Study of MLN8237 in Participants With Advanced Hematological Malignancies: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Study of MLN8237 in Participants With Advanced Hematological Malignancies (NCT00697346), a non-randomized phase 1 study. The primary outcome is Number of Participants With Dose-Limiting Toxicity (DLT).

2. Clinical Question

The primary endpoint is a safety and dose-tolerability measure rather than a direct comparison of clinical benefit. In this study, dose-limiting toxicity was defined using specified adverse-event criteria and was assessed from the first dose through 30 days after the last dose, up to 422 days. The endpoint focuses on events considered related to alisertib and serious enough, or persistent enough, to meet the protocol definition.

The groups differed in dose, formulation, schedule, and disease population. Part 1 included powder-in-capsule and enteric-coated tablet dose escalation; Part 2 enrolled participants with peripheral T-cell lymphoma. Because allocation was nonrandomized, differences in their outcomes could reflect those design and population differences as well as treatment exposure. The registry record does not report the number of participants who met the primary endpoint, so the study's observed DLT count cannot be described from the posted results.

3. Trial Design

This was a phase 1, nonrandomized, parallel study of alisertib in participants with advanced hematological malignancies. The registry reports enrollment of 58 participants and a completed status. The recorded start date is 2008-07-11, and the primary completion date is 2016-10-01. The sponsor was Millennium Pharmaceuticals, Inc. The study summary describes an open-label, multicenter study for participants with limited standard treatment options.

The treatment program had three named groups: Part 1 PIC Dose Escalation, Part 1 ECT Dose Escalation, and Part 2 PTCL. Part 1 used powder-in-capsule or enteric-coated tablet formulations at varying doses and schedules. The listed schedules included daily dosing for 21 days followed by a 7-day recovery period, daily dosing for 14 days followed by a 14-day recovery period, and twice-daily dosing for 7 days followed by a 14-day recovery period. The registry also lists 21-day and 28-day cycles. Part 2 enrolled participants with peripheral T-cell lymphoma and used alisertib 50 mg enteric-coated tablets twice daily for 7 days followed by a 14-day recovery period in 21-day cycles, until disease progression or unacceptable toxicity, up to 2 cycles.

The registry records allocation as nonrandomized and the model as parallel. It also records masking as quadruple, while the official title and study summary describe the study as open-label. The registry record does not reconcile these descriptions, so the masking status is not clear from the record alone. Regardless of the masking entry, the groups were not randomized and do not constitute a randomized control comparison.

The primary endpoint was the number of participants with dose-limiting toxicity. Secondary outcomes included investigator-assessed best overall response rate, duration of response, and participant counts with specified UGT1A1 and Aurora A kinase polymorphisms. The registry lists 41 outcome measures overall, but the posted results do not provide the primary endpoint count or statistical analyses for it.

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 with a diagnosis of advanced hematological malignancies were enrolled 1 of 3 treatment groups, Part 1:alisertib powder-in capsule (PIC) 25 to 90 mg dose escalation cohort, Part 1: alisertib 30 to 50 mg enteric-coated tablet (ECT) dose escalation cohort, or Part 2: alisertib ECT 50 mg participants with peripheral T-cell lymphoma (PTCL).

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: DLT evaluable population included all participants who received at least 75% of their planned alisertib doses for their first cycle of treatment (unless interrupted by DLT) and had sufficient follow up data to allow the investigators and sponsor to determine whether DLT occurred.

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: Safety population is defined as all participants who received any amount of alisertib.

6. Primary Endpoint

The registry lists the primary outcome as Number of Participants With Dose-Limiting Toxicity (DLT), measured over From first dose of study drug to 30 days after the last dose (up to 422 days).

The primary endpoint counts participants with dose-limiting toxicity (DLT). It is a participant-level safety measure: a participant is counted if an event meets the study's DLT definition during the specified observation window. That window began with the first study-drug dose and continued until 30 days after the last dose, up to 422 days.

The definition used NCI CTCAE version 3.0 and required the event to be related to alisertib. It included specified severe or persistent blood-count abnormalities, fever or infection with severe neutropenia, low platelet count, significant nausea, vomiting or diarrhea despite supportive treatment, other qualifying nonhematologic toxicity, prolonged treatment delay, and certain toxicities that led the investigator to require dose reduction or discontinuation. Some exceptions were explicitly listed, including any grade of alopecia and brief grade 3 fatigue.

As a general statistical concept, DLT counts are used in early-phase studies to describe tolerability under the tested doses and schedules. They can inform understanding of the safety profile observed in the study, but a count alone does not describe the full range, timing, or severity of every adverse event. Nor does it establish comparative safety against a control group when allocation is nonrandomized.

