Phase II Copanlisib in Relapsed/Refractory Diffuse Large B-cell Lymphoma (DLBCL): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Phase II Copanlisib in Relapsed/Refractory Diffuse Large B-cell Lymphoma (DLBCL) (NCT02391116), a na phase 2 study. The primary outcome is ORR by CD79b Status Based on Investigator Assessment.

2. Clinical Question

The study also listed duration of response, progression-free survival and overall survival among its secondary outcomes, including analyses by CD79b status or DLBCL/cell-of-origin subtype for some outcomes. Listing an outcome describes what was intended to be assessed; it does not establish the size or direction of a finding. Those estimates are not reported in the registry record. In particular, the available record does not show how many participants in each biomarker group responded or the precision of any such estimate.

3. Trial Design

This was an open-label, single-arm Phase II study of copanlisib in people with relapsed or refractory DLBCL. The registry reports enrollment of 67 and a completed study. Treatment was single-agent copanlisib, given as an intravenous infusion. Because there was no randomized allocation or control group, the study can describe outcomes among enrolled participants but cannot directly estimate a comparative treatment effect against another group within the trial.

Open-label means the study was not masked. The primary outcome was assessed by investigators, according to the registry record. Investigator assessment is an important feature when interpreting response results: the recorded outcome follows the investigators' application of the stated response criteria, rather than an assessment described as independently blinded in this record. The record does not report whether any separate review process was used.

The primary endpoint was ORR by CD79b status, based on investigator assessment. The endpoint definition counts participants with at least one post-baseline complete or partial response during study conduct, according to the Lugano Classification, 2014, assessed by CT/MRI/PET-CT. The stated timeframe runs from the start of study treatment through 24 weeks after the last participant fully evaluable for the primary endpoint started treatment, described as about 12 months.

The listed primary analysis populations are the FAS and PPS. The registry record does not give the inclusion rules for these populations or the number assessed in each. It also does not report a detailed statistical analysis plan, how biomarker categories were defined, or how missing endpoint assessments were handled. These omissions limit interpretation of which participants contributed to each reported analysis; the available record does not include the actual primary endpoint estimates.

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: A total of 91 participants were screened, of which 67 were assigned to study treatment and also started the treatment, and 24 were screened but never assigned to treatment. Altogether 27 participants were excluded from the per protocol set, which comprised 40 participants.

Treatment

Reasons for not completing this period:

Safety Follow-up

Reasons for not completing this period:

Active Follow-up

Reasons for not completing this period:

Survival Follow-up

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Full analysis set (FAS) and per protocol set (PPS)

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 ORR by CD79b Status Based on Investigator Assessment, measured over From start of study treatment assessed up to 24 weeks after the last participant fully evaluable for the primary endpoint started treatment (about 12 months).

Objective response rate is the percentage of participants who have a qualifying response after baseline. In this study, a qualifying response was complete response (CR) or partial response (PR), determined under the Lugano Classification, 2014. Response was assessed using CT/MRI/PET-CT and investigator assessment. The primary endpoint was specified by CD79b status, making the biomarker relationship part of the central question rather than an incidental subgroup description.

As a general statistical concept, ORR compresses response information into a proportion: it indicates how often a defined response was recorded in the population being analyzed. It is more interpretable when the numerator, denominator, assessment rules and uncertainty are available. The registry record identifies percentage of participants as the unit and a 90% confidence interval as the dispersion measure. However, it does not report the observed rates, response counts, subgroup denominators, or confidence interval limits in the record.

Response rate is useful for describing the occurrence of tumor response, but it does not by itself describe how long a response lasted, whether disease later progressed, or overall survival. The registry lists duration of response, progression-free survival and overall survival as secondary outcomes. Their inclusion reflects different questions; their results are not reported in the registry record. The endpoint definition therefore clarifies what was intended to count as response, but cannot establish the response rate or its relationship to CD79b without the associated results.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

Several statistical ideas help interpret the design, even when the registry record does not contain the numerical primary outcome results. The explanations below describe general concepts and do not fill gaps in the trial's reported analyses.

What does a single-group design estimate?

A single-group study follows participants receiving the study intervention and describes outcomes in that group. A response proportion can summarize the observed frequency of a defined response, provided the numerator and denominator are specified. The design does not create a concurrent counterfactual group: it cannot directly show what would have happened to comparable participants under another intervention. That is why descriptive response findings and comparative treatment effects are different statistical quantities.

In this study, allocation is listed as not applicable and the model as single group. There is no randomized control arm. If an estimate is available, it should be read as an outcome observed in the study population, not as a randomized contrast. The record does not report an adjusted comparison with an external group, and such a comparison should not be inferred.

What does a confidence interval add?

A point estimate alone does not convey how much statistical uncertainty surrounds it. A confidence interval gives a range of values compatible with the estimate under a specified statistical procedure and its assumptions. A narrow interval generally indicates a more precise estimate than a wide one. The interval does not eliminate possible bias from selection, measurement, or missing data, and it is not a guarantee about an individual participant's outcome.

