CAMPFIRE: A Study of Abemaciclib (LY2835219) in Participants With Ewing's Sarcoma: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for CAMPFIRE: A Study of Abemaciclib (LY2835219) in Participants With Ewing's Sarcoma (NCT05440786), a randomized phase 2 study. The primary outcome is Progression Free Survival (PFS) Assessed by Blinded Independent Review Committee (BIRC) by Bayesian Analysis.

2. Clinical Question

The registry identifies the study as CAMPFIRE, a randomized, open-label, phase 2 study in participants with relapsed or refractory Ewing’s sarcoma. Its stated purpose is to measure the benefit of adding abemaciclib to irinotecan and temozolomide. The clinical question is therefore about the comparative outcome of these regimens in this trial population, not about abemaciclib in every setting or for every person with Ewing’s sarcoma.

The primary comparison uses progression-free survival rather than response alone. This endpoint can capture either documented tumor progression or death, so it reflects more than whether a tumor shrinks on assessment. The registry also lists response, duration of response, and overall survival among secondary outcomes. Those outcomes address related but distinct questions and should not be treated as interchangeable with the primary endpoint.

3. Trial Design

CAMPFIRE is a randomized, open-label, parallel-group phase 2 study sponsored by Eli Lilly and Company. The registry lists enrollment of 46 participants. Participants were assigned to abemaciclib plus irinotecan and temozolomide or to irinotecan and temozolomide. Because the study is open-label, the treatment assignment was not masked to participants or investigators. The primary progression-free survival assessment is nevertheless attributed to a blinded independent review committee, a feature that separates this endpoint assessment from the open-label treatment setting.

The registry lists the study start as 2022-09-20 and primary completion as 2025-01-15. Its current status is active, not recruiting. The trial is part of the CAMPFIRE master protocol, described in the registry as a platform intended to speed development of treatments for children and young adults with cancer. The record does not provide enough detail to infer how this particular comparison relates operationally to other parts of that platform.

The two regimens share irinotecan and temozolomide, with abemaciclib added in the experimental group. The arm descriptions state that treatment continues until disease progression, unacceptable toxicity, or a discontinuation criterion is met. The full dose instructions are not reported in the registry record available for these descriptions, so no complete dosing schedule should be inferred.

Randomization supports a comparison of assigned treatment strategies by reducing systematic differences between groups at allocation, although it does not guarantee that all characteristics are balanced in a trial of this size. The available record identifies the randomized groups and enrollment but does not report the randomization procedure, allocation ratio rationale, stratification variables, or a sample-size justification. These design details 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.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: All randomized participants (including the censored participants). Number of participants censored in Abemaciclib + Irinotecan +Temozolomide =8, Irinotecan +Temozolomide = 3.

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: All randomized participants.

6. Primary Endpoint

The registry lists the primary outcome as Progression Free Survival (PFS) Assessed by Blinded Independent Review Committee (BIRC) by Bayesian Analysis, measured over From Date of Randomization until Disease Progression or Death Due to Any Cause (Up to 22.36 months).

The primary endpoint is progression-free survival assessed by a blinded independent review committee. It is measured from randomization until the first documented disease progression under RECIST v1.1 criteria or death from any cause in the absence of documented progression, whichever occurs first. Participants known to be alive without progression are censored at their last adequate tumor assessment or randomization date, whichever is later.

Progression is defined in the registry using changes in target lesions, unequivocal progression of non-target lesions, or the appearance of new lesions. This endpoint is used because it combines tumor progression and death into a time-to-event outcome. It can therefore capture disease worsening even when a participant has not died, while retaining death as an event rather than treating it as a separate outcome.

The primary outcome is described as a posterior mean with an 80% credible interval, estimated using Bayesian analysis. The registry also reports a Bayesian hierarchical model comparison as a posterior mean hazard ratio. The posted arm-level progression-free survival means and their intervals are missing, so the primary comparison cannot be reconstructed from arm-level summaries in the record. The reported hazard ratio is the available comparative estimate.

Time-to-event outcomes require attention to censoring: a censored participant contributes follow-up information up to the censoring time but is not recorded as having experienced the event during observed follow-up. This approach depends on assumptions about censoring and follow-up. The registry gives the censoring definition and counts, but further details about missing assessments and the handling of other missing data are not reported in the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary result is a Bayesian time-to-event comparison. It is reported as a posterior mean hazard ratio from a Bayesian hierarchical model, together with an 80% credible interval. A hazard ratio compares the modeled event hazards between groups over follow-up. Here the event is progression or death. A value below 1 means the estimated hazard is lower in the abemaciclib combination group relative to chemotherapy alone, according to the stated comparison direction. It is not a ratio of median survival times and does not directly describe the probability that a particular participant will experience an event.

