Recombinant Interleukin-15 in Combination With Checkpoint Inhibitors Nivolumab and Ipilimumab in People With Refractory Cancers: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Recombinant Interleukin-15 in Combination With Checkpoint Inhibitors Nivolumab and Ipilimumab in People With Refractory Cancers (NCT03388632), a non-randomized phase 1 study. The primary outcome is Number of Dose-limiting Toxicities (DLT) Possibly, Probably, or Definitely Related to Study Drugs.

2. Clinical Question

3. Trial Design

This was a phase I, non-randomized, unmasked, single-group study sponsored by the National Cancer Institute (NCI). The registry records a start date of 2018-02-05, primary completion on 2024-11-27, and a completed status. Enrollment is recorded as 31. The listed conditions are metastatic solid tumors and treatment-refractory cancers.

The brief description states that participants were assigned to rhIL-15 with nivolumab, rhIL-15 with ipilimumab, or rhIL-15 with both checkpoint inhibitors. The arm descriptions further distinguish lead-in doublets for initial safety evaluation and triplet groups for dose escalation. The rhIL-15 doses named in the adverse-event group titles include 0.5mcg/kg, 1mcg/kg, and 2mcg/kg. The schedule described in the record gives rhIL-15 by subcutaneous injection and the other drugs intravenously. The record’s brief summary describes four 6-week cycles, with rhIL-15 limited to the first 4 cycles.

Because allocation was not randomized and there was no stated control group, the regimens should be read as distinct safety and dose groups rather than treatment arms in a comparative efficacy trial. The registry does not report the allocation procedure, detailed eligibility criteria beyond the brief description, or a statistical analysis plan. It also does not report a numerical result for the primary endpoint.

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.

Only the first 6 of 9 groups are shown.

Lead-In Doublet A

Lead-In Doublet B

Triplet C

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: This analysis includes participants (pts) from multiple groups, as different groups received different doses of rhIL-15. Includes pts from group 4 (n=2), group 5 (n=9), group 6 (n=2), group 7 (n=1), group 8 (n=2), and group 9(n=1). Of the 18 pts in group 5, 9 pts were enrolled in the dose escalation phase to determine MTD \& an additional 9 pts were enrolled in an expansion cohort at that dose after the MTD was determined. Pts in this expansion cohort were not included in the MTD determination.

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.

Only the first 6 of 10 groups are shown.

6. Primary Endpoint

The registry lists the primary outcome as Number of Dose-limiting Toxicities (DLT) Possibly, Probably, or Definitely Related to Study Drugs, measured over Cycle 1 (42 days).

The primary endpoint was the number of dose-limiting toxicities (DLTs) possibly, probably, or definitely related to study drugs during Cycle 1 (42 days). A DLT was defined as a related adverse event meeting prespecified criteria. The description includes conditions requiring long-term corticosteroid treatment or permanent discontinuation of an agent among those criteria. The detailed criteria are not reported in the registry record.

A DLT count is used in early-phase research to describe treatment-limiting toxicity during a defined assessment period. In general, this kind of endpoint can inform evaluation of dose levels and tolerability; it does not by itself establish comparative benefit or characterize every aspect of safety. The registry specifies that DLTs were assessed in the lead-in doublet arms and triplet escalation phase groups only. It does not state that the endpoint was assessed across all enrolled participants.

The posted record identifies the endpoint but does not give its observed count or a group-by-group DLT result. Consequently, no estimate of the frequency of DLTs, uncertainty around that frequency, or comparison among the dose groups can be drawn from the primary endpoint information recorded.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The design determines what statistical comparisons are justified. This study is recorded as non-randomized and single-group, with no masking. In general, random allocation helps balance measured and unmeasured characteristics between treatment groups. Without it, groups may differ for reasons other than the regimens being studied. Here, the doublet and triplet groups also represent different combinations and doses, so a numerical contrast between them would not be a clean estimate of a treatment effect.

The primary endpoint is a count of DLTs in Cycle 1 (42 days), not a continuous measurement such as a laboratory value. A DLT required a specified relationship to study drugs and meeting prespecified criteria. In general, a count endpoint records how many events occur within a defined assessment scope; it is not interchangeable with the number of participants who experience any adverse event. A participant could have more than one event, and the registry’s endpoint wording is about toxicities. The record does not report the DLT count, so no rate or uncertainty estimate for this primary outcome can be described.

The registry does identify the population for DLT assessment: the lead-in doublet arms and triplet escalation phase groups. This scope matters when interpreting an endpoint. A result from specified dose-escalation groups would not automatically describe all enrolled participants, nor would it necessarily apply to all safety outcomes. The record does not provide further detail on the analysis population, how participants with incomplete Cycle 1 assessment were handled, or whether the endpoint was analyzed by dose group.

