Milademetan Tosylate and Low-Dose Cytarabine With or Without Venetoclax in Treating Participants With Recurrent or Refractory Acute Myeloid Leukemia: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Milademetan Tosylate and Low-Dose Cytarabine With or Without Venetoclax in Treating Participants With Recurrent or Refractory Acute Myeloid Leukemia (NCT03634228), a non-randomized phase 1, phase 2 study. The primary outcome is Participants With a Response.

2. Clinical Question

3. Trial Design

This was a phase I/II, parallel-group, non-randomized study sponsored by M.D. Anderson Cancer Center. The registry records enrollment of 16 and a terminated status. The study start date is 2018-12-17, and the primary completion date is 2022-04-03. Those dates describe the record; they do not explain why the study was terminated.

In phase I, participants received low-dose cytarabine and milademetan tosylate. The registry describes cytarabine given subcutaneously twice daily on days 1–10 and milademetan orally once daily on days 8–14, 8–21, or 5–7 and 15–17. In phase II, the described regimen also included oral venetoclax once daily on days 1–14. Cycles were described as repeating every 28 days in the absence of disease progression or unacceptable toxicity. These are regimen descriptions, not evidence that every participant received every schedule.

The record specifies no masking. Because allocation was non-randomized, the phase cohorts should not be treated as equivalent groups assigned to provide a controlled treatment contrast. Differences could reflect cohort composition, regimen, timing, or other factors, and the registry does not report an adjusted comparative analysis. Overall survival and event-free survival are listed as secondary outcomes, but their results are not reported in the registry record.

The record does not report the reason for termination, the dose-selection decision, a recruitment flow, or a statistical analysis plan. The listed enrollment is not a substitute for the number assessed for each endpoint, and those endpoint-specific analysis populations are not reported in the registry record.

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

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

6. Primary Endpoint

The registry lists the primary outcome as Participants With a Response, measured over Up to 3 years, 4 months.

The primary endpoint is the number of participants with a response. The registry defines response as complete response (CR), complete response with incomplete blood count recovery (CRi), partial response (PR), or morphologic leukemia-free state (MLFS). Combining these categories gives a broad summary of whether participants met one of several specified response criteria. It also means that the endpoint is not restricted to complete response with full blood count recovery.

The registry’s definitions use bone marrow findings and, for some categories, blood count recovery. For example, CR includes bone marrow blasts below 5%, absence of circulating blasts and blasts with Auer rods, absence of extramedullary disease, and specified neutrophil and platelet recovery. CRi follows the CR criteria except for the stated recovery requirements. PR requires a decrease in bone marrow blast percentage to the stated range and specified blood count recovery. MLFS uses marrow and disease findings without requiring hematologic recovery.

Response is used in trials of this kind as a clinically interpretable measure of tumor response; as a general statistical point, a count or proportion of participants meeting a defined endpoint can be easier to interpret than a continuous laboratory measure. However, the posted results do not report the response count, the denominator evaluated, or an estimate of response frequency. The registry lists a timeframe of up to 3 years, 4 months. It does not report the timing of individual assessments or the handling of participants with missing response assessments.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The statistical interpretation depends first on distinguishing a response endpoint from a treatment comparison. The primary endpoint is a participant count classified using specified response criteria. A count is descriptive until it is paired with an analysis population and an appropriate comparison or uncertainty estimate. Neither the primary response count nor its endpoint-specific denominator is reported in the registry record.

What does a response count tell us?

For a binary endpoint, each participant is classified as having or not having the defined event. In this study, the event is meeting any of the listed response categories. A response count alone does not express the chance of response unless the number assessed is also known and the approach to missing assessments is clear. A proportion would describe the observed frequency in the analyzed group, while an interval estimate would convey the uncertainty around that estimate. The record does not report these results for the primary endpoint.

