Safety, Tolerability, and Efficacy of Doxorubicin and Pembrolizumab for Sarcoma: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Safety, Tolerability, and Efficacy of Doxorubicin and Pembrolizumab for Sarcoma (NCT03056001), a na phase 2 study. The primary outcome is Number of Participants With at Least One Severe or Life-Threatening Adverse Event.

2. Clinical Question

The registry also lists overall survival, progression-free survival, objective response, and duration of response as secondary outcomes. These outcomes address disease course and tumor response, while the primary endpoint is a safety measure. The record’s stated purpose includes evaluating safety, tolerability, and efficacy, but the clinical question that can be interpreted most directly from the primary endpoint is whether participants experienced a qualifying adverse event after beginning the combination. The available record does not establish what would have occurred under another treatment or without treatment. The primary endpoint’s definition also matters: a general serious-event count should not automatically be treated as identical to the endpoint, because the endpoint additionally requires treatment relatedness and clinical significance as judged by the Sponsor-Investigator.

3. Trial Design

This is a completed phase 2, single-group study sponsored by Wake Forest University Health Sciences. The registry lists enrollment as 30. The study began on 2017-05-01 and its primary completion date is 2021-11-16. The registry describes the population as patients with metastatic or unresectable soft tissue sarcoma and lists both adult and child soft tissue sarcoma among the conditions.

All participants were assigned to the combination of pembrolizumab and doxorubicin. The arm description states that pembrolizumab was given by IV infusion and doxorubicin by IV injection on Day 1 of every 21 (+/- 3) days. The design is not randomized, and masking is listed as NONE. The allocation field is NA. Thus, the record describes treatment in a single group, not a randomized comparison between the combination and a control treatment.

The primary endpoint is the number of participants with at least one severe or life-threatening adverse event meeting the registry’s definition. The observation period runs from enrollment to at least 90 days following cessation of study treatment. The registry reports a median time on treatment of 5.8 months. These details define the period and the event criteria relevant to the endpoint, but the record does not report additional details about visit schedules, exposure management, or how events were adjudicated beyond the Sponsor-Investigator’s stated judgments.

The registry lists overall survival, progression-free survival, objective response, and duration of response as secondary outcomes. The record reports 5 outcome measures and 5 analyses. The statistical plan, including the specific analyses attached to those outcomes, is not reported in the registry record. The completion status describes the study’s registry status; it does not by itself establish that every outcome has a posted numerical result. Because there is no control arm, results from this design are descriptive of the treated group and cannot isolate the contribution of pembrolizumab from that of doxorubicin or other influences.

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: The evaluable population will consist of all subjects who begin study therapy.

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 Participants With at Least One Severe or Life-Threatening Adverse Event, measured over From enrollment to at least 90 days following cessation of study treatment. The median time on treatment was 5.8 months..

The primary endpoint counts participants who experienced at least one severe or life-threatening adverse event satisfying the registry’s full definition. An event must be a serious adverse event per CFR 21 Part 312, be considered related to study treatment by the Sponsor-Investigator, and be considered clinically significant by the Sponsor-Investigator. Relatedness may be attributed to pembrolizumab, doxorubicin, or both. The endpoint is therefore a participant-level binary measure: each subject is classified according to whether any qualifying event occurred, rather than by the total number of events.

The observation window extends from enrollment to at least 90 days following cessation of study treatment. That window is important because safety events can occur during treatment or after it ends. The registry also reports a median time on treatment of 5.8 months, which provides context for exposure but is not itself a safety outcome or a follow-up duration for every participant.

A participant-level endpoint is used to describe how many people experienced at least one event meeting the prespecified definition. It avoids treating multiple events in the same person as if they came from different people. However, the count alone does not convey event timing, duration, recurrence, severity within the stated categories, or individual clinical circumstances. Nor should an aggregate serious adverse-event count be assumed to match the primary endpoint automatically: the primary definition additionally requires treatment relatedness and clinical significance. Separate endpoint-specific results are not reported in the registry record beyond the listed safety figures, so the extent to which those figures satisfy every primary-endpoint criterion is not clear from the record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

A participant-level event count is a binary outcome: each participant either experienced at least one event meeting the endpoint definition or did not. The resulting count is easy to describe, but it does not retain the number of qualifying events per participant or their timing. A person with repeated events contributes to the participant count in the same way as a person with a single qualifying event. The registry’s primary endpoint is defined this way, using whether or not a participant experienced at least one qualifying event.

