A Study of Pembrolizumab (MK-3475) in Relapsed or Refractory Classical Hodgkin's Lymphoma (rrcHL) or Relapsed or Refractory Primary Mediastinal Large B-cell Lymphoma (rrPMBCL) (MK-3475-B68): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study of Pembrolizumab (MK-3475) in Relapsed or Refractory Classical Hodgkin's Lymphoma (rrcHL) or Relapsed or Refractory Primary Mediastinal Large B-cell Lymphoma (rrPMBCL) (MK-3475-B68) (NCT04875195), a na phase 2 study. The primary outcome is Objective Response Rate (ORR) Per Lugano Classification as Assessed by Investigator.

2. Clinical Question

3. Trial Design

This was a Phase 2, open-label, single-group study of pembrolizumab in participants with relapsed or refractory classical Hodgkin's lymphoma or primary mediastinal large B-cell lymphoma. The registry record lists no random allocation and no masking. Pembrolizumab was administered by intravenous infusion at 400 mg every six weeks, starting on Day 1, for up to 18 doses.

The two disease groups were analyzed as separate cohorts. The study's primary objective was to evaluate objective response rate by cohort, based on investigator assessment under the 2014 Lugano classification criteria. The stated response-assessment timeframe was up to approximately 30 months. For this primary endpoint, the analysis population included participants who received at least one dose of study treatment.

The registry record lists 66 participants enrolled. The safety summaries identify 60 participants in the rrcHL cohort and 6 in the rrPMBCL cohort. These cohort counts describe a non-randomized study and should not be read as a treatment-versus-control comparison. Because there was no control group, changes in outcomes cannot be compared with outcomes under another assigned treatment within this study.

The registry record gives a study start date of 2021-06-07, a primary completion date of 2023-12-14, and a completed status. Secondary outcomes included central-review objective response, duration of response assessed by investigators and by blinded independent central review, and pembrolizumab pharmacokinetic measures. The record lists 13 outcome measures, but does not report a statistical analysis plan or outcome analysis results.

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: The analysis population consisted of all participants who received ≥1 dose of study treatment.

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) Per Lugano Classification as Assessed by Investigator, measured over Up to approximately 30 months.

The primary endpoint was objective response rate (ORR), defined as the percentage of participants with a complete response or partial response. Investigators assessed response using the 2014 Lugano classification criteria, which combine anatomic imaging by computed tomography with metabolic imaging by FDG-PET. The primary assessment was reported separately for the rrcHL and rrPMBCL cohorts, with a stated timeframe of up to approximately 30 months.

A complete response requires complete metabolic response, meaning no or minimal FDG uptake, along with radiologic response. Target lesions must regress to at most 1.5 cm in longest transverse diameter, and there must be no new lesions. A partial response requires partial metabolic response, described as moderate or high FDG uptake, and radiologic response. The radiologic criteria include at least a 50% decrease in the sum of products of diameters across multiple lesions, no increase in lesions, and spleen regression exceeding 50% in length beyond normal.

ORR is used to describe the proportion of the analyzed group meeting a defined response threshold. It is a binary outcome at the participant level: each participant is classified as having a complete or partial response, or not meeting either response category. An estimate of that proportion would describe the observed response frequency, while a confidence interval would convey its statistical uncertainty. The registry record does not report the primary ORR estimates or confidence intervals for either cohort.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry record describes the clinical endpoint and study structure, but does not report the analysis methods in enough detail to reconstruct the primary analysis. The following points explain the statistical ideas relevant to interpreting the stated endpoint; general statistical explanations are identified as such.

What does an objective response proportion represent?

ORR is a proportion: the numerator is the number of participants classified as having a complete or partial response, and the denominator is the analysis population used for the endpoint. The registry defines that population as all participants who received at least one dose of study treatment. The record does not provide the response counts or cohort-specific denominators for the primary efficacy analysis, so the proportion cannot be recovered from the registry information.

