Pembrolizumab (Keytruda) in Advanced Hepatocellular Carcinoma: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Pembrolizumab (Keytruda) in Advanced Hepatocellular Carcinoma (NCT02658019), a na phase 2 study. The primary outcome is Disease Control Rate (DCR) in Study Participants.

2. Clinical Question

The study asks whether disease control was observed during treatment with pembrolizumab among participants with advanced, unresectable hepatocellular carcinoma. The primary endpoint was the disease control rate (DCR), defined using RECIST 1.1 as the percentage of participants whose best overall response was complete response, partial response or stable disease maintained for at least 8 weeks.

This is a question about the proportion meeting a prespecified response-based definition within the treated group. Without a comparator, the study cannot directly estimate how outcomes would have differed under another treatment or under no study treatment. The registry lists progression-free survival, overall survival, objective response rate and duration of response as secondary outcomes. Their results are not reported in the registry record.

3. Trial Design

The registry describes a phase II, single-group study of pembrolizumab in advanced hepatocellular carcinoma. It lists an enrollment of 29 and records the study as completed. The study start date is 2016-05-06, and the primary completion date is 2019-11-10. The sponsor is listed as Lynn Feun, MD.

Allocation is recorded as not applicable, consistent with the absence of a randomized comparison, and masking is listed as none. All participants were assigned to the pembrolizumab group described in the registry. The treatment schedule specifies intravenous pembrolizumab at 200 mg on day 1 of each 3-week cycle. Therapy was described as continuing until consent withdrawal, disease progression and/or unacceptable toxicity, whichever occurred first. These entries describe the planned treatment and stopping conditions; they do not establish how long any individual participant was treated.

The primary objective in the registry summary is to assess therapeutic efficacy in participants with unresectable hepatocellular carcinoma. The primary outcome is DCR at 8 weeks, defined by response categories assessed according to RECIST 1.1. The registry also names progression-free survival, overall survival, objective response rate and duration of response as secondary outcomes. It records six outcome measures overall, but detailed outcome information beyond the listed entries is not reported in the registry record.

A single-group design can describe outcomes among participants who received the study treatment. It does not provide a concurrent control group for estimating a between-group treatment effect. Accordingly, any observed disease control proportion would need to be interpreted as a result within this study population, not as a randomized comparison. The registry record does not report participant flow, baseline characteristics, the number included in each outcome analysis, or the study’s analysis plan.

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:

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 Disease Control Rate (DCR) in Study Participants, measured over 8 weeks.

The primary endpoint is the disease control rate (DCR), expressed as the percentage of participants whose best overall response was complete response, partial response or stable disease maintained for at least 8 weeks. Response is defined using RECIST 1.1, with target lesions assessed by MRI or CT. The description defines complete response as disappearance of all target lesions and partial response as a decrease of at least 30% in the sum of the longest diameters of target lesions.

DCR combines tumor shrinkage categories with stable disease. In general, this kind of endpoint captures whether participants meet a defined threshold for response or disease stability over the specified assessment period. It is different from an endpoint that counts only complete or partial responses, because stable disease can also meet the DCR definition if maintained for the required duration. The time requirement is part of the endpoint: a response category alone is not sufficient unless it is maintained for at least 8 weeks.

The registry record does not report the number of participants meeting the DCR definition, the number assessed, the resulting percentage, or the stated 95% confidence interval. It also does not report details of assessment timing or handling of participants without an evaluable response. The endpoint definition therefore clarifies what was intended to be measured, but the posted record does not establish the observed result.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry identifies a percentage-based primary endpoint and names a 95% confidence interval as its dispersion measure. It does not report the statistical method used to estimate the DCR or construct that interval. The discussion below describes general statistical concepts; it does not assert that a particular method was used in this study.

What does a percentage endpoint represent?

For a binary outcome such as meeting or not meeting a response-based definition, a percentage summarizes the share of an analysis group that meets the definition. Its interpretation depends on both the count meeting the endpoint and the denominator included in the analysis. In this study, the registry states the endpoint definition and an enrollment of 29, but does not report the DCR count, analysis denominator or participant-level outcome status. Enrollment should not be assumed to equal the denominator used for the primary analysis.

DCR is a composite endpoint: a participant can meet it through complete response, partial response or stable disease maintained for at least 8 weeks. Such a combined measure can capture more than tumor shrinkage alone, but it also means that the overall percentage does not reveal the separate contribution of each response category. The registry record does not provide those category-specific counts.

What does a confidence interval add?

In general, a confidence interval gives a range of values compatible with the data under the interval’s statistical method. It communicates precision, not certainty that the true value lies within that particular interval. The interval’s width depends on features including the amount of information available and the method used. The primary outcome specifies a 95% confidence interval, but the interval estimate and its calculation method are not reported in the registry record. No conclusion about the precision of the primary result can therefore be drawn from the stated confidence level alone.

