A Study of REGN2810 and Ipilimumab in Patients With Lung Cancer: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study of REGN2810 and Ipilimumab in Patients With Lung Cancer (NCT03430063), a randomized phase 2 study. The primary outcome is Objective Response Rate (ORR) in Participants Whose Tumors Express Programmed Cell Death Ligand 1 (PD-L1) in <50% of Tumor Cells.

2. Clinical Question

The clinical question concerns whether the response proportions differ across the randomized treatment groups in this defined population. A complete or partial response counted toward the primary outcome when it was the participant’s best overall response, as determined by a blinded independent review committee using RECIST 1.1. The endpoint therefore summarizes the proportion of participants in the efficacy analysis set who met either response category during the specified follow-up period.

The posted arm-level results describe response proportions and 95% confidence intervals, but the registry does not report a formal statistical comparison, an effect estimate comparing groups, or a p-value. The outcome values can be read as descriptive evidence about responses in these participants; they do not by themselves establish a statistically supported difference between regimens. The study was randomized and open-label, with no masking, and its registry status is terminated.

3. Trial Design

This was a randomized, open-label, parallel-group Phase 2 study sponsored by Regeneron Pharmaceuticals. It evaluated three regimens in second-line treatment of advanced non-small cell lung cancer: cemiplimab 350 mg every three weeks, the same cemiplimab regimen combined with ipilimumab 50 mg every six weeks, and cemiplimab 1050 mg every three weeks. The stated primary objective focused on objective response rate in participants whose tumors expressed PD-L1 in less than 50% of tumor cells.

The registry lists enrollment as 28. For the primary outcome, the arm-level participant counts are 8 for cemiplimab 350 mg every three weeks, 11 for the combination, and 8 for cemiplimab 1050 mg every three weeks. The primary outcome description says the analysis population included randomized participants who received any study drug and was based on treatment received. The arm counts should be understood in that context, rather than assumed to represent every randomized participant.

Randomization is intended, in general, to balance measured and unmeasured participant characteristics across groups; with small groups, chance imbalances can still occur. Open-label means the trial was not masked. The primary response assessment, however, is described as based on review by a blinded independent review committee. That distinction concerns assessment of response and does not mean the treatment assignment was masked throughout the trial.

The registry gives a primary outcome timeframe from randomization up to 41 months. It lists a start date of 2018-05-29 and a primary completion date of 2021-10-27. The study status is terminated; a reason for termination is not reported in the registry record. Details about randomization procedures, allocation concealment, planned enrollment rationale, and the specific statistical comparison method are not reported in the registry record.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Before assignment: Of the 27 participants treated, four participants completed the study. The most common reason for study discontinuation was death (8 participants) followed by withdrawal of consent (6 participants), and progressive disease (4 participants).

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: SAF included all randomized participants who received any study drug; it is based on the treatment received (as treated).

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.

Population: Safety analysis set (SAF) included all randomized participants who received any study drug; it is based on the treatment received (as treated).

6. Primary Endpoint

The registry lists the primary outcome as Objective Response Rate (ORR) in Participants Whose Tumors Express Programmed Cell Death Ligand 1 (PD-L1) in <50% of Tumor Cells, measured over From date of randomization up to 41 months.

The primary endpoint is objective response rate, expressed as the percentage of participants with a best overall response of complete response or partial response. The registry defines the rate as the number of participants with either response divided by the number of participants in the efficacy analysis set. The best overall response was determined by a blinded independent review committee using RECIST 1.1, from randomization until the first objectively documented progression or subsequent anti-cancer therapy, whichever came first.

Under the stated definitions, complete response means disappearance of all target lesions, with pathological lymph nodes reduced to less than 10 mm in short axis. Partial response means at least a 30% decrease in the sum of target-lesion diameters relative to the baseline sum. These criteria make the outcome a defined assessment of tumor response rather than a measure of survival or symptom change.

