A Pilot Study of Combined Immune Checkpoint Inhibition in Combination With Ablative Therapies in Subjects With Hepatocellular Carcinoma (HCC) or Biliary Tract Carcinomas (BTC): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Pilot Study of Combined Immune Checkpoint Inhibition in Combination With Ablative Therapies in Subjects With Hepatocellular Carcinoma (HCC) or Biliary Tract Carcinomas (BTC) (NCT02821754), a non-randomized phase 2 study. The primary outcome is 6 Month Progression Free Survival (PFS).

2. Clinical Question

3. Trial Design

This was a completed phase 2 study sponsored by the National Cancer Institute. The registry lists 54 enrolled participants, a non-randomized allocation, no masking and a parallel model. The recorded study start was 2016-07-05, and primary completion was 2021-02-17. These design details describe how the trial was organized; they do not establish that every enrolled participant contributed data to every outcome.

The four arm descriptions distinguish durvalumab plus tremelimumab alone from the same drug combination with TACE, RFA or cryoablation. The brief summary states that the HCC objective included the drugs alone or with cryoablation, TACE or RFA, while the BTC objective included the drugs alone or with cryoablation or RFA. The record does not report the number of participants assigned to each arm, or how the disease types were distributed across arms.

Because treatment allocation was non-randomized, comparisons between regimens are not equivalent to randomized comparisons. Differences in observed outcomes could reflect differences in the participants or clinical circumstances as well as differences between regimens. The registry record does not report the allocation criteria, any adjustment strategy, or a statistical plan that would address this issue.

The primary outcome was six-month progression-free survival, described as median time in months. The record also lists the number of grades 1–5 adverse events related to tremelimumab and durvalumab as a secondary outcome. It records three outcome measures but no analysis entries. Numerical primary-outcome results and the analysis population 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.

Before assignment: Two participants were enrolled but not treated. With the exception of baseline data, no other data were collected for these participants.

Overall Study

Reasons for not completing this period:

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 6 Month Progression Free Survival (PFS), measured over At 6 months.

Progression-free survival (PFS) is a time-to-event outcome: it tracks time without disease progression, with the relevant event defined in the registry description by assessment using the Response Evaluation Criteria in Solid Tumors (RECIST). Progression is defined as the appearance of one or more new lesions and/or unequivocal progression of existing non-target lesions. The stated endpoint is median PFS in months, assessed at six months.

In general, a median summarizes the midpoint of an outcome's time distribution. For time-to-event data, it can describe when the estimated proportion remaining free of the event reaches the midpoint, provided the data support that estimate. It is not the same as the proportion of participants who are progression-free at a particular time, nor does it describe every participant's experience. Censoring and the method used to estimate the survival curve matter when interpreting such a median.

PFS is used in oncology trials to characterize how long participants remain alive without documented progression. It can capture progression events before death, rather than focusing only on survival. That makes it a different outcome from overall survival and from a response rate. The registry record does not report the numerical median, a range of observed values, a survival estimate at six months, or details about censoring and endpoint assessment. The primary endpoint's result is therefore not available for quantitative interpretation from this record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does a median PFS represent?

In general, a median is a measure of the center of a distribution. For time-to-event outcomes, the median is interpreted through the estimated probability of remaining free of the event over time. Here, the event definition is based on RECIST-assessed progression: new lesions and/or unequivocal progression of existing non-target lesions. The endpoint is described as median PFS in months, with a timeframe of at six months.

Time-to-event analysis can use information from participants whose event has not been observed by the time their follow-up ends. Such observations are commonly treated as censored: they contribute information up to their last known event-free time, rather than being treated as if an event occurred then. This is a general explanation of survival analysis. The registry record does not report censoring rules, follow-up completeness, or the method used to estimate the median. It also does not provide a numerical median or confidence interval.

Why does non-randomized allocation matter?

In general, random assignment helps balance measured and unmeasured characteristics across treatment groups, on average. Non-randomized allocation does not provide that protection. Participants receiving different regimens may differ in disease characteristics, prior care, suitability for a procedure or other factors related to prognosis. Those factors could influence PFS as well as the regimen received.

Therefore, a difference between observed outcomes in these arms would be an association, not automatically a causal treatment effect. Statistical adjustment can sometimes address measured differences, but it cannot guarantee control of unmeasured confounding. The registry record does not report why particular regimens were assigned, baseline balance, covariate adjustment, or an analysis plan for comparisons. No randomized, like-for-like control comparison is described.

What can a confidence interval add?

Generally, a confidence interval expresses uncertainty around an estimate under the assumptions of the statistical method. A narrower interval indicates greater statistical precision than a wider interval, while neither interval alone establishes that an estimate is unbiased or clinically important. For a time-to-event endpoint, interval estimates can communicate uncertainty about a median, a survival probability or a relative event rate, depending on the analysis.