What does a Bayesian credible interval express?

In general, a Bayesian credible interval describes a range containing a specified share of the posterior probability for a parameter, conditional on the model, data, and prior assumptions. The primary result’s 80% interval runs from 0.317 to 0.783. Its placement below 1 indicates that the posterior estimate is concentrated on hazard ratios favoring the combination, under the fitted analysis. The interval does not remove uncertainty, and its interpretation depends on the Bayesian model and priors. The prior distributions and model diagnostics are not reported in the registry record.

This differs from the usual frequentist confidence interval, which is interpreted through the long-run behavior of an interval-generating procedure under repeated sampling. The registry reports 95% confidence intervals for some secondary frequentist outcomes. Those intervals should be interpreted in that frequentist framework, not as posterior probability statements. In general, intervals describe estimation uncertainty; a wider interval indicates less precision than a narrower one, all else equal.

Why use a time-to-event endpoint?

Progression-free survival records the time until progression or death rather than classifying each participant only as having or not having an event by a fixed assessment. Participants can have different lengths of observation. The registry accounts for participants who remain alive without documented progression by censoring them at their last adequate assessment or randomization date, whichever is later. Censoring retains information from observed follow-up without treating an unobserved future event as if it had occurred. Valid interpretation requires assumptions about the relationship between censoring and the event process; the registry does not report additional analyses of those assumptions.

For a time-to-event comparison, a hazard ratio summarizes relative event rates over time under the model. It can be difficult to interpret as a single constant contrast if the relative hazards change during follow-up. The record does not report a proportional-hazards assessment or further details on how the Bayesian model handles time-varying effects. The posted hazard ratio should therefore be understood as the analysis’s model-based summary, not as proof that the relative hazard is constant throughout follow-up.

How do the secondary frequentist analyses differ?

The investigator-assessed progression-free survival and overall survival comparisons use a log-rank method and report hazard ratios with 95% confidence intervals and p-values. The log-rank test compares event-time experience between groups, while the hazard ratio expresses the estimated relative hazard. The reported p-value evaluates compatibility with a specified null hypothesis under the test’s assumptions; it is not the probability that the null hypothesis is true. A confidence interval that includes 1 indicates that the data are compatible with the null hazard ratio as well as a range of effects in either direction.

The response outcome is binary: participants are classified by whether they achieved a complete or partial response under the stated assessment. The registry reports proportions and 95% confidence intervals, along with a recalculated unadjusted risk difference and other unadjusted measures. A risk difference compares proportions by subtraction, while a risk ratio compares them by division. These measures summarize the observed proportions but do not adjust for covariates or account for the trial’s full analysis plan. The recalculated p-value is an unadjusted calculation from posted summary information, not a replacement for a sponsor-reported analysis.

Why can over-dispersion matter?

In general, over-dispersion occurs when observed variability is greater than a model assumes. If a model treats outcomes as less variable than they are, uncertainty can be understated, which can make intervals too narrow and tests too optimistic. This concept is relevant to statistical modeling broadly, but the registry does not report that over-dispersion was assessed or identified in this trial. It should not be assumed to have affected the posted results.

The sample-size illustrations use the observed effect and specified planning assumptions: a two-sided significance level of 0.05, equal allocation, and a normal approximation. For time-to-event outcomes, the illustration concerns total events; for other outcomes, it concerns participants per arm. These are planning calculations, not observed power and not evidence about the trial’s original design. The registry does not report the trial’s own sample-size justification or whether these calculations were used in planning.

What do the missing summaries prevent?

The primary endpoint is specified as a posterior mean with an 80% credible interval, but arm-level values are missing. Without those summaries, readers cannot inspect the primary endpoint separately within each group from the posted record. The hazard ratio remains the reported comparative estimate. Missing arm-level summaries do not change the reported model result, but they limit independent checks and make it harder to understand the absolute event-time experience in each group.

10. Analyses Reported by the Sponsor

The registry’s primary progression-free survival comparison is the sponsor-reported Bayesian hierarchical model analysis. It gives the posterior mean hazard ratio and 80% credible interval, and no p-value is reported for this primary analysis. The arm-level primary progression-free survival means are missing, so there is no posted pair of arm summaries from which to make an independent reconstruction of the primary comparison.

Some secondary outcomes include sponsor-reported frequentist analyses, such as log-rank comparisons for investigator-assessed progression-free survival and overall survival. The recalculated binary response comparison is an unadjusted calculation from the posted response summaries, rather than a sponsor-reported inferential analysis. It does not account for a covariate-adjusted model or a broader multiplicity strategy.