Response rate is used in general as a binary endpoint: each participant is classified as having or not having a qualifying response. It is straightforward to express as a proportion and can summarize antitumor activity during a specified assessment period. It does not describe how long a response lasts, whether disease control persists, or the overall balance of benefits and harms. The registry lists overall survival, progression-free survival, and safety outcomes as secondary measures, but they answer different questions from the primary response-rate endpoint.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is a binary response measure. Each participant is classified according to whether the best overall response met the definition of complete or partial response. The arm-level percentage is the number of participants meeting either definition divided by the number in the efficacy analysis set. In general, a binary endpoint reduces a complex clinical course to a clearly defined event status over a specified period. It is useful for estimating response proportions, but it does not retain all details about timing, duration, or changes that fall short of the response criteria.

What does a confidence interval express?

The registry posts a 95% confidence interval for each arm’s response rate. In general, a confidence interval describes the statistical uncertainty around an estimate under the assumptions of the method used. Repeated use of a method with the stated confidence level would cover the underlying parameter at that rate over repeated samples, assuming the method’s conditions hold. It is not a probability statement about the fixed true response rate after observing this particular interval.

The interval width is relevant because a point estimate alone can make a small study appear more definitive than it is. Here, the participant counts are 8, 11, and 8 across the arms, and each estimate has substantial uncertainty as expressed by its interval. The registry does not identify the confidence-interval construction method, so it is not appropriate to infer whether it used a particular exact, score-based, or other procedure. The posted limits should be read as registry results, without assigning an undocumented method.

Why is an arm-level result different from a treatment comparison?

An arm-level response rate estimates the proportion in one group. A randomized treatment comparison asks whether the response probability differs between groups and quantifies that contrast with an estimate and uncertainty. A comparison might use a difference in proportions, a relative measure, or a test suited to the design, but the choice and result are not reported in the registry record. No new contrast can be inferred solely from the separate response percentages and intervals.

In general, the precision of a comparison depends on the number of observations and on the event frequencies in both groups. Small groups can leave considerable uncertainty even when the observed arm percentages look far apart. A formal analysis would also need to follow the study’s planned estimand, analysis population, and handling of incomplete outcome data. Those details are not reported in the registry record. The lack of posted analyses means there is no basis for attributing a particular inferential procedure or adjustment to the trial.

What is the role of the analysis population?

The primary outcome description states that the population included all randomized participants who received any study drug and was based on treatment received. This is an as-treated description: classification follows the treatment participants received, rather than necessarily retaining every randomized participant in the group originally assigned. In general, an analysis based on treatment received can differ from an analysis that preserves randomized assignment. Randomization supports comparisons when the assigned groups are maintained for analysis; departures from that principle can affect comparability. The registry does not report whether participants switched treatment or describe any such departures, so their presence or impact should not be assumed.

The outcome record lists the participant counts separately for each regimen. It does not report additional details about exclusions from the efficacy analysis set, the timing of individual assessments, or how participants without an evaluable response were classified. Those details can matter for a binary outcome because the denominator and the treatment of missing assessments determine the interpretation of a response proportion.

How should planning calculations be interpreted?

Any sample-size planning illustration associated with these results is not observed power. The stated illustration assumptions use a two-sided significance level of 0.05, equal allocation between two arms, a normal approximation, and an effect to detect defined as a fraction of the observed effect. For time-to-event outcomes, the described planning quantity is a total event count; for other outcomes, it is participants per arm. These are planning assumptions, not evidence that the original trial was designed with those assumptions or that a target sample size was achieved.

In general, power is a design-stage probability of detecting an effect under specified assumptions if that effect is true. A calculation using an effect estimated from the study results is especially sensitive to uncertainty in that estimate. It cannot establish that the observed result had a particular probability of being correct, and it should not be presented as the study’s observed power. The original sample-size justification and statistical design rationale are not reported in the registry record.

Why can over-dispersion matter for other outcomes?

Over-dispersion is a general issue for count outcomes when observed variability exceeds what a simple model assumes. If such a model is used without accounting for extra variability, standard errors can be too small and intervals or tests may overstate precision. This concept does not change the interpretation of the primary binary response measure directly. The registry lists survival, progression-free survival, adverse events, and laboratory abnormalities among secondary outcomes, but it does not post statistical analyses that would show which models were used or whether over-dispersion was evaluated for any outcome.

