Decitabine and Gemtuzumab Ozogamicin in Acute Myelogenous Leukemia and High-risk Myelodysplastic Syndrome: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Decitabine and Gemtuzumab Ozogamicin in Acute Myelogenous Leukemia and High-risk Myelodysplastic Syndrome (NCT00968071), a na phase 2 study. The primary outcome is Number of Participants With Complete Response (CR).

2. Clinical Question

In statistical terms, the central question was whether CR was observed among participants receiving the combination during the stated assessment period, alongside a description of safety. A single-group result can describe outcomes in the enrolled study population, but cannot isolate the treatment’s contribution from other possible explanations for those outcomes. The registry record does not report enough outcome information to quantify the primary response result.

3. Trial Design

This was a Phase II, completed, single-group study sponsored by M.D. Anderson Cancer Center. The registry record lists an enrollment of 71, a start date of 2008-02, and a primary completion date of 2012-08. Allocation is recorded as NA, and masking as none. These design features mean that the record does not describe random assignment or blinding, nor does it identify a separate control arm.

The intervention was a combination of decitabine and gemtuzumab ozogamicin. The regimen description specifies decitabine 20 mg/m^2 intravenously over an hour and half daily for 5 days and gemtuzumab ozogamicin 3 mg/m^2 intravenously on day 5. The description says response was evaluated after each treatment course, identified as 5-6 weeks, for up to 6 cycles. The primary endpoint timeframe is stated as up to 36 weeks.

Enrollment is not necessarily the same as the number included in an endpoint analysis. The analysis population, the number with an evaluable response, treatment received by each participant, and handling of participants who discontinued or lacked follow-up are not reported in the registry record. The record lists one primary outcome measure and no analyses. Thus, it identifies the endpoint and broad treatment structure, but does not supply a statistical analysis plan or a reported primary outcome estimate.

The distinction between a study design and a planning illustration matters. The registry describes a single-group trial. Any planning calculations based on a hypothetical two-arm comparison do not make the clinical study a two-arm trial and do not create a control group. They should not be interpreted as observed power or as evidence of a treatment difference.

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 Number of Participants With Complete Response (CR), measured over Up to 36 weeks.

The primary endpoint was the number of participants with complete response (CR), with assessment up to 36 weeks. CR was defined using findings from both peripheral blood and bone marrow: normalization with 5% or fewer blasts, a peripheral ANC of at least 1 * 10^9 /l, and a platelet count of at least 100 * 10^9 /l. The registry description says evaluations occurred after each treatment course, described as 5-6 weeks, up to 6 cycles.

A response endpoint of this kind records whether participants meet a prespecified clinical and laboratory definition. It turns several measurements into a binary classification for each participant: CR or not CR, according to the stated criteria. Counting participants meeting the definition can summarize how often the endpoint was attained in the study population. The count alone, however, does not describe how consistently the result was assessed, how many participants had missing assessments, or how precisely the underlying response frequency is estimated.

General statistical context: a defined response endpoint can make results easier to interpret because the criteria specify what qualifies as a response. Its interpretation still depends on details such as the analysis population, assessment rules, and handling of missing or unevaluable results. Those details are not reported in the registry record. Although the endpoint is named and defined, its observed participant count is not reported in the registry record, so the primary response result cannot be quantified from the posted information.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The statistical methods that can be described from the registry record are limited. It identifies one primary outcome measure, a participant count for complete response, and no analyses. A named endpoint is not the same as a reported analysis: the record does not state an estimator, confidence interval method, hypothesis test, adjustment procedure, or analysis population. The explanations below distinguish general statistical principles from methods actually documented for this study.

What kind of endpoint is complete response?

The endpoint is a binary outcome at the participant level. A participant either meets the stated CR definition or does not, based on blood and bone marrow findings and the specified thresholds. The primary measure is described as the number of participants with CR. A count is directly interpretable as a tally, but comparisons or response rates require a clearly defined denominator. The registry record gives total enrollment but does not report how many participants had an evaluable primary endpoint or what population was used in an analysis.

