A Study of Tazemetostat in Adult Participants With Soft Tissue Sarcoma: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study of Tazemetostat in Adult Participants With Soft Tissue Sarcoma (NCT02601950), a na phase 2 study. The primary outcome is Cohorts 1, 3, 4, 5, 6 and 7: Objective Response Rate (ORR).

2. Clinical Question

3. Trial Design

This is a completed Phase II study with a registry enrollment of 267. Its structured design is identified as single group, with allocation listed as not applicable and masking listed as none. The intervention entry names tazemetostat, and the arm description states that all cohorts received 800 mg orally twice a day in continuous 28-day cycles. The record gives a start date of 2015-12-22 and a primary completion date of 2024-02-26.

The study spans several tumor cohorts rather than one clinically uniform population. The primary objective response endpoint is assigned to Cohorts 1, 3, 4, 5, 6 and 7. Cohorts 2 and 8 have objective response rate listed as a secondary outcome. Other listed secondary outcomes include duration of response, progression-free survival, overall survival, and disease control rate. These are named outcomes, not reported results.

There is a notable difference between the brief summary and the structured design entries. The brief summary describes a first part examining tazemetostat with doxorubicin and a later comparison of that combination against doxorubicin plus placebo. In contrast, the structured design describes a single group and lists only tazemetostat as an intervention. The registry record does not explain whether the brief summary describes an intended plan, a different study phase, or a design change. It is therefore not possible to treat the brief-summary description as evidence that a randomized comparator was conducted.

The registry lists an overall enrollment of 267, while the cohort-level safety denominators vary. Enrollment does not, by itself, establish how many participants were included in each efficacy analysis. The record does not report a cohort-by-cohort disposition of participants or explain the differences between overall enrollment and the reported safety denominators.

The primary endpoint observation period is described as beginning with screening and continuing with tumor assessments every 8 weeks beginning at Cycle 3 Day 1, until disease progression or the start of subsequent anti-cancer therapy, whichever occurred first, up to 5.75 years. The registry record does not report the protocol’s detailed analysis plan, allocation procedures beyond the not-applicable entry, or any statistical justification for the chosen enrollment.

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: Participants were enrolled into 1 of the 8 cohorts based on tumor type. A total of 267 participants were enrolled in the study.

Only the first 6 of 8 groups are shown.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: ITT population included all participants who received at least 1 dose of tazemetostat.

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: Intent-to-treat (ITT) population included all participants who received at least 1 dose of tazemetostat.

Only the first 6 of 9 groups are shown.

6. Primary Endpoint

The registry lists the primary outcome as Cohorts 1, 3, 4, 5, 6 and 7: Objective Response Rate (ORR), measured over Tumor assessment performed at screening and every 8 weeks beginning at Cycle 3 Day 1, until disease progression or the start of subsequent anti-cancer therapy, whichever occurred first, up to 5.75 years (cycle duration: 28 days).

The primary endpoint is objective response rate, expressed as the percentage of participants with a confirmed complete response or partial response. It is based on investigator review from treatment start until progression, subsequent anti-cancer therapy, or cancer-related surgery, whichever occurred first. The record specifies RANO criteria for primary brain tumors and RECIST version 1.1 for other solid tumors.

In the registry definition, complete response means disappearance of all target and non-target lesions, with pathological lymph nodes required to be <10 mm in short axis. Partial response means at least 30% reduction in the sum of target-lesion diameters relative to baseline. The definition of progression includes at least a 20% increase relative to the smallest sum recorded since treatment began, together with an absolute increase from nadir of 5 mm.

As a general statistical concept, response rate reduces complex tumor assessments to a binary outcome for each participant: response or no confirmed response. This can make the endpoint easier to communicate as a proportion, while discarding information about the size or duration of changes that do not meet response criteria. A confidence interval is useful because it conveys uncertainty around an estimated proportion; its width depends on the amount of information and the observed response pattern.

