Mature B-Cell Lymphoma And Leukemia Study III: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Mature B-Cell Lymphoma And Leukemia Study III (NCT01046825), a non-randomized phase 2, phase 3 study. The primary outcome is Gene Differential Expression Profiling of Burkitt Lymphoma (BL) vs. Non-BL in the US and Other Selected Geographic Regions of the World.

2. Clinical Question

The study description also frames the trial as prospective research into tumor biology and late effects among patients treated with a modified regimen derived from LMB-96. The registry record does not report a treatment-effect comparison between randomized groups, and the primary gene-expression question should not be read as one.

3. Trial Design

This is a phase 2 and phase 3 study with nonrandomized allocation, no masking, and a parallel model. The registry lists enrollment of 128 and describes a risk-adapted treatment approach. The record does not explain how the enrollment figure relates to the number who received treatment, completed follow-up, or contributed to each analysis.

Group assignment is based on disease characteristics rather than random allocation. Group A is described as completely resected stage I or completely resected abdominal stage II lesions and is assigned COPAD for two cycles. Group B includes cases not eligible for Group A or Group C, described as Murphy Stage III and non-CNS Stage IV. Its listed sequence is a COP pre-phase, COPAD M3 induction for two cycles, and CYM consolidation for two cycles. Group C includes CNS involvement and/or bone marrow involvement of at least 25% blasts. The complete Group C treatment description is not reported in the registry record.

Because treatment is selected according to risk group, the groups are not interchangeable comparison arms. Differences in outcomes between them could reflect differences in the disease characteristics used to assign treatment, as well as other factors. The registry record does not describe a randomized comparison or a control group receiving a single alternative regimen. No causal contrast between the listed treatment groups can be inferred from their labels alone.

The study description emphasizes prospective biological and late-effect research alongside treatment. The primary measure specifically concerns gene-expression profiling, while the record also lists eight outcome measures and three analyses. Details of those other measures and analyses are not given in the available record. The study is listed as active, not recruiting; its recorded start is 2010-09-09 and primary completion is 2023-08-31. Those registry dates do not, by themselves, establish that every planned analysis or follow-up activity is complete.

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: No data was generated for this measure due to the PI leaving St. Jude, COVID-19 disruptions, stricter regulations on international tumor material shipping, and unavailability of the required assay platform. One sample was received from Singapore, none from Egypt. No participant is analyzed, and no gene expression data is generated because the Affymetrix U133A GeneChip is no longer available. No participant will be analyzed, and no gene expression data will be generated in the future.

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: Of the 87 participants that started in Group B, one was found to be ineligible and is excluded from all analysis, including adverse events.

6. Primary Endpoint

The registry lists the primary outcome as Gene Differential Expression Profiling of Burkitt Lymphoma (BL) vs. Non-BL in the US and Other Selected Geographic Regions of the World, measured over 1 year after the participant is enrolled.

The primary measure is gene differential-expression profiling of Burkitt lymphoma (BL) versus non-BL in the United States and other selected geographic regions. It is intended to compare gene-expression levels, using approximately 22,000 probesets on the Affymetrix U133A GeneChip and a two-factor analysis of variance model for each gene. The stated timeframe is one year after a participant is enrolled.

Gene-expression profiling is used in research to characterize patterns of activity across many genes. In general, comparing such patterns can help investigate biological distinctions between tumor groups or geographic settings. That research purpose is distinct from assessing whether a treatment improves a clinical outcome.

The registry reports that no data were generated for this measure and no participant was analyzed. It attributes this to the principal investigator leaving St. Jude, COVID-19 disruptions, stricter regulations on international tumor material shipping, and unavailability of the required assay platform. One sample was received from Singapore and none from Egypt, but the record states that no participant was analyzed. Consequently, there is no posted gene-expression estimate, uncertainty interval, or statistical test result for this primary measure. The measure's planned method and timeframe describe an intended analysis, not a completed result.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does the planned gene-level model do?

The primary measure describes a two-factor analysis of variance model for each gene. In general, analysis of variance compares outcome averages across categories defined by factors in a model. A two-factor model includes two explanatory factors; depending on its specification, it may also assess whether their effects interact. The registry record does not identify the factors or explain the model's specific terms, so their identities and interpretation are not reported in the registry record.

The outcome for each model is a gene-expression level measured by a probeset on the Affymetrix U133A GeneChip. The plan covers approximately 22,000 probesets, with a model applied gene by gene. This is a high-dimensional analysis: many outcomes are considered, not just one prespecified measurement. In general, evaluating many outcomes raises the chance of obtaining apparently notable findings by chance if results are judged one at a time. Methods for addressing multiplicity, such as controlling an overall error rate or the expected proportion of false discoveries, are general statistical approaches. The multiplicity procedure for this measure is not reported in the registry record.

Why does the absence of analyzed samples matter?

