Phase II Study of TAK228 in Relapsed Lymphoma: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Phase II Study of TAK228 in Relapsed Lymphoma (NCT02727777), a non-randomized phase 1, phase 2 study. The primary outcome is Response Assessment (RA).

2. Clinical Question

3. Trial Design

NCT02727777, titled “Phase II Study of TAK228 in Relapsed Lymphoma,” was registered as a Phase I/Phase II study sponsored by M.D. Anderson Cancer Center. It was a non-randomized, parallel-group study with no masking. The registry lists enrollment of 4 and a terminated status. The listed study start date is 2017-03-01, and the primary completion date is 2018-07-10.

The planned Phase II groups divided participants by lymphoma category: aggressive non-Hodgkin lymphoma, indolent non-Hodgkin lymphoma, and Hodgkin lymphoma. The registry describes daily TAK-228 dosing during 28-day cycles. For the indolent and Hodgkin groups, treatment was described as continuing until disease progression or a maximum of 12 months. The aggressive-group description is incomplete in the record, so its stated maximum treatment duration is not available. Blood sugar testing was specifically described for the Hodgkin group.

This design does not support a randomized treatment-versus-control comparison. The groups were organized by disease category, and the record does not describe a control intervention. The study's purpose was to investigate whether TAK-228 could help control relapsed lymphoma and to study safety, but the registry does not provide a primary response analysis. It reports that two participants failed screening and two participants had disease progression before the analysis could be performed. The registry record does not report a statistical analysis plan, sample-size justification, or interim-analysis plan.

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: No significant events in the study occurred after enrollment, prior to assignment to arm/group

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: No analysis could be performed due to 2 participants failed screening and 2 participants disease progressed before analysis could be performed

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: Dr. Westin (Principal Investigator) was only participating in the DLBCL arm.

6. Primary Endpoint

The registry lists the primary outcome as Response Assessment (RA), measured over Time frame for response assessment was from Baseline to end of treatment or progression of disease up to 1 year..

The primary outcome was “Response Assessment (RA),” defined using Lugano criteria. It was intended to classify disease status from baseline through the end of treatment or progression, with follow-up for response assessment up to 1 year. The description bases assessment on CT scans obtained at screening and after completion of every 2 cycles of therapy.

In general, radiographic response criteria provide a structured way to characterize changes in measurable disease. The registry's description distinguishes complete radiographic response, partial response, stable disease, and progressive disease. These categories use changes in target lesions and other disease findings to determine whether the imaging pattern meets a response or progression definition. Applying a prespecified set of criteria can make imaging assessments more consistent than an informal judgment alone.

In this study, the endpoint's intended value was to describe response during treatment in relapsed lymphoma. But the registry explicitly states that no analysis could be performed because two participants failed screening and two had disease progression before analysis could be performed. It does not report response counts, an estimated response rate, or an uncertainty interval. The endpoint definition therefore describes what the investigators intended to assess; it is not evidence of an observed response result.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does a response analysis usually estimate?

In general, an endpoint analysis turns participant-level observations into a summary that addresses the study question. For a categorical response outcome, a summary might report how many participants fall into each response category or estimate the proportion meeting a response definition. A confidence interval can describe the uncertainty around an estimate: it gives a range of values compatible with the data and the method's assumptions. A hypothesis test, when planned and appropriate, assesses how compatible the observed data are with a specified null hypothesis. These tools do not replace clinical interpretation, and neither can be calculated meaningfully when the required outcome analysis was not performed.

For this study, the registry defines response using Lugano criteria and describes imaging assessments at screening and after every 2 cycles. However, it states that no analysis could be performed because two participants failed screening and two participants had disease progression before analysis could be performed. No response estimate, confidence interval, or p-value is reported. The absence of these statistics is not a result of zero effect; it reflects the absence of an analyzed primary outcome in the registry record.

Why does the non-randomized design matter?

In general, random allocation helps balance both measured and unmeasured participant characteristics between treatment groups. Without randomization, group membership may be related to the characteristics that also influence outcomes. Here, the listed groups are defined by lymphoma subtype, not by assignment to TAK228 versus a control treatment. The record therefore does not describe a randomized comparison from which a treatment contrast can be estimated. It is important to distinguish a parallel structure from a controlled comparison: parallel groups can receive treatment during the same study without serving as exchangeable treatment arms.

The study is also described as unmasked. In general, masking can reduce the possibility that knowledge of treatment assignment affects care, outcome assessment, or reporting. The registry does not report whether any outcome assessment safeguards were used beyond the stated imaging criteria, so such procedures cannot be assumed.

