade 3-5 Adverse Events.

2. Clinical Question

3. Trial Design

This is a Phase II, nonrandomized, single-group feasibility study, identified in the registry as the AB7 Trial. It evaluates atezolizumab plus bevacizumab in patients with advanced or metastatic hepatocellular carcinoma and Child-Pugh B7 or B8 liver disease who had received no prior systemic therapy. The listed condition is unresectable hepatocellular carcinoma. The enrollment is 6.

There is no randomized allocation or masking: allocation is listed as NA, masking as none, and the model as single group. All participants in the study-treatment arm were assigned to the combination described in the record. Atezolizumab 1,200 mg IV and bevacizumab 15 mg/kg IV were scheduled on day 1 of each 21-day cycle. Treatment was to continue until disease progression or unacceptable toxicity. The registry record does not report a control regimen or a protocol for comparing the combination with another treatment.

The primary purpose was feasibility and safety assessment, with the primary outcome defined as participants experiencing Grade 3-5 treatment-related adverse events. The record gives an adverse-event observation window from consent until 30 days after study-drug discontinuation or until a new anti-cancer treatment starts, whichever occurs first, up to a maximum of 10 months. The analysis population for adverse events is described as all patients receiving at least one dose, regardless of dosage.

The registry lists overall response rate, disease control rate, duration of response, overall survival and progression-free survival as secondary outcomes. It does not report numerical results for these outcomes. The study status is completed; the listed start date is 2021-03-22 and the primary completion date is 2024-06-19. Further details of the protocol, including a sample-size justification and a detailed 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.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: In accordance with the Statistical Analysis Plan, the analysis population for Adverse Events was defined as all patients receiving at least one dose of study treatment regardless of the dosage.

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 Grade 3-5 Adverse Events, measured over Adverse Events have been recorded from the time of consent until 30 days after treatment discontinuation of study drugs or until a new anti-cancer treatment starts, whichever occurs first, up to a maximum of 10 months.

The primary endpoint counts participants with a Grade 3-5 treatment-related adverse event. Events are to be described by CTCAE v5 term and grade. The registry defines the adverse-event analysis population as all patients who received at least one dose of study treatment, regardless of dosage. The observation period begins at consent and ends 30 days after treatment discontinuation or when a new anti-cancer treatment starts, whichever occurs first, with a maximum of 10 months.

In general, a safety endpoint is used to characterize the types and severity of harms observed during treatment and the specified follow-up period. Counting participants with an event answers a different question from counting every event: one participant may experience more than one event, while a participant-level count records whether the endpoint occurred for that person. The registry's primary endpoint is specifically restricted to events described as treatment-related and Grade 3-5. That definition should not be conflated with all serious adverse events or with all adverse events.

The registry record does not report the numerical result for the primary endpoint itself. It does report separate summaries of serious and other adverse events, but those summaries do not establish the number meeting the primary endpoint's treatment-related and Grade 3-5 criteria. The endpoint result is therefore not reported in the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does a participant count represent?

The primary endpoint is a count of participants with at least one qualifying Grade 3-5 treatment-related adverse event. In general, participant-level event counts describe how many people met a definition at least once during the stated observation period. They are not necessarily counts of all events: a person can experience multiple events, but a participant count does not count each occurrence separately. The registry's endpoint description specifies classification by CTCAE v5 term and grade. It also specifies that the analysis population includes patients who received at least one dose, regardless of dosage.

The primary result itself is not reported in the registry record. The safety summary separately reports serious adverse events and other adverse events. Those categories cannot be assumed to reproduce the primary endpoint, because the primary definition additionally requires an event to be treatment-related and Grade 3-5. The posted serious-event terms and affected-participant count should therefore be read as safety summaries, not substituted for the primary endpoint result.

What is the role of a confidence interval?

In general, a confidence interval accompanies an estimate to convey its statistical precision. It represents a range of values compatible with the data and specified method, rather than a guarantee about where a true value lies. Wider intervals indicate less precision than narrower intervals, all else being equal. The registry record does not report a confidence interval for the primary endpoint or the serious-event summaries, nor does it report the underlying primary endpoint count. Consequently, the precision of that endpoint cannot be assessed from the registry information.

What does the single-group design permit?

The model is listed as single group, with no randomization and no masking. In general, random allocation can help create comparable groups for estimating treatment contrasts, while masking can reduce some forms of assessment or behavior bias. Neither creates a comparison group when a study has only one treatment arm. Here, the record describes the combination but does not provide a concurrent control arm. Its observed counts can describe events in this study group; they cannot by themselves isolate the effect of the combination from disease, other care, or other influences.

The record lists allocation as NA and describes the study as nonrandomized. There is no randomized between-group contrast in the posted record. Any interpretation should therefore remain descriptive and should not assign a causal explanation to the safety counts.

What can be inferred from the listed safety summaries?