A confidence interval also does not resolve design limitations. Even a precise estimate of the proportion in a single treated group would not supply the missing randomized comparison. Precision about an observed within-group proportion and evidence about a comparative treatment effect are different statistical questions.

What can be inferred from the analysis record?

The registry record reports no analysis entries. It does not identify a statistical model, a hypothesis test, a p-value, a confidence-interval procedure, or a planned treatment-effect contrast. Those methods should not be inferred from the endpoint definition. The record also does not report whether any interim analyses were planned or performed.

The study’s sample-size assumptions are described as a planning illustration based on an observed effect, using a two-sided significance level of 0.05, equal allocation between two arms and a normal approximation. That framework is not the study’s design: this trial is recorded as single-group, with no randomized comparator. Any such illustration is hypothetical and is not observed power, nor does it show that the trial was designed with those assumptions. The registry does not report a study-specific sample-size justification.

In general, over-dispersion matters when the variability in observed outcomes exceeds what a statistical model assumes. If a model underestimates variability, its uncertainty estimates may be too narrow. The registry does not report an over-dispersion assessment or a model for the primary endpoint, so this issue cannot be evaluated for this study.

Why does the analysis population matter?

An analysis population defines which participants contribute to an outcome estimate. For a response endpoint, exclusions or missing assessments can affect both the numerator and denominator, depending on the prespecified rules. The registry record does not identify an intention-to-treat, per-protocol or other analysis population, and it does not state how missing or unevaluable response assessments were handled. Consequently, the population to which a reported DCR would apply cannot be established from this record.

In general, prespecified definitions and transparent reporting allow readers to distinguish the planned analysis from choices made after observing data. Here, the endpoint description is available, but the detailed analysis plan and results are not. That limits assessment of reproducibility, precision and potential analytic flexibility.

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("Pembrolizumab in Advanced HCC")
n      <- c(28)
events <- c(13)
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  Pembrolizumab in Advanced HCC  28  13  46.4  27.5  66.1  29.5  64.2  29.4  64.3

11. Multiplicity, Missing Data and Analysis Populations

The registry names a primary outcome, DCR, and secondary outcomes of progression-free survival, overall survival, objective response rate and duration of response. It records six outcome measures overall. The detailed information for any other measure, beyond the named outcomes, is not reported in the registry record.

When a study evaluates multiple outcomes or performs multiple statistical tests, the chance of at least one apparently notable result can increase. In general, a prespecified testing hierarchy or multiplicity adjustment can help define how such results should be interpreted. The registry record does not report a multiplicity procedure, a hierarchy among outcomes, or a statistical analysis plan. It does not report analysis entries or inferential results, so it is not possible to determine whether multiplicity was addressed in the study’s analyses.

The primary DCR timeframe is 8 weeks. Timeframes and analysis details for the named secondary outcomes are not reported in the registry record. Any additional outcome timeframes, assessment schedules or planned interim analyses are likewise not reported in the registry record.

The record does not describe the analysis population for any outcome. It also does not state how missing, incomplete or unevaluable data were handled, whether participants who stopped treatment remained in an outcome analysis, or how follow-up was treated for time-to-event outcomes. In general, missing-data rules can influence which observations contribute to an estimate and how uncertainty is represented. Without those rules and the relevant participant counts, the potential effect of missingness cannot be assessed.

These omissions do not establish that a particular procedure was or was not used outside the registry record. They mean that readers cannot verify the approach from the posted record. The reported enrollment of 29 should not be treated as the analysis denominator for every outcome.

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 0% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: blood bilirubin increased (6.9%), abdominal pain (3.4%), aspartate aminotransferase increased (3.4%), atrial fibrillation (3.4%), diarrhea (3.4%).

The registry lists 29 participants at risk for serious adverse events and 7 affected. For other adverse events, it lists 29 at risk and 28 affected. These are counts in separate registry categories; they do not describe the number of distinct events, their duration or severity beyond the category labels, or whether an event was judged related to pembrolizumab.

The listed serious adverse event terms and percentages are blood bilirubin increased, 6.9%; abdominal pain, 3.4%; aspartate aminotransferase increased, 3.4%; atrial fibrillation, 3.4%; diarrhea, 3.4%; and fatigue, 3.4%. The registry also records 9 deaths. The timing of those deaths, their causes and their relationship to treatment are not reported in the registry record.