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("Cemiplimab 350 mg Q3W", "Cemiplimab 350 mg Q3W + Ipilimumab 50 mg Q6W", "Cemiplimab 1050 mg Q3W")
n      <- c(8, 11, 8)
events <- c(0, 5, 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  Cemiplimab 350 mg Q3W  8  0  0.0  0.0  36.9  0.0  32.4  0.3  33.6
ARM  Cemiplimab 350 mg Q3W + Ipilimumab 50 mg Q6W  11  5  45.5  16.7  76.6  21.3  72.0  21.1  72.3
ARM  Cemiplimab 1050 mg Q3W  8  0  0.0  0.0  36.9  0.0  32.4  0.3  33.6

11. Multiplicity, Missing Data and Analysis Populations

The registry identifies objective response rate as the primary outcome and lists secondary outcomes covering overall response rate, overall survival, progression-free survival, treatment-emergent adverse events, serious treatment-emergent adverse events, adverse events resulting in death, and laboratory abnormalities of grade 2 or higher. It does not report a multiplicity strategy, such as an adjustment to account for testing several outcomes or comparisons. As a result, the interpretation of any statistical significance across multiple tests cannot be assessed; no such significance analyses are posted for the primary outcome in any case.

Missing-data handling is not reported in the registry record. The primary outcome definition gives the response criteria and timeframe, but does not state how missing scans, unevaluable responses, withdrawal, or incomplete follow-up were handled. For a response proportion, these choices can affect who contributes to the denominator and how an unknown response status is treated. The posted endpoint results should therefore be interpreted with the stated analysis population, while recognizing that the record does not explain whether all relevant assessments were available.

The primary outcome population is described as randomized participants who received any study drug, analyzed according to treatment received. This is not enough information to reconstruct participant flow or establish how the analysis set differs from all randomized participants. The record lists enrollment as 28 and gives arm-level primary outcome counts, but it does not provide a detailed accounting of exclusions or outcome availability. Reasons for termination and the original statistical analysis plan are not reported in the registry record.

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: pneumonia (27.3%), brain stem embolism (12.5%), cerebrovascular accident (12.5%), hemiparesis (12.5%), bradypnoea (12.5%).

The safety results are counts of participants, not estimates of comparative risk adjusted for follow-up or exposure. Serious adverse events were reported for 1 of 8 participants receiving cemiplimab 350 mg every three weeks, 7 of 11 receiving the combination, and 4 of 8 receiving cemiplimab 1050 mg every three weeks. Other treatment-emergent adverse events were reported for 7 of 8, 11 of 11, and 8 of 8 participants, respectively. The registry also lists 3 deaths in the first group, 3 in the combination group, and 4 in the high-dose group.

These counts do not, by themselves, show that a treatment caused an event or that rates are directly comparable. The record does not provide exposure-adjusted rates or detailed follow-up by arm in the listed safety figures. The serious-event terms listed include Pneumonia at 27.3%, Brain stem embolism at 12.5%, Cerebrovascular accident at 12.5%, Hemiparesis at 12.5%, Bradypnoea at 12.5%, and Dyspnoea at 12.5%. The registry entry does not provide sufficient context in these figures to assign each listed term to a particular treatment group. These findings describe reported events and should not be treated as a causal comparison.

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

The posted results show that the observed objective response rate was 45.5% in the cemiplimab-plus-ipilimumab group and 0.0% in each cemiplimab-alone group. The corresponding confidence intervals are wide, and the arm-level participant counts are small. This is a description of the results in the registry’s defined population, assessed by the stated response criteria and analysis population.

The results do not establish that one regimen is statistically superior to another. No formal comparison, p-value, or confidence interval for a contrast is posted. Nor do separate arm-level confidence intervals answer the comparison question on their own. The planned primary objective was comparative, but the registry record does not include the comparative analysis needed to evaluate that objective statistically.

Response rate does not answer every clinically relevant question. It does not indicate how long responses lasted, whether participants lived longer, or how progression-free survival differed. Those outcomes are listed as secondary measures, but their results are not included in the registry record described for the primary endpoint. The safety counts provide a separate description of adverse events and deaths; they do not establish treatment causation or a comparative safety profile.

