Pilot Study With CY, Pembrolizumab, GVAX, and IMC-CS4 (LY3022855) in Patients With Borderline Resectable Adenocarcinoma of the Pancreas: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Pilot Study With CY, Pembrolizumab, GVAX, and IMC-CS4 (LY3022855) in Patients With Borderline Resectable Adenocarcinoma of the Pancreas (NCT03153410), a na early/phase 1 study. The primary outcome is Number of Patients With a Treatment-related Immunologic Effect.

2. Clinical Question

In statistical terms, the endpoint is a participant count within a defined analysis population, rather than a direct comparison between randomized groups. It addresses whether the specified tissue-based immune change was observed among eligible patients who had the required samples and surgery. It does not, on its own, identify whether the combination caused that change or how outcomes would compare with another approach.

3. Trial Design

This was a completed early-phase study with a single-group design. The registry lists the phase as EARLY PHASE1, the model as SINGLE GROUP, allocation as NA, and masking as NONE. Enrollment was 11. The sponsor was Sidney Kimmel Comprehensive Cancer Center at Johns Hopkins. The study start date was 2018-09-27, and the primary completion date was 2022-03-09.

The intervention group received cyclophosphamide, GVAX Pancreas Vaccine, pembrolizumab, and IMC-CS4. The registry describes the combination as neoadjuvant immunotherapy for patients with borderline resectable pancreatic adenocarcinoma. Because the design has one group and no listed comparator, the treatment outcomes are descriptive for the participants represented in the record. They cannot be read as a randomized treatment contrast.

The primary outcome was the number of patients with a treatment-related immunologic effect, assessed at 8 weeks. The registry also lists overall survival, disease free survival, immune-related objective response rate, surgical resectability rate, and pathologic response rate as secondary outcomes. It reports 7 outcome measures and 0 analyses. The registry record does not report a statistical analysis plan, the planned handling of multiple outcomes, or the timing and methods of any interim analyses.

The study is marked COMPLETED. That status describes the registry record and does not mean that every outcome has a posted result. In particular, the primary endpoint definition and analysis population are recorded, but the observed primary endpoint count is not reported in the registry record. The trial design therefore supports describing what was intended to be measured and what outcome records are present, but not estimating a treatment effect from 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.

Before assignment: 2 participants were excluded from analysis because they did not receive the study drug.

Overall Study

Analysis population for the primary outcome, as defined in the registry: 8/9 patients were included in the analysis because obtaining pre-treatment biopsy was not possible in one patient.

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 Number of Patients With a Treatment-related Immunologic Effect, measured over 8 weeks.

The primary endpoint counts patients who met the study’s definition of a treatment-related immunologic effect at 8 weeks. The definition requires an 80% or greater increase in CD8+ T cells in surgically resected tumor tissue compared with the baseline biopsy, and an increase of at least 1.8 times the baseline median absolute deviation. The assessment applies to subjects who received at least 1 dose of neoadjuvant combination immunotherapy and underwent an R0, R1, or R2 surgical resection.

This is a binary participant-level outcome: each eligible patient either meets the stated definition or does not. Reporting a count can describe how often the specified tissue change was observed in the analysis population. The baseline comparison and variability requirement are intended to make the threshold refer to a change relative to each patient’s pretreatment measurement, rather than to a post-treatment tissue value alone.

The registry says 8/9 patients were included because a pre-treatment biopsy could not be obtained in one patient. It does not report the observed count meeting the endpoint definition. Consequently, the endpoint’s result and its precision cannot be assessed from the registry record. The endpoint is also a biological measure; it is not itself a direct comparison of clinical outcomes between treatment groups.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is a count of participants meeting a binary definition. In general, a binary endpoint classifies each eligible participant into one of two categories, such as meeting or not meeting a prespecified criterion. A count can be interpreted alongside its denominator; without both an outcome count and a clearly defined analysis population, the frequency and precision of the outcome cannot be assessed. The registry record gives the endpoint definition and analysis population but does not report the observed count.

What does the endpoint definition do?

The endpoint uses a pretreatment biopsy as a reference and a post-treatment tissue assessment in a surgically resected tumor. The stated threshold is an 80% or greater increase in CD8+ T cells, together with a requirement that the increase be at least 1.8 times the baseline median absolute deviation. In general, including a measure of baseline variability can help distinguish a change that is large relative to the patient’s baseline variation from a change that only meets a fixed percentage threshold. The registry does not report the distribution of the tissue measurements or the number meeting the definition.

The endpoint is restricted to participants who received at least 1 dose of neoadjuvant combination immunotherapy and underwent an R0, R1, or R2 surgical resection. This makes the population used for the endpoint narrower than a general description of all people enrolled could imply. In general, conditioning an analysis on having a required post-treatment procedure or specimen can affect which participants contribute data. The registry does not report enough information to characterize how the endpoint population compares with all enrolled participants.

How should a count be read?