The defined analysis population was not necessarily every enrolled participant. It included participants who received at least 75% of planned alisertib doses during the first cycle, unless a DLT interrupted treatment, and who had enough follow-up for investigators and the sponsor to determine whether a DLT occurred. The registry record does not report how many participants met this population definition or how many had a DLT.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does the primary count represent?

The endpoint is a count of participants meeting the study's DLT definition, rather than a continuous measurement or a time-to-event summary. The event definition is composite: several different clinical events can qualify, provided the specified severity, duration, treatment relationship, and other criteria are met. A composite endpoint can summarize several kinds of clinically important events in one measure, but its overall count does not reveal which component drove the result. In this registry record, the component-specific primary outcome counts and the overall DLT count are not reported.

The observation window is tied to treatment: it starts at the first dose and extends to 30 days after the last dose, up to 422 days. As a general concept, a defined window helps establish which events contribute to an endpoint. For a treatment-related toxicity endpoint, the window and attribution rule matter because the count depends on whether an event occurs in the period and is judged related to treatment. The registry provides the window and the relationship requirement, but does not give participant-level timing or attribution details.

Who contributes to the primary endpoint?

The DLT-evaluable population included participants who received at least 75% of planned alisertib doses during their first cycle, unless a DLT interrupted treatment, and who had sufficient follow-up to determine whether DLT occurred. This definition makes the evaluable population narrower than the overall enrolled group in principle. The registry does not report the number included in that population, so the denominator for the primary count cannot be recovered from the posted record.

As a general statistical concept, an evaluable-population rule can make an endpoint easier to interpret in relation to exposure, but exclusions after enrollment can affect who contributes to the estimate. This is especially relevant when early discontinuation, insufficient follow-up, or toxicity itself influences whether a participant is evaluable. The exception for a DLT-related interruption is part of the stated rule. The registry does not describe participant-level exclusions, missing outcome handling, or sensitivity analyses.

Why are the treatment groups not comparative arms?

Allocation is recorded as nonrandomized. The groups also differ by formulation, dose, schedule, and, for Part 2, disease population. In a randomized comparison, random assignment is intended to balance prognostic factors on average across groups. That protection does not apply. As a general statistical principle, observed differences between nonrandomized groups may reflect differences in who entered each group, the treatment regimens, follow-up, or other factors, rather than treatment alone. The registry itself notes that these arms are not a like-for-like control comparison.

The parallel model means the study lists distinct groups rather than a crossover structure in which participants receive multiple treatments in sequence. It does not make the groups equivalent. The phase 1 dose-escalation context and the different schedules mean that group-level counts should be described, not interpreted as causal estimates of comparative risk.

What does a confidence interval convey?

As a general statistical explanation, a confidence interval expresses uncertainty around an estimated quantity under the assumptions of the analysis. A narrow interval indicates greater statistical precision than a wide interval, while neither width nor coverage can be assessed without an estimate and its interval. The registry record reports no confidence interval for the DLT endpoint, and the underlying DLT count and denominator are not reported. No precision claim about the primary estimate is therefore possible.

Similarly, a p-value is a measure arising from a specified statistical test under its assumptions; it is not a measure of clinical importance or a substitute for an effect estimate and its uncertainty. The record reports no primary endpoint test or p-value. Because the study is nonrandomized and has no listed control comparison, an unreported test should not be inferred from the group counts.

How should safety counts be read?

The registry gives at-risk and affected counts for serious and other adverse events by study group. These are descriptive counts. As a general concept, a count depends on its denominator, the event definition, and the period during which participants were observed. If follow-up differs, or if the populations and regimens differ, simple group counts do not isolate a treatment effect. The record provides no comparative model, person-time analysis, or adjustment for such differences.

Over-dispersion is a general issue for some count-data models: variability can be greater than the model's simplest distribution assumes, which can make standard errors and intervals too small if not handled appropriately. This study's primary outcome is a participant count, but no count-model analysis is reported, and the record does not provide enough information to determine whether such a model was used. No model choice should be imputed from the endpoint format.

What do planning illustrations mean?

As a general explanation, sample-size planning evaluates how many observations might be needed to detect a specified effect under stated assumptions. The record's planning note describes an illustration based on the observed effect, two-sided significance level 0.05, equal allocation, and a normal approximation; it states that the effect to detect is a specified fraction of the observed effect. Such planning figures are not observed trial outcomes and are not observed power. They also do not establish that these assumptions guided the original design. The registry record does not report the original sample-size justification or a protocol-based statistical plan.