The primary outcome identifies a 90% confidence interval. The registry record does not report its bounds or the observed ORR, so neither the estimate nor its precision can be evaluated. The confidence level is not itself a measure of clinical importance; interpretation also depends on the estimate, the interval, the endpoint, and the study design.

Why separate biomarker results?

The primary question concerns ORR by CD79b status. Subgroup analysis separates participants according to a characteristic and estimates outcomes within those groups. To interpret such an analysis, readers need the group definitions, number analyzed in each group, response counts, estimates and uncertainty. A statistical test of association, if used, would also need to be described. Those analysis details and results are not reported in the registry record.

Subgroup estimates can be less precise than an overall estimate because they use fewer observations. This is a general statistical consideration, not a finding about the CD79b groups in this study. A biomarker relationship also differs from evidence that the biomarker predicts a differential treatment effect. In a single-arm study, outcomes across biomarker groups do not by themselves establish that treatment effects differ by biomarker, since there is no control group to estimate the treatment contrast within each group.

What do the analysis populations mean?

The registry names the full analysis set (FAS) and per protocol set (PPS) as primary endpoint populations. Broadly, an analysis population defines which participants contribute to a particular estimate. Different population rules can lead to different denominators and potentially different estimates. The exact FAS and PPS definitions are not reported in the registry record, so it is not possible to determine from this record who qualified for each set or how the sets differed.

Missing outcome information matters because participants without an evaluable response may affect the denominator or require a missing-data approach. Different approaches can answer different questions and rely on different assumptions. The registry record does not report the amount or handling of missing primary endpoint data. No assumption about those participants' response status should be made.

How should the planning illustration be read?

The registry record describes a planning illustration based on an observed effect, using 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 stated fraction of the observed effect. Time-to-event illustrations are described as total event counts using the Schoenfeld method; other illustrations are described as participants per arm. Such calculations are hypothetical planning exercises, not results from the trial's single-group design and not observed power.

The illustration's assumptions do not establish the trial's original design rationale. The registry record does not report the study's sample-size justification or a detailed statistical plan. In particular, hypothetical equal allocation between two arms should not be confused with this study, which had one group and no concurrent control. The illustration cannot recover the missing ORR estimates or establish that the study had a particular probability of detecting an effect.

Finally, over-dispersion is a general issue for some count-data models: when observations vary more than a model assumes, standard errors may be too small and uncertainty understated if the model does not account for that extra variation. The primary outcome is a percentage of participants, and the registry record does not report an over-dispersion model or any such adjustment. This general concept should not be mistaken for a description of the trial's actual analysis.

10. Multiplicity, Missing Data and Analysis Populations

The primary endpoint is ORR by CD79b status. The registry also lists secondary outcomes for duration of response, progression-free survival and overall survival, with some analyses by CD79b status or DLBCL/COO subtype. It reports 21 outcome measures and 8 analyses. Multiple outcomes, subgroup analyses and repeated statistical tests can increase the chance of at least one apparently unusual result arising by chance. Whether and how that concern was addressed depends on the planned testing strategy.

The registry record does not report a multiplicity adjustment, a hierarchy of hypotheses, or a detailed statistical analysis plan. It is therefore not possible to determine whether the analyses were confirmatory, exploratory, adjusted, or interpreted under another prespecified approach. The existence of several listed outcomes does not by itself establish that multiplicity was controlled or that any particular result should be treated as definitive.

Missing-data handling is also not reported. The record names the FAS and PPS, but does not define them, state the number contributing to each endpoint analysis, or describe how unevaluable assessments were handled. These details can affect the denominator and interpretation of a response percentage. The stated primary endpoint timeframe describes follow-up through 24 weeks after the last participant fully evaluable for the primary endpoint started treatment, described as about 12 months; it does not explain the missing-data method.

Consequently, the record leaves uncertainty about the analysis population, missingness and multiplicity strategy, in addition to not reporting the primary endpoint estimates. These are reporting limitations, not proof that a particular analysis choice was made or that any result is biased.

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 4% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: general physical health deterioration (13.4%), hyperglycaemia (7.5%), pyrexia (6.0%), abdominal pain (4.5%), pneumonitis (4.5%).

The registry record lists 67 participants at risk for both serious and other adverse events in the copanlisib group. It records 44 as affected by serious adverse events and 64 as affected by other adverse events. It also lists 53 deaths. These are separate registry figures; they should not be added together because the record does not establish that the categories are mutually exclusive. The death figure alone does not establish cause or attribution to treatment.

The listed serious adverse event terms include general physical health deterioration (13.4%), hyperglycaemia (7.5%), pyrexia (6.0%), abdominal pain (4.5%), pneumonitis (4.5%) and death (3.0%). These are reported event-term percentages, not comparative risks against a control group. In a single-arm study, there is no concurrent control estimate in this record to distinguish background events from events related to treatment.

