Pembrolizumab With Rituximab or Obinutuzumab in Treating Patients With Relapsed or Refractory Follicular Lymphoma or Diffuse Large B Cell Lymphoma: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Pembrolizumab With Rituximab or Obinutuzumab in Treating Patients With Relapsed or Refractory Follicular Lymphoma or Diffuse Large B Cell Lymphoma (NCT03401853), a non-randomized phase 2 study. The primary outcome is Overall Response Rate.

2. Clinical Question

The clinical question concerns whether response is observed when pembrolizumab is administered with rituximab or obinutuzumab in these lymphoma groups. The combinations are not compared across an equivalent randomized population: the rituximab arms include different lymphoma diagnoses, and the obinutuzumab arm is described for FL. Differences between arm results, if available, could therefore reflect differences in diagnosis or participant characteristics as well as treatment. The registry explicitly cautions against treating the arms as a like-for-like control comparison.

The study also lists serious or drug-related adverse events, progression-free survival (PFS), and overall survival (OS) as secondary outcomes. Their presence indicates the range of outcomes registered for assessment, but does not by itself establish that numerical results are available. The registry record does not report numerical efficacy results for the primary endpoint, and it contains no analysis records. The clinical question can therefore be described from the endpoint definition and assigned regimens, but the size and precision of any response estimate cannot be assessed from this record.

3. Trial Design

This is a phase II, non-randomized, open-label study with a parallel model. The registry lists the sponsor as the University of Washington and enrollment as 18. The recorded start date is 2018-03-23, the primary completion date is 2023-12-19, and the study status is terminated. The registry record does not explain the reason for termination.

The treatment groups are organized by lymphoma diagnosis and antibody combination. Arm I is described for patients with relapsed or refractory FL receiving pembrolizumab and rituximab. Arm II is described for patients with relapsed or refractory DLBCL receiving pembrolizumab and rituximab. Arm III is described for patients with relapsed or refractory FL receiving pembrolizumab and obinutuzumab. The record does not describe an obinutuzumab arm for DLBCL.

For induction, the arm descriptions state that pembrolizumab is given intravenously over 30 minutes on day 1, repeating every 3 weeks for 4 cycles in the absence of disease progression or unacceptable toxicity. Rituximab is listed on days 1, 8, and 15 of cycle 1 and day 1 of cycle 2 in the rituximab arms. Obinutuzumab is listed on those same days in the obinutuzumab arm. The registry text also refers to treatment beyond induction, but the available arm descriptions do not give complete details of that portion of the regimen. Further treatment schedules are not reported in the registry record.

Because allocation is non-randomized and masking is absent, participants and study personnel are not described as blinded, and treatment assignment does not create randomized groups. The record cautions that the arms are not a like-for-like control comparison. This matters because the arms differ in diagnosis and, for the FL groups, in the anti-CD20 antibody. An observed difference between groups would not isolate the effect of changing the antibody or adding pembrolizumab. The registry record does not report a protocol rationale for the group structure, a sample-size justification, or details of any planned interim analyses.

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

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 Overall Response Rate, measured over Up to 90 days after the last dose of pembrolizumab.

The primary endpoint is overall response rate, expressed as the percentage of participants with a complete or partial response. The registry defines response using computed tomography criteria based on Lugano 2014 and sets the assessment period as up to 90 days after the last dose of pembrolizumab. The primary measure is therefore a binary outcome for each participant—response or no response—summarized as a proportion within the relevant analysis population.

In general, an overall response rate is used to describe how often a predefined tumor response occurs during a specified assessment period. Combining complete and partial responses gives a single summary of tumor response under the stated criteria. It does not, on its own, describe how long a response lasts, whether disease later progresses, or how long participants live. Those are distinct questions, reflected in the registered secondary outcomes of PFS and OS.

A response proportion is most interpretable when the assessment rules, denominator, and handling of participants without an evaluable assessment are clear. In general, the denominator should correspond to a defined analysis population, and the response count should be reported alongside it so readers can judge the estimate’s precision. The registry record does not report arm-specific response counts, response percentages, confidence intervals, or the analysis population used for the primary endpoint. As a result, the endpoint’s definition is available, but its observed result and precision are not reported in the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry record identifies a binary percentage endpoint and gives a response definition, but does not report an analysis method or any analysis records. The methods below explain general statistical concepts relevant to this kind of endpoint; they are not claims about methods used in this study unless explicitly stated in the registry record.