In general, adverse-event counts describe events recorded under the applicable reporting categories; they do not establish causation. A participant may experience more than one event, so counts of event terms should not be interpreted as counts of unique participants unless the record explicitly says so. The record does not provide enough detail to determine whether the listed term percentages refer to unique participants, event occurrences or a particular analysis population. Nor does it describe exposure duration or the observation period used for the safety summaries.

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

The record establishes that a phase II, single-group study of pembrolizumab in advanced, unresectable hepatocellular carcinoma was completed, with enrollment listed as 29. It also defines the primary outcome as DCR at 8 weeks and describes the response categories used to determine whether a participant met that endpoint.

It does not report the DCR estimate, the number meeting the definition, the analysis denominator or the stated 95% confidence interval. Thus, the posted record cannot establish how often disease control was observed or how precisely that proportion was estimated. The listed secondary outcomes also have no reported results in this record. Completion status is not a substitute for outcome results.

Even if a within-group DCR result were available, the single-group design would not provide a randomized comparison. It could describe outcomes among participants in the study, but could not by itself distinguish the contribution of treatment from other influences on those outcomes. The absence of masking is recorded, but the registry does not report assessment procedures that would allow its implications for response classification to be evaluated.

The safety entries document affected-participant counts and selected serious adverse event terms, as well as deaths. They do not, on their own, establish that treatment caused any event or death. The registry record does not supply the timing, attribution or detailed context needed for those judgments.

These conclusions are limited to what the registry record reports. They do not establish whether further results were published or recorded elsewhere. The statistical meaning of the study’s efficacy and safety findings cannot be fully assessed from this record without outcome estimates, uncertainty measures, analysis populations and methods.

14. Why This Trial Matters Statistically

This trial is statistically informative as an example of the distinction between a clearly defined endpoint and a reported endpoint result. The registry specifies DCR, its response categories, the use of RECIST 1.1 and the requirement that stable disease be maintained for at least 8 weeks. Those details make the intended measure interpretable. But a definition alone does not show how many participants met it, what denominator was analyzed or how uncertain the estimate was.

The single-group design also makes the estimand important. A disease-control proportion in the treated group would answer a descriptive question about outcomes in that group. It would not be a randomized estimate of a contrast with a control treatment. Statistical precision in a single group cannot replace a comparator when the question concerns between-group differences.

Enrollment of 29 gives context for the study’s scale, but the registry does not report the primary analysis denominator or the DCR result. In general, a limited number of observations can leave an outcome proportion imprecisely estimated; the actual precision must be judged from the reported interval, which is absent. This illustrates why a point estimate and its uncertainty should be read together.

The record also illustrates the importance of reporting analysis populations, missing-data rules, multiplicity procedures and statistical methods. Without these details, readers cannot fully evaluate whether the endpoint was analyzed as intended or how secondary outcomes should be interpreted. The listed safety counts likewise require context about attribution, timing and exposure to support a fuller statistical reading.

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 main limitation for interpreting efficacy is that the registry record does not report the primary DCR result or its 95% confidence interval. It gives no count of participants meeting the endpoint and no analysis denominator. The enrollment of 29 therefore cannot be assumed to represent the number included in the primary analysis.

The study has no control group: it is recorded as single-group, with allocation not applicable and no masking. A within-group outcome, if reported, would not directly estimate a comparative treatment effect. The registry does not report baseline participant characteristics, detailed eligibility criteria or participant flow, limiting assessment of whom the results would describe.

The record does not specify the statistical analysis method, the analysis population, handling of missing or unevaluable assessments, or a multiplicity plan. It records no analysis entries. The primary endpoint definition is available, but response assessment schedules and the handling of participants who discontinue or lack an assessment are not reported in the registry record. The named secondary outcomes lack reported results and detailed analysis information.

For safety, the record lists affected counts, selected serious adverse event terms and deaths, but does not report attribution, timing, exposure duration or detailed event-level context. These limitations restrict interpretation of the posted safety information and prevent causal conclusions from the counts alone.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Pembrolizumab (Keytruda) in Advanced Hepatocellular Carcinoma — NCT02658019.
• Registry record last updated 2020-11-12; status last verified 2020-10.
• Results first posted 2020-11-12.
• 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

The registry describes a completed phase II, single-group study of pembrolizumab in advanced, unresectable hepatocellular carcinoma. Enrollment is listed as 29. The primary outcome is DCR at 8 weeks, based on complete response, partial response or stable disease maintained for at least 8 weeks under RECIST 1.1. Pembrolizumab is described as 200 mg intravenously on day 1 of each 3-week cycle.

The registry lists progression-free survival, overall survival, objective response rate and duration of response as secondary outcomes. It records serious and other adverse-event counts and 9 deaths. The primary efficacy estimate, its confidence interval, secondary outcome results, analysis methods, analysis populations and missing-data procedures are not reported in the registry record.