The registry does not flag a discrepancy between a sponsor analysis and a recalculation. Where a sponsor’s analysis is reported, that analysis is the relevant trial analysis; a simple recalculation can serve only as a check based on the available summary statistics and assumptions. It cannot reproduce analyses that depend on participant-level data, model specifications, or adjustment procedures not reported in the registry record.

11. Multiplicity, Missing Data and Analysis Populations

The registry lists a primary endpoint and multiple secondary outcomes, including several versions of response and duration outcomes assessed by different reviewers. When multiple hypotheses are tested, the chance of at least one apparently favorable result can increase. Statistical plans often address this with a prespecified testing sequence or another multiplicity adjustment. Such a strategy is not reported in the registry record, so the secondary p-values should not be interpreted as if family-wise error or another error rate were controlled.

The primary analysis is Bayesian and reports a posterior mean hazard ratio with an 80% credible interval. The record does not report prior distributions, sensitivity analyses, or a broader decision rule for interpreting that posterior result. It also does not report interim analyses or stopping rules. These details are not reported in the registry record, which prevents evaluation of how the final inference relates to any planned monitoring or decision process.

The primary endpoint population is described as all randomized participants, including censored participants. The registry reports censoring counts of 8 in the abemaciclib combination group and 3 in the chemotherapy group. Censoring is part of the defined time-to-event analysis, not equivalent to excluding participants. The registry explains when participants without documented progression are censored, but further information about missing tumor assessments, reasons for missingness, and sensitivity analyses is not reported in the registry record.

The registry does not provide a complete account of the analysis populations for every secondary outcome. It also does not report how missing values were handled for those outcomes. Readers should not assume that all outcomes use identical populations or missing-data procedures. The primary population description is clear about inclusion of censored participants, but that does not establish the analysis set for every other endpoint.

The trial’s own sample-size justification is not reported in the registry record. The posted planning illustrations are based on observed effects and stated assumptions; they are not evidence of observed power, and they do not establish that the trial was designed around those assumptions. The record also does not report whether the analysis plan was amended or how any amendments were handled.

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 0% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: pyrexia (13.3%), abdominal pain (10.0%), diarrhoea (10.0%), abdominal pain upper (6.7%), catheter site infection (6.7%).

The registry reports serious adverse events in 12 of 30 participants at risk in the abemaciclib combination group and 5 of 15 at risk in the chemotherapy group. It also reports other adverse events affecting 30 of 30 and 15 of 15 participants at risk, respectively. These counts use the adverse-event denominators in the record; they should not be substituted with the randomized group sizes or interpreted as adjusted comparisons.

The registry lists pyrexia at 13.3%, abdominal pain at 10.0%, diarrhoea at 10.0%, abdominal pain upper at 6.7%, catheter site infection at 6.7%, and clostridial infection at 6.7% among the leading serious adverse-event terms. The record does not provide an arm-specific allocation for these listed percentages in the information, so they should not be attributed to one group.

Deaths are reported as 15 in the abemaciclib combination group and 8 in the chemotherapy group. These are counts, not a stated assessment of treatment-related deaths. The registry information does not establish cause or attribution, and the figures alone do not permit a causal safety comparison. Adverse-event counts should be interpreted with their denominators, follow-up, severity definitions, and attribution rules in mind; additional details are not reported in the registry record.

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

The primary result is an estimated treatment-group contrast for progression-free survival under a Bayesian hierarchical model. Its posterior mean hazard ratio and credible interval favor the abemaciclib combination on the defined endpoint. That is a statistical finding from this randomized phase 2 study, not a guarantee of benefit for every participant or a direct measure of how long an individual will live without progression.

The result does not establish an effect on overall survival. Overall survival is a separate secondary endpoint, with its own estimate and uncertainty. Nor does the primary endpoint alone establish that tumors respond more often, that responses last longer, or that treatment is safer. Those outcomes require their own evidence and are not uniformly aligned with the primary comparison in the posted record.

A randomized design strengthens the comparison of assigned regimens, but interpretation remains bounded by the trial population, follow-up, outcome definitions, and uncertainty. The study concerns relapsed or refractory Ewing’s sarcoma and compares regimens that share irinotecan and temozolomide. It does not test abemaciclib as a standalone treatment in this record, and it does not answer questions about other diseases or treatment settings.