How should planning calculations be understood?

In general, sample-size calculations are planning tools. They describe how many participants might be needed under stated assumptions about an effect, variability, significance level, power, and design. They are not a measure of what a completed study observed. A planning illustration based on a fraction of an observed effect is especially dependent on having an interpretable effect estimate. In this record, no primary response analysis was performed, so a primary response effect is not reported as an observed result. Any planning illustration should therefore be kept separate from the trial's actual outcome evidence.

The registry's planning assumptions describe a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation, with the effect to detect expressed as a fraction of the observed effect. Those assumptions are not the same as a reported protocol design justification. The record does not report the trial's own sample-size justification, and the actual design was non-randomized with disease-category groups rather than two like-for-like treatment arms. It would be inappropriate to read planning quantities as observed power or as evidence that the study had a particular chance of detecting an effect.

Why does over-dispersion matter in general?

For count outcomes, some statistical models assume a particular relationship between the average count and its variability. Over-dispersion means that the observed variability is greater than the model expects. If it is ignored, standard errors can be too small and confidence intervals can be too narrow, making results appear more precise than the data warrant. This is a general issue for count-data analyses; the registry record does not report an over-dispersion assessment or a count-model analysis for this study.

What do the safety counts establish?

In general, adverse-event counts are descriptive summaries of events recorded during a study. The number at risk identifies the group denominator used in the registry's safety summary, while the affected count indicates how many participants had the event category recorded. Counts do not by themselves establish causality, severity beyond the category definition, timing, recurrence, or comparative risk. Those interpretations require details and analyses that may not be present in a registry summary.

The registry reports small group counts, including serious adverse events and deaths in the aggressive group, but does not provide a comparative treatment group or a statistical analysis of these outcomes. The counts should therefore be read as reported observations, not as adjusted estimates or proof of a causal relationship. No additional inference about timing, cause, or preventability is supported by the recorded figures.

10. Multiplicity, Missing Data and Analysis Populations

The primary outcome was response assessment, but no analysis could be performed. The registry lists 3 outcome measures and 0 analyses. It does not report a multiplicity strategy describing how statistical error would be controlled across outcomes, response categories, disease groups, or repeated assessment times. Without a reported analysis plan, it is not possible to determine whether any hierarchy or adjustment was intended.

Missing outcome information is a central limitation. The registry says two participants failed screening and two participants experienced disease progression before analysis could be performed. It does not report a primary response result for these participants or describe an analysis population in which response was evaluated. The record does not report how screening failures, early progression, or other missing observations would have been handled in an efficacy analysis. No imputation method, censoring rule, or sensitivity analysis is reported.

In general, missing-data methods rely on assumptions about why measurements are absent and how the absence relates to outcome. Different assumptions can lead to different estimates, so transparent reporting of the analysis population and missing-data approach is important. Here, the registry does not provide enough information to assess those assumptions or reproduce an analysis.

The registry record also does not report a statistical analysis plan, prespecified interim analyses, or a sample-size justification. Because the study was non-randomized and the groups were distinct lymphoma categories, the missing inferential details cannot be filled in by treating the groups as conventional randomized arms. The absence of a reported analysis should be interpreted as unavailable outcome evidence, not as a null finding.

11. 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: cellulitis (50.0%).

The registry's safety summary is organized by lymphoma group. In the aggressive non-Hodgkin lymphoma group, 2 participants were at risk for serious adverse events and 1 was affected. For other adverse events, 2 were at risk and 2 were affected. The same group has 2 deaths reported. The indolent non-Hodgkin lymphoma and Hodgkin lymphoma groups each report zero participants at risk and zero affected in the displayed safety categories, with zero deaths.

These are descriptive counts, not a comparison between TAK228 and another treatment. The registry does not report an attribution analysis linking an event or death to study treatment, nor does it provide enough information to interpret timing or cause. Counts from a group with very few participants are especially limited as a basis for judging how often an event might occur in a broader population.

Cellulitis is listed among the serious adverse-event terms at 50.0%. This term-level figure should be interpreted in the context of the reported group counts, not as a general estimate of risk. The record does not report an inferential safety analysis. The figures describe what was recorded in the registry and do not by themselves establish that TAK228 caused an adverse event or death.

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

The registry record establishes that this was a small, terminated, non-randomized study of TAK228 in relapsed lymphoma, with intended response assessment using Lugano criteria. It also establishes that no primary response analysis could be performed: two participants failed screening and two had disease progression before analysis could be performed. As a result, the record does not supply an analyzed estimate of response or its precision.