How a response percentage is formed

In general, a binary endpoint records whether each participant meets a prespecified criterion. For overall response rate, the registry’s criterion is complete or partial response under CT-based Lugano 2014 criteria. The percentage is a summary of the proportion meeting that criterion in a defined group. To interpret it, readers need both the numerator—the number meeting the response definition—and the denominator—the number included in the analysis. The registry record does not report these values for the primary endpoint.

The denominator is important because it defines whose outcomes contribute to the estimate. An analysis might use all enrolled participants, only participants who received treatment, or only participants with an evaluable assessment. These choices can lead to different summaries when some participants have no assessment or leave the study before response evaluation. The registry record does not state the primary analysis population or how participants without an evaluable response were handled. Those details should not be inferred from the endpoint wording.

Why uncertainty matters

In general, a response percentage is an estimate of an underlying response probability, not a certainty about what would occur in another group of similar participants. A confidence interval describes sampling uncertainty around that estimate, conditional on the statistical method and analysis assumptions. Narrower intervals indicate greater precision; wider intervals indicate less precision. The registry record does not report a response estimate or confidence interval, so the precision of the endpoint cannot be evaluated.

Small studies often produce imprecise estimates because each participant has a substantial influence on the proportion. This is a general statistical consideration, not a calculated result for this study. Although the registry lists enrollment as 18, the primary endpoint denominator and arm-specific participant counts are not reported in the record. It would therefore be inappropriate to infer how many participants contributed to the response estimate or to calculate a confidence interval from the enrollment listing alone.

What the non-randomized structure allows

In general, random allocation helps balance measured and unmeasured participant characteristics across treatment groups, on average. Without random allocation, group differences may be related to the reasons participants received a particular regimen or to differences in diagnosis and prognosis. This study is explicitly non-randomized, and the registry cautions that the arms are not a like-for-like control comparison. Arm I and Arm III concern FL with different anti-CD20 antibodies, while Arm II concerns DLBCL with rituximab. Consequently, a simple difference between arm response percentages would not by itself identify a causal treatment effect.

Adjustment for measured characteristics can sometimes address observed imbalances, but it cannot guarantee removal of confounding and depends on suitable data and a prespecified method. The registry record does not report an adjustment plan, covariates, or an analysis strategy for between-arm comparisons. It also contains zero analysis records. No adjusted or unadjusted comparison should be attributed to the registry results.

What the planning illustration means

The registry’s planning description specifies a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. It describes the effect to detect as a stated fraction of the observed effect. For time-to-event outcomes, the illustration is framed in terms of a total number of events; for other outcomes, it is framed as participants per arm.

In general, a sample-size calculation asks how much information a design would need to detect a specified effect under stated assumptions. Its output is a planning illustration, not a measure of what happened in the study. It is not observed power, and it does not establish that an effect was found. The record does not report numerical sample-size illustration results or a trial-specific sample-size justification. The observed effect on which an illustration depends is also not reported in the registry record, so the calculation cannot be independently interpreted from the information available.

Multiplicity and analysis choices

In general, testing several outcomes or several comparisons can increase the chance of obtaining at least one apparently notable result by chance. A statistical plan may address this through a prespecified testing hierarchy or other multiplicity strategy. The registry record lists a primary endpoint and secondary outcomes, but does not report a multiplicity adjustment, a testing sequence, or the inferential procedures used. It would be speculative to assume that any particular approach was planned or applied.

Analysis populations, missing-data rules, confidence-interval methods, and any interim analyses are also not reported in the registry record. These omissions limit the ability to reproduce or evaluate a numerical endpoint analysis. The registry’s endpoint definition and design description remain useful for identifying the question under study, but they do not substitute for an analysis plan or posted numerical results.