General statistical context: binary outcomes can be summarized by counts and proportions. A proportion places the count in relation to an eligible denominator. Uncertainty around that proportion depends on the number of observed outcomes and the denominator, as well as the chosen statistical method. Since the primary CR count and analysis denominator are not reported, neither a response proportion nor its uncertainty can be described from the posted results.

What can a single-group design estimate?

The design is listed as single-group, with allocation NA and no masking. It does not contain a registry-listed randomized comparator. A single-group analysis can describe outcomes within the participants studied. It cannot directly estimate a between-group treatment contrast, because there is no separate arm in the study record against which to compare the combination. Differences from other studies would also be difficult to interpret as treatment effects without addressing differences in populations, assessment, follow-up, and other factors.

General statistical context: random allocation, when used, helps balance measured and unmeasured prognostic factors between groups on average. A single-group design does not obtain that protection through randomization. This is a design limitation for causal comparisons, not a reason to disregard descriptive observations. The registry record does not report an alternative method that would support a causal comparison.

How should uncertainty be understood?

General statistical context: a confidence interval communicates the statistical uncertainty around an estimate under its assumptions. A narrower interval generally indicates greater precision, while a wider interval indicates less precision. An interval does not establish that an estimate is unbiased, and it does not correct for missing data, selection, or lack of a control group. No confidence interval or other uncertainty estimate for the primary endpoint is reported in the registry record. Without an outcome estimate and its analysis denominator, precision cannot be evaluated.

A p-value, if reported, would address compatibility of the data with a specified statistical hypothesis under the assumptions of the test. It would not measure the probability that a treatment works, nor would it by itself establish clinical importance. The registry record lists no analyses and no p-value for the primary endpoint. No hypothesis test or inferential conclusion should be attributed to this record.

What do the planning assumptions describe?

The planning illustration is described as using a two-sided significance level of 0.05, equal allocation between two hypothetical arms, and a normal approximation, with the effect to detect set as a stated fraction of the observed effect. For time-to-event outcomes, the stated convention is a total event count using the Schoenfeld method; for other outcomes, the convention is participants per arm. These are planning assumptions, not evidence that the clinical trial itself had two arms. The trial design is single-group.

General statistical context: sample-size planning asks how much information a future study might need under selected assumptions. Results depend on the assumed effect, outcome variability or event rate, allocation, significance level, and analysis method. A planning illustration is not an observed result, and it is not observed power. In particular, calculations that posit equal allocation between two arms do not describe the observed design of this single-group study. The registry record does not report the study's own sample-size justification or whether these illustrative assumptions were used in its design.

Why does the outcome scale matter?

General statistical context: methods should match the outcome structure. A count of participants meeting a binary definition differs from a continuous laboratory measure or a time-to-event endpoint. For a binary count, relevant issues include the denominator, completeness of assessment, and the rule used to classify each participant. Methods for time-to-event outcomes would require information about event times and censoring, neither of which is part of the stated primary endpoint. The registry lists no time-to-event primary measure.

Over-dispersion is a general concern when count data vary more than a model assumes, which can make model-based uncertainty estimates too small if ignored. It is not possible to determine whether over-dispersion is relevant to an analysis of this primary endpoint because no model or analysis is reported; the endpoint is a participant count rather than a reported repeated-event count. No model choice, diagnostic assessment, or variance adjustment is described in the registry record.

In sum, the endpoint definition and planning assumptions provide some context, but the record does not document a completed statistical analysis of CR. The observed result, denominator, analysis method, uncertainty estimate, and treatment comparison are not reported in the registry record.

