Allogeneic Hematopoietic Stem Cell Transplantation (AlloSCT) Initial Salvage Therapy for Induction Failure Acute Myeloid Leukemia (AML): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Allogeneic Hematopoietic Stem Cell Transplantation (AlloSCT) Initial Salvage Therapy for Induction Failure Acute Myeloid Leukemia (AML) (NCT02441803), a non-randomized phase 2 study. The primary outcome is Overall Response (OR) Post Transplant.

2. Clinical Question

The study's objectives included assessing the safety and feasibility of allogeneic transplantation in this clinical setting, as well as response after DCIA salvage chemotherapy and transplant. Other stated objectives concerned eligibility for transplant after induction chemotherapy, treatment-related mortality, marrow blasts before transplant, relapse, overall survival, and event-free survival.

The primary endpoint was the count of participants classified as having an overall response post transplant, with assessment specified as 4 months after initial treatment/3 months after transplant. The registry record does not report the observed response count. Therefore, the clinical question can be described from the protocol objectives and endpoint definition, but the posted record does not give a numerical primary result with which to assess response in the enrolled participants.

3. Trial Design

This was a phase 2, non-randomized, unmasked, parallel-group study sponsored by M.D. Anderson Cancer Center. The registry record lists the condition as leukemia and enrollment as 11. The study started on 2015-09-14, had a primary completion date of 2020-09-22, and is listed as completed.

The planned treatment sequence began with DCIA salvage chemotherapy and proceeded to allogeneic stem-cell transplantation. The regimen description lists decitabine on Days 1–5, followed by cytarabine and idarubicin on Days 6–10 and clofarabine on Days 6–9. Transplantation was organized into matched sibling donor, haploidentical donor, and matched unrelated donor groups. The described transplant interventions include busulfan and other conditioning or supportive treatments, including fludarabine, clofarabine, total body irradiation, thymoglobulin, cyclophosphamide, tacrolimus, and stem-cell infusion. Some group-specific regimen descriptions in the registry record are incomplete.

Because allocation was non-randomized, the donor groups should not be interpreted as randomized comparisons of treatment strategies. Donor availability and other factors may differ across such groups, but the registry record does not describe those factors or how group assignment occurred. It also does not report a control group receiving an alternative salvage strategy.

The registry record gives the study objectives and outcome definitions but does not report the statistical analysis plan, any formal sample-size justification, or the observed primary endpoint result. Although overall survival is listed as a secondary outcome, no survival estimates or analysis details are included in the registry record. The design information therefore supports a description of the intended study question and treatment structure, but not a comparative estimate of efficacy.

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:

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 Overall Response (OR) Post Transplant, measured over 4 months after initial treatment/3 months after transplant.

The primary endpoint is the number of participants classified as having an overall response after transplant. The response category combines complete remission (CR), CR without platelet recovery (CRp), and CR with insufficient hematological recovery (CRi). The registry record specifies marrow blast and blood-count criteria for these categories: each requires less than 5% blasts in bone marrow, with platelet and absolute neutrophil count thresholds distinguishing the response subtypes.

Combining these categories creates a broader measure of response than counting complete remission alone. It can include participants whose marrow findings meet the stated blast criterion but whose platelet or neutrophil recovery is incomplete. That broader definition may be useful when evaluating response in a setting where blood-count recovery can vary, but the endpoint should be understood as the registry-defined composite rather than as a single uniform state of blood-count recovery.

The assessment timeframe is listed as 4 months after initial treatment/3 months after transplant. The endpoint is a participant count, not a time-to-event measure, and the record does not provide the observed count or a denominator for an analysis population. It also does not report how missing assessments, early deaths, or participants who did not proceed to transplant were handled. Without those details and the observed result, the definition clarifies what was intended to be counted but does not establish how many participants met the criteria.