Confidence intervals are a general statistical tool for expressing uncertainty around an estimate. A wider interval usually indicates less precision, while a narrower interval indicates greater precision under the assumptions of the method. No confidence interval for the primary endpoint is reported in the registry record. Given the small counts in the listed safety groups, any group-level estimate would also need to be interpreted with attention to its uncertainty; the registry does not provide inferential estimates for those counts.

In general, over-dispersion is relevant when observed count variation exceeds what a statistical model assumes. It can make model-based standard errors and confidence intervals too small if ignored. The posted primary endpoint is a count, but the registry record does not report a model, a dispersion assessment, or an analysis of that endpoint. It is therefore not possible to say whether over-dispersion was evaluated or whether a count model was used.

The record identifies four outcome measures and no analyses. It does not report a statistical analysis plan, a hypothesis-testing procedure, or a multiplicity adjustment. Multiplicity refers generally to the increased chance of a false-positive finding when many statistical tests are conducted without appropriate control. Without registered analysis details or reported test results, no conclusion can be drawn about how multiplicity was handled. Nor is there a reported basis for interpreting any nominal significance level or p-value for the primary outcome.

The record describes the study as phase I and states that its objective was to test the effects and maximum dose of the drug combinations. The detailed dose-escalation rules, dose-selection criteria, stopping rules, and any interim analyses are not reported in the registry record. The primary endpoint definition notes prespecified criteria, but their full content is not given. Thus, the dose-escalation procedure cannot be reconstructed from the available record.

The sample-size material is a planning illustration based on an observed effect, with a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. It is not the trial’s own design or an observed-power calculation. The registry record does not provide a sample-size justification for this phase I study. Because the study is non-randomized and its groups are not like-for-like controls, any planning illustration involving a two-arm comparison should not be treated as evidence that the trial was designed to estimate a causal difference between these regimens.

In general, an analysis population should be defined before results are interpreted, particularly when the endpoint is assessed only in specified groups or within a specified period. The registry identifies the DLT assessment groups and timeframe, but does not report how missing observations, withdrawals, or incomplete treatment were handled. No adjustment variables, model specifications, or sensitivity analyses are reported. These omissions limit assessment of how robust any unreported endpoint estimate would be.

10. Multiplicity, Missing Data and Analysis Populations

The registry record lists four outcome measures and no analyses. It does not report a statistical analysis plan, a hierarchy of outcomes, a multiplicity adjustment, or the number and purpose of planned hypothesis tests. In general, multiplicity arises when several outcomes or comparisons are tested; without a stated approach, the chance of false-positive findings across tests can be difficult to assess. No inference about multiplicity control can be made from the registry information.

The primary endpoint has a defined assessment window, Cycle 1 (42 days), and is specified for the lead-in doublet arms and triplet escalation phase groups. The registry does not report the number analyzed for this endpoint, the DLT counts, or whether participants with incomplete observation were included. The handling of missing endpoint data, withdrawals, and treatment discontinuation is not reported in the registry record. There is also no reported analysis population definition beyond the stated DLT assessment scope.

In general, missing data can affect an estimate when the people with observed outcomes differ from those without observed outcomes. The direction and size of any such effect depend on why data are missing and how the analysis handles them. Since the record does not describe missingness or its handling, the impact cannot be assessed.

Likewise, no analysis model, covariate adjustment, sensitivity analysis, or interim-analysis procedure is reported. The record does not state whether any formal comparisons were planned among dose groups. This absence is particularly relevant because the study is non-randomized and its groups differ by regimen and dose; descriptive differences should not be interpreted as adjusted or causal comparisons.

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 0% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: adrenal insufficiency (100.0%), alanine aminotransferase increased (100.0%), colitis (100.0%), dehydration (100.0%), diarrhea (100.0%).

The registry lists serious and non-serious adverse-event counts by group. The group entries give the number at risk and the number affected. For serious events, the affected counts are 2, 0, 1, 2, 13, 2, 1, 2, and 1 across the listed groups; the corresponding at-risk counts are 2, 1, 2, 2, 18, 2, 1, 2, and 1. For non-serious events, the affected counts are 2, 1, 2, 2, 18, 2, 1, 2, and 1, with the same at-risk counts as listed for those groups.

These are descriptive group-level entries, not randomized comparisons. The regimens and doses differ, and the groups are small. The record also lists 1 death in the group titled Arm C Triplet 2-Recombinant Human Interleukin(rHIL-15) 1mcg/kg/Nivolumab 240mg/Ipilimumab 1mg/kg; it does not report whether that death was attributed to treatment. The serious-event terms displayed as 100.0% include adrenal insufficiency, alanine aminotransferase increased, colitis, dehydration, diarrhea, and glucose intolerance. Their percentage denominator is not reported in the registry record, so those percentages should not be interpreted as incidence among all participants.