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("Arm I Follicular Lymphoma (Pembrolizumab, Rituxima", "Arm II DLBCL (Pembrolizumab, Rituximab)", "Arm III Follicular Lymphoma (Pembrolizumab, Obinut")
n      <- c(7, 8, 3)
events <- c(2, 0, 0)
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  Arm I Follicular Lymphoma (Pembrolizumab, Rituxima  7  2  28.6  3.7  71.0  8.2  64.1  8.5  65.1
ARM  Arm II DLBCL (Pembrolizumab, Rituximab)  8  0  0.0  0.0  36.9  0.0  32.4  0.3  33.6
ARM  Arm III Follicular Lymphoma (Pembrolizumab, Obinut  3  0  0.0  0.0  70.8  0.0  56.1  0.6  60.2

11. Multiplicity, Missing Data and Analysis Populations

The registry lists overall response rate as the primary outcome and serious or drug-related adverse events, PFS, and OS as secondary outcomes. It does not report a strategy for multiplicity, such as a testing hierarchy, adjustment across outcomes, or rules for interpreting several comparisons. In general, when multiple outcomes or comparisons are examined, a prespecified approach helps clarify whether statistical findings are confirmatory or exploratory. No such approach should be assumed for this study.

The record contains zero analysis records. It does not report an analysis population for the primary endpoint, whether analyses were planned by assigned treatment or treatment received, or whether participants were grouped in another way. The response endpoint is defined, but the number assessed and the denominator used are not reported. The registry record also does not state how missing or unevaluable response assessments were handled, or how incomplete follow-up was treated for PFS and OS.

In general, missing-data rules matter because participants without an outcome assessment may differ from those with one. Treating missing outcomes as non-response, excluding them, or using a model to account for them can produce different summaries. These are possible general approaches, not methods reported for this trial. The statistical analysis plan, confidence-interval procedures, and any interim analyses are not reported in the registry record. Consequently, neither the handling of missing information nor the strength of control over multiple statistical tests can be assessed.

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: heart failure (14.3%), pleural effusion (14.3%), pneumonia (14.3%), diarrhea (12.5%), hypokalemia (12.5%).

The registry reports adverse-event groups with separate counts for participants at risk and affected. For serious events, Arm I lists 7 at risk and 2 affected; Arm II lists 8 at risk and 5 affected; Arm III lists 3 at risk and 0 affected. For other adverse events, the corresponding figures are 7 at risk and 7 affected, 8 at risk and 8 affected, and 3 at risk and 3 affected. The record also lists deaths as 2, 5, and 1 for Arms I, II, and III, respectively.

These are descriptive registry figures, not a randomized comparison. The arm groups differ by diagnosis and regimen, and the record cautions that they are not like-for-like controls. The counts alone do not establish that a treatment caused an event or death, and they do not describe event duration, severity beyond the serious-event classification, or timing. The registry record does not provide enough information to interpret the listed deaths as treatment-related.

The record lists Heart Failure, Pleural Effusion, and Pneumonia at 14.3%, and Diarrhea, Hypokalemia, and Arthralgia at 12.5%, among the top serious terms. The event counts and the precise denominators underlying these term percentages are not reported in the registry record. These percentages should therefore be read as posted descriptive figures, not as comparative risks.

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

The record establishes that a phase II study was registered to examine pembrolizumab with rituximab or obinutuzumab in relapsed or refractory FL and DLBCL. It defines overall response rate as the primary endpoint and specifies CT-based Lugano 2014 criteria and an assessment period up to 90 days after the last pembrolizumab dose. It also records secondary outcomes concerning adverse events, PFS, and OS, as well as adverse-event counts by arm.

Those details describe the study question and the scope of outcomes, but they do not demonstrate the frequency of response or the comparative effect of either combination. Numerical primary endpoint results, confidence intervals, and analysis populations are not reported in the registry record. No conclusion about the magnitude or precision of response can be drawn from the endpoint definition alone. The record’s zero analysis records further means that no statistical analysis results are posted in the analysis section.

The study does not provide a randomized, like-for-like comparison. Arm assignment is non-randomized, the arms differ in lymphoma diagnosis, and the FL arms use different anti-CD20 antibodies. Even if numerical arm results were available, between-arm differences would need cautious interpretation and would not alone show that a treatment component caused the difference.