Response is used as an early measure of disease status following the treatment sequence. It does not by itself summarize survival, relapse, treatment toxicity, or duration of response. Those outcomes address different aspects of the study objectives and are not interchangeable with the primary response count.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry record identifies a participant-count endpoint and a non-randomized parallel-group design. It does not report the statistical methods used to analyze the outcomes. The explanations below are general statistical context, not a description of an analysis confirmed to have been used in this study.

What does a participant count tell us?

A count records how many participants meet a defined outcome. To interpret a count as a response frequency, readers also need to know the denominator: which participants were eligible for the analysis, whether everyone enrolled was assessed, and how participants without an assessment were handled. A count alone does not show the proportion of a population with the outcome. The registry record gives the overall enrollment but does not report the primary response count or the response analysis denominator.

A response definition also determines exactly what is counted. Here, the primary endpoint combines CR, CRp, and CRi after transplant. Those categories share the stated marrow blast criterion but differ in platelet and neutrophil recovery criteria. A composite can capture several clinically distinct response states in one endpoint. Interpretation should therefore retain the component definitions rather than treating every counted response as identical in all respects.

Why does a confidence interval matter?

General: A confidence interval communicates uncertainty around an estimate. For a response frequency, it would help readers understand the range of values compatible with the observed participant count and denominator under a specified method. The interval's width reflects both the amount of data and the method used. When enrollment is limited, estimates may be imprecise; a point estimate by itself can conceal that uncertainty. The registry record does not report a confidence interval for the primary response endpoint, so its precision cannot be evaluated from the posted information.

A p-value, if used, would address how compatible the data are with a specified statistical hypothesis under the assumptions of the test. It would not measure the size or clinical importance of an effect, nor would it establish causation. No primary endpoint p-value or other analysis result is reported in the registry record.

What does non-randomized allocation imply?

General: Random allocation is one way to balance, on average, measured and unmeasured characteristics between comparison groups. In a non-randomized study, group membership can be associated with participant characteristics, clinical circumstances, or treatment availability. Observed differences between groups may therefore reflect those differences as well as treatment. This is why the donor groups in this study should not be treated as randomized arms or as a like-for-like control comparison. The registry record does not report adjustment methods that might address measured differences.

The donor-group categories describe the transplant donor type, not necessarily distinct randomized treatment assignments. The study's stated purpose includes assessment of treatment safety and feasibility and response, but group-level counts alone cannot determine whether one donor type caused a better or worse outcome than another. Such a claim would require suitable design and analysis information, neither of which is reported in the registry record.

How should the planning illustrations be read?

General: Sample-size calculations use assumptions about the target effect, statistical test, allocation, and desired power to illustrate how much information a future study might need. They are planning aids, not observed outcomes and not evidence that a completed study had a particular probability of detecting an effect. The registry record's planning assumptions describe a two-sided significance level of 0.05, equal allocation between two arms, a normal approximation, and an effect specified as a fraction of the observed effect. It also distinguishes calculations for time-to-event outcomes from other outcomes. These assumptions are described as a planning illustration based on the observed effect, not as the trial's own design.

Because the study is non-randomized and comprises donor groups, the assumed equal allocation between two arms should not be confused with its actual design. The registry record does not report a prospective sample-size justification or the underlying observed effect used for the illustration. Planning estimates must not be interpreted as observed power, nor do they repair the absence of a reported primary response result.

What information is needed for time-to-event outcomes?

General: Overall survival, relapse, and event-free survival are time-to-event outcomes. Their interpretation typically depends on the time origin, event definition, follow-up duration, and handling of participants whose event status is not observed during follow-up. A survival curve can describe the estimated probability of remaining event-free over time, while a hazard ratio, if used, compares event rates under model assumptions. Neither a curve nor a model estimate is reported for these outcomes in the registry record. The record lists overall survival as a secondary outcome but does not specify the analysis method or provide estimates.

For time-to-event analyses, participants without an observed event by their last known follow-up are often treated as censored. General: Censoring methods rely on assumptions about the relationship between follow-up ending and the chance of an event. The registry record does not report follow-up completeness or censoring information, so no assessment of these assumptions is possible.

