PIPAC Nab-pac for Stomach, Pancreas, Breast and Ovarian Cancer: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for PIPAC Nab-pac for Stomach, Pancreas, Breast and Ovarian Cancer (NCT03304210), a randomized phase 1 study. The primary outcome is Maximally Tolerated Dose (MTD) of Abraxane.

2. Clinical Question

3. Trial Design

This is a phase I study of PIPAC with Abraxane for participants with peritoneal carcinomatosis associated with several advanced cancers. The registry lists a planned enrollment of 20. It identifies University Hospital, Ghent as the sponsor, gives a start date of 2017-09-16 and a primary completion date of 2020-05-06, and marks the study completed.

Five dose levels are listed: 35 mg/m², 70 mg/m², 90 mg/m², 112.5 mg/m², and 140 mg/m². Each arm description specifies administration every 4 weeks for 3 cycles. The posted primary outcome is the maximally tolerated dose (MTD), defined as the highest dose administered 3 times using PIPAC that does not cause unacceptable side effects. The assessment window is within 14 weeks of treatment start.

The registry labels allocation randomized and masking double, but labels the study model single group. It does not describe a separate control treatment, the randomization procedure, or how masking was maintained across dose groups. Those design fields therefore do not establish that this was a conventional randomized comparison of treatments. The record describes use of a time-to-event continual reassessment model (TITE-CRM) for dose optimization and says an initial design was followed until the first dose-limiting toxicity (DLT) occurred. It also says conservative prior DLT estimates were used for the original dose-escalation scheme.

The secondary outcomes include surgical morbidity, plasma concentration and area under the curve for Abraxane, histological response by peritoneal regression grading, and treatment-related neutropenia and decreased platelets. The registry lists six secondary outcomes. It does not report a separate statistical analysis for these outcomes in the posted results.

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

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 Maximally Tolerated Dose (MTD) of Abraxane, measured over Within 14 weeks of the start of the treatment.

The primary endpoint is the maximally tolerated dose of aerosolized Abraxane delivered by PIPAC. It is defined as the highest dose administered 3 times without unacceptable side effects, assessed within 14 weeks from the start of treatment. The endpoint is reported as a count of participants, rather than as a continuous measurement.

The registry defines DLT in advance using clinical and surgical criteria. These include any grade 3 or 4 non-hematologic toxicity, with stated exceptions; grade 4 thrombocytopenia; grade 4 neutropenia lasting more than 7 days or associated with fever; inability to perform more than one PIPAC because of toxicity; and a surgical complication of Dindo-Clavien grade IIIB or higher. These criteria give the dose-finding question a safety-based definition rather than relying on a laboratory measurement alone.

In general, a dose-finding endpoint is used to guide which dose level merits further study, not to establish a comparative treatment effect. Its interpretation depends on the DLT definition, observation window, and dose-assignment method. The registry record describes the endpoint and the TITE-CRM approach, but the posted primary outcome shows zero participants for every dose group and does not provide an MTD estimate. It is therefore not possible to identify a selected dose from the posted primary outcome alone.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does the TITE-CRM description mean?

The registry says the study used a time-to-event continual reassessment model (TITE-CRM) to optimize the balance between safety and efficacy. In general, continual reassessment methods use accumulating toxicity information to guide dose selection as a study proceeds. The time-to-event adaptation can account for incomplete observation: a participant who has not yet experienced a DLT may still be within the period in which one could occur. That is why follow-up time can matter in a dose-finding study. The record describes a 14-week DLT assessment window, but does not post the participant-level follow-up, the model inputs, or the fitted dose-toxicity estimates.

The registry also states that an initial design was followed until the first DLT occurred and that conservative a priori DLT estimates were used to calculate the original dose-escalation scheme. These are descriptions of the intended method. They do not provide the final model output, the dose assignments over time, or the rationale and numerical values for the prior estimates. Those details are not reported in the registry record.

Why define DLT criteria in advance?