The adverse-event summary reports 4 participants affected by serious adverse events among 6 at risk, and 6 affected by other adverse events among 6 at risk. It lists 4 deaths. In general, a count indicates how many participants were recorded under a particular summary definition, not why an event occurred. Seriousness and severity are different concepts in adverse-event reporting: seriousness commonly concerns outcomes or required interventions, whereas severity describes intensity. A serious event is not automatically a Grade 3-5 event, and a Grade 3-5 event is not necessarily categorized as serious. The registry record does not provide enough detail in these summaries to map each serious event to the primary endpoint's grade and treatment-relatedness criteria.

The term-specific percentages are reported as 16.7% for each listed serious-event term. These entries indicate the registry's recorded percentages for the terms; they do not establish whether a participant had more than one listed term. No additional total across terms should be inferred from the displayed term entries. The record also does not describe the cause or timing of the listed deaths.

What do the planning assumptions mean?

The record describes sample-size calculations as planning illustrations based on the observed effect, not as the trial's own design. It specifies a two-sided significance level of 0.05, equal allocation between two arms, a normal approximation, and effect targets expressed as a stated fraction of the observed effect. For time-to-event outcomes, the stated planning quantity is the total number of events; for other outcomes, it is participants per arm. These assumptions do not describe the actual single-group study design, which has no two-arm randomized comparison.

In general, a planning calculation explores the size of a hypothetical future study under chosen assumptions. It is not an observed result and is not observed power. The registry record does not report the trial's sample-size justification or an inferential analysis using these assumptions. The planning illustrations cannot fill in the missing primary endpoint result or provide evidence of comparative treatment effects.

What analysis details remain unknown?

The registry record lists no statistical analyses. Beyond the stated adverse-event analysis population, it does not report the statistical methods used for the primary endpoint, handling of repeated events, confidence intervals, missing-data procedures, or adjustments for multiple outcomes. The methods for the secondary outcomes are not reported in the registry record. Without those details and the endpoint result, readers cannot reconstruct an inferential analysis or determine whether particular statistical assumptions were evaluated.

10. Multiplicity, Missing Data and Analysis Populations

The record lists one primary safety outcome and secondary outcomes of overall response rate, disease control rate, duration of response, overall survival and progression-free survival. The numerical results for these outcomes are not reported in the registry record. The record lists 6 outcome measures and no statistical analyses, but does not describe a multiplicity procedure, an ordering of hypotheses, or an adjustment to significance thresholds. Whether and how the several outcomes were handled together is not reported in the registry record.

Multiplicity matters in general when several outcomes or comparisons are examined: the chance of at least one apparently unusual result can change as the number of statistical tests increases. A prespecified testing strategy can clarify which findings are confirmatory and how error rates are controlled. The registry information does not report such a strategy, so no claim about multiplicity control can be made.

The adverse-event analysis population is defined as all patients receiving at least one dose of study treatment, regardless of dosage. The record does not report the number of participants in that analysis population as a separate figure, the handling of missing adverse-event information, or whether follow-up was complete for each participant. It also does not report missing-data methods for secondary outcomes. Missing-data handling can affect estimates when outcome availability differs across participants; in general, interpretation depends on why data are missing and on the assumptions used to address them. Those details are not reported in the registry record.

The endpoint observation window is described as beginning at consent and ending 30 days after study-drug discontinuation or when a new anti-cancer treatment starts, whichever occurs first, up to a maximum of 10 months. The record does not provide participant-level follow-up, endpoint-specific missingness, or a numerical result for the primary outcome. These omissions prevent assessment of how missing or incomplete observation may have affected the posted safety interpretation.

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 0% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: acute kidney injury (16.7%), anemia (16.7%), esophageal varices hemorrhage (16.7%), gastrointestinal disorders - other, specify (16.7%), generalized muscle weakness (16.7%).

The registry safety summary reports 6 participants at risk for serious adverse events and 4 affected. For other adverse events, it reports 6 at risk and 6 affected. It lists 4 deaths. The serious-event terms shown are acute kidney injury, anemia, esophageal varices hemorrhage, gastrointestinal disorders - other, specify, generalized muscle weakness, and sepsis; each term is reported as 16.7%.

These figures are descriptive and should be kept separate from the primary endpoint. The primary endpoint counts Grade 3-5 treatment-related adverse events, whereas the posted summary is organized into serious and other adverse events. In general, seriousness and severity are not interchangeable classifications. The record does not establish that each serious event met the primary endpoint definition, nor does it report the primary endpoint's numerical result.

The registry record does not state the cause of the deaths or establish that they were treatment-related. The listed term percentages do not show whether individual participants experienced multiple terms, and should not be added together. The safety summaries describe events recorded in this study group; with no control group, they do not estimate a treatment-attributable difference in risk.

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

The registry record identifies a Phase II feasibility study focused on the safety profile of atezolizumab plus bevacizumab in patients with hepatocellular carcinoma and Child-Pugh B7 or B8 liver disease who had received no prior systemic therapy. It records enrollment of 6 and describes the study as completed. Its design is single-group and nonrandomized, so the posted results describe the study-treatment group rather than a comparison between randomized treatments.