Adverse event counts describe events recorded during the study; they do not alone establish causation, severity for every participant, timing, or the relationship between multiple events in the same participant. The registry record does not report the detailed definitions or attribution for these figures. They should therefore be read as descriptive safety information, not as a causal comparison.

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

The record establishes that a completed, open-label, single-arm Phase II study evaluated single-agent copanlisib in relapsed or refractory DLBCL. It defines the primary endpoint as investigator-assessed ORR by CD79b status, with response based on complete or partial response according to the Lugano Classification, 2014. It also records a 90% confidence interval as the intended measure of dispersion and names the FAS and PPS as analysis populations.

Those design details do not reveal the primary result. The observed ORR, its confidence interval, response counts, subgroup denominators and the relationship between response and CD79b status are not reported in the registry record. The listed secondary outcomes likewise do not establish results for duration of response, progression-free survival or overall survival. Without the estimates and their uncertainty, the magnitude and precision of efficacy outcomes cannot be assessed from this record.

The single-arm design also does not support a randomized comparison with another treatment. Any outcome observed in the study group would be descriptive of those participants in the study context. It would not, on its own, establish how outcomes compare under an alternative treatment or isolate the effect of copanlisib from other influences.

The safety figures provide descriptive information about adverse events and deaths recorded in the group. They do not by themselves establish that events were caused by the study drug, and there is no concurrent control group for a comparative safety estimate. A record of deaths does not specify their cause.

Thus, the registry record supports a clear account of the study question, structure and selected adverse event figures, but not a conclusion about the size of the primary response outcome or a biomarker association. Missing estimates should not be replaced by assumptions based on the study title, endpoint definition, or outcome list.

13. Why This Trial Matters Statistically

This study is statistically notable for the alignment between its primary endpoint and biomarker question: ORR was specified by CD79b status rather than only as an overall response proportion. That framing makes clear that the investigators sought to assess both response and its relationship to a potentially predictive biomarker. It also highlights why subgroup denominators and uncertainty are essential when interpreting biomarker analyses.

The single-arm structure matters just as much. A response proportion can describe outcomes in the treated group, but the design does not provide a concurrent randomized comparison. This limits causal and comparative interpretation, regardless of how clearly the endpoint is defined. The open-label design and investigator assessment are also relevant context for understanding how the response outcome was obtained.

Statistical completeness is important for this question. The registry specifies a 90% confidence interval, FAS and PPS populations, and a broad set of secondary outcomes. However, the estimates, interval limits, population definitions, missing-data approach and multiplicity strategy are not reported in the registry record. Those details would be needed to assess precision, determine which participants contributed to each analysis, and interpret the biomarker findings.

The record's adverse event counts and selected terms are useful as descriptive safety information, but the absence of a control group means they cannot provide a comparative risk estimate. Overall, the study illustrates why a stated statistical objective is not the same as a reported statistical result: design and endpoint definitions explain the intended analysis, while estimates and uncertainty are needed to understand what the data show.

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 principal trial-specific limitation for interpreting efficacy is that the registry record does not report the observed primary endpoint results. It defines ORR by CD79b status but gives no response estimates, counts, subgroup denominators or confidence interval limits. The stated 90% confidence interval therefore cannot be used to assess precision. The relationship between CD79b status and response remains unquantified in the record.

The single-arm design provides no concurrent control group. Outcomes in the copanlisib group cannot be read as a randomized comparison with another treatment. The open-label structure and investigator assessment are additional context for the response endpoint; the record does not say whether an independent assessment was also used.

The analysis populations are named as FAS and PPS, but their definitions and analysis-specific denominators are not reported. Nor does the record describe how missing or unevaluable assessments were handled. The statistical plan, multiplicity strategy, subgroup definitions and sample-size justification are also not reported in the registry record. As a result, the extent to which the analyses were prespecified or how uncertainty across multiple outcomes was addressed cannot be determined.

Safety interpretation is limited because the registry provides descriptive counts and selected event terms without a concurrent comparison. The record does not establish causality for the listed events or explain the causes of deaths. Secondary outcomes are listed, but their estimates are not reported in the available record. These limits prevent conclusions about comparative efficacy, biomarker prediction, and comparative safety from being drawn from the record alone.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Phase II Copanlisib in Relapsed/Refractory Diffuse Large B-cell Lymphoma (DLBCL) — NCT02391116.
• Registry record last updated 2019-01-04; status last verified 2018-12.
• Results first posted 2018-01-08.
• 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

NCT02391116 is a completed, open-label, single-arm Phase II study of copanlisib in relapsed or refractory DLBCL, with enrollment of 67. Its primary endpoint was investigator-assessed ORR by CD79b status, defined by complete or partial response under the Lugano Classification, 2014. The registry specifies a 90% confidence interval and identifies the FAS and PPS as analysis populations.

The available registry record does not report the primary ORR estimates, subgroup results, confidence interval limits, or secondary outcome estimates. It lists 44 participants affected by serious adverse events, 64 by other adverse events, and 53 deaths, alongside selected serious event terms and percentages. These safety figures are descriptive and do not establish causation or a comparison with another treatment.