The composite definition combines several response states. As a general statistical point, combining categories can capture a broader range of clinically relevant outcomes and increase the number of events available for analysis. It can also make interpretation less specific: a participant counted as a responder may meet one category but not another, and the categories need not represent identical degrees of response. The registry supplies the component definitions but not the count in each component.

Why does non-randomized allocation matter?

Random allocation is a design method intended to balance prognostic factors between treatment groups on average. Here, allocation is recorded as non-randomized, and masking is absent. Cohort membership may therefore be associated with participant characteristics, treatment timing, or clinical decisions. A difference between cohorts cannot automatically be attributed to a treatment difference. This is especially important when considering phase I and phase II regimens, because the phase II regimen includes venetoclax while the phase I description does not.

Parallel groups mean that the record describes separate cohorts rather than a crossover sequence in which each participant receives multiple regimens. That feature does not itself create a valid control comparison. The posted cohort descriptions and adverse-event counts can be read as cohort-level observations, but the registry does not report a randomized comparison, an adjusted estimate, or a statistical model to account for differences among groups.

How should uncertainty be understood?

As a general statistical explanation, a confidence interval gives a range of values compatible with the data under the chosen model and method. It communicates precision, not certainty that the true value lies within a particular range. Counts from small cohorts generally allow less precise estimates than counts from larger cohorts, particularly when events are uncommon or frequent. No confidence interval for the primary response endpoint is reported in the registry record, so its precision cannot be assessed from the posted result.

For comparisons, a p-value addresses how compatible the data are with a specified null hypothesis under the assumed analysis. It is not a measure of clinical importance, and it does not correct a non-randomized design’s potential confounding. The registry reports no analyses, so no primary-endpoint p-value or comparative test is available. A lack of a reported test is not evidence that groups are the same.

What can planning calculations establish?

Sample-size calculations are planning tools. In general, they estimate how much information a future study would need under stated assumptions about the target effect, significance level, allocation, and statistical method. They are not observed outcomes and do not establish that an existing study had adequate power. The record describes a planning illustration based on an observed effect, using a two-sided significance level of 0.05, equal allocation, and a normal approximation; the target effect is a stated fraction of the observed effect. It says time-to-event calculations use a total event count, while other calculations use participants per arm.

Such an illustration should not be confused with the trial’s original design rationale. The registry does not report the study’s own sample-size justification, and the primary response result needed to identify an observed effect is not reported. The registry record also does not report a target power, original effect assumption, or the role of any planning calculation in trial conduct. Planning figures, where included, are illustrative rather than evidence of observed power.

Why are event-time outcomes different?

Overall survival and event-free survival are listed as secondary outcomes. As a general statistical point, time-to-event analyses use both whether an event occurred and when it occurred, and may account for participants whose event has not been observed by the end of follow-up. Such analyses depend on definitions of the event, censoring, follow-up, and modeling choices. These details and the results are not reported in the registry record. The presence of outcome names alone is not enough to reconstruct an analysis.

Finally, adverse-event counts and response counts answer different questions. Safety counts describe recorded events or affected participants, whereas the primary endpoint describes response classifications. Neither type of count establishes causation or a comparative effect without an appropriate design and analysis. The record’s non-randomized allocation and absent outcome analyses limit what can be inferred from any descriptive differences between cohorts.

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("Phase I (Low Dose Cytarabine, MDM2 Inhibitor DS-30", "Phase I (Low Dose Cytarabine, MDM2 Inhibitor DS-30", "Phase I (Low Dose Cytarabine, MDM2 Inhibitor DS-30")
n      <- c(3, 6, 7)
events <- c(0, 1, 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  Phase I (Low Dose Cytarabine, MDM2 Inhibitor DS-30  3  0  0.0  0.0  70.8  0.0  56.1  0.6  60.2
ARM  Phase I (Low Dose Cytarabine, MDM2 Inhibitor DS-30  6  1  16.7  0.4  64.1  3.0  56.4  3.7  57.9
ARM  Phase I (Low Dose Cytarabine, MDM2 Inhibitor DS-30  7  1  14.3  0.4  57.9  2.6  51.3  3.2  52.7