What does the endpoint definition add?

The criteria narrow the outcome beyond an unqualified record of adverse events. The event must be serious under the cited regulatory definition, judged related to one or both study drugs by the Sponsor-Investigator, and considered clinically significant by the Sponsor-Investigator. Those judgments are part of how the outcome is defined; they are not equivalent to a blinded, independent causal assessment unless such a process is reported. The registry record does not describe an independent adjudication process.

Because the endpoint includes investigator judgments, interpretation depends on consistent application of the criteria. A report of all serious adverse events can be informative, but it does not by itself show how many satisfy the added relatedness and clinical-significance requirements. The record lists serious-event counts, but does not provide a separate breakdown showing whether the primary endpoint’s complete definition was met for each event. This distinction is especially relevant when comparing an endpoint label with a broader safety summary.

What can a single-group result estimate?

In a single-group study, an event count describes what was observed among participants receiving the study combination. It can characterize the treated group over the registry’s stated observation window. It cannot directly estimate the difference between treatment and control, because there is no concurrent control group. Even a carefully measured event frequency could reflect the treatment, the underlying disease, participant characteristics, other care, or a combination of influences.

Random allocation, when used, helps balance measured and unmeasured characteristics between comparison groups on average. This study is not randomized and has no like-for-like control comparison in the registry record. Therefore, the posted safety figures should be read as descriptive rather than as a causal estimate of the combination’s effect on adverse events. The record does not report a comparative estimate for the primary endpoint.

How should uncertainty be understood?

A confidence interval is a general statistical method for expressing uncertainty around an estimate. It gives a range of values compatible with the data and the assumptions of the analysis; it is not a guarantee that the true value lies within that range. The registry entries used for this interpretation do not report a confidence interval for the primary safety result. Consequently, the count alone does not quantify its statistical precision. A reader should not infer precision from the number of participants or from a percentage without an interval or other uncertainty analysis.

For event counts, uncertainty can be substantial when information is limited, and the relevant interval depends on the analysis method and assumptions. The registry record does not state which method, if any, was used to calculate uncertainty for these safety figures. The event terms and percentages are descriptive summaries, not substitutes for a confidence interval or a comparative model.

What do the planning assumptions mean?

The registry entries describe a sample-size planning illustration based on the observed effect, not the trial’s own design. Its assumptions are a two-sided significance level of 0.05, equal allocation between two arms, a normal approximation, and an effect to detect defined as a stated fraction of the observed effect. Time-to-event illustrations use the total number of events under the Schoenfeld method; other illustrations express participants per arm. These are planning assumptions, not outcomes from the study and not evidence about the probability that the observed result is correct.

The planning setup is distinct from this study’s actual single-group design. Equal allocation between two arms is not the design described in the registry, which lists a single treatment group and no randomized comparison. Planning illustrations can show how assumptions about effect size, significance level, allocation, and analysis influence a hypothetical future study. They do not show that the present study was planned with those assumptions, nor do they establish that its enrollment was adequate for a particular endpoint. The sample-size justification and protocol-level statistical plan are not reported in the registry record.

Why can event data be harder to interpret than a count?

Adverse events may vary in timing, duration, seriousness, and relationship to study treatment. A participant-level count compresses these features into whether at least one event occurred. A time-to-event analysis can incorporate when an event occurs and can account for follow-up that ends before an event, but the registry entries do not specify that such an analysis was used for the primary safety endpoint. General statistical methods should not be mistaken for methods documented in the record.

For time-to-event outcomes such as overall survival or progression-free survival, censoring occurs when a participant’s event time is not observed through the end of follow-up. General methods for those outcomes require assumptions about censoring and event definitions. The registry lists these as secondary outcomes, but the analysis details, censoring rules, and numerical results are not reported in the registry record. No conclusions about those methods or outcomes can be drawn from their names alone.

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("Pembrolizumab + Doxorubicin")
n      <- c(30)
events <- c(1)
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  Pembrolizumab + Doxorubicin  30  1  3.3  0.1  17.2  0.6  16.7  0.8  16.7

11. Multiplicity, Missing Data and Analysis Populations

The registry lists 5 outcome measures and 5 analyses, including the primary safety endpoint and secondary outcomes for overall survival, progression-free survival, objective response, and duration of response. When several outcomes or analyses are assessed, the chance of at least one apparently noteworthy result can increase if each is tested separately without accounting for the family of tests. General approaches to multiplicity include prespecifying a primary outcome, adjusting significance thresholds, or treating secondary analyses as exploratory. The appropriate approach depends on the trial’s analysis plan.