The primary endpoint uses a prespecified definition of DLT within 14 weeks of the first PIPAC. As a general statistical principle, a definition set before outcomes are assessed makes the endpoint more reproducible and reduces the scope for selectively classifying events after results are known. Here, the criteria cover particular grades of non-hematologic toxicity, thrombocytopenia and neutropenia, inability to continue PIPAC because of toxicity, and serious surgical complications. The stated exclusions for some non-hematologic symptoms are part of that definition. A toxicity that appears clinically important is not automatically a DLT under this endpoint; whether it qualifies depends on the recorded criteria and assessment.

The outcome description defines MTD as the highest dose administered 3 times that does not cause unacceptable side effects. That definition links the dose level to tolerability across repeated treatment, rather than to a single administration. The posted result, however, is a zero participant count for every dose group, without an accompanying MTD value or explanation. As a result, the defined endpoint and the posted numerical values cannot by themselves be reconciled into a dose recommendation.

What can counts tell us?

The primary outcome is specified as a count of participants. A count is a straightforward description of how many participants meet an outcome definition, but it does not by itself describe the uncertainty around a risk estimate or distinguish between groups unless a clear denominator and comparison are also given. The arm entries provide group sizes, but no confidence intervals or model-based estimates are posted for the primary endpoint. In general, uncertainty intervals help show which values remain plausible given the observed data and the analysis method. Their absence means that precision cannot be assessed from the primary outcome entries.

The adverse-event record includes counts of participants at risk and counts affected. These are descriptive safety summaries. They are not interchangeable with the DLT endpoint because the adverse-event categories may not match the DLT definition, and the registry does not report a mapping between them. A serious adverse event may be relevant to dose selection, but it is not necessarily a DLT unless it meets the prespecified criteria.

How should comparisons across dose groups be understood?

The registry identifies five dose groups, but does not post an inferential comparison among them. In general, a dose-ranging pattern can be assessed by estimating event probabilities across ordered dose levels, with uncertainty that reflects the amount of data at each level. The TITE-CRM is described as the dose-finding approach, but its estimates are not available in the posted results. Comparing raw counts alone would not reproduce that model or account for differences in group size and follow-up.

The design fields also require care: allocation is labeled randomized, masking is labeled double, and the model is labeled single group. The registry does not explain these labels or document a treatment comparator. The dose groups should therefore be read as the listed exposure levels, not as evidence of a specific randomized comparison. A claim about relative effects would require a defined contrast and an analysis, neither of which is posted for the primary endpoint.

What do planning calculations mean?

In general, sample-size planning illustrations describe the number of participants that might be needed under stated assumptions about an effect, significance level, allocation, and statistical approximation. Such calculations are planning exercises, not observed trial results and not evidence that a study had a particular probability of detecting an effect. For this study, the registry record does not report the original sample-size justification. The posted primary outcome and model description also do not supply a completed estimate of the MTD or the uncertainty needed to interpret a planning calculation as a study finding.

The record lists an enrollment of 20 and dose-specific arm counts, but it does not explain any differences between those entries. No participant flow, analysis population definition, or account of missing outcome data is posted. These omissions constrain interpretation of both the primary endpoint and the safety summaries.

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("Abraxane 35 mg/m²", "Abraxane 70 mg/m²", "Abraxane 90 mg/m²", "Abraxane 112.5 mg/m²", "Abraxane 140 mg/m²")
n      <- c(2, 2, 3, 3, 10)
events <- c(0, 0, 0, 0, 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  Abraxane 35 mg/m²  2  0  0.0  0.0  84.2  0.0  65.8  0.8  70.8
ARM  Abraxane 70 mg/m²  2  0  0.0  0.0  84.2  0.0  65.8  0.8  70.8
ARM  Abraxane 90 mg/m²  3  0  0.0  0.0  70.8  0.0  56.1  0.6  60.2
ARM  Abraxane 112.5 mg/m²  3  0  0.0  0.0  70.8  0.0  56.1  0.6  60.2
ARM  Abraxane 140 mg/m²  10  0  0.0  0.0  30.8  0.0  27.8  0.2  28.5

11. Multiplicity, Missing Data and Analysis Populations

The registry lists six secondary outcomes alongside the primary MTD outcome. They cover surgical morbidity, Abraxane maximum plasma concentration and area under the curve, histological response by peritoneal regression grading, and treatment-related neutropenia and decreased platelets. The record lists 7 outcome measures and 6 secondary outcomes, but it does not describe a multiplicity adjustment, a hierarchy for testing outcomes, or which secondary outcomes were intended to support confirmatory conclusions. No secondary-outcome analyses are posted in the analysis record.