The posted safety summaries report 4 participants affected by serious adverse events among 6 at risk, 6 affected by other adverse events among 6 at risk, and 4 deaths. Serious adverse-event terms are listed with their reported percentages. These summaries document events in the registry record, but they do not show that the events were caused by treatment. The record does not state the causes of deaths or provide enough information to classify the listed serious events according to the primary endpoint's treatment-relatedness and Grade 3-5 criteria.

The primary outcome is the number of participants with Grade 3-5 treatment-related adverse events. Its numerical result is not reported in the registry record. Therefore, the posted serious-event figures cannot be used as a replacement estimate for the primary endpoint. The record also lists secondary response and survival outcomes, but does not report their numerical results.

In general, a single-group study can describe observed events and provide information relevant to feasibility, but it cannot show how outcomes would compare under an alternative treatment. The absence of randomization and a concurrent control group limits comparative interpretation. The small enrollment also means that the posted counts describe a limited group; they do not, by themselves, establish the frequency of events in a broader population. No claim that the combination works or does not work follows from these safety summaries.

The record does not report confidence intervals, a detailed statistical analysis, a sample-size justification, or methods for missing data and multiplicity. The study's stated safety aim is clear, but the missing primary endpoint result and analysis details limit how fully that aim can be evaluated from the registry record.

13. Why This Trial Matters Statistically

This trial is statistically notable because its primary aim is safety assessment in a specifically described population with hepatocellular carcinoma and Child-Pugh B7 or B8 liver disease. The combination was evaluated in a nonrandomized, single-group Phase II feasibility study, rather than in a randomized comparison. That distinction determines what the posted counts can answer: they describe recorded events in the study group, but do not estimate a contrast against another treatment.

The endpoint definition also matters. The primary outcome is not simply all adverse events or all serious adverse events; it is a count of participants with Grade 3-5 treatment-related adverse events. The registry's separate serious-event summary cannot be assumed to answer that primary question. Because the numerical primary endpoint result is not reported in the registry record, the central safety outcome cannot be quantified from the posted record.

The record illustrates why event definitions, analysis populations and observation windows are essential to interpreting clinical-trial safety data. It specifies an adverse-event analysis population based on receipt of at least one dose and a follow-up window tied to treatment discontinuation or new anti-cancer treatment, with a maximum of 10 months. At the same time, it does not report participant-level follow-up, missing-data handling, or a detailed statistical plan. Those omissions limit assessment of precision and completeness.

Finally, the registry includes planning illustrations but describes them as distinct from the study's own design. Such illustrations can address hypothetical future-study requirements under stated assumptions; they do not add an observed comparison to this single-group trial and are not observed power. Reading the design and endpoint definitions together helps distinguish what the safety summaries document from what remains unanswered.

14. Limitations

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

The most direct limitation is that the numerical result for the primary endpoint, Grade 3-5 treatment-related adverse events, is not reported in the registry record. The serious and other adverse-event summaries do not resolve that omission because their definitions are not the same as the primary endpoint's stated criteria. The listed deaths also lack reported causes and cannot be attributed to the study treatment from this information.

The study is single-group and nonrandomized, with no concurrent control treatment. It therefore cannot provide a randomized estimate of how outcomes differ between the combination and another regimen. The enrollment is 6, limiting the amount of safety information recorded in this study. The registry's listed serious-event terms and percentages remain descriptive; the record does not provide confidence intervals or a statistical analysis that would quantify uncertainty.

The record does not report numerical results for the listed secondary outcomes, methods for missing data, details of participant-level follow-up, or a multiplicity strategy. It also does not report the trial's sample-size justification or a detailed statistical analysis plan. These omissions prevent a full assessment of outcome completeness, precision, and the planned interpretation of multiple outcomes. The adverse-event window is described, but individual observation durations and whether all participants completed the relevant follow-up are not reported in the registry record.

Finally, serious adverse-event terms should not be taken as interchangeable with Grade 3-5 treatment-related events. The registry does not provide enough detail to map each listed term to the primary endpoint's grade or treatment attribution. Interpretation should stay within the definitions and descriptive information recorded, without inferring comparative effects, causation, or unreported outcome results.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: A Study of Atezolizumab and Bevacizumab in Hepatocellular Carcinoma — NCT04829383.
• Registry record last updated 2026-06-16; status last verified 2026-06.
• Results first posted 2026-06-16.
• 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

The registry record describes NCT04829383, a completed Phase II, nonrandomized, single-group feasibility study of atezolizumab plus bevacizumab for unresectable hepatocellular carcinoma in patients with Child-Pugh B7 or B8 liver disease and no prior systemic therapy. Enrollment is 6. The listed study dates are 2021-03-22 for start and 2024-06-19 for primary completion.

The primary outcome is the number of participants with Grade 3-5 treatment-related adverse events, classified by CTCAE v5 term and grade. Its numerical result is not reported in the registry record. The safety summary reports 4 of 6 affected by serious adverse events, 6 of 6 affected by other adverse events, and 4 deaths; causes of death are not reported. Listed secondary outcomes include response, disease control, response duration, overall survival and progression-free survival, with numerical results not reported in the registry record. The record lists no statistical analyses.