No confidence interval or numerical PFS estimate is reported in the registry record. The record also contains no analysis entries. Thus there is no registry-reported interval from which to assess precision, and no sponsor-reported hazard ratio in the analysis information. A hazard ratio, in general, compares event rates over follow-up under the assumptions of a time-to-event model; it is not a direct measure of the difference between median times.

How should sample-size planning be read?

General sample-size illustrations estimate the amount of information a future study might need under stated assumptions. The planning assumptions associated with the posted illustrations are a two-sided significance level of 0.05, equal allocation between the two arms and a normal approximation; the effect to detect is expressed as a fraction of the observed effect. For time-to-event outcomes, the stated approach gives a total number of events. These are planning illustrations based on an observed effect, not the trial's own design or evidence of the achieved power of this study.

Sample-size planning depends on assumptions about the target effect and the outcome's variability or event rate. If those assumptions are uncertain, the resulting planning figure is uncertain as well. A planning illustration does not show that a study was adequately powered, nor does a study's result become more or less credible merely because a planning number is displayed. The trial-specific sample-size justification and design assumptions are not reported in the registry record. Numerical planning outputs are not reported in the registry record's details described for this analysis.

What does the safety summary measure?

The registry lists a secondary outcome concerning grades 1–5 adverse events related to tremelimumab and durvalumab, and it also records serious-event summaries. Adverse events can be counted as participants with an event, as event episodes, or by severity and attribution categories. Those measures answer different questions. A count of affected participants is not interchangeable with a count of all event episodes.

In general, a percentage needs a clear denominator and definition to be interpreted. Comparisons also need comparable follow-up and event ascertainment, especially when treatment groups differ in size or time under observation. The registry record does not report all such details for the serious-event terms. The recorded figures should therefore be read as descriptive safety information, not as adjusted comparisons or proof of causation.

10. Multiplicity, Missing Data and Analysis Populations

Multiplicity arises when a study evaluates several outcomes, groups, time points or statistical comparisons. In general, examining many hypotheses can increase the chance of at least one apparently notable result by chance unless the analysis plan accounts for that process. The registry lists three outcome measures and four treatment-arm descriptions, but no analysis entries. It does not report which outcomes were treated as confirmatory, how comparisons were selected, or whether any multiplicity adjustment was planned. The record therefore does not support a claim that a particular approach to multiple comparisons was used.

Missing-data handling is also important for a time-to-event endpoint. Participants may have incomplete follow-up, and their last known progression-free time can affect how survival estimates are constructed. This is a general consideration, not a description of this trial's procedures. Rules for censoring, missing assessments, withdrawals and incomplete outcome data are not reported in the registry record.

The registry record does not report the primary analysis population, whether all enrolled participants were included, or whether any participants were excluded from outcome analyses. Enrollment is recorded as 54, but this does not establish the number analyzed for PFS or safety. Group-specific enrollment and analysis counts are not reported in the registry record.

These omissions limit assessment of how robust an analysis would be to alternative outcome definitions, missing-data assumptions or population choices. They also prevent evaluation of whether reported safety counts and percentages used consistent denominators. No statistical analysis plan or detailed analysis population is reported in 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 0% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: aspartate aminotransferase increased (18.8%), fever (18.8%), abdominal pain (12.5%), blood bilirubin increased (12.5%), pleural effusion (12.5%).

The registry's serious adverse-event summary separates entries for Radiofrequency Ablation/Trans-arterial Catheter Chemoembolization (RFA/TACE) and Radiofrequency Ablation/Cryoablation (RFA/CA). For RFA/TACE, the summary records 36 at risk and 20 affected for serious events, 36 at risk and 36 affected for other events, and 26 deaths. For RFA/CA, it records 16 at risk and 7 affected for serious events, 16 at risk and 16 affected for other events, and 15 deaths. These are recorded counts; the registry record does not explain all definitions or analysis details behind these fields.

The listed serious-event terms and percentages are aspartate aminotransferase increased, 18.8%; fever, 18.8%; abdominal pain, 12.5%; blood bilirubin increased, 12.5%; pleural effusion, 12.5%; and sepsis, 12.5%. These are descriptive entries. The record does not provide enough detail to infer that an event was caused by treatment, compare event rates reliably across all four arms, or determine how many event episodes occurred. The percentages should not be interpreted without their precise denominator and event-counting definitions, which are not reported in the registry record.