The registry record does not report whether multiplicity was addressed, which outcomes were tested confirmatorily, how analyses were prioritized, or whether any adjustment was used. The existence of multiple listed outcomes is not itself evidence that an adjustment was or was not performed. Without the statistical plan, the strength of any inference across outcomes cannot be assessed from the registry information alone. No inferential claim should be made from the number of outcomes or analyses alone.

The registry defines the primary endpoint’s evaluable population as all subjects who begin study therapy. It does not report the number who began therapy, the number included in the endpoint analysis, or participant-level missingness for this outcome. The registry record also does not describe how missing safety assessments, incomplete follow-up, withdrawals, or events occurring after treatment cessation were handled. The stated observation window extends to at least 90 days following cessation of study treatment, but the record does not give the completeness of follow-up for that window.

The population definition does not resolve whether every enrolled participant began treatment or whether all who began therapy had complete endpoint data. Nor does it state how exclusions, if any, were handled. The record does not report an intention-to-treat analysis population or other analysis-population definitions for the secondary outcomes. These omissions limit assessment of potential bias from missing data and from differences between enrolled, treated, and analyzed participants.

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: febrile neutropenia (16.7%), nausea (13.3%), vomiting (10.0%), lung infection (10.0%), death nos (6.7%).

The registry lists 19 participants with serious adverse events among 30 at risk in the Pembrolizumab + Doxorubicin group. It also lists other adverse events for 30 participants among 30 at risk, and reports 24 deaths. These categories should be read as separate posted safety summaries. The record does not clarify in the registry record whether categories overlap, so they should not be combined or treated as mutually exclusive.

The serious-event terms listed are febrile neutropenia, nausea, vomiting, lung infection, Death NOS, and anemia, with percentages of 16.7%, 13.3%, 10.0%, 10.0%, 6.7%, and 3.3%, respectively. These are event-term percentages, not percentages to be summed into a total. The registry does not state whether a participant could contribute to more than one term. The term Death NOS does not specify cause or establish treatment relatedness.

The primary endpoint has additional criteria beyond seriousness: treatment relatedness and clinical significance must be judged by the Sponsor-Investigator. The listed serious-event count should therefore not be assumed to be the primary endpoint count without a separate endpoint-specific result. These safety figures describe recorded events in a single treated group; they do not establish comparative risk or causation.

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

The record describes a completed phase 2, single-group study of pembrolizumab with doxorubicin in patients with metastatic or unresectable soft tissue sarcoma. Its primary endpoint focuses on participants who experience at least one serious, treatment-related, clinically significant adverse event under the registry’s definition. The posted safety information gives counts for serious events, other adverse events, and deaths, and lists several serious-event terms with associated percentages. These figures provide a descriptive account of safety events recorded in the treated group.

The figures do not, by themselves, establish that the combination caused any particular event. The primary endpoint definition includes a treatment-relatedness judgment, but the listed serious-event count is not separately broken down by all endpoint criteria in the registry information. Deaths are reported, but the record does not identify the cause of each death or establish that deaths were related to treatment. Event-term percentages also do not reveal whether terms occurred in the same participants or how severe, prolonged, or recurrent each event was.

The design does not provide a randomized control group. As a result, the study cannot make a direct like-for-like estimate of how the observed event frequency compares with another treatment or with no study treatment. The same limitation applies to attributing outcomes to either drug individually: both drugs were administered together, so the contribution of each cannot be isolated from these data. The absence of a comparator does not make the descriptive findings meaningless, but it restricts the questions they can answer.

The registry lists secondary outcomes concerning survival and tumor response, but numerical results for those outcomes are not included in the registry record. Their listing does not establish a response or survival effect. Likewise, the stated study purpose and completed status are not evidence of efficacy. The registry record does not report the statistical analysis plan, multiplicity handling, uncertainty intervals, or missing-data procedures. Those omissions make it difficult to assess how precisely the endpoint was estimated and how inferential claims should be calibrated.

