Treo/Flu/TBI With Donor Stem Cell Transplant for Patients With Myelodysplastic Syndrome or Acute Myeloid Leukemia: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Treo/Flu/TBI With Donor Stem Cell Transplant for Patients With Myelodysplastic Syndrome or Acute Myeloid Leukemia (NCT01894477), a randomized phase 2 study. The primary outcome is Number of Participants That Did Not Progress Within 6 Months.

2. Clinical Question

The central statistical contrast is between Arm B, which adds low-dose total-body irradiation to the treosulfan and fludarabine regimen, and Arm A, which uses treosulfan and fludarabine without irradiation. The endpoint counts participants who did not progress by the stated time point. The comparison therefore describes the observed proportions meeting that definition in the posted result; it does not, on its own, describe the severity or timing of relapse, other causes of illness, or all outcomes after transplantation.

The registry record identifies the trial as randomized and phase II, but it does not report the detailed randomization procedure, planned analysis population, or complete protocol-level rationale for the endpoint. The posted comparison gives a risk difference and additional relative measures. Those estimates describe the contrast between the two arms as recorded, with uncertainty that is important to consider alongside the point estimate.

3. Trial Design

This is a randomized, parallel-group phase II trial. Participants were assigned to one of two conditioning regimens before allogeneic hematopoietic cell transplantation. The registry lists no masking, so the study is described as open label. The sponsor is Fred Hutchinson Cancer Center. Enrollment is listed as 102. The study start is recorded as 2013-11, and primary completion is recorded as 2017-01-09. Current study status is listed as unknown.

Arm A received treosulfan and fludarabine phosphate. The regimen description specifies intravenous treosulfan over 2 hours on days -6 to -4 and intravenous fludarabine phosphate over 30 minutes on days -6 to -2. Arm B received the same treosulfan and fludarabine regimen and low-dose total-body irradiation on day 0. Both arms are described in the context of donor stem cell transplantation. The intervention list includes allogeneic bone marrow transplantation and peripheral blood stem cell transplantation, but the record does not specify the distribution of transplant sources in the primary outcome groups.

The primary endpoint is the number of participants who did not progress within 6 months after transplant, with progression defined as relapse. The posted primary outcome has 35 participants in Arm A and 65 in Arm B. These are the denominators for the primary result, while the overall enrollment is listed separately as 102. The registry record does not explain the difference between enrollment and the primary outcome denominators, and it does not report a detailed account of exclusions or missing endpoint observations.

Secondary outcomes include acute and chronic graft-versus-host disease, relapse or progression, non-relapse mortality, overall survival, and gene-expression measures. Seven secondary outcomes are listed as posted. The record does not provide the detailed analysis methods for those outcomes in the information summarized by the primary comparison. Randomization supports a comparative trial design, but interpretation still depends on the completeness of outcome follow-up, the analysis population, and the methods used to handle missing data; those details 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

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 Number of Participants That Did Not Progress Within 6 Months, measured over At 6 months post-transplant.

The primary endpoint counts participants who did not experience progression within 6 months after transplant. The registry defines progression as relapse and gives the timeframe as 6 months post-transplant. The outcome is reported as a count in each arm, with each count paired with the number of participants in that arm. This makes the endpoint a binary status for each participant at the stated time: either progression was recorded by that point or it was not.

A fixed-time outcome can provide a direct summary of the proportion meeting a clinically relevant event definition at a specified follow-up point. In this trial, the stated endpoint focuses on relapse by 6 months; it is not the same as overall survival, relapse-free time, or a composite of relapse and death. The registry lists overall survival and other outcomes separately, but their results are not part of the primary comparison.

The endpoint is useful for comparing the observed frequency of the defined status between the conditioning arms. It does not convey when within the follow-up interval relapse occurred, and a count of participants without recorded progression should not be interpreted as evidence about outcomes beyond the specified timepoint. Detailed rules for determining endpoint status, handling participants without complete follow-up, or classifying other events are not reported in the registry record.

7. Results Posted in the Registry

• Arm B (Treosulfan, Fludarabine Phosphate, TBI): 73.8% (48/65) versus 57.1% (20/35) on control; risk difference 16.7 percentage points (95% CI -2.9 to 36.3), risk ratio 1.29 (0.94–1.78); unadjusted two-proportion z-test p = 0.088.

8. Statistical Interpretation

9. Statistical Methods Explained

The main comparison is between two randomized arms and a binary endpoint: whether each participant had not progressed within 6 months after transplant. The registry reports the count meeting that endpoint and the total number in each arm. A proportion is the natural summary for a binary outcome because it expresses the number meeting the definition relative to the number assessed. The comparison then expresses how those observed proportions differ between the arms.

What does a risk difference express?

A risk difference subtracts one arm’s event probability from the other arm’s event probability. Here the event is defined as not progressing within 6 months, and Arm B is compared with Arm A. The posted risk difference is 0.167. A positive value indicates a higher observed proportion meeting the endpoint in Arm B. Unlike a relative measure, a risk difference expresses the contrast on an absolute probability scale. It is useful for understanding how far apart the observed proportions are, but its estimate is subject to sampling variation and should be read with its confidence interval.