In general, a proportion summarizes how frequently a defined event or status occurs in an analyzed group. Its interpretation depends on the outcome definition, the population included, and how observations that cannot be classified are handled. Those details matter because changing the denominator or response-evaluation rules can change the estimate. The endpoint definition is stated, but the registry does not report how missing or unevaluable response assessments were handled.

What would a confidence interval add?

The primary endpoint specifies a 95% confidence interval. In general, a confidence interval communicates uncertainty around an estimate under the statistical method used to construct it. It is not a guarantee that the true value lies within the particular interval, nor does it resolve limitations in study design or measurement. A narrower interval usually indicates greater statistical precision under the method's assumptions; a wider interval indicates more uncertainty. Since neither the ORR estimates nor their intervals are reported in the registry record, precision cannot be judged for either cohort.

The cohort sizes differ, as reflected in the safety summaries: 60 participants at risk for rrcHL and 6 for rrPMBCL. In general, estimates based on fewer observations tend to be less precise, all else being equal. That principle does not supply a numerical interval or establish the actual precision. The primary analysis denominators and response counts are not reported, so even those descriptive quantities cannot be confirmed for the efficacy analysis.

What does the single-group design allow?

A single-group design can describe outcomes observed among participants receiving the study treatment. It does not include a concurrently randomized control group for estimating a between-group treatment contrast. Therefore, an observed response proportion, if reported, would not by itself distinguish the contribution of treatment from other influences on outcomes. Those influences may include the natural course of disease, participant selection, and assessment processes; this is a general limitation of uncontrolled designs, not a claim about what occurred in this study.

The registry lists allocation as not applicable and masking as none. These design features are consistent with the stated single-group structure. The endpoint was assessed by investigators, and the study also listed central-review ORR as a secondary outcome. The record does not report whether, or how, investigator and central-review assessments agreed. Nor does it report operational details such as the timing and frequency of assessments beyond the primary timeframe.

How should the safety counts be understood?

The safety summaries give numbers at risk and numbers affected for serious and other adverse events, separately by cohort. These are descriptive summaries, not estimates of comparative risk. General statistical interpretation of event proportions depends on the event definitions, follow-up time, and whether participants could experience repeated events. The record does not provide event-level timing, recurrent-event counts, exposure-adjusted rates, or attribution details in the information reported for these summaries.

The serious adverse event terms are listed with percentages of 1.7% each. Those percentages describe the registry's presentation of the named terms; they do not establish that the terms exhaust all serious events or that events were caused by pembrolizumab. The record gives a separate serious-event affected-participant count by cohort, but does not describe the method used to generate the term-level percentages. No denominator should be inferred for those percentages beyond what is explicitly reported.

What is the role of planning calculations?

General sample-size calculations estimate how many participants might be needed to detect a chosen effect under stated assumptions. They are not observed trial outcomes and are not observed power. The registry record does not report a trial-specific sample-size justification. The planning assumptions associated with the study calculations describe a two-arm illustration, with a two-sided significance level of 0.05, equal allocation, a normal approximation, and effects expressed as fractions of an observed effect. Such assumptions are not the study's single-group design and should not be confused with its actual enrollment or analysis.

The record does not report the numerical planning illustrations or the observed effect used for them. Without those results, there is no basis to state a required sample size or connect a hypothetical planning calculation to the completed study. In general, planning calculations depend on the target effect, variability or event frequency, significance criterion, allocation, and analysis method; changing those assumptions can change the calculated requirement.

Finally, the registry record lists no analysis results. It does not state the statistical method used to calculate the confidence interval, any adjustment for covariates, or any sensitivity analyses. Those omissions limit reproducibility: the endpoint is defined, but the estimate, uncertainty, and analytic decisions needed to evaluate it are not reported.