In practical statistical terms, the record supports a description of the study design, treatment schedule, endpoint definition, follow-up period, and posted adverse-event summaries. It does not support a comparative safety conclusion, a causal attribution for all events, or an efficacy conclusion. Interpretation should remain tied to the specific definitions and limits of this single-group registry report.

14. Why This Trial Matters Statistically

This study illustrates why a safety endpoint’s exact definition matters. The primary outcome is not simply any adverse event or even any serious adverse event. It requires seriousness, a Sponsor-Investigator judgment of treatment relatedness, and a judgment of clinical significance. A general safety count may be useful context, but it cannot automatically stand in for a more narrowly defined endpoint. Statistical interpretation depends on matching the reported result to the outcome that was actually specified.

The design also highlights the difference between describing events and estimating comparative effects. A single-group study can record the safety experience of participants receiving a combination, but it lacks a randomized comparator that would help separate treatment effects from other influences. This distinction applies even when the event count is clearly reported: a count is evidence about what was recorded in the group, not a direct estimate of what would have happened under an alternative.

The listed secondary outcomes show that the registry considers safety alongside survival and response questions. But the presence of outcome names does not supply their results or indicate how multiple analyses were handled. Without the analysis plan, readers cannot determine from the registry record whether multiplicity adjustments, missing-data procedures, or prespecified analysis populations were used. Those details affect the interpretation and reliability of statistical claims.

The registry’s sample-size information is also useful as a reminder to distinguish planning illustrations from the actual design. The stated illustration assumes two-arm equal allocation, while the study is single-group. Its hypothetical planning assumptions should not be read as a sample-size justification for this trial or as observed power. Clear separation of design details, analysis assumptions, and outcome results is essential when translating a registry record into statistical conclusions.

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.

Several limitations are specific to interpreting this trial from the registry record. The design is single-group, with no randomized comparator. Consequently, the safety counts describe participants treated with the combination, but do not estimate comparative safety against another regimen. Because pembrolizumab and doxorubicin were administered together, the record cannot separate the contribution of either drug to observed events.

The primary endpoint includes criteria beyond the serious-event category. Although the registry defines the endpoint carefully, it does not report a separate endpoint-specific count demonstrating how many participants met every criterion. The available safety summary therefore cannot be assumed to match the primary endpoint exactly. Sponsor-Investigator judgments of relatedness and clinical significance are part of the definition, but the registry does not report additional adjudication procedures or reproducibility information.

The observation period is described as extending from enrollment to at least 90 days after treatment cessation, and median time on treatment is reported as 5.8 months. The record does not state the completeness of follow-up for each participant, the handling of missing event information, or whether participants had different observation durations. A participant-level count cannot convey event timing, recurrence, duration, or severity beyond the stated classifications.

The registry record does not report the primary endpoint’s confidence interval, a detailed statistical analysis plan, multiplicity handling, or the sample-size justification. It defines an evaluable population as subjects who begin study therapy, but does not report how many were included in that analysis or how missing outcomes were handled. Numerical results for the listed secondary outcomes are not reported in the information used. These gaps limit assessment of statistical precision, selection effects, and the strength of conclusions beyond the descriptive safety summaries.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Safety, Tolerability, and Efficacy of Doxorubicin and Pembrolizumab for Sarcoma — NCT03056001.
• Registry record last updated 2023-11-18; status last verified 2023-10.
• Results first posted 2023-02-08.
• 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

NCT03056001 is a completed phase 2, single-group study sponsored by Wake Forest University Health Sciences. The registry lists enrollment as 30 and describes patients with metastatic or unresectable soft tissue sarcoma. Participants received pembrolizumab IV infusion and doxorubicin IV injection on Day 1 of every 21 (+/- 3) days. The study began on 2017-05-01, with primary completion on 2021-11-16.

The primary endpoint counts participants with at least one severe or life-threatening adverse event that meets the registry’s seriousness, treatment-relatedness, and clinical-significance criteria. Its observation period extends to at least 90 days following cessation of study treatment; median time on treatment was 5.8 months. The posted safety figures list 19 participants with serious adverse events, 30 with other adverse events, and 24 deaths. The study has no randomized control group, and the registry information does not report a separate endpoint-specific count showing that the serious-event figures meet every primary-endpoint criterion.

Secondary outcomes include overall survival, progression-free survival, objective response, and duration of response. Their numerical results and the detailed statistical plan are not reported in the registry record. The safety figures are descriptive and do not establish comparative effects or causation.