The endpoint is specified for Cohorts 1, 3, 4, 5, 6 and 7, and its population is defined as participants who received at least 1 dose of tazemetostat. The observed response percentage, the number of responders, and the 95% confidence interval are not reported in the registry record. The endpoint definition alone cannot establish how many participants responded or how precise an estimate would be.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

Objective response rate is a binary proportion: each participant is classified according to whether a confirmed complete or partial response was documented under the specified criteria. The proportion is calculated from the number classified as responders and the relevant analysis denominator. The registry record names the endpoint and defines its population, but does not report the responder count, denominator, or resulting proportion. No numerical response estimate should therefore be reconstructed from the enrollment figure or from unrelated safety counts.

What does a confidence interval describe?

In general, a confidence interval expresses sampling uncertainty around an estimate. For a response proportion, the interval gives context for how much the observed percentage might vary across comparable samples under the assumptions of the interval method. A narrower interval generally indicates greater precision, while a wider interval indicates less precision. It does not mean that every value within the interval is equally plausible, nor does it guarantee that the true value lies inside a particular calculated interval.

The primary endpoint lists a 95% confidence interval as its dispersion measure. However, the numerical interval and the method used to calculate it are not reported in the registry record. Different methods for proportions can behave differently, particularly when the number of observations is limited or the observed proportion is near a boundary. The record does not state which method was prespecified or used, so a reader cannot evaluate that choice or reproduce the interval.

Why does the analysis population matter?

The endpoint population is defined as all participants who received at least 1 dose of tazemetostat. That definition establishes a treatment-exposed population for the endpoint, but it is not the same as having a complete accounting of everyone enrolled. The record does not state how many participants received a dose in each primary endpoint cohort, whether any treated participants lacked an evaluable tumor assessment, or how missing response assessments were handled. Those details matter because the denominator and classification rules determine the estimated response percentage.

As a general principle, missing outcome information can affect a proportion if participants without an assessment are excluded, counted as nonresponders, or handled by another rule. Each approach answers a somewhat different question. The registry record does not report which approach was used for this endpoint. It also does not describe sensitivity analyses that might show whether conclusions change under alternative assumptions about missing assessments.

What does a single-group design allow?

The structured design is single group, with allocation listed as not applicable and masking listed as none. A single-group response percentage is a description of outcomes observed among participants treated in the study. It is not a randomized estimate of how outcomes differ between treatments. General causal interpretation requires a comparison that can address what would have happened under an alternative treatment; the structured design entries do not identify such a comparator. The registry record does not report an adjusted comparison or a statistical model that would create one.

Investigator response assessment is defined using RANO criteria for primary brain tumors and RECIST version 1.1 for other solid tumors. These criteria specify how lesion changes and response categories are assessed. Using defined criteria can make outcome classification more systematic, but the registry lists no masking and does not report independent central review. It does not state how disagreements, confirmation requirements, or unevaluable assessments were handled beyond describing confirmed responses and investigator review.

Why are the cohorts important?

The primary endpoint combines specified cohorts that include different tumor categories. A proportion across several cohorts can conceal variation among them, while separate cohort estimates can be less precise when the cohort has fewer participants. The registry record does not report the primary endpoint results by cohort or an overall result, so neither the pooled pattern nor variation across cohorts can be assessed. The total enrollment figure should not be treated as the number contributing to this endpoint.

General statistical methods for binary outcomes can also account for design features such as clustering or repeated observations when those features are relevant. The primary endpoint, however, is defined as a participant-level response proportion, and the record does not report a model or adjustment for such features. No particular statistical model, test, or covariate adjustment should be inferred from the endpoint definition alone.

Finally, sample-size planning and observed precision are distinct. Planning calculations use assumptions about a target effect and design to illustrate the amount of information sought; they do not establish the observed result or its precision. The registry record does not report the study’s sample-size justification, and it does not give the primary response estimate needed to assess precision from the observed data. Any planning illustration should not be interpreted as observed power or as evidence of the study’s findings.