In general, when several outcomes or comparisons are examined, the chance of finding at least one apparently notable result can increase. Statistical approaches may control that risk, or outcomes may be treated as exploratory and interpreted cautiously. The registry record does not state which approach, if any, was used. Because the study is a phase I dose-finding trial, descriptive safety and dose-selection objectives may be central, but the record does not provide a detailed statistical analysis plan to establish how those objectives were prioritized.

Missing-data handling is not reported in the registry record. There is no posted account of participants who did not complete the treatment cycles or the 14-week assessment window, and no explanation of how incomplete follow-up would enter the TITE-CRM. The record does not define analysis populations or describe exclusions from the outcome summaries. These details matter because dose-finding methods can depend on the timing and completeness of follow-up, while simple counts depend on clear outcome ascertainment and denominators.

The registry’s design fields state randomized and double masked, while the model field states single group. The record does not explain the relationship among these labels or provide a prespecified analysis plan clarifying how the dose groups were to be handled. Accordingly, the posted information does not establish a multiplicity strategy, missing-data approach, or analysis population for interpreting the outcomes.

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: progression peritoneal disease leading to death (100.0%), redness (50.0%), intestine obstruction (50.0%), thrombocytopenia grade 3 (ctcae) (33.3%), surgical wound infection (20.0%).

The registry reports serious adverse-event counts by dose group, with both the number at risk and the number affected. The entries are 2 affected among 2 at risk at 35 mg/m²; 2 among 2 at 70 mg/m²; 2 among 3 at 90 mg/m²; 1 among 3 at 112.5 mg/m²; and 6 among 10 at 140 mg/m². These are descriptive counts in the registry record, not adjusted comparisons between doses.

The record separately lists deaths by dose group: 1 at 35 mg/m², 2 at 70 mg/m², 1 at 90 mg/m², 1 at 112.5 mg/m², and 1 at 140 mg/m². It does not establish from these counts whether a death was caused by treatment. The listed serious-event terms include progression of peritoneal disease leading to death, reported as 100.0%; redness, 50.0%; intestine obstruction, 50.0%; grade 3 thrombocytopenia, 33.3%; surgical wound infection, 20.0%; and anaphylactic shock during surgery, 10.0%. The record does not provide dose-specific denominators for those term percentages in the listed summary.

Serious adverse events and DLTs are not the same endpoint. The primary DLT definition has specific criteria and a defined observation window; the adverse-event summaries do not state which events met those criteria or how event timing related to treatment. These counts describe reported safety experience but do not, on their own, establish causation or an MTD.

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

The study addresses a dose-finding question: which dose of aerosolized Abraxane, given repeatedly by PIPAC, meets the registry’s definition of the maximally tolerated dose. The record describes a phase I design, five dose levels, a 14-week DLT window, and use of a TITE-CRM. It also reports adverse-event and death counts. These elements show the type of safety and dose-selection information the study was intended to collect.

The posted primary endpoint does not identify an MTD. It records zero participants in each dose group for the primary outcome, without explaining those values or providing a final dose-selection estimate. There is no posted primary analysis or uncertainty interval. It is therefore not possible to determine from these entries which dose, if any, met the stated MTD definition. Nor can the zero values be interpreted as evidence that no DLTs occurred: the registry does not clarify what the outcome values represent.

The adverse-event counts add descriptive context, but do not resolve the primary-outcome uncertainty. The record does not link individual serious events to the DLT criteria, establish attribution to treatment, or provide an analysis of safety by dose. Likewise, it does not post secondary-outcome analyses for pharmacokinetics, histological response, surgical morbidity, neutropenia, or decreased platelets.