11. Multiplicity, Missing Data and Analysis Populations

The registry lists a primary response endpoint and two secondary outcomes: overall survival and event-free survival. It does not report how many statistical hypotheses were formally tested, whether the secondary outcomes were confirmatory or exploratory, or whether any adjustment for multiple comparisons was planned. Without that information, the strength of any hypothetical finding across several outcomes could not be assessed on a common error-control basis. No outcome analyses are reported in the registry record.

The record does not report how missing response assessments were handled, how participants who discontinued treatment were included, or which analysis population was used for the primary endpoint. It also does not give the number assessed for response or provide participant-level follow-up completeness. These omissions matter because response classifications require an assessment at a defined time, and exclusion or classification of participants without an assessment can change the interpretation of a response frequency.

The registry does not report an intention-to-treat, treated, evaluable, or other defined analysis population. It also does not describe imputation, sensitivity analyses, endpoint adjudication, or handling of protocol deviations. Those methods should not be inferred from the phase I/II label or the regimen descriptions. The listed enrollment is not a substitute for endpoint-specific denominators.

Overall survival and event-free survival require follow-up and event-handling information, including how censoring was defined. Those details are not reported in the registry record. The record likewise does not report interim analyses or stopping rules. The terminated status alone does not establish whether a formal interim analysis occurred or why the study ended.

12. Safety Results

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

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 0% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: diarrhea (66.7%), sepsis (66.7%), lung infection (50.0%), chest pain (33.3%), dyspnea (33.3%).

The registry reports serious adverse event information by phase I cohort. Cohort 0 lists 3 participants at risk and 3 affected; cohort 1 lists 6 at risk and 6 affected; cohort 2 lists 7 at risk and 7 affected. The separate other-adverse-event figures list 3 at risk and 2 affected in cohort 0, 6 at risk and 0 affected in cohort 1, and 7 at risk and 0 affected in cohort 2. Death counts are listed as 2, 4, and 1 for the respective cohorts. These are descriptive entries, not adjusted comparisons.

The registry’s leading serious adverse event terms include diarrhea and sepsis at 66.7% each, lung infection at 50.0%, and chest pain, dyspnea, and fatigue at 33.3% each. These percentages should be read as registry-reported summaries; the record does not explain their event-specific denominators or provide uncertainty intervals. The term list does not show that events were caused by treatment, and it does not establish differences between regimens.

Safety counts can reflect more than one event per participant, while an affected-participant count refers to people with an event. The record does not provide enough detail to reconstruct event recurrence, timing, severity grading beyond the serious-event grouping, or attribution. The phase I cohort figures should not be treated as randomized comparisons with one another or with phase II.

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

The record establishes that a phase I/II study evaluated milademetan with low-dose cytarabine, with venetoclax added in the described phase II regimen, in an AML study whose brief summary focuses on recurrent or refractory disease. It also specifies a composite response endpoint and lists safety observations by phase I cohort. These details explain the study question and the types of outcomes intended to be assessed.

However, the registry does not report the primary response count, the denominator assessed, or an analysis of response. It therefore does not show the observed response frequency or its precision. Although overall survival and event-free survival are listed as secondary outcomes, their results are not reported in the registry record. No comparative treatment-effect estimate is available.

The non-randomized design is central to interpretation. Phase I and phase II cohorts differ in the described regimen, and the record does not establish exchangeability between them. Any apparent cohort differences would not by themselves demonstrate that adding venetoclax or changing a dosing schedule caused an outcome difference. The posted safety counts are descriptive, and the record does not establish treatment attribution.

The terminated status is a recorded administrative status, not a clinical result. The reason for termination is not reported in the registry record, so it should not be used to infer lack of activity, unacceptable toxicity, or any other explanation. Similarly, enrollment alone cannot establish how much information was available for the primary endpoint.