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("Rhabdoid Tumors: Cohort 1", "Other INI1-negative: Cohort 3", "RMC: Cohort 4", "ES: Cohort 5", "ES Paired Biopsy: Cohort 6", "Chordoma: Cohort 7")
n      <- c(32, 32, 14, 62, 69, 18)
events <- c(3, 3, 0, 10, 11, 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  Rhabdoid Tumors: Cohort 1  32  3  9.4  2.0  25.0  3.2  24.2  3.4  24.3
ARM  Other INI1-negative: Cohort 3  32  3  9.4  2.0  25.0  3.2  24.2  3.4  24.3
ARM  RMC: Cohort 4  14  0  0.0  0.0  23.2  0.0  21.5  0.2  21.8
ARM  ES: Cohort 5  62  10  16.1  8.0  27.7  9.0  27.2  9.1  27.3
ARM  ES Paired Biopsy: Cohort 6  69  11  15.9  8.2  26.7  9.1  26.3  9.2  26.4
ARM  Chordoma: Cohort 7  18  1  5.6  0.1  27.3  1.0  25.8  1.3  26.0

11. Multiplicity, Missing Data and Analysis Populations

The registry lists 10 outcome measures, spanning response, duration of response, progression-free survival, overall survival, and disease-control outcomes. Objective response rate is designated as primary for Cohorts 1, 3, 4, 5, 6 and 7. Objective response rate is also listed as a secondary outcome for Cohorts 2 and 8. The record does not describe a multiplicity strategy, such as adjustment of significance thresholds or a hierarchy that would control the interpretation of multiple outcomes. It does not report a statistical analysis plan that explains how primary and secondary results were to be interpreted together.

The registry lists 0 analyses. This does not establish which analyses were performed outside the registry; it indicates that analysis entries are not reported in the registry record. No primary endpoint estimate, confidence interval, p-value, or cohort-specific response analysis is reported. Nor is a method for combining or separating results across the tumor cohorts described.

Missing-data handling is also not reported in the registry record. The primary endpoint population includes participants who received at least 1 dose, but the record does not say how participants without an adequate post-treatment tumor assessment were classified or whether any response assessments were missing. It does not describe a rule for unevaluable participants, imputation, censoring, or sensitivity analyses for missing response data.

These omissions matter because multiplicity concerns how many statistical questions are evaluated, while missing-data rules concern who contributes to each estimate and how incomplete outcomes are treated. Without those details and the results themselves, readers cannot assess the endpoint’s analysis population, reproducibility, or the effect of alternative analysis choices. No protocol-specific multiplicity or missing-data method should be assumed.

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: dyspnoea (21.4%), death (21.4%), pleural effusion (14.3%), pulmonary embolism (14.3%), respiratory failure (14.3%).

The registry reports cohort-level denominators and counts for serious adverse events, other adverse events, and deaths. For serious adverse events, the at-risk and affected counts are: Rhabdoid Tumors, Cohort 1, 32 and 18; Synovial Sarcoma, Cohort 2, 33 and 13; Other INI1-Negative, Cohort 3, 32 and 11; RMC, Cohort 4, 14 and 8; ES, Cohort 5, 62 and 25; ES Paired Biopsy, Cohort 6, 69 and 19; Chordoma, Cohort 7, 18 and 7; and ES 1600 mg, Cohort 8, 7 and 2. The other-adverse-event counts are also reported by cohort, as are death counts.

The record lists dyspnoea and death at 21.4% among selected serious-event terms, and pleural effusion, pulmonary embolism, respiratory failure, and pneumonia staphylococcal at 14.3% each. The denominator and event-term calculation details are not reported in the registry record. These percentages should not be assumed to use the full trial enrollment as their denominator.

These entries are descriptive counts, not estimates of comparative risk. The design information identifies no control group, and the registry record does not establish that an event was caused by study treatment. Death counts in the safety listing should not be interpreted as treatment-related deaths without attribution information, which is not reported in the registry record.

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

The registry record identifies a study question and a defined primary endpoint: the percentage of participants in specified cohorts with a confirmed complete or partial response. It also specifies tumor-response criteria, assessment timing, and an endpoint population based on receipt of at least 1 dose. These details clarify what the intended measure represents.

What the record does not report is the primary response result. There is no responder count, response percentage, or 95% confidence interval in the posted results. As a result, this record cannot establish the observed frequency of response, the precision of that frequency, or how the estimate varied among cohorts. It also does not report a p-value or any other inferential result for the primary endpoint.