These results do not establish comparative efficacy. The record does not describe an untreated or active comparator and does not post an efficacy comparison. The design labels of randomized and double masked do not, without an explanation of the study model and assignment process, establish a conventional randomized treatment comparison. The record also does not provide enough information to assess missing data, participant flow, or the precision of the dose-specific safety estimates.

In short, the registry documents the trial’s aim and several descriptive outcomes, but the posted primary result does not answer the MTD question in an interpretable way. The findings should be read as registry-reported counts and event summaries, not as proof of a selected dose, a treatment effect, or a causal safety conclusion.

14. Why This Trial Matters Statistically

Early-phase dose-finding studies pose a different statistical problem from trials designed primarily to compare clinical outcomes. Their central challenge is to learn about tolerability across dose levels while treatment and follow-up are accumulating. The registry’s use of TITE-CRM reflects that challenge: as a general class of approach, time-to-event dose-finding methods can incorporate the timing of toxicity and incomplete observation rather than requiring all participants to finish the full assessment window before every decision.

The trial also illustrates why the analysis method and posted results need to be read together. A description of a sophisticated dose-finding model is not itself evidence of the final dose selected. To understand the conclusion, readers need the model output, the observed DLT information, and an account of how incomplete follow-up was handled. Those details are not reported in the registry record, while the primary outcome lists zero participants at all five dose levels without an explanatory estimate.

Another statistical issue is the difference between an adverse-event summary and the endpoint used for dose selection. Serious adverse events are important safety observations, but a DLT endpoint is defined by explicit criteria and a specified observation window. Treating those summaries as interchangeable could misstate what the study measured. The record presents both types of information but does not connect them analytically.

The record also shows why consistency among design descriptors matters. Randomized allocation and double masking appear alongside a single-group model, and the available description does not explain how the labels fit together. Clear reporting of assignment, analysis populations, missing data, and dose-selection results would help readers distinguish a dose-escalation design from a comparative treatment trial. These reporting details are particularly important when the primary result is not self-explanatory.

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 interpretation is that the posted primary outcome does not yield an identifiable MTD. The registry records zero participants for each dose group but does not explain the meaning of those values or provide a final model-based estimate. There are no posted confidence intervals or other uncertainty measures for the primary endpoint. The record lists 0 analyses, so the stated TITE-CRM method cannot be evaluated from the posted analysis results.

The registry does not report participant flow, dose-specific follow-up duration, or how participants with incomplete follow-up were handled. Those omissions are important for a time-to-event dose-finding method and for interpreting event counts within the 14-week assessment window. The record also does not describe analysis populations or the handling of missing outcome data.

Interpretation of safety is limited by the lack of a reported link between serious adverse events and the prespecified DLT criteria. The record does not establish causal attribution for the reported deaths or provide a dose-specific analysis of the listed adverse-event terms. The percentages attached to the serious-event terms are reported, but their denominators are not stated alongside those terms.

The design description is also incomplete for comparative interpretation. Allocation is labeled randomized and masking double, while the model is labeled single group; the registry does not explain the assignment procedure or identify a comparator treatment. The original sample-size justification, detailed statistical analysis plan, multiplicity approach, and missing-data method are not reported in the registry record. These gaps limit reproducibility and make it difficult to distinguish planned analyses from descriptive reporting.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: PIPAC Nab-pac for Stomach, Pancreas, Breast and Ovarian Cancer — NCT03304210.
• Registry record last updated 2023-11-07; status last verified 2023-01.
• Results first posted 2023-11-07.
• 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 completed phase I study record concerns Abraxane administered by PIPAC for participants with peritoneal carcinomatosis associated with advanced ovarian, breast, stomach, or pancreatic cancer. It lists five dose groups, with treatment described every 4 weeks for 3 cycles, and a planned enrollment of 20. The primary objective was to identify the maximally tolerated dose within 14 weeks of treatment start, using stated DLT criteria and a TITE-CRM dose-finding approach.

The posted primary outcome records zero participants in each dose group, but does not explain the values or report an MTD estimate. Serious adverse events and deaths are listed by dose group, alongside several serious-event terms. The registry lists six secondary outcomes but no analysis results. Missing-data handling, analysis populations, multiplicity procedures, and the final dose-selection analysis are not reported in the registry record.