That missing estimate is not equivalent to a finding of no response, no benefit, or no effect. It is also not evidence of a positive response. The primary outcome remains unquantified in the registry record. The intended endpoint definition can explain what was to be measured, but it cannot substitute for participant outcome data or a completed analysis.

The safety counts provide a limited descriptive account. They identify reported serious adverse events, other adverse events, and deaths in the aggressive group, while the other listed groups show zero counts and zero participants at risk in those categories. They do not establish why the events occurred, whether they were caused by treatment, or how the observed counts compare with another treatment. There is no control group described for a treatment comparison.

Because the study was non-randomized and its groups were based on lymphoma category, differences between groups would not provide a clean estimate of the treatment effect. The registry also does not report the statistical analysis plan, sample-size justification, interim-analysis plan, missing-data methods, or a multiplicity strategy. The record is therefore most informative about the study's design, planned outcome, and the limited safety observations posted—not about the efficacy or comparative safety of TAK228.

13. Why This Trial Matters Statistically

This study matters statistically because it illustrates the difference between a clearly defined endpoint and an available endpoint result. The registry describes a structured radiographic response assessment using Lugano criteria, including CT assessment at screening and after every 2 cycles. Yet the record states that no analysis could be performed. A detailed measurement definition improves clarity about the intended question, but it does not generate an estimate when the outcome has not been analyzed.

It also highlights the importance of design when interpreting group labels. Parallel groups can be organized around different disease categories, as in this record, without providing a randomized comparison of treatment strategies. Calling groups “arms” does not make their outcomes directly comparable as treatment effects. The non-randomized allocation and absence of a comparator treatment constrain causal interpretation.

The small enrollment and the reported screening and progression outcomes make transparent reporting especially consequential. Safety counts can be described, but they cannot support precise generalization or a treatment comparison on their own. Likewise, planning calculations are not observed evidence and should not be confused with completed outcome analyses.

Finally, the record demonstrates why a missing analysis should be reported and interpreted distinctly from a null result. The registry gives a reason that the primary analysis could not be performed, but does not provide response estimates or uncertainty measures. Readers can identify the intended endpoint and the limits of the posted data without overinterpreting an unavailable result.

14. Limitations

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

The primary limitation is that no response analysis was performed, despite response assessment being the primary outcome. The registry attributes this to two participants failing screening and two participants experiencing disease progression before analysis could be performed. It does not report response counts, an estimate, or an uncertainty interval.

Enrollment was 4, and the record is terminated. The reported safety counts are confined to the aggressive non-Hodgkin lymphoma group, with zero participants at risk in the other listed groups. The registry does not provide a control treatment or randomized allocation, so the design cannot support a like-for-like treatment comparison. Disease-category groups should not be interpreted as interchangeable treatment arms.

The aggressive-group treatment description is truncated in the registry record, so its maximum treatment duration is not reported in the available description. The registry does not report a statistical analysis plan, sample-size justification, interim-analysis plan, multiplicity strategy, missing-data method, or analysis population details. It also does not report an attribution analysis for the safety observations. These gaps limit assessment of how outcomes would have been analyzed and how safety events were interpreted.

The posted safety counts are descriptive and arise from very small group totals. They do not establish causality or provide a stable estimate for a broader population. The reported cellulitis term and death counts should be read as registry observations, not as comparative or causal findings.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Phase II Study of TAK228 in Relapsed Lymphoma — NCT02727777.
• Registry record last updated 2020-03-12; status last verified 2020-03.
• Results first posted 2020-03-12.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

18. Record Summary

NCT02727777 was a terminated, non-randomized Phase I/Phase II parallel-group study sponsored by M.D. Anderson Cancer Center. It enrolled 4 participants and studied TAK228 in relapsed lymphoma, with planned groups for aggressive non-Hodgkin lymphoma, indolent non-Hodgkin lymphoma, and Hodgkin lymphoma. The registry lists response assessment using Lugano criteria as the primary outcome, with imaging from baseline through end of treatment or progression, up to 1 year.

No primary response analysis could be performed because two participants failed screening and two participants had disease progression before analysis could be performed. The registry reports limited descriptive safety counts, including serious adverse events, other adverse events, and deaths in the aggressive group. It does not report a control treatment or the statistical analysis details needed for further inference. The posted record describes the study's aims and design but does not provide an analyzed primary response result.