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

The record documents a completed phase 2 study of durvalumab and tremelimumab, with or without TACE, RFA or cryoablation, in advanced HCC or BTC. It identifies six-month PFS as the primary outcome and defines progression using RECIST. It also records adverse-event information, including serious-event summaries and a secondary outcome concerning grades 1–5 events related to the two drugs.

Those details describe the study's scope and the outcomes it intended to evaluate. They do not provide a numerical primary-endpoint result. In particular, the registry record does not report median PFS, an estimate of PFS at six months, uncertainty intervals, or a treatment comparison. Without these results, the magnitude and precision of any efficacy finding cannot be assessed from the record.

The study's non-randomized allocation further limits causal interpretation. Even if group-specific outcome estimates were available, differences might reflect participant or disease characteristics associated with regimen assignment. The registry record does not report allocation criteria or adjustment methods that would clarify how such differences were handled.

The safety entries can describe reported event counts and listed serious-event terms, but they do not establish causation. Nor do they alone provide a complete basis for comparing safety among regimens: denominator definitions, follow-up, event ascertainment and detailed attribution are not reported in the registry record.

Accordingly, the posted information is useful for understanding the question, design and selected safety reporting, but it does not establish whether the regimens changed PFS or provide a quantitative estimate of benefit or harm. No medical recommendation follows from these registry entries.

13. Why This Trial Matters Statistically

This study raises a statistical challenge common to early clinical research: evaluating a complex treatment strategy across clinically distinct cancers and several regimens. The record includes HCC and BTC, the immune checkpoint inhibitor combination, and ablative procedures. A single overall summary could conceal differences among disease groups or treatment combinations, while subgroup comparisons may be imprecise. The registry record does not report subgroup estimates or the planned approach to those comparisons.

The non-randomized design also makes the interpretation of comparisons especially important. Treatment groups may differ for reasons connected to both the selected regimen and prognosis. General statistical methods can adjust for measured characteristics, but they cannot remove all concern about unmeasured differences. Knowing the allocation criteria, baseline characteristics and adjustment plan would help readers judge how far a comparison can support causal interpretation; those details are not reported in the registry record.

The endpoint adds another layer: PFS is a time-to-event measure, so interpretation depends on event definitions, follow-up and handling of censored observations. A reported median without uncertainty or analysis details would be incomplete; in this record, the numerical endpoint result itself is not reported. The safety summary likewise illustrates why counts require clear denominators and definitions before they can be compared.

These issues make transparent reporting essential. A reader needs group-specific outcome estimates, precision measures, analysis populations and methods to understand what the data show. The registry record documents a completed study and its intended endpoint, but its available analysis information is limited. That distinction between a study question and an interpretable treatment-effect estimate is statistically important.

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 record does not report a numerical result for the primary PFS endpoint, confidence intervals, group-specific PFS estimates or a sponsor-reported hazard ratio. No analysis entries are listed. Consequently, the efficacy result cannot be evaluated quantitatively from the registry record.

Non-randomized allocation limits causal comparisons, and the registry record does not report the reasons for assigning regimens, baseline characteristics by arm, or adjustment methods. It also does not report arm-specific enrollment or the population analyzed for each outcome. The recorded enrollment of 54 should not be treated as an analysis count.

Additional endpoint details are absent, including the method for estimating median PFS, censoring rules, follow-up completeness, handling of missing assessments and any independent review procedures. The statistical analysis plan and sample-size justification are not reported in the registry record. The planning assumptions associated with sample-size illustrations are not evidence of the trial's original design or achieved power.

Safety interpretation is limited by the lack of detailed denominator definitions, event-collection methods and information about follow-up and attribution. The recorded serious-event counts and percentages are descriptive and do not establish causation. The registry record also does not describe how multiple outcomes or comparisons were handled. These limitations restrict both efficacy and safety comparisons among the regimens.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: A Pilot Study of Combined Immune Checkpoint Inhibition in Combination With Ablative Therapies in Subjects With Hepatocellular Carcinoma (HCC) or Biliary Tract Carcinomas (BTC) — NCT02821754.
• Registry record last updated 2023-03-28; status last verified 2023-03.
• Results first posted 2022-03-24.
• 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

NCT02821754 is a completed phase 2, non-randomized, parallel study sponsored by the National Cancer Institute. It enrolled 54 participants and evaluated durvalumab plus tremelimumab, alone or with TACE, RFA or cryoablation, in advanced HCC or BTC. The recorded study start was 2016-07-05, and primary completion was 2021-02-17.

The primary outcome was median PFS in months, with progression assessed using RECIST. The registry record does not report a numerical primary-outcome result or analysis entries. It also records a secondary adverse-event outcome and serious-event summaries, which are descriptive and do not establish treatment causation. Non-randomized allocation and missing analysis details constrain interpretation of comparisons between regimens.