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("Relapsed or Refractory Classical Hodgkin's Lymphom", "Relapsed or Refractory Primary Mediastinal Large B")
n      <- c(60, 6)
events <- c(40, 3)
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  Relapsed or Refractory Classical Hodgkin's Lymphom  60  40  66.7  53.3  78.3  54.1  77.3  54.0  77.3
ARM  Relapsed or Refractory Primary Mediastinal Large B  6  3  50.0  11.8  88.2  18.8  81.2  18.4  81.6

11. Multiplicity, Missing Data and Analysis Populations

The registry record lists a primary objective of evaluating ORR by cohort and 13 outcome measures. Secondary outcomes include central-review ORR, duration of response by investigator and central review, and pharmacokinetic measures. The record does not report a multiplicity plan describing how statistical error across outcomes, cohorts, or repeated analyses was controlled, nor does it report a hierarchy or testing sequence. Without that information, the role of any secondary analyses in confirmatory inference cannot be determined. No analysis results are listed in the registry record.

The primary analysis population is defined as participants who received at least one dose of study treatment. The record does not explain how participants with missing, incomplete, or unevaluable response assessments were handled, whether follow-up was censored for any reason, or how missing safety information was treated. It also does not report the number of participants with evaluable primary endpoint data. These details affect how an observed proportion is interpreted because the analysis denominator and outcome classification rules determine which participants contribute to the estimate.

The record does not report whether sensitivity analyses were planned or conducted to assess the impact of missing data, alternative response classifications, or differences between investigator and central review. It also does not report interim analyses or stopping rules. These methods are not inferable from the study's phase, endpoint, or completed status. Consequently, readers cannot assess the robustness of an efficacy estimate or the handling of multiple outcomes from the registry information available.

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 5% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: febrile neutropenia (1.7%), gastric stenosis (1.7%), haemophagocytic lymphohistiocytosis (1.7%), peritonitis (1.7%), pneumonia (1.7%).

The registry reports serious adverse events by cohort, giving participants at risk and participants affected. In the rrcHL cohort, 60 participants were at risk and 6 were affected by serious adverse events. In the rrPMBCL cohort, 6 were at risk and 0 were affected. For other adverse events, the corresponding counts were 37 of 60 in rrcHL and 4 of 6 in rrPMBCL. Deaths were reported for 15 participants in rrcHL and 3 in rrPMBCL.

The record lists febrile neutropenia, gastric stenosis, haemophagocytic lymphohistiocytosis, peritonitis, pneumonia, and urinary tract infection as serious adverse event terms, each reported as 1.7%. These term-level entries should not be confused with the number of participants who experienced any serious adverse event. The record does not state whether each event was considered related to treatment, provide timing or duration, or describe causes of death. The counts are descriptive and do not establish comparative risk or causation.

The cohorts differ substantially in size, so direct comparison of their counts is not informative by itself. The registry does not report exposure-adjusted rates or fuller event-level detail in these summaries. These figures should therefore be read as selected safety descriptions, not as a complete account of tolerability.

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

The study addresses a defined statistical question: the proportion of participants in each lymphoma cohort who had a complete or partial response under investigator assessment using the 2014 Lugano criteria. The treatment schedule and endpoint definition are specified, and the record identifies the treated population for the primary analysis. However, the primary response estimates and their 95% confidence intervals are not reported in the registry record. Without those quantities, the observed frequency of response and its precision cannot be evaluated from this record.

Even if a cohort-specific ORR were available, an uncontrolled study would describe outcomes in participants receiving pembrolizumab rather than establish a contrast against another treatment. The study did not randomize participants to a control group. Consequently, a response proportion alone could not quantify how outcomes would differ under an alternative treatment or separate treatment effects from other factors. This is a general interpretive limitation of single-group studies, not a statement about the unreported efficacy result.

The safety counts describe participants affected by serious and other adverse events, as well as deaths, within each cohort. They do not by themselves establish causation, comparative safety, or the circumstances surrounding an event. The registry's term-level serious event listings add descriptive detail, but the record does not provide enough information to assess event timing, attribution, or rates relative to exposure.

The record can therefore support understanding of the study's design, dose schedule, population definition, endpoint criteria, and selected adverse-event summaries. It cannot support a quantitative conclusion about the primary response outcome, its statistical precision, or comparative treatment effects. The missing results and methods also prevent a full assessment of analysis choices, missing data, and the relationship between investigator and central-review assessments.