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("Matched Sibling Donor Group", "Haploidentical Donor Group", "Matched Unrelated Donor Group")
n      <- c(1, 8, 2)
events <- c(1, 6, 0)
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  Matched Sibling Donor Group  1  1  100.0  2.5  100.0  20.7  100.0  15.8  98.7
ARM  Haploidentical Donor Group  8  6  75.0  34.9  96.8  40.9  92.9  40.0  92.5
ARM  Matched Unrelated Donor Group  2  0  0.0  0.0  84.2  0.0  65.8  0.8  70.8

11. Multiplicity, Missing Data and Analysis Populations

The study description includes multiple objectives and outcomes: post-transplant overall response, treatment safety and feasibility, transplant eligibility, treatment-related mortality, marrow blasts before transplant, relapse, overall survival, and event-free survival. Considering multiple outcomes can raise multiplicity concerns, because testing many hypotheses can increase the chance of at least one apparently notable result. Whether and how that issue was addressed depends on the analysis plan. A multiplicity adjustment or a hierarchy of outcomes is not reported in the registry record.

The primary endpoint is identified, but the registry record does not provide a statistical analysis plan or report whether other outcomes were treated as exploratory, secondary confirmatory outcomes, or descriptive measures. It is therefore not possible to determine which comparisons, if any, were prespecified for formal testing or how interpretation across multiple outcomes was intended.

Missing-data handling is also not reported. In particular, the record does not specify how participants without an endpoint assessment were treated, whether outcome status was available for every enrolled participant, or how participants who did not proceed to transplant were accounted for in the post-transplant endpoint. These details matter because an analysis limited to participants who reach transplant may answer a different question from an analysis of everyone enrolled for initial salvage treatment.

The analysis populations are not described in the registry record. It is not stated whether the primary endpoint was analyzed among all enrolled participants, those who received salvage chemotherapy, those who underwent transplant, or another defined group. Without those definitions, the response endpoint's denominator and relationship to the overall enrollment cannot be established.

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: skin gvhd (12.5%), tacrolimus induced pres (12.5%), secondary graft failure due to gvhd (12.5%), fungal infections (12.5%), viral infections (12.5%).

The registry record reports serious adverse-event summaries separately for the three donor groups. For the matched sibling donor group, 1 participant was listed as at risk and 0 as affected by serious adverse events. For the haploidentical donor group, 8 were listed as at risk and 5 as affected. For the matched unrelated donor group, 2 were listed as at risk and 0 as affected. These are group-specific registry counts; they should not be interpreted as a randomized comparison of donor strategies.

The record also lists deaths as 1 in the matched sibling donor group, 7 in the haploidentical donor group, and 0 in the matched unrelated donor group. It does not specify in this summary the timing, cause, or relationship of these deaths to treatment, or how death relates to the serious adverse-event counts. The numbers should therefore be read as reported summaries, not as a complete mortality analysis.

The listed serious terms are Skin GvHD, Tacrolimus induced PRES, Secondary Graft Failure due to GvHD, Fungal Infections, and Viral Infections, each with a reported percentage of 12.5%. The registry record does not explain the denominator underlying these percentages or whether participants could contribute more than one event. It also does not provide event-level timing or duration. Counts of participants affected and percentages for listed terms are different summaries and should not be assumed to have the same denominator.

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

The registry record shows that the study was designed to examine allogeneic transplantation as initial salvage treatment for patients with primary induction failure acute myeloid leukemia, using a DCIA chemotherapy regimen before transplant. It defines the primary outcome as a post-transplant overall response count that includes CR, CRp, and CRi. The record also gives serious adverse-event and death counts by donor group.

Those details do not establish the primary response frequency, because the observed response count and analysis denominator are not reported. Nor do they establish the precision of any response estimate: no confidence interval or statistical analysis is presented for that endpoint. The listed overall survival outcome has no reported survival estimate in the registry record.