A count describes the number of participants with the outcome in the population being analyzed. For a binary outcome, an uncertainty interval can describe how imprecisely the underlying outcome frequency is estimated. In general, intervals are wider when information is limited and narrower when estimates are more precise. The registry record does not report the primary outcome count or an interval, so neither the observed frequency nor its precision can be evaluated.

Counts for this single-group study are not treatment contrasts. A treatment contrast requires a comparison between groups or conditions, with the relevant groups and analysis defined. The registry lists a single-group model and no comparator. Thus, even if a primary endpoint count were reported, it would describe observations in the study group rather than quantify a difference attributable to the combination.

What can planning illustrations tell readers?

Planning illustrations are hypothetical calculations about the sample size that might be needed for a future comparison under stated assumptions. The registry description says the illustrations use the observed effect, a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation; the effect to detect is a stated fraction of the observed effect. For time-to-event outcomes, the illustrations give a total number of events; for other outcomes, they give participants per arm.

These are planning assumptions, not the design of this single-group trial. They do not show that the trial had a particular probability of detecting an effect, and they are not observed power. Their relevance depends on the assumed effect and model being suitable for a future study. Because the primary result is not reported in the registry record, the observed primary effect cannot be checked against those assumptions. A hypothetical equal-allocation comparison also differs from the registry’s single-group design.

What is not documented?

The registry record does not report the statistical analysis plan, the method used to estimate uncertainty for the primary outcome, or a prespecified approach to missing data. It also does not report whether covariate adjustment was planned or used, whether any interim analyses took place, or the criteria for stopping the study. These details should not be inferred from the endpoint definition or the study’s completed status.

The record lists 7 outcome measures and 0 analyses. This indicates that an analysis record is not listed in the registry information; it does not establish what analyses may have been performed outside the registry. Likewise, the absence of a reported primary estimate means the statistical properties of that estimate cannot be assessed. General statistical methods can explain how counts, intervals, missingness, and planning calculations are usually interpreted, but they cannot replace an unreported result or an unreported analysis plan.

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("Cyclophosphamide, GVAX, Pembrolizumab and IMC-CS4")
n      <- c(8)
events <- c(6)
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  Cyclophosphamide, GVAX, Pembrolizumab and IMC-CS4  8  6  75.0  34.9  96.8  40.9  92.9  40.0  92.5

11. Multiplicity, Missing Data and Analysis Populations

The registry lists multiple outcomes: the primary immunologic-effect count and secondary outcomes covering overall survival, disease free survival, immune-related objective response rate, surgical resectability rate, and pathologic response rate. It reports 7 outcome measures and 0 analyses. The record does not report how the outcomes were prioritized beyond identifying the primary endpoint, whether the secondary outcomes were formally tested, or whether any adjustment for multiple testing was planned or used.

In general, testing multiple outcomes can increase the chance that at least one result appears notable by chance. A prespecified multiplicity strategy can clarify how evidence is to be interpreted across outcomes. No such strategy is reported in the registry record, so it would be inappropriate to assume that the outcomes were adjusted, unadjusted, or interpreted under a particular statistical plan. The listed outcomes should be understood as registered measures, not as evidence that results were reported or that confirmatory testing occurred.

The primary endpoint population is described as 8/9 patients because a pre-treatment biopsy was not possible in one patient. Since the endpoint requires a baseline comparison, lack of that biopsy prevents the stated assessment for that patient. The registry does not report other missing-data methods, whether participants were excluded for other reasons, or how missing observations were handled for secondary outcomes. It does not report whether any sensitivity analyses were conducted.

In general, missingness can affect an estimate when participants with unavailable measurements differ from those with available measurements. The record does not provide enough detail to assess whether that concern applies to this endpoint or to quantify its effect. Nor does it report the full relationship between enrollment and each outcome-specific analysis population. These gaps limit assessment of how representative the analyzed participants are of everyone enrolled.

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: diarrhea (11.1%), abdominal pain (11.1%), alanine aminotransferase increased (11.1%), atrial fibrillation (11.1%), biliary anastomotic leak (11.1%).

The registry lists serious adverse events for the treatment group as 5 affected among 9 at risk, and other adverse events as 9 affected among 9 at risk. It also records 7 deaths. The record does not state in these figures whether the deaths were considered treatment-related, nor does it give their timing or cause. They should therefore be read as recorded counts, not as evidence that the combination caused the events or deaths.

The listed serious-event terms are diarrhea, abdominal pain, alanine aminotransferase increased, atrial fibrillation, biliary anastomotic leak, and fever. Each is accompanied by 11.1%. The registry record does not provide further detail in the listed information about the denominator or participant-level overlap for these terms. The percentages should not be used to infer that the terms occurred in separate people or to reconstruct unreported counts.

These safety figures are descriptive and come from a single-group study. Without a comparison group, the registry record does not show how the event counts compare with another treatment or background risk. It also does not provide the event timing, severity details beyond the serious-event classification, duration, or full attribution information needed for a fuller safety assessment.