The study is marked terminated, but its reason for termination is not reported in the registry record. The record also does not report the original sample-size justification, missing-data procedures, or a multiplicity plan. Those omissions limit interpretation of the planned comparative question and of the full body of outcomes. The most defensible conclusion is therefore descriptive: the registry reports the stated response proportions and intervals for these small treatment groups, while leaving comparative inference and several aspects of study conduct unresolved.

14. Why This Trial Matters Statistically

This trial illustrates why a response percentage should be read together with its denominator and uncertainty. The primary outcome counts were 8, 11, and 8, and the 95% confidence intervals span appreciable ranges. The point estimates alone—45.5% in the combination group and 0.0% in each single-agent group—do not convey the full uncertainty of estimates from groups of this size.

It also highlights the distinction between randomized design and reported analysis. Randomization is a design feature intended, in general, to support fair comparisons, but the primary outcome population is described as based on treatment received. The registry does not report a formal comparison or explain how the analysis set was formed in detail. Readers therefore need to distinguish a study’s comparative objective from the evidence available to evaluate that objective.

The endpoint definition adds another important layer. An objective response was determined using prespecified tumor-response criteria and independent review, which makes the outcome more explicit than an undefined judgment of improvement. At the same time, response proportion is not a substitute for duration of response, survival, or safety assessment. Those are distinct outcomes and require their own results and interpretation.

Finally, the registry illustrates the limits of a record that contains outcome values without a posted analysis plan or comparative test. A randomized trial can generate informative descriptive data, but transparent reporting of analysis methods, missing-data handling, and the reason for termination is important for judging how far those data support the trial’s original question. In this case, those details are not reported in the registry record.

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 primary endpoint results come from small groups, with participant counts of 8, 11, and 8. The associated confidence intervals reflect substantial uncertainty, so the estimates should not be treated as precise response rates for a broader population. The record does not report a formal comparison of the groups, despite the stated objective of comparing the experimental regimens with standard-dose cemiplimab.

The primary analysis population is described as randomized participants who received study drug, analyzed according to treatment received. Detailed participant flow, exclusions, treatment changes, and the relationship between that population and all randomized participants are not reported in the registry record. Missing response assessments and their handling are also not reported, limiting assessment of possible bias in the observed proportions.

The study was open-label, although the primary response assessment is described as conducted by a blinded independent review committee. The registry does not give enough detail to assess other safeguards against bias, including allocation concealment or the specific randomization procedure. The statistical analysis plan, confidence-interval method, sample-size justification, and multiplicity approach are not reported in the registry record.

The study is listed as terminated, but the reason is not reported in the registry record. Follow-up is described as extending from randomization up to 41 months, but the arm-specific duration and completeness of follow-up are not stated in the primary outcome details. Finally, the findings concern participants with advanced non-small cell lung cancer whose tumors expressed PD-L1 in less than 50% of tumor cells. They should not be generalized to populations outside that stated group without further evidence.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: A Study of REGN2810 and Ipilimumab in Patients With Lung Cancer — NCT03430063.
• Registry record last updated 2022-11-03; status last verified 2022-10.
• Results first posted 2022-11-03.
• 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 randomized, open-label, parallel-group Phase 2 study of cemiplimab regimens, with or without ipilimumab, for second-line treatment of advanced non-small cell lung cancer. The primary objective concerned objective response rate among participants whose tumors expressed PD-L1 in less than 50% of tumor cells. The study is listed as terminated, with enrollment of 28; the reason for termination is not reported in the registry record.

Posted primary response rates were 0.0% with cemiplimab 350 mg every three weeks, 45.5% with that regimen plus ipilimumab 50 mg every six weeks, and 0.0% with cemiplimab 1050 mg every three weeks. The registry reports 95% confidence intervals for each arm, but no formal between-group analysis. The primary outcome population is described as participants who received any study drug, analyzed according to treatment received.

The findings are descriptive and imprecise in these small groups. Comparative analyses, missing-data procedures, multiplicity handling, the original sample-size rationale, and the reason for termination are not reported in the registry record.