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

The registry record supports a description of the study question and design: a phase I, non-randomized evaluation of rhIL-15 with nivolumab, ipilimumab, or both in people with refractory cancers. The primary endpoint was DLT count during Cycle 1 (42 days), with an attribution requirement and specified assessment groups. The posted record does not give the numerical primary endpoint result.

That absence means the record cannot answer how many DLTs occurred or how the primary safety endpoint varied across the named groups. It also does not provide an estimate of precision for that endpoint. The adverse-event entries do provide descriptive counts by group, but they answer a different question and do not replace a reported DLT result. The recorded safety counts should be interpreted alongside the different regimens, doses, and group sizes, not as evidence that one regimen is safer or less safe than another.

The design also limits what can be inferred about effectiveness. There was no randomized control group, and the listed groups were not like-for-like comparisons. The registry record does not report a comparative efficacy analysis or a numerical response outcome. Accordingly, the record does not establish a comparative benefit, a causal treatment effect, or how these regimens would perform in a broader population.

These limits do not make the registry information meaningless. It documents a clinical investigation of dose and safety questions, identifies its primary endpoint and assessment window, and records adverse-event counts. But a complete interpretation of the primary question would require the missing DLT results and relevant analysis details. The listed safety counts should remain descriptive, with no assumption that recorded events were caused by treatment unless the record explicitly says so.

13. Why This Trial Matters Statistically

This study raises statistical issues common to early-phase work: the endpoint is a safety count, the assessment period is defined, and participants are grouped by regimen and dose without random allocation. Reading those elements together is important. The primary endpoint is not a general count of all adverse events; it is a count of DLTs meeting attribution and prespecified criteria during Cycle 1 (42 days), assessed in specified lead-in and triplet groups.

The record also illustrates why endpoint definitions and results must be considered separately. A detailed definition can clarify what would count as a DLT, but it cannot show how often DLTs occurred. The registry gives the definition but not the numerical result. The separately listed serious and non-serious adverse-event counts contribute descriptive safety information, yet they are not interchangeable with the DLT endpoint.

Non-randomized dose groups present another statistical challenge. In general, differences in outcomes may reflect differences in participant characteristics, treatment combinations, or dose as well as treatment effects. The posted groups are not randomized controls, so direct contrasts do not isolate the effect of any drug or dose. Small group counts also mean that descriptive numbers should not be treated as stable estimates of risk without uncertainty information.

Finally, the record lists no analyses and does not report a statistical analysis plan, missing-data approach, or multiplicity procedure. Those omissions restrict the ability to assess how results were intended to be interpreted. The record is therefore most useful as a description of design and reported safety entries, rather than as a basis for inferential comparisons.

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 most consequential limitation for interpreting the primary question is that the registry record identifies the DLT endpoint but does not report its numerical result. It also does not report confidence intervals, a group-specific DLT breakdown, or the analysis population beyond stating which phases and arms were assessed. The full prespecified DLT criteria are not included, so the endpoint cannot be independently reconstructed from the registry description.

The study was non-randomized, unmasked, and single-group, and the treatment groups differ in combination and dose. There is no like-for-like randomized control comparison. The safety counts are therefore descriptive and cannot establish causal differences among regimens. The listed group counts are small, and the record gives no precision estimates for their adverse-event results.

The adverse-event entries do not explain whether events were considered related to study drugs, apart from the primary DLT definition’s attribution language. The record does not explain the denominator behind the serious-event term percentages or whether the listed death was treatment-related. It does not report details of missing observations, withdrawals, analysis populations, dose-escalation rules, interim analyses, or the statistical plan. These limitations leave important questions about how the outcomes were collected and analyzed unresolved.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Recombinant Interleukin-15 in Combination With Checkpoint Inhibitors Nivolumab and Ipilimumab in People With Refractory Cancers — NCT03388632.
• Registry record last updated 2025-12-18; status last verified 2025-12.
• Results first posted 2025-12-03.
• 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.

18. Record Summary

NCT03388632 is a completed phase I, non-randomized, unmasked, single-group study sponsored by the National Cancer Institute (NCI). It evaluated recombinant human interleukin-15 with nivolumab, ipilimumab, or both in people with refractory cancers. Enrollment is recorded as 31; the study start date is 2018-02-05 and primary completion is recorded as 2024-11-27.

The primary endpoint was the number of DLTs possibly, probably, or definitely related to study drugs during Cycle 1 (42 days), assessed in the lead-in doublet and triplet escalation groups. The numerical primary endpoint result is not reported in the registry record. Group-level serious and non-serious adverse-event counts are recorded, but they are descriptive and do not provide a randomized comparison. The registry lists four outcome measures and no analyses.