A statistical model can estimate differences and their uncertainty only when the relevant outcome data are available for analysis. Here, the registry states that no data were generated for the measure and no participant was analyzed. It gives operational and regulatory reasons for the lack of an analysis, including the unavailable assay platform and barriers affecting international tumor material. The record also notes one sample received from Singapore and none from Egypt. Sample receipt is not the same as an analyzed participant or a completed expression result.

Because no participants were analyzed, the planned model does not produce reported gene-level estimates. There are no posted standard errors, confidence intervals, test statistics, or p-values for this endpoint. A confidence interval, in general, describes a range of values compatible with the data and a specified procedure; without an estimate based on analyzed data, there is no interval to interpret. A missing result is not evidence that the groups have identical expression, nor is it evidence of a difference.

What can be said about treatment comparisons?

The allocation is nonrandomized and treatment is risk-adapted. Randomization, in general, aims to make treatment groups comparable on average with respect to measured and unmeasured baseline characteristics. When treatment is assigned according to disease features, groups can differ for reasons that also relate to outcomes. A direct comparison of Group A, Group B, and Group C therefore cannot be treated as a randomized estimate of regimen effects. The registry record does not report a statistical adjustment strategy that would address such differences.

The parallel model indicates that groups are followed in parallel rather than in a crossover sequence. The record lists no masking. It does not report allocation procedures beyond the nonrandomized designation, a detailed protocol for the clinical outcome analyses, or an inferential treatment-comparison plan. Those details should not be inferred from the intervention names or group descriptions.

How should planning calculations be understood?

Planning illustrations based on an observed effect are hypothetical calculations, not trial outcomes. The stated assumptions for those illustrations are a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation; the target effect is a stated fraction of the observed effect. For time-to-event outcomes, the stated planning quantity is the total number of events; for other outcomes, it is participants per arm. These assumptions describe the illustration, not the trial's own design or a reported sample-size justification. The registry record does not report a sample-size justification for this study.

In general, planning calculations depend on the effect size selected, the variability or event rate, the analysis method, and the desired power. A calculation based on an observed effect can be especially uncertain when that estimate is unstable or when the study design differs from the assumptions. It does not measure the probability that a study result is true, and it is not observed power. No planning figure can substitute for the unavailable gene-expression analysis or establish that a treatment effect was detected.

What remains unknown about the statistical plan?

The registry describes the gene-level model but does not report the identities of its two factors, the way the model handles potential interaction, covariate adjustment, transformation or normalization of expression values, or procedures for multiple testing. It also does not report a detailed missing-data approach, analysis populations, interim analyses, or a full statistical analysis plan. Each of these details can influence how estimates are produced and interpreted. In their absence, the planned method should be described only at the level stated: a two-factor analysis of variance model for each gene across approximately 22,000 probesets.

10. Multiplicity, Missing Data and Analysis Populations

Multiplicity is a central methodological issue for the primary measure because the plan calls for gene-by-gene analyses across approximately 22,000 probesets. In general, testing many outcomes can increase the chance of false-positive findings when each test is interpreted independently. Statistical approaches can adjust for the number of tests or control other error criteria, but the procedure planned for this study is not reported in the registry record. No gene-level results are available to assess how any such procedure would have been applied.

Missingness is not a matter of a few unreported values within an otherwise completed primary analysis: the registry states that no data were generated for this measure and no participant was analyzed. It attributes the lack of analysis to the principal investigator leaving St. Jude, COVID-19 disruptions, stricter international shipping regulations for tumor material, and unavailability of the required assay platform. One sample was received from Singapore and none from Egypt. How those circumstances affected the broader study population or other outcome measures is not reported in the registry record.

The record does not report analysis populations for the clinical outcomes, or specify whether those analyses include all enrolled participants, treated participants, or another defined set. It also does not describe a missing-data method, interim analyses, or a complete statistical analysis plan. Those omissions limit assessment of how participants and incomplete observations were handled across outcomes. They do not alter the explicit status of the primary gene-expression measure: there is no analyzed participant for it. Any interpretation must distinguish that lack of data from a measured null result.

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 non-serious adverse event terms

The registry reports serious adverse event counts and other adverse event counts by risk group. For serious events, the record lists 5 at risk and 0 affected in Group A, 86 at risk and 0 affected in Group B, and 22 at risk and 0 affected in Group C. The Unknown Risk row lists 0 at risk and 0 affected. No serious adverse event terms are listed.

For other adverse events, the affected counts are 3 in Group A, 77 in Group B, 21 in Group C, and 0 in Unknown Risk. The listed deaths are 0 in Group A, 1 in Group B, 1 in Group C, and 1 in Unknown Risk. These are descriptive registry counts, not comparative estimates of risk. The record does not provide event-level terms for the serious events, details about timing or duration, or a statistical comparison between groups.

The Unknown Risk row's death count of 1 appears alongside 0 at risk and 0 affected in the adverse-event categories. The registry record does not explain or reconcile this combination, so it should be reported as recorded rather than interpreted as a rate. The risk groups are also nonrandomized and differ by disease characteristics, which prevents simple group counts from serving as a like-for-like safety comparison.