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("Alisertib 25mg PIC QD 21D", "Alisertib 35 mg PIC QD 21D", "Alisertib 35 mg PIC QD 14D", "Alisertib 45 mg PIC QD 14D", "Alisertib 65 mg PIC QD 14D", "Alisertib 90 mg PIC QD 14D")
n      <- c(6, 4, 3, 6, 6, 2)
events <- c(1, 2, 0, 1, 2, 2)
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  Alisertib 25mg PIC QD 21D  6  1  16.7  0.4  64.1  3.0  56.4  3.7  57.9
ARM  Alisertib 35 mg PIC QD 21D  4  2  50.0  6.8  93.2  15.0  85.0  14.7  85.3
ARM  Alisertib 35 mg PIC QD 14D  3  0  0.0  0.0  70.8  0.0  56.1  0.6  60.2
ARM  Alisertib 45 mg PIC QD 14D  6  1  16.7  0.4  64.1  3.0  56.4  3.7  57.9
ARM  Alisertib 65 mg PIC QD 14D  6  2  33.3  4.3  77.7  9.7  70.0  9.9  71.0
ARM  Alisertib 90 mg PIC QD 14D  2  2  100.0  15.8  100.0  34.2  100.0  29.2  99.2

11. Multiplicity, Missing Data and Analysis Populations

The registry lists 41 outcome measures, including the primary DLT count and secondary outcomes for response, duration of response, and polymorphisms. When a study assesses multiple outcomes, multiple analyses, or multiple time points, the chance of at least one apparently notable result can increase if each is tested independently. As a general statistical concept, multiplicity procedures can address that issue by controlling a chosen error rate or by defining a hierarchy of analyses. The registry record does not report a multiplicity strategy, endpoint hierarchy beyond the primary and secondary labels, or a prespecified testing sequence.

The registry reports no statistical analyses for the posted outcomes. It gives no p-values, confidence intervals, model specifications, or analysis population results. Whether any multiplicity adjustment was planned or performed is not reported in the registry record. The list of outcomes alone cannot establish which outcomes were formally tested or how results were selected for reporting.

Missing-data handling is also not described. The primary endpoint definition does specify a DLT-evaluable population: participants needed to receive at least 75% of planned first-cycle doses, unless a DLT caused interruption, and to have sufficient follow-up for determination. However, the number meeting those requirements, the number excluded, the reasons for any exclusions, and the handling of uncertain DLT status are not reported in the registry record. The adequacy of follow-up is part of the population definition, but no further missing-outcome method is stated.

As a general principle, excluding participants with incomplete outcome information can affect interpretation if missingness is related to treatment exposure or outcome risk. This does not establish that such bias occurred in this trial; it identifies why the missingness details matter. The posted record does not permit assessment of the direction or magnitude of any possible effect.

12. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

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: staphylococcal bacteraemia (50.0%), herpes zoster (50.0%), delirium (50.0%), lung infection (50.0%), abdominal pain (50.0%).

The registry reports serious adverse-event counts by group. In Part 1 PIC Dose Escalation, 12 participants were affected among 28 at risk; in Part 1 ECT Dose Escalation, 15 were affected among 28 at risk; and in Part 2 PTCL, 2 were affected among 2 at risk. For other adverse events, the affected and at-risk counts are 28 and 28 for each Part 1 group, and 2 and 2 for Part 2. These figures describe the listed groups and should not be treated as randomized comparisons.

The serious-event terms listed with a frequency of 50.0% are Staphylococcal bacteraemia, Herpes zoster, Delirium, Lung infection, Abdominal pain, and Acute respiratory failure. The posted record does not give term-specific participant counts or clarify the denominator attached to these percentages. They therefore do not establish how many participants experienced each term. The record also does not report death counts.

As a general statistical point, adverse-event counts are interpretable only alongside their definitions, denominators, observation periods, and the treatment context. The DLT endpoint has its own event definition and evaluable population; it should not be equated with the broader serious or other adverse-event groupings. The registry does not report a statistical comparison of safety between groups, and the groups differ in regimen and population. No causal or comparative conclusion follows from these descriptive counts alone.

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

The registry record establishes that alisertib was studied in a phase 1, nonrandomized program involving dose-escalation groups and a PTCL group. It specifies the primary endpoint as the number of participants with DLT, defines the toxicity criteria, and states the period during which events were assessed. It also describes the population considered evaluable for that endpoint. These details clarify what the study intended to count.

The posted results do not give the observed DLT count, its denominator, or an uncertainty estimate. They therefore do not show the numerical primary endpoint result. The record also does not provide statistical analyses for the listed outcomes. Although it lists secondary response and genetic outcomes, their results are not reported in the posted information. This limits what can be concluded about outcomes beyond the descriptive safety counts.

The serious and other adverse-event counts provide a record of events by study group, and several serious-event terms are listed at 50.0%. Those observations must be interpreted in context: the groups were not randomized, had differing regimens, and included different populations. The serious-event term percentages lack accompanying term-specific counts and a stated denominator in the posted information. The counts do not support a controlled treatment comparison.