The non-randomized donor groups do not provide a controlled comparison of donor types or treatment strategies. Differences in their counts cannot be interpreted as causal effects, particularly because group assignment procedures and adjustment methods are not reported. The adverse-event summaries are useful as descriptions of the registry entries, but the record does not provide enough information about definitions, timing, follow-up, or denominators to support broader safety comparisons.

The stated planning assumptions are illustrations rather than the study's own sample-size justification. They do not provide evidence about efficacy, observed power, or whether the study was adequately sized for its objectives. Overall, the registry entry describes what the investigators intended to evaluate, but it does not report the outcome results and analysis detail needed to determine what the enrolled participants experienced on the primary endpoint or how outcomes compare with another approach.

14. Why This Trial Matters Statistically

This study is statistically notable for the gap between a clearly stated clinical question and the limited outcome information in the registry record. The primary endpoint is defined in detail, including the response categories and assessment timeframe, yet the observed response count is not reported. As a result, readers can identify what was meant to be measured but cannot estimate the response frequency or its uncertainty from this record.

The small enrollment of 11 makes clear reporting of individual outcome counts, denominators, and uncertainty especially important. In a small study, each participant can substantially affect an estimate, so a count without its analysis population or an interval can be difficult to interpret. This is a general statistical consideration, not a claim about an unreported result in this study.

The non-randomized design also shapes what questions can be answered. Donor groups can support descriptive summaries, but they are not randomized control arms. Without details on group assignment and adjustment, numerical differences between groups cannot isolate the effect of donor type from other factors. The design is therefore not a like-for-like comparative trial of those groups.

Finally, the record lists several safety and longer-term outcomes but does not report their estimates or analysis methods. Distinguishing objectives from results is essential: an outcome being named in a registry does not mean its result is available. This case illustrates why complete reporting should connect endpoint definitions to analysis populations, observed counts, uncertainty, missing-data handling, and methods suitable to the design.

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 main limitation for interpretation is that the registry record does not report the observed primary endpoint count, its denominator, or an uncertainty estimate. It also does not provide outcome estimates for the secondary survival objective or the other outcomes described in the study objectives. Consequently, the record supports no quantitative conclusion about response, survival, relapse, or event-free survival.

Enrollment was 11, and the study was non-randomized. The donor groups are not like-for-like controls, and the registry record does not describe the allocation process or statistical adjustment for differences between groups. Thus, even if group-level outcomes were available, comparisons would require caution and could not automatically be interpreted as causal effects.

The record does not report the statistical analysis plan, the definitions of analysis populations, or how missing endpoint assessments and participants not proceeding to transplant were handled. It also does not state how multiplicity across the several objectives and outcomes was addressed. These omissions limit evaluation of potential selection, analysis, and reporting choices.

The adverse-event information is a limited summary. The record does not clarify the denominator for the listed 12.5% serious-event terms, whether events could overlap within participants, or the timing and ascertainment of the events. The death counts are not accompanied by timing or cause details in the summary. Some donor-specific treatment descriptions are incomplete, further limiting a precise reconstruction of the treatment received in each group.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Allogeneic Hematopoietic Stem Cell Transplantation (AlloSCT) Initial Salvage Therapy for Induction Failure Acute Myeloid Leukemia (AML) — NCT02441803.
• Registry record last updated 2024-02-22; status last verified 2024-02.
• Results first posted 2024-01-17.
• 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

NCT02441803 is a completed phase 2, non-randomized, unmasked, parallel study sponsored by M.D. Anderson Cancer Center. It enrolled 11 participants and evaluated a plan of DCIA salvage chemotherapy followed by allogeneic hematopoietic stem-cell transplantation for primary induction failure acute myeloid leukemia. The transplant groups were matched sibling, haploidentical, and matched unrelated donors.

The primary endpoint was the participant count achieving post-transplant overall response, defined to include CR, CRp, or CRi, at 4 months after initial treatment/3 months after transplant. The registry record does not report the observed primary result, statistical analysis, analysis population, or missing-data approach. It lists overall survival as a secondary outcome without reporting an estimate. Serious adverse-event and death counts are reported by donor group, but key details about denominators, timing, and event handling are not given.