14. Why This Trial Matters Statistically

This study is statistically notable for asking the same broad response question in two distinct lymphoma cohorts while using a single-group design. Reporting cohort-specific estimates would be important because the populations are defined by different diagnoses and the cohort sizes differ. The registry's primary objective explicitly distinguishes rrcHL and rrPMBCL, but the primary estimates and confidence intervals are not reported in the registry record. That gap prevents statistical evaluation of the central endpoint, regardless of how clearly the endpoint itself is defined.

The design also illustrates the difference between describing an outcome and estimating a comparative effect. An investigator-assessed response proportion could describe the observed experience of treated participants, but without a control group it would not provide a randomized contrast. This distinction is central when interpreting outcomes from a single-group study and when considering what additional evidence would be needed to compare treatment strategies.

The study lists central-review response and duration-of-response outcomes alongside pharmacokinetic measures. Those outcomes address different statistical questions: response confirmation, time spent in response, and drug concentration over time. Their presence in the registry does not supply their results, and the record does not report how analyses across multiple outcomes were handled. Clear reporting of estimates, uncertainty, analysis populations, and missing-data methods is needed to make these different outcome types interpretable.

The reported safety summaries further show why event counts need context. Cohort-specific counts are descriptive, and the smaller rrPMBCL group limits the amount of information those summaries can convey. Statistical interpretation depends not only on counts but also on definitions, follow-up, exposure, and attribution. The record's selected safety figures do not provide all of those details.

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 principal limitation for statistical interpretation is that the registry record does not report the primary ORR estimates or confidence intervals. Although the endpoint, assessment criteria, and analysis population are described, the response counts and efficacy denominators are absent. The magnitude and precision of the primary outcome therefore cannot be assessed.

The study was single group, with no randomized comparator, and masking was listed as none. These features limit comparative interpretation and leave the investigator assessment open to assessment effects that masking might otherwise help address. A secondary central-review response outcome is listed, but its results are not reported, so it cannot be used to evaluate agreement with investigator assessment.

The registry does not report a trial-specific sample-size justification, detailed statistical analysis plan, multiplicity strategy, interim analyses, or sensitivity analyses. It also does not describe how missing or unevaluable response assessments were handled, or report the efficacy analysis denominator. These omissions prevent assessment of whether the analysis choices were prespecified and how robust any estimate would be to alternative assumptions.

The safety information is limited to selected group counts and listed serious adverse event terms. The record does not provide event timing, treatment attribution, exposure-adjusted rates, or causes of death in these summaries. The cohorts differ in size, and the smaller cohort offers limited descriptive information. The listed counts should not be interpreted as complete safety rates or as evidence of causation.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: A Study of Pembrolizumab (MK-3475) in Relapsed or Refractory Classical Hodgkin's Lymphoma (rrcHL) or Relapsed or Refractory Primary Mediastinal Large B-cell Lymphoma (rrPMBCL) (MK-3475-B68) — NCT04875195.
• Registry record last updated 2026-10-27; status last verified 2026-10.
• Results first posted 2025-02-10.
• 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

NCT04875195 is a completed Phase 2, open-label, single-group study of pembrolizumab in relapsed or refractory classical Hodgkin's lymphoma and relapsed or refractory primary mediastinal large B-cell lymphoma. The regimen was 400 mg by intravenous infusion every six weeks, beginning on Day 1, for up to 18 doses. The registry lists 66 participants enrolled and gives separate cohort summaries for rrcHL and rrPMBCL.

The primary endpoint was investigator-assessed ORR under the 2014 Lugano classification criteria, with a stated timeframe of up to approximately 30 months. The analysis population was participants who received at least one dose. The primary ORR estimates and 95% confidence intervals are not reported in the registry record. Selected safety counts and secondary outcomes are listed, but the record does not provide a statistical analysis plan, multiplicity details, or missing-data methods.