A Phase II Study of Crenolanib in Relapsed/Refractory Acute Myeloid Leukemia Patients With FLT3 Activating Mutations: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Phase II Study of Crenolanib in Relapsed/Refractory Acute Myeloid Leukemia Patients With FLT3 Activating Mutations (NCT01657682), a non-randomized phase 2 study. The primary outcome is Response Rate of Patients Receiving Crenolanib Therapy.

2. Clinical Question

3. Trial Design

This is a Phase II study of crenolanib besylate in relapsed or refractory AML with FLT3 activating mutations. Its allocation is listed as non-randomized, its masking as none, and its model as parallel. Enrollment is recorded as 56. The study status is completed; the record gives a start date of 2012-10 and a primary completion date of 2019-04.

The brief description identifies two treatment-history cohorts: patients whose leukemia recurred after chemotherapy that did not include a FLT3 TKI, and patients whose leukemia progressed after prior treatment with a FLT3 TKI. The arm descriptions name Cohort A as having no prior FLT3 TKI exposure and Cohort B as having prior FLT3 TKI therapy. These are clinical-history groups, not randomized treatment arms. The registry record does not describe a concurrent control group or a comparison treatment.

The safety results also list Cohort C, described as having an antecedent hematological disorder. The record does not explain how this group relates to the stated Cohort A and Cohort B descriptions, or whether membership in the groups could overlap. The safety counts should therefore be read under the labels recorded, rather than as evidence of mutually exclusive study groups.

The primary outcome is response rate, assessed from first dose through the end of protocol treatment. Secondary outcomes are duration of overall survival and study drug exposure. The registry record does not report a detailed statistical analysis plan, the criteria for analysis populations, or a sample-size justification. The enrollment entry alone does not establish how many participants contributed data to each outcome.

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 Response Rate of Patients Receiving Crenolanib Therapy, measured over From the date of first dose to the end of protocol treatment..

The primary endpoint is the response rate of patients receiving crenolanib. It is a binary outcome at the participant level: each participant is assessed against protocol response categories, and the response rate describes the proportion meeting the specified response definition. The registry record identifies the outcome unit as participants and gives the assessment period as first dose through the end of protocol treatment.

The definition distinguishes complete remission (CR), CR with incomplete blood count recovery (CRi), partial response (PR), blast reduction (BR), and resistant disease (RD). CR requires a post-baseline bone marrow biopsy or aspiration with blasts <5%, an absolute neutrophil count >1×10^9/L, and a platelet count >100×10^9/L. CRi meets the CR criteria except that platelet or absolute neutrophil count recovery has not occurred. PR requires a decrease of ≥50% in bone marrow blasts from baseline to a post-baseline result between 5% and 25%. BR is a decrease of ≥50% in blasts. RD is defined by the absence of CR, CRi, CRp, PR, or MLFS.

A response endpoint is used to describe whether participants meet prespecified disease-response criteria during treatment. The categories make the clinical meaning of the outcome dependent on the stated laboratory and marrow definitions. The registry record does not clarify which response categories were combined to calculate the primary response rate, and it does not give the observed response count or rate. Thus the definition can be described, but the endpoint result cannot be reconstructed from this record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry record gives a participant-count response endpoint and a non-randomized, unmasked, parallel study design. It does not report the statistical analysis plan, model, estimator, confidence-interval method, hypothesis test, or analysis population used for the primary outcome. The methods discussion below explains general statistical principles relevant to reading this kind of study; it does not identify methods as having been used unless the registry record states them.

What does a response rate describe?

For a binary endpoint, each participant is classified according to whether the defined event occurred. A response rate is a proportion: the numerator is participants meeting the response definition, and the denominator is the participants included in the analysis. The endpoint definition matters because choices about which categories count as a response change the numerator. Here, the registry describes several response categories, but does not specify which categories were combined for the reported primary response rate. It also does not give the response count, denominator, or rate.

A proportion is an estimate from the participants assessed, rather than a guarantee that the same proportion would be observed in another patient group. A confidence interval is a statistical range expressing uncertainty around an estimated quantity under a specified method and assumptions. It helps readers judge the precision of an estimate. No confidence interval for response is reported in the registry record, so its precision cannot be evaluated from the posted outcome information.

Why does non-randomized allocation matter?

Random allocation is intended to make treatment groups comparable, on average, with respect to measured and unmeasured characteristics. In a non-randomized study, treatment or cohort membership can be related to patient characteristics, disease history, or prior treatment. Those factors may also relate to outcomes. This creates the possibility of confounding: an observed outcome pattern may reflect differences between groups as well as the treatment under study.

This study groups patients partly by prior FLT3 TKI history. That makes the groups useful for describing response and safety in different clinical settings, but does not make them a randomized treatment comparison. The record does not report adjusted estimates or the covariates that might have been considered in an adjustment. No causal comparison between cohorts can be inferred from the design description alone.

What do the planning assumptions mean?