How should relative measures be read?

The registry also reports a risk ratio and an odds ratio. A risk ratio compares the probabilities of the endpoint in the two arms by division. An odds ratio compares the odds, where odds express the probability of the endpoint relative to the probability of not meeting it. These measures use different scales, so their numerical values need not match. A ratio of 1 represents no relative difference; values on either side describe the direction and size of the relative contrast. The reported intervals for both measures include 1, so the posted estimates are not precise enough to exclude the null relative contrast.

Odds ratios can differ noticeably from risk ratios when an outcome is not rare. This is a general statistical property: an odds ratio is not interchangeable with a risk ratio, and it can appear farther from 1. For that reason, interpretation should use the measure’s name and scale rather than treating all relative estimates as though they were probabilities. The risk difference is also distinct from the ratio measures: it describes an absolute contrast and has zero, rather than 1, as its no-difference value.

What does the confidence interval add?

A confidence interval gives a range of values compatible with the estimate under the statistical method and its assumptions. It conveys uncertainty that the point estimate alone does not show. In this trial, the risk-difference interval crosses zero, so the posted data do not rule out a null difference. The reported intervals for the ratios cross 1, their corresponding null value. The intervals also include values pointing in different directions, which means the data do not provide a sharply determined estimate of the contrast. A confidence interval is not a guarantee that the true value lies within its bounds, and it does not show the probability that any particular value is true.

How should the p-value be understood?

The registry reports an unadjusted p-value of 0.088 for the comparison. A p-value describes how unusual data at least as incompatible with a specified null model would be, if that model and the assumptions of the test were correct. It is not the probability that the null hypothesis is true, and it does not measure the size or clinical importance of the observed contrast. The registry record does not report the detailed test specification or any multiplicity adjustment. The p-value should therefore be read as the posted unadjusted result, alongside the effect estimate and interval, rather than as a standalone decision rule.

“Unadjusted” indicates that the posted comparison is not described as adjusted for other variables. The registry record does not report an adjusted model, so it is not possible to say whether adjustment for prognostic characteristics was planned or performed. In a randomized study, an unadjusted arm comparison is a straightforward way to summarize the randomized groups, but the exact analysis population and implementation details remain relevant. The registry does not provide those details for this result.

What is over-dispersion, and does it apply?

Over-dispersion is a general concern for count outcomes when variability exceeds what a chosen statistical model assumes. It is particularly relevant for counts that arise from repeated events or clustered observations. The primary outcome is reported as the number of participants meeting a binary endpoint, rather than as a repeated-event count. The registry does not report a model or dispersion parameter for this endpoint. There is therefore no basis in the record to claim that an over-dispersion correction was used or needed. The general concept should not be confused with uncertainty in a binary proportion, which is already reflected in the uncertainty interval for the comparison.

What is not specified?

The registry record does not report the original sample-size justification, the detailed allocation procedure, the protocol’s statistical analysis plan, the planned handling of missing data, or the analysis population definition for the primary endpoint. It also does not describe interim analyses. These details matter because they can clarify which methods were prespecified and how participants with incomplete follow-up were treated. Their absence from the registry record should not be filled in by assumption. The posted result can be described as an unadjusted comparison with the reported estimates and uncertainty, while the underlying protocol-level decisions remain unknown.

10. Confidence Intervals and Precision

Precision concerns how narrowly the data estimate the arm contrast. The risk-difference estimate is 0.167, while the reported interval extends from -0.029 to 0.363. The interval includes zero, the value corresponding to no difference on the risk-difference scale. It therefore leaves open the possibility of little or no difference as well as effects in either direction. Its breadth also indicates that the precise size of the contrast is uncertain.

The relative measures tell a similar story about uncertainty. The risk ratio is 1.292, with an interval from 0.937 to 1.782. The odds ratio is 2.118, with an interval from 0.889 to 5.047. Both intervals include 1, the no-relative-difference value. These ranges should not be read as competing conclusions; they are different ways to express the same underlying arm comparison, each on its own scale.

The numbers of participants in the primary outcome groups are 35 in Arm A and 65 in Arm B. The unequal denominators and the overall enrollment of 102 are relevant context when considering precision and completeness. The registry record does not explain the primary outcome denominators or report whether any endpoint information was missing. It is not possible to determine from the posted information alone whether differences in follow-up or exclusions affected the estimate.

A confidence interval describes uncertainty under the method used to produce it; it does not incorporate every source of uncertainty, such as limitations in outcome ascertainment or unreported analysis choices. The methods behind the reported interval are not fully described in the registry record. Thus, the interval is informative about sampling uncertainty as presented, but the available record does not support a complete evaluation of all possible biases or sources of imprecision.

11. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The sample-size figures are planning illustrations based on the observed effect, not the trial’s original design calculation. They are reported as participants per arm for 80% power and participants per arm for 90% power, not as total events. The assumptions listed for this illustration are a two-sided significance level of 0.05, equal allocation between the two arms, a normal approximation, and an effect to detect equal to the stated fraction of the observed effect. The trial’s own original sample-size assumptions are not reported in the registry record.