The adverse-event counts are descriptive. They do not establish causation, and the listed deaths should not be interpreted as treatment-related without supporting information. The registry does not report why the study was terminated. In general, a terminated status may have different explanations; the record does not identify which applies.

Thus, the record supports a description of the registered design, interventions, endpoint, and posted safety counts. It does not support a conclusion that either regimen did or did not produce a particular response rate, nor a conclusion about comparative effectiveness or safety. Such conclusions would require numerical outcome analyses and sufficient detail about the populations, missing data, and methods used.

14. Why This Trial Matters Statistically

This study raises a useful statistical distinction between describing an intervention and establishing a comparative treatment effect. A phase II response endpoint can summarize whether participants met a defined response criterion, but a response proportion does not by itself establish how the outcome compares with another treatment or what caused it. That distinction is particularly important in a non-randomized study whose arms differ in diagnosis and treatment combination.

The registry gives a precise primary endpoint definition: complete or partial response by CT criteria (Lugano 2014), assessed up to 90 days after the last pembrolizumab dose. This helps identify what the investigators intended to measure. Yet the record does not report response counts, percentages, or confidence intervals. The difference between a clear endpoint definition and a reportable estimate matters: statistical interpretation requires both a defined measure and results with their denominator and uncertainty.

The record also illustrates why registered secondary outcomes and descriptive safety counts should not be mistaken for complete evidence. PFS and OS are listed, but numerical analyses are not reported. Adverse-event counts are present, but the non-randomized arm structure limits direct comparisons, and counts alone do not establish causation. A reader needs outcome definitions, follow-up information, analysis populations, and uncertainty estimates to assess how much the data can support.

Finally, the stated planning framework is not a substitute for a trial-specific design justification or observed analysis. Planning illustrations based on an observed effect can help convey assumptions, but they do not represent observed power and cannot replace transparent reporting of the effect estimate and its precision. The registry record does not report numerical illustration results or the primary endpoint analysis. Its statistical importance lies in making visible both the intended question and the limits of what can be concluded from the posted 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 main limitation for interpreting this study is the absence of numerical primary endpoint results in the registry record. There are no reported response counts, denominators, percentages, or confidence intervals, and the record contains zero analysis records. The registered definition does not reveal how many participants had an evaluable CT assessment or what proportion met the complete or partial response criteria.

The non-randomized design and the structure of the arms also constrain comparisons. The registry cautions that the arms are not a like-for-like control comparison. The arm descriptions place DLBCL in the rituximab group and FL in both the rituximab and obinutuzumab groups. Differences between groups therefore cannot be interpreted as an isolated effect of the antibody choice or of pembrolizumab.

The record lists enrollment as 18, but does not report arm-specific participant counts for the primary endpoint analysis or describe the analysis population. It does not report how missing or unevaluable outcomes were handled, which statistical procedures were applied, or whether covariate adjustment was planned. These details are needed to assess the reliability and reproducibility of an outcome analysis.

The study is listed as terminated, but the reason is not reported in the registry record. The record also does not give a trial-specific sample-size justification or numerical values for the planning illustrations. Safety figures are descriptive, and details about event timing, attribution, and the denominators for the listed event-term percentages are not reported. These limitations prevent a full assessment of efficacy, safety, and the reasons the study ended.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Pembrolizumab With Rituximab or Obinutuzumab in Treating Patients With Relapsed or Refractory Follicular Lymphoma or Diffuse Large B Cell Lymphoma — NCT03401853.
• Registry record last updated 2025-07-22; status last verified 2025-07.
• Results first posted 2025-04-18.
• 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

NCT03401853 is a phase II, non-randomized, open-label, parallel study sponsored by the University of Washington. It examines pembrolizumab with rituximab or obinutuzumab for relapsed or refractory FL and DLBCL. The registry lists enrollment as 18, a start date of 2018-03-23, a primary completion date of 2023-12-19, and status as terminated.

The primary endpoint is overall response rate, defined as complete plus partial responses by CT criteria (Lugano 2014) up to 90 days after the last pembrolizumab dose. Secondary outcomes include serious or drug-related adverse events, PFS, and OS. The registry record does not report numerical primary endpoint results or analysis records. It cautions that the non-randomized arms are not a like-for-like control comparison.