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

The registry record establishes that the study evaluated a combination of cyclophosphamide, GVAX, pembrolizumab, and IMC-CS4 in patients with borderline resectable pancreatic adenocarcinoma. It identifies a tissue-based primary endpoint at 8 weeks and explains the threshold used to classify a treatment-related immunologic effect. It also states that 8/9 patients were included in the primary analysis because a pre-treatment biopsy was not possible in one patient.

However, the registry record does not report how many patients met the primary endpoint definition. Without that result, readers cannot determine the observed frequency of the specified immunologic effect or how precisely it was estimated. The secondary outcome results are also not reported in the registry record for this analysis. A completed status does not fill those gaps.

The single-group design also limits what any outcome can establish. A result in the treated group would be an observation, not a comparison against a control group. It would not show what would have happened to similar patients under a different treatment or no study intervention. Because the study used a multi-agent combination, a group-level outcome would not identify the effect of any single component.

The recorded adverse-event and death figures are important descriptive information, but they do not establish causation. The registry information does not provide a comparative safety estimate, and the selected event terms lack further context in the listed record. Interpretation should remain tied to the reported counts and their limitations.

Accordingly, the record can be used to understand the trial question, design, endpoint definition, analysis population, and some reported safety counts. It cannot support a conclusion about the primary endpoint result, a comparative benefit, or a causal explanation of the listed safety events. Planning illustrations, where available, address hypothetical future study requirements and should not be mistaken for observed clinical outcomes or observed power.

14. Why This Trial Matters Statistically

This trial is statistically informative as an example of how a narrowly defined biological endpoint depends on both the measurement rule and the population in which it can be assessed. The primary outcome is not simply a general immune response: it requires a baseline biopsy, a surgically resected tumor sample, and a specified increase in CD8+ T cells relative to baseline variability. Such definitions make the analysis population and availability of specimens central to interpretation.

The record makes that issue visible by reporting 8/9 patients in the primary analysis and noting that a pre-treatment biopsy could not be obtained in one patient. This illustrates a general challenge in studies where an endpoint requires paired biological samples: missing baseline material can prevent outcome classification, even when a participant is otherwise part of the study. The record does not explain all differences between enrollment and the endpoint population, so readers must keep the stated denominators distinct.

The design also illustrates the difference between describing an outcome and estimating a comparative effect. With a single group and no comparator, observed counts cannot provide a randomized contrast. This is especially important for a combination intervention, because the study does not separate the roles of its individual components.

Finally, the registry distinguishes an endpoint definition from an endpoint result. A detailed definition can clarify what investigators intended to measure, but it cannot substitute for a reported count and measure of uncertainty. The absence of the primary result means that statistical interpretation must focus on what can be established from the design and record, rather than on an unreported treatment effect.

15. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

The main limitation for interpreting the primary endpoint is that the registry record does not report the observed number of patients meeting its definition. Without that result, neither the outcome frequency nor its precision can be assessed. The record states an analysis population of 8/9 and explains that one patient could not have a pre-treatment biopsy, but it does not report further details about exclusions or missing measurements.

The registry lists enrollment of 11 but does not describe the full relationship between enrollment and the primary analysis population. It is not possible to infer why the other participants are not represented in the stated endpoint denominator. The record also does not report outcome-specific populations for the secondary measures or how missing data were handled for them.

The single-group design has no like-for-like control comparison. This prevents a comparative estimate from the registered design and limits causal interpretation of both efficacy-related outcomes and adverse events. The use of four intervention components further means that outcomes cannot be attributed to any one component based on this study record.

Additional statistical details are not reported in the registry record, including the analysis plan, uncertainty method, approach to multiplicity, and whether interim analyses occurred. The listed safety information also lacks event timing, duration, detailed attribution, and context for the deaths. These limitations do not invalidate the recorded definitions and counts, but they constrain what can be concluded from them.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Pilot Study With CY, Pembrolizumab, GVAX, and IMC-CS4 (LY3022855) in Patients With Borderline Resectable Adenocarcinoma of the Pancreas — NCT03153410.
• Registry record last updated 2025-01-09; status last verified 2024-12.
• Results first posted 2024-10-15.
• 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 record describes a completed EARLY PHASE1, single-group study in patients with borderline resectable adenocarcinoma of the pancreas. Enrollment was 11. The intervention combined cyclophosphamide, GVAX Pancreas Vaccine, pembrolizumab, and IMC-CS4. Allocation was listed as NA and masking as none.

The primary endpoint was the number of patients with a treatment-related immunologic effect at 8 weeks, defined using CD8+ T-cell change in resected tumor tissue relative to a baseline biopsy. The primary analysis included 8/9 patients because obtaining a pre-treatment biopsy was not possible in one patient. The observed primary endpoint count is not reported in the registry record.

The record lists secondary outcomes and selected adverse-event counts, but does not provide the results of the secondary outcomes in the registry record for this analysis. It reports 7 outcome measures and 0 analyses. The single-group design does not provide a control comparison, and the recorded safety counts do not by themselves establish causation.