The primary arm-level summaries are missing, and the record does not report several details needed to fully assess modeling assumptions, missing data, multiplicity, and sample-size planning. These limitations do not invalidate the reported estimate, but they constrain independent evaluation and the scope of conclusions. The primary result is best understood as evidence about the specified comparison and endpoint, with uncertainty, rather than as a complete account of the treatment’s overall effects.

No medical decision should be based on the statistical summary alone. The registry results describe group-level outcomes under the study’s design and analysis; they do not predict an individual outcome. The available evidence also does not support a claim of equivalence when secondary intervals are broad or include the null. Uncertainty is not proof of no effect, just as a favorable estimate is not proof of a definitive effect in every context.

14. Why This Trial Matters Statistically

This trial raises a statistical question that is important in small phase 2 studies: how to interpret a primary signal when different outcomes and analysis methods provide a less uniform picture. The primary endpoint uses a Bayesian hierarchical model and a blinded independent review committee, while some secondary time-to-event comparisons use frequentist log-rank analyses and investigator assessment. These approaches answer related questions, but their estimates and uncertainty summaries are not interchangeable.

The primary analysis’s interval is entirely below the null hazard ratio, while secondary analyses have broader uncertainty and do not provide a consistently favorable pattern. That contrast makes transparency about the primary model, analysis populations, censoring, and multiplicity especially valuable. The registry record provides the primary estimate and endpoint definition but does not report the priors, model diagnostics, multiplicity strategy, or full missing-data approach. Those omissions limit how fully readers can assess robustness.

The trial also illustrates why endpoint assessment matters. Open-label treatment does not mean every outcome must be assessed without blinding: the primary progression-free survival endpoint is assigned to a blinded independent review committee, while other endpoints include investigator assessments. Differences in assessment can affect how outcomes are determined and should be kept in view when comparing results.

Finally, statistical significance and clinical importance are distinct questions. A model-based hazard ratio describes a relative event hazard, not the absolute experience of participants in each group. The missing primary arm-level summaries make that distinction particularly relevant. The study’s findings can inform further evaluation of the regimen, but the registry record alone does not settle the broader clinical question.

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 outcome is missing its arm-level posterior means and credible intervals. The reported comparative hazard ratio remains available, but the absence of group-specific summaries limits interpretation of absolute progression-free survival experience and independent checking of the primary comparison. The registry also does not report the Bayesian priors, full model specification, diagnostics, or sensitivity analyses.

The study is open-label, although the primary progression-free survival assessment is attributed to a blinded independent review committee. The record does not describe the committee’s procedures or the blinding process. Other outcomes include investigator assessment, which takes place in an open-label design. Potential assessment differences should be considered without assuming that bias occurred.

Several design and analysis details are not reported in the registry record: the sample-size justification, randomization procedure, allocation rationale, stratification factors, multiplicity strategy, interim analyses, stopping rules, and detailed missing-data procedures. The record also does not explain whether the hazard ratio assumptions were assessed or how potential changes in relative hazards over time were handled.

The secondary results are heterogeneous in endpoint, assessor, and method. The registry does not report a multiplicity adjustment, making secondary p-values less suitable as confirmatory evidence. The recalculated response comparison is unadjusted and based on summary data; it cannot reproduce participant-level or adjusted analyses. The planning illustrations likewise rely on stated assumptions and observed effects, and they do not represent the trial’s design justification or observed power.

Safety information is presented as counts and selected terms, with limited detail in the record about attribution, duration, or event timing. The death counts do not identify cause. These constraints mean that efficacy and safety conclusions should remain specific to the reported outcomes and available follow-up rather than being generalized beyond the study.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: CAMPFIRE: A Study of Abemaciclib (LY2835219) in Participants With Ewing's Sarcoma — NCT05440786.
• Registry record last updated 2026-10-20; status last verified 2026-10.
• Results first posted 2026-03-12.
• 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

CAMPFIRE is a randomized, open-label, parallel-group phase 2 study of abemaciclib added to irinotecan and temozolomide versus irinotecan and temozolomide for relapsed or refractory Ewing’s sarcoma. The registry lists enrollment of 46, sponsor Eli Lilly and Company, and status active, not recruiting.

The primary endpoint is BIRC-assessed progression-free survival analyzed using a Bayesian approach. The reported posterior mean hazard ratio is 0.530, with an 80% credible interval from 0.317 to 0.783. Primary arm-level summaries are missing. Secondary endpoints include investigator-assessed progression-free survival, overall survival, response, duration of response, and disease control. Their results do not form a uniformly favorable pattern.

The record does not report the multiplicity strategy, sample-size justification, full Bayesian model specification, or detailed missing-data procedures. The findings therefore describe the posted outcomes and analyses, with important limits on independent assessment and generalization.