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

The registry record describes a prospective, risk-adapted study of mature B-cell lymphoma with biological and late-effect research, but the primary gene-expression measure has no analyzed participants. The stated analysis would compare BL and non-BL expression across selected geographic regions using approximately 22,000 probesets and a two-factor analysis of variance model for each gene. No expression data were generated for this measure, and no participant was analyzed. Therefore, it does not provide evidence for or against a difference in gene expression.

The record's explanation for the missing analysis is specific: the principal investigator left St. Jude, COVID-19 disruptions occurred, regulations on shipping international tumor material became stricter, and the required assay platform was unavailable. One sample was received from Singapore and none from Egypt. Those details explain why the primary analysis is absent; they do not reveal what the results would have been if the planned assay had been completed.

The nonrandomized, risk-adapted structure also limits what can be concluded about treatment. Different disease features define the groups and their assigned regimens, so comparisons between groups would not be randomized treatment comparisons. The record does not establish that one regimen is better or worse than another. The listed safety counts are descriptive and are not a comparative analysis.

The study's stated purpose includes further study of tumor biology and prospective late effects. That aim is distinct from a direct efficacy conclusion. The registry record lists eight outcome measures and three analyses, but it does not provide full details for all of them in the material described. Conclusions about measures other than the primary endpoint require their own posted results and analysis details. For the primary endpoint, the scientifically accurate conclusion is that no result was generated for interpretation, rather than that the groups showed no difference.

13. Why This Trial Matters Statistically

This study illustrates why the stated research question, design, and available results must be read together. Its primary endpoint is biological: gene-expression profiling of BL versus non-BL across selected regions. The registry gives a specific planned assay and a gene-level statistical model, but reports that no participant was analyzed. The contrast between a detailed planned method and an unavailable result is important for readers interpreting clinical trial records.

It also demonstrates how a risk-adapted design differs from a randomized treatment comparison. Groups are defined by disease features and assigned different regimens, rather than formed by random allocation. This design can support prospective observation in clinically defined groups, but it does not make the groups exchangeable for causal comparisons. That distinction matters when interpreting outcome or safety counts.

The gene-expression plan raises a further general statistical consideration: thousands of gene-level outcomes create a multiplicity problem. A clear report of the model and the approach to multiplicity is important for evaluating a completed analysis. Here, the multiplicity procedure is not reported in the registry record, and the absence of analyzed data means that no gene-level findings can be evaluated.

Finally, the registry's reasons for the missing primary analysis show that statistical interpretation depends on more than a model specification. Assay availability, specimen movement, and study disruptions can determine whether the intended analysis is possible. Recognizing these limits prevents a missing result from being mistaken for a negative result, and prevents descriptive group counts from being treated as evidence of a randomized treatment effect.

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 the absence of data and analyzed participants for the gene-expression endpoint. The registry names several reasons for this, but no gene-level estimates or uncertainty measures are reported. The listed one sample from Singapore and none from Egypt do not change the record's statement that no participant was analyzed for the measure.

The trial is nonrandomized, and its groups are determined by disease characteristics. The groups therefore are not like-for-like controls, and their counts cannot be interpreted as unbiased treatment comparisons. The registry record does not describe an adjustment strategy that would address differences arising from risk-based assignment.

Important elements of the analysis plan are not reported in the registry record. These include the identities of the two factors in the gene-level model, how multiplicity would be handled across approximately 22,000 probesets, the analysis populations for clinical outcomes, and the approach to missing data. The full Group C treatment description is also not reported in the registry record. The record lists eight outcome measures and three analyses, but further details for them are not available in the information.

The safety data provide group counts but not detailed serious-event terms, event timing, or comparative analyses. The Unknown Risk row lists 0 at risk and 1 death, a combination the registry does not explain. The enrollment figure of 128 does not by itself establish the number treated, followed, or analyzed. These limitations constrain interpretation of both the unavailable primary endpoint and the descriptive clinical and safety information.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Mature B-Cell Lymphoma And Leukemia Study III — NCT01046825.
• Registry record last updated 2026-06-02; status last verified 2026-05.
• Results first posted 2025-05-30.
• 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

NCT01046825 is a phase 2 and phase 3, nonrandomized, unmasked, parallel study of risk-adapted treatment for mature B-cell lymphoma. The registry lists enrollment of 128 and describes Group A, Group B, and Group C according to disease features. Treatment differs by group; the full Group C regimen is not reported in the registry record.

The primary measure is gene-expression profiling of BL versus non-BL in the United States and selected other geographic regions. The planned approach uses approximately 22,000 probesets and a two-factor analysis of variance model for each gene. No data were generated and no participant was analyzed for that measure. The record attributes this to investigator departure, COVID-19 disruptions, stricter international tumor-material shipping regulations, and an unavailable assay platform.

Safety counts are descriptive, and the nonrandomized design does not support a like-for-like comparison among risk groups. The study is listed as active, not recruiting, with primary completion recorded as 2023-08-31.