The registry’s sample-size assumptions describe a planning illustration based on an observed effect, not the trial’s own design. They specify a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation; the effect to detect is stated as a fraction of the observed effect. For non-time-to-event outcomes, the convention is participants per arm for 80% power or another stated power level; time-to-event planning uses total events. These are general planning concepts, not observed trial results.

A planning calculation asks what sample size would be needed under specified assumptions to detect a target effect with a chosen probability, if the assumptions hold. It does not show that the completed study had that probability of detecting an effect, and it is not observed power. The registry record does not provide a sample-size justification for the study’s enrollment, and no response effect estimate is reported to support an independent planning interpretation. Any planning values should therefore be read as illustrative rather than as an account of the design rationale.

How should uncertainty and significance be read?

Statistical significance is assessed using a test statistic and a prespecified threshold under a particular analysis plan. A p-value, when reported, describes how compatible the observed data are with a specified null hypothesis and test assumptions. It does not measure the probability that a treatment works, nor does it quantify clinical importance. The registry record reports no p-value or inferential analysis for the primary response endpoint.

Precision and statistical significance answer different questions. An interval can show the range of estimates reasonably compatible with the data under its method, while a test addresses a particular null hypothesis. Neither can be reconstructed without the relevant outcome data and analysis details. The absence of those details prevents assessment of sampling uncertainty, model assumptions, or whether any reported analysis was adjusted.

Why can over-dispersion matter?

General explanation: over-dispersion occurs when variability in count data exceeds what a statistical model assumes. If a model underestimates variability, its uncertainty estimates may be too narrow and its tests may appear more decisive than warranted. This issue is most relevant to count outcomes; the primary outcome is recorded as participants with an event, rather than a count of repeated events. The registry record does not report an over-dispersion assessment or model, so it is not possible to say whether this issue arose in any analysis.

The safety fields report affected-participant counts, but do not give the event-level data or analysis methods needed to assess repeated events, exposure time, or differences in risk. Descriptive counts and term percentages should therefore remain descriptive. The record does not report an adjusted safety comparison, a time-to-event analysis, or methods for handling censoring.

10. Multiplicity, Missing Data and Analysis Populations

The registry lists response rate as the primary outcome and duration of overall survival and study drug exposure as secondary outcomes. It does not report a statistical testing hierarchy, multiplicity adjustment, or a prespecified rule for interpreting several endpoints together. When multiple outcomes are tested, the chance of at least one apparently positive result can increase unless the analysis accounts for multiplicity or interprets results within a prespecified framework. This is a general statistical concern; the record does not establish that formal testing was performed.

Missing-data handling is not reported in the registry record. For a participant-level response endpoint, the denominator could depend on how participants without an assessable post-baseline response are handled. Possible approaches include excluding unassessable cases, classifying them under a prespecified outcome rule, or using a model-based method. Those approaches can yield different estimates, so the missing-data rule is important. The record does not identify the approach used or give the number of participants with missing response assessments.

The record also does not define the primary analysis population. It is not clear whether the response analysis included all enrolled participants, only participants who received treatment, or another defined set. No protocol deviations, discontinuation handling, or sensitivity analyses are described. The listed enrollment therefore should not be treated as the response denominator.

Subgroup definitions and subgroup analysis methods are not reported in detail. Cohorts are described by prior FLT3 TKI exposure, while the safety listing additionally includes Cohort C with an antecedent hematological disorder. The relationship among these labels is not explained. Without clear population definitions and an analysis plan, subgroup counts cannot be assumed to be mutually exclusive or to constitute a formal comparison.

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 0% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: pneumonia (53.8%), febrile neutropenia (30.8%), disease progression (16.7%), atrial fibrillation (15.4%), gastric hemorrhage (15.4%).

The safety record lists all patients with 56 at risk, 52 affected by serious adverse events, 56 affected by other adverse events, and 53 deaths. These are registry counts. They do not by themselves establish that crenolanib caused an event or death. Seriousness is a regulatory classification and is not the same as severity or causal attribution.

The cohort-specific serious adverse-event counts are listed as 10 affected among 13 at risk in Cohort A, 30 among 30 in Cohort B, and 12 among 13 in Cohort C. The record lists other adverse events for all participants at risk in each of those groups. The group labels and denominators are reproduced as recorded; the record does not clarify whether the cohort categories overlap or how Cohort C relates to the treatment-history cohorts.

The registry lists pneumonia at 53.8%, febrile neutropenia at 30.8%, disease progression at 16.7%, and atrial fibrillation, gastric hemorrhage, and nausea at 15.4% each among the top serious terms. The detailed denominator and counting convention for these percentages are not reported in the registry record. They should not be interpreted as comparative risks, and the listed terms do not indicate causality. The record does not provide a control group for estimating excess risk attributable to treatment.

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

The registry record establishes that a Phase II study of crenolanib was conducted in relapsed or refractory AML with FLT3 activating mutations, with response rate named as the primary outcome. It provides detailed response definitions and records enrollment and safety counts. However, it does not report the primary response count or rate, a confidence interval, a p-value, or results for the listed secondary outcomes. The central efficacy question is therefore not answered quantitatively by the outcome information in the record.