As a general interpretation principle, absence of a result in a registry record is not evidence that the event did not occur. It means the record does not report the value needed to quantify it. Likewise, a study's completed status does not, by itself, establish the size, precision, or direction of any treatment effect.

The registry's planning note describes sample-size calculations as illustrations based on the observed effect and stated assumptions. Such planning illustrations are not trial outcomes, do not constitute observed power, and do not establish the original design rationale. The original sample-size justification and full statistical plan are not reported in the registry record.

Accordingly, the strongest interpretation is descriptive: the registry identifies the study structure, endpoint definition, and certain safety counts, while leaving the primary DLT result and inferential results unreported. No conclusion about comparative efficacy or safety can be drawn from the nonrandomized group counts alone.

14. Why This Trial Matters Statistically

This trial is statistically informative as an example of why endpoint definitions and analysis populations matter in early-phase research. The primary outcome is not simply any adverse event: it is a participant-level DLT count governed by detailed severity, duration, treatment-relatedness, and management criteria. Those rules determine which events contribute to the endpoint and help distinguish it from broader adverse-event summaries.

The DLT-evaluable population also makes the denominator a substantive part of the question. Participants had to receive a specified amount of first-cycle treatment, subject to an exception for DLT interruption, and have enough follow-up to determine outcome status. Without the number in that population and the observed count, the primary estimate cannot be interpreted or assessed for precision. This illustrates why a clear endpoint definition alone is not a substitute for reporting the result and denominator.

The nonrandomized, parallel design is another important statistical feature. Several dose levels, formulations, schedules, and disease populations are represented. As a general principle, such groups are useful for describing experience within the enrolled cohorts, but their outcomes do not form a clean treatment contrast. Differences in counts cannot be attributed to treatment alone when allocation is not randomized and the groups are not alike in regimen and population.

The record also shows why registry completeness matters. It identifies 41 outcome measures and gives selected safety information, but does not report the primary endpoint value, analyses, uncertainty intervals, or multiplicity approach. These omissions constrain reproducibility and prevent readers from assessing precision or the relationship between planned and reported analyses. Statistical interpretation depends on both the design and the completeness of outcome reporting.

15. Limitations

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

The most important limitation of the posted record for statistical interpretation is that it does not report the primary DLT count or the size of the DLT-evaluable population. The endpoint definition is detailed, but without its numerator and denominator there is no primary estimate to interpret. No confidence interval or other measure of precision is reported.

The study was nonrandomized, and the registry explicitly cautions that its arms are not a like-for-like control comparison. Dose, formulation, schedule, and disease group varied across the listed parts. Group-level safety counts therefore cannot establish comparative effects. The record does not provide an adjusted analysis or information sufficient to account for differences between groups.

Several aspects of outcome reporting are not described: the handling of missing or indeterminate DLT status, the number of participants excluded from the evaluable population, and the extent of follow-up by group. The record does not report statistical models, hypothesis tests, uncertainty intervals, or a multiplicity plan. It lists secondary outcomes but does not include their results in the posted information.

Safety interpretation has additional limits. The serious-event counts are reported by group, but the record does not provide term-specific participant counts alongside the 50.0% frequencies, or clarify their denominator. Death counts are not reported in the registry record. The broad other-adverse-event counts should not be substituted for the DLT endpoint, which has a separate definition.

Finally, the registry describes the study as open-label while recording quadruple masking. That inconsistency is not reconciled in the registry record. The original sample-size justification and statistical plan are also not reported, so the planning illustration should not be mistaken for the trial's own design rationale.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Study of MLN8237 in Participants With Advanced Hematological Malignancies — NCT00697346.
• Registry record last updated 2019-05-31; status last verified 2019-02.
• Results first posted 2019-05-31.
• 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

NCT00697346 is a completed, nonrandomized phase 1 study of alisertib in participants with advanced hematological malignancies. The registry reports enrollment of 58 participants and lists dose-escalation groups using powder-in-capsule and enteric-coated tablet formulations, plus a PTCL group.

The primary endpoint was the number of participants with dose-limiting toxicity, assessed under specified criteria from first dose through 30 days after the last dose, up to 422 days. The primary population required a defined level of first-cycle dosing or DLT-related interruption and sufficient follow-up. The observed primary endpoint count and evaluable-population size are not reported in the registry record.

Serious and other adverse-event counts are reported by group, but the study was not randomized and the groups are not like-for-like comparisons. The registry lists secondary response and polymorphism outcomes, but does not report statistical analyses or their results. The original sample-size justification and statistical plan are not reported in the registry record.