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("Decitabine Gemtuzumab Ozogamicin")
n      <- c(71)
events <- c(3)
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  Decitabine Gemtuzumab Ozogamicin  71  3  4.2  0.9  11.9  1.4  11.7  1.5  11.7

11. Multiplicity, Missing Data and Analysis Populations

The registry record lists one outcome measure and no analyses. It does not report a multiplicity strategy, such as whether any adjustment was planned for testing multiple outcomes, time points, or subgroups. Because no inferential analyses are listed, there is no reported set of tests from which to assess how multiplicity was handled. The fact that one primary measure is listed does not establish what other analyses may have been planned or conducted.

Missing-data handling is also not reported in the registry record. The primary endpoint was assessed over treatment courses up to 6 cycles, but the record does not describe how participants without an assessment, those who stopped treatment, or those who could not be classified were handled. It does not state whether the analysis used all enrolled participants, only evaluable participants, or another population. The response denominator therefore cannot be identified from the posted results.

General statistical context: missing outcome data can affect both the denominator and the interpretation of a response estimate. Excluding participants without an assessment can produce a different summary from counting them as nonresponders, and either approach depends on assumptions about why data are missing. A transparent analysis typically identifies the population analyzed and explains the treatment of missing observations. Those details are not reported in the registry record.

The record also does not state an estimand or a statistical analysis plan. General context: an estimand clarifies the treatment effect or outcome quantity a study seeks to describe, including how events such as discontinuation are treated. These omissions prevent assessment of alignment between the planned question, the analysis population, and the posted results. They do not establish that a particular method was or was not used outside the registry record.

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: death (21.1%).

The safety entries list 15 participants affected by serious adverse events, with 71 at risk, and 44 affected by other adverse events, also with 71 at risk. The top serious adverse event term is Death, listed as 21.1%. These figures are counts and a listed percentage as recorded; they should be read as descriptive information, not as estimates of a comparative treatment effect.

The deaths field has no reported value in the registry record. The listed term Death at 21.1% should not be converted into a death count or assumed to represent the value of that separate field. The record does not explain the relationship between the term listing and the deaths field, nor does it provide additional detail to clarify that relationship.

General statistical context: serious adverse events are a category of events, while adverse-event terms describe the event types recorded. Participant counts and event-term percentages may summarize different aspects of safety reporting. Without definitions of the counting rules, event-level detail, and attribution information, they should not be combined or interpreted as causal evidence. The registry record does not report a comparator, event timing, or a treatment-attribution analysis. Thus, the listed figures describe what the record reports among those at risk, but do not by themselves establish that the intervention caused an event or indicate how these counts compare with another treatment.

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

The available record establishes the study's broad design and its stated clinical question: a Phase II evaluation of decitabine plus gemtuzumab ozogamicin in AML or high-risk MDS, with complete response as the primary endpoint and safety also under study. It specifies a CR definition and describes when response assessments were intended to occur. It also reports enrollment and selected safety counts.

It does not report the number of participants who achieved CR, the denominator used for response analysis, or any analysis of that endpoint. As a result, the posted information cannot show the observed frequency of CR or how precisely it was estimated. It also cannot establish whether outcomes differed from those under another treatment, because the registry lists a single-group design and no comparator.

The safety figures are likewise limited. The registry lists serious and other adverse-event participant counts and a serious-event term with a percentage, but provides no comparative safety estimate or reported attribution analysis. The deaths field has no reported value. Those entries should not be used to draw conclusions beyond their descriptive content.

General statistical context: a study can produce useful descriptive information even when it does not support a causal comparison. But a treatment-effect claim requires a suitable contrast and analysis, and an estimate is easier to judge when its denominator and uncertainty are reported. The registry record does not include those elements for the primary endpoint. This means the missing result cannot be replaced by assumptions based on the intervention description, enrollment, or safety data.

Planning illustrations answer a different question from observed trial results. They can explore hypothetical sample-size requirements under stated assumptions, but do not report what happened to participants and are not observed power. Their assumptions about hypothetical arms do not alter the single-group design recorded for this trial. Overall, the record supports an account of the study structure and the outcomes it intended to examine, but leaves the principal response result and its statistical interpretation unresolved.