The structured design is single group, so its listed entries do not support a randomized treatment comparison. A response proportion among treated participants is not equivalent to a difference between tazemetostat and another treatment. Although the brief summary describes a planned comparison involving doxorubicin and placebo, the registry record does not reconcile that account with the structured single-group design. The existence, conduct, and results of a comparator analysis are therefore not established by these entries.

The safety information offers counts for adverse-event categories and deaths, but it does not by itself show whether those events were caused by treatment or whether their frequency differed from what would have occurred under another treatment. The adverse-event denominators also differ by cohort, and selected event-term percentages have no stated denominator in the registry record.

In general, a study result should be interpreted alongside its analysis population, missing-data rules, precision, and design. Here, the definitions are available, but important results and analysis details are not reported in the registry record. The appropriate conclusion is limited to describing what was registered: the study’s endpoint and design are specified, while the numerical primary efficacy finding and its uncertainty are not.

14. Why This Trial Matters Statistically

This study illustrates why a registry’s design description and its results need to be read together. A named endpoint may appear straightforward, but an interpretable estimate also requires the number of responders, the denominator, and a measure of uncertainty. The registry record defines objective response rate and specifies a 95% confidence interval, yet reports neither the response estimate nor the interval. Without those results, statistical precision cannot be judged.

The multi-cohort structure adds an important interpretive issue. Participants have several different tumor types, and the primary endpoint is assigned to a defined subset of cohorts. An overall trial enrollment does not tell readers how many participants informed the primary response result, and it cannot replace cohort-specific denominators. Cohort-level reporting helps distinguish a result driven by one group from a pattern shared across groups; such endpoint results are not reported.

The contrast between the brief summary and structured design also shows why design details matter. The brief summary describes a planned comparison, while the structured entries identify a single group and no comparator. Statistical claims about treatment differences depend on knowing which design was actually conducted and how comparisons were analyzed. The registry record does not reconcile these descriptions.

Finally, endpoint definitions, analysis populations, and missing-data methods are not merely technical details. They determine which observations contribute to an estimate and what the estimate can reasonably represent. This record states the endpoint population but does not report missing-data handling or the primary result. The statistical value of a registry entry therefore depends not only on what it names, but also on whether the numerical results and methods needed to interpret them are reported.

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 specific to this registry record constrain interpretation. The primary objective response rate is defined, but its observed value, responder count, denominator, and 95% confidence interval are not reported in the registry record. The listed overall enrollment cannot substitute for the missing analysis denominator, particularly because the primary endpoint covers only specified cohorts.

The structured design is single group and lists no comparator. The brief summary describes a treatment comparison involving doxorubicin and placebo, but the registry record does not clarify how that description relates to the structured design. Therefore, the record does not support a clear interpretation of a comparative analysis.

The population includes multiple tumor cohorts, and the primary endpoint uses different response criteria for primary brain tumors and other solid tumors. The record does not report cohort-specific efficacy results, independent review, or details on handling unassessable responses. It also does not report the statistical analysis plan, sample-size justification, multiplicity procedures, or missing-data methods.

The safety entries provide cohort-specific counts, but event attribution and the denominator for the selected serious-event percentages are not reported in the registry record. These limitations prevent a full assessment of efficacy, comparative effects, and causal interpretation of safety findings from the registry entries alone.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: A Study of Tazemetostat in Adult Participants With Soft Tissue Sarcoma — NCT02601950.
• Registry record last updated 2025-06-22; status last verified 2025-06.
• Results first posted 2025-03-11.
• 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 describes a completed Phase II, single-group study of tazemetostat in adult participants across several tumor cohorts. Overall enrollment is 267; the listed regimen is 800 mg orally twice a day in continuous 28-day cycles. The primary endpoint is objective response rate for Cohorts 1, 3, 4, 5, 6 and 7, assessed by investigator review using RANO or RECIST version 1.1 criteria.

The record lists secondary response, duration, progression-free survival, overall survival, and disease-control outcomes. It also reports cohort-level adverse-event and death counts. However, the primary response estimate and its 95% confidence interval are not reported in the registry record. The structured design does not list a comparator, and the discrepancy with the brief summary’s planned comparison is not explained.