In practical statistical terms, the record supports an account of study structure, endpoint definitions, regimen descriptions, and certain safety counts. It does not support a conclusion that a regimen was effective or ineffective, nor does it provide a valid randomized estimate comparing regimens. Those questions would require outcome results, denominators, follow-up information, and an analysis suited to the design.

14. Why This Trial Matters Statistically

This study is statistically notable because it combines early-phase dose and safety work with a response question in a non-randomized design. The registry describes different regimens across phases, with venetoclax added in phase II. That structure makes it important to separate the purpose of describing cohorts from the stronger claim of estimating a causal treatment contrast. Cohort differences are not equivalent to randomized evidence.

The primary endpoint is a composite that includes CR, CRi, PR, and MLFS. Such an endpoint can summarize several response states, but its interpretation depends on the component-specific results and on knowing how participants were assessed. Since the response count and component breakdown are not reported in the registry record, the intended endpoint cannot be translated into an observed response estimate.

The safety entries also illustrate why denominators and context matter. Counts and percentages can describe what was recorded, but their precision and comparability depend on clear risk periods, event definitions, attribution, and cohort composition. The available figures do not provide an adjusted or randomized safety comparison.

Finally, the terminated status and listed enrollment should not be mistaken for statistical findings. Without a reported reason for termination, analysis populations, missing-data methods, or outcome analyses, the record has important limits for evaluating evidence. Recognizing those limits is itself useful: it prevents descriptive study details from being overinterpreted as proof of a treatment effect.

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 principal limitation is that the registry record does not report the primary response result. Without the number of participants meeting the response criteria, the denominator assessed, or an uncertainty estimate, the primary endpoint cannot be interpreted quantitatively. The record also does not report the results of the listed secondary outcomes.

The design is non-randomized and unmasked. The phase I and phase II regimens are not identical, and the phase II description includes venetoclax. There is no like-for-like randomized control comparison in the record. Differences among cohorts could reflect treatment, patient characteristics, treatment timing, follow-up, or other factors that are not described. No adjusted analysis is reported.

There is some variation in how the population is described: the brief summary emphasizes recurrent or refractory AML, while the official title refers to newly diagnosed or relapsed/refractory AML. The registry does not resolve this wording difference with a detailed participant eligibility or enrollment description. The record also does not report cohort-specific enrollment flow or endpoint analysis populations.

The registry records termination but does not report its reason. It does not report a statistical analysis plan, sample-size justification, missing-data methods, interim analyses, multiplicity procedures, or outcome-specific follow-up details. The serious adverse event summary lacks reported uncertainty intervals and does not clarify event-level denominators or attribution. These omissions constrain both efficacy and safety interpretation.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Milademetan Tosylate and Low-Dose Cytarabine With or Without Venetoclax in Treating Participants With Recurrent or Refractory Acute Myeloid Leukemia — NCT03634228.
• Registry record last updated 2023-06-15; status last verified 2023-05.
• Results first posted 2023-06-15.
• 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

NCT03634228 is a phase I/II, parallel, non-randomized, unmasked study sponsored by M.D. Anderson Cancer Center. The registry lists enrollment of 16 and status as terminated. The phase I regimen combined milademetan tosylate and low-dose cytarabine; the phase II description added venetoclax. The primary endpoint was the count of participants with CR, CRi, PR, or MLFS, with a stated timeframe of up to 3 years, 4 months.

The registry lists overall survival and event-free survival as secondary outcomes and reports phase I cohort safety counts, serious adverse event terms, and deaths. It does not report the primary response result, secondary outcome results, or an analysis of treatment effects. The termination reason, analysis populations, missing-data strategy, and statistical analysis plan are not reported in the registry record. The non-randomized design does not support treating the phase cohorts as equivalent controls.