14. Why This Trial Matters Statistically

This trial is statistically notable because the registry identifies a clinically defined binary primary endpoint but does not report its observed count or analysis. That distinction illustrates why an endpoint description alone is not enough to judge a study result. For a response count to be interpretable, readers also need the number assessed, the analysis population, and a clear account of participants whose outcomes were unavailable. Those details are not reported in the registry record.

The design also makes the boundary between description and comparison important. A single-group study can record outcomes among participants receiving the combination, but it does not supply a randomized control group for estimating a comparative treatment effect. A response result, if reported, would need to be interpreted within that design rather than as a direct comparison with another treatment.

The record's safety entries offer another statistical lesson: distinct types of information should not be conflated. Counts of participants affected, an adverse-event term percentage, and a deaths field are not interchangeable measures. Their interpretation depends on definitions and reporting rules, some of which are not reported in the registry record. Descriptive figures are valuable, but they do not by themselves establish causation or comparative risk.

Finally, planning calculations should be kept separate from observed evidence. Assumptions that posit equal allocation between two arms can support a hypothetical planning exercise, but they do not describe this single-group trial. Recognizing that distinction helps avoid treating a design illustration as either a trial result or proof of study power. The central statistical takeaway is that sound interpretation depends not only on what outcome was chosen, but also on the analysis and reporting details needed to quantify it.

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 primary limitation for interpreting outcomes is that the registry record does not report the primary CR count or a statistical analysis. Although enrollment is listed as 71, the number of participants with an evaluable response and the analysis denominator are not reported. The precision of any response estimate therefore cannot be assessed.

The single-group design has no registry-listed control arm and no random allocation. It can support descriptive reporting about the study group, but it does not provide a like-for-like randomized comparison. The record does not report adjustment methods or another comparative approach that would address this design limitation.

Further details needed to evaluate the result are not reported in the registry record, including the analysis population, missing-data rules, endpoint assessment completeness, and statistical analysis plan. The response definition is stated, but the record does not explain how participants who were not assessed or could not be classified were handled. It also does not report uncertainty estimates for the primary endpoint.

Safety interpretation is limited by the descriptive nature of the posted counts. The record gives the numbers at risk and affected, but does not report event-level detail, timing, attribution, or a control group. The separate deaths field has no reported value, while Death appears among the listed serious adverse-event terms at 21.1%; the relationship between these entries is not explained in the registry record.

Finally, the record does not report the study's sample-size justification, interim analyses, or multiplicity plan. Planning assumptions listed for illustrative calculations should not be mistaken for the trial's design or a documented analysis plan. These gaps restrict both assessment of statistical precision and interpretation of the results.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Decitabine and Gemtuzumab Ozogamicin in Acute Myelogenous Leukemia and High-risk Myelodysplastic Syndrome — NCT00968071.
• Registry record last updated 2013-03-08; status last verified 2013-02.
• Results first posted 2013-03-08.
• 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

The record describes a completed Phase II, single-group study of decitabine and gemtuzumab ozogamicin for AML and high-risk MDS, sponsored by M.D. Anderson Cancer Center. Enrollment was 71; the listed start date is 2008-02 and primary completion date is 2012-08. The described regimen combines decitabine 20 mg/m^2 intravenously over an hour and half daily for 5 days with gemtuzumab ozogamicin 3 mg/m^2 intravenously on day 5.

The primary endpoint was the number of participants with CR, assessed up to 36 weeks. The registry record defines CR and states that safety was also studied. It lists 15 participants with serious adverse events and 44 with other adverse events, each among 71 at risk, and lists Death at 21.1% among serious adverse-event terms. The deaths field has no reported value.

The CR result, endpoint analysis, and analysis population are not reported in the registry record. With no randomized comparator listed, the record does not provide a comparative treatment-effect estimate.