The term “completed” describes the recorded study status. It does not mean that results for every outcome are present in the registry record. Likewise, an enrollment entry is not necessarily the denominator for every analysis. Without the analysis population, endpoint-specific denominators, and handling of missing assessments, an outcome estimate cannot be reconstructed.

The non-randomized design also limits what comparisons could establish. Prior treatment history defines the stated cohorts, and the record does not describe a randomized control group. Differences between groups, if reported elsewhere, could reflect differences in participant history or prognosis rather than a treatment effect. The registry record does not provide an adjusted comparative analysis that would address this issue.

Safety counts are informative as recorded descriptions of events among participants in the study. They do not establish whether events were caused by crenolanib, whether risk was higher than with another treatment, or how safety compares across settings. The deaths field is not accompanied by a cause-of-death breakdown in the record.

Accordingly, the record supports a description of the study’s question, design, endpoint definitions, and listed safety data. It does not support a conclusion that crenolanib did or did not produce a particular response rate, or that it did or did not change survival or safety outcomes. Such conclusions would require outcome estimates and methods not reported in the registry record.

13. Why This Trial Matters Statistically

This study is statistically notable as a non-randomized Phase II evaluation in a population defined by relapsed or refractory AML and FLT3 activating mutations. Its cohorts distinguish prior FLT3 TKI exposure, a clinically relevant feature of treatment history. Such a design can describe outcomes in specific groups, but it does not provide the protections against confounding that random allocation is intended to provide.

The response endpoint uses explicit bone marrow and blood count criteria. A well-defined endpoint is important because a response rate depends on how response is classified, when it is assessed, and which participants enter the denominator. The registry record supplies category definitions and an assessment timeframe, but does not specify the aggregation rule for the primary rate or report its result. This illustrates why endpoint definitions and analysis rules need to be read together.

The safety record also highlights the distinction between descriptive event counts and causal comparisons. Serious adverse events, other adverse events, and deaths are separately recorded, but there is no randomized control group described for estimating comparative risks. A list of frequent serious terms can characterize what was recorded during the study, yet it cannot establish treatment attribution on its own.

Finally, the record’s sample-size assumptions are identified as planning illustrations based on an observed effect, rather than the trial’s own design justification. This distinction matters: a planning sample size is conditional on assumptions and is not evidence of the study’s achieved precision or power. With no primary response estimate or inferential results reported, readers should separate the design and safety descriptions from claims about efficacy or comparative treatment effects.

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.

A key limitation of the registry record is that it does not report the primary response result. The response count, rate, denominator, confidence interval, and statistical test are not reported in the registry record. The definitions describe how response categories are characterized, but do not clarify which categories were combined for the primary endpoint. This prevents independent assessment of the endpoint estimate and its precision.

The study is non-randomized and unmasked, with no control treatment described. Cohorts are based on prior therapy, and the safety record includes an additional cohort label for antecedent hematological disorder. The record does not clarify how Cohort C relates to Cohorts A and B, whether group membership overlaps, or whether the cohort structure was changed during the study. This limits interpretation of between-group descriptions.

Statistical methods for the outcome analyses are not reported in the registry record. Missing-data rules, analysis populations, adjustment for baseline characteristics, handling of multiple outcomes, and sensitivity analyses are also not reported. The enrollment entry cannot substitute for outcome-specific denominators.

Safety counts are presented without a detailed account of event attribution, time at risk, or the denominator and counting convention for the listed serious-event percentages. The record lists deaths but does not provide a cause-of-death breakdown. These limitations preclude comparative safety conclusions and treatment attribution from the registry entries alone. Results for duration of overall survival and study drug exposure are not reported in the registry record.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: A Phase II Study of Crenolanib in Relapsed/Refractory Acute Myeloid Leukemia Patients With FLT3 Activating Mutations — NCT01657682.
• Registry record last updated 2023-11-30; status last verified 2023-11.
• Results first posted 2023-11-30.
• 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

NCT01657682 is a completed Phase II, non-randomized, unmasked, parallel study of crenolanib besylate in relapsed or refractory AML with FLT3 activating mutations. Enrollment is recorded as 56. The described treatment-history cohorts concern prior FLT3 TKI exposure; the safety listing also includes a cohort with an antecedent hematological disorder.

The primary outcome is response rate from first dose through the end of protocol treatment, using bone marrow and blood count response criteria. Duration of overall survival and study drug exposure are listed as secondary outcomes. The registry record does not report numerical results for these outcomes or the statistical analysis plan.

Safety entries include serious adverse events, other adverse events, deaths, and percentages for listed serious-event terms. These are descriptive registry data, without a randomized comparator or reported causal attribution. The record supports an account of the study design and recorded safety information, but does not provide a quantitative primary response result or comparative treatment-effect estimate.