“Participants per arm for 80% power” means the approximate number in each of two equally allocated groups that the illustration indicates would be needed to have an 80% chance of detecting the specified effect under the stated assumptions, if that effect were the true one. The 90% power column describes a higher target chance of detecting that same specified effect, again under the illustration’s assumptions. These are planning quantities, not descriptions of the number actually randomized or analyzed in this trial.

The rows consider planning for 100% of the observed effect, 75% of the observed effect, and 50% of the observed effect. The listed participants per arm for 80% power are 126, 231, and 533, respectively; for 90% power they are 169, 309, and 713, respectively. Smaller effects require more participants because a less pronounced difference is harder to distinguish from random variation. The curve represents the same general relationship between the size of the effect targeted and the planning sample size.

These calculations are not observed power. They do not tell how likely the completed trial was to find a difference, nor do they replace the confidence interval or the posted result. Because the illustration uses the observed effect as its starting point, it is a way to explore hypothetical planning under stated assumptions, not evidence about the trial’s original design or a guarantee of future results.

12. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

n is the number of participants in each arm, p₀ and p₁ are the control and treatment proportions, and p̄ is their average.

Worked calculation for this trial

1. Control proportion p₀ = 0.571; treatment proportion p₁ = 0.738; average p̄ = 0.655.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.6723 = 1.3177.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.6618 = 0.5570.
4. n = (1.3177 + 0.5570)² / (0.1670)² = 126.0, rounded up to 126 participants per arm for 80% power.

The smaller-effect rows in the table above repeat the same calculation with the effect multiplied by 0.75 and 0.5 (for ratios, applied to the logarithm of the ratio).

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
p0 <- 0.571428571428571                          # control proportion
p1 <- 0.738461538461539                          # treatment proportion
n_per_arm <- function(power, frac = 1) {
  q1 <- p0 + (p1 - p0) * frac               # smaller effect = fraction of the observed difference
  pbar <- (p0 + q1) / 2
  z_b <- qnorm(power)
  ceiling(round((z_a * sqrt(2 * pbar * (1 - pbar)) + z_b * sqrt(p0 * (1 - p0) + q1 * (1 - q1)))^2 / (q1 - p0)^2, 6))
}
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: n per arm = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(n_per_arm(0.80, f), big.mark = ","), format(n_per_arm(0.90, f), big.mark = ",")))

# Check against R's built-in function (unrounded n per arm)
cat("power.prop.test:", round(power.prop.test(p1 = p0, p2 = p1, power = 0.80)$n, 1), "\n")

# Simulation: how often does the z-test reject at the 80%-power sample size?
set.seed(1894477)
n    <- n_per_arm(0.80)
reps <- 20000
x0 <- rbinom(reps, n, p0); x1 <- rbinom(reps, n, p1)
ph <- (x0 + x1) / (2 * n)
z  <- ((x1 - x0) / n) / sqrt(ph * (1 - ph) * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(z) > z_a, na.rm = TRUE), reps))

Output from running the code

100% of observed effect: n per arm = 126 (80% power), 169 (90% power)
 75% of observed effect: n per arm = 231 (80% power), 309 (90% power)
 50% of observed effect: n per arm = 533 (80% power), 713 (90% power)
power.prop.test: 126 
Simulated power at n = 126 per arm: 0.805 (20000 replicates)

The simulation draws many pairs of event counts at the planned sample size and counts how often the two-proportion z-test rejects the null. It should land close to the target power; at very small required sample sizes it can fall below it, because the normal approximation is rough there.

13. Multiplicity, Missing Data and Analysis Populations

The registry lists secondary outcomes including acute graft-versus-host disease, chronic graft-versus-host disease, relapse or progression, non-relapse mortality, overall survival, and gene-expression outcomes. Seven secondary outcomes are listed as posted. The primary result discussed for the 6-month endpoint is only one of several outcome questions in the study. When multiple outcomes or analyses are examined, the chance of at least one apparently unusual result can increase. Statistical approaches can address multiplicity, but whether any adjustment was planned or applied is not reported in the registry record.

The posted p-value for the primary comparison is described as unadjusted. The registry record does not report a multiplicity strategy, a hierarchy of endpoints, or whether the primary endpoint was analyzed under a prespecified confirmatory plan. Without those details, the p-value should be read as the reported unadjusted comparison rather than as a result known to account for all outcome testing in the study.

The primary endpoint denominators are 35 in Arm A and 65 in Arm B, while total enrollment is listed as 102. The record does not explain the difference, give a participant flow account, or state how missing endpoint observations were handled. It is therefore not possible to determine whether participants were excluded from the endpoint analysis, lacked follow-up, or were treated under another rule. No assumptions should be made about the reasons.

The record also does not specify the primary analysis population, such as whether it included all randomized participants or another defined group. Nor does it report imputation methods, sensitivity analyses for missingness, or outcome adjudication procedures. These omissions limit assessment of how robust the posted estimate is to incomplete data and analysis choices. They do not negate the reported counts; they identify information needed to understand the full analysis context.

The record lists no detailed analyses in the registry analysis summary. The existence of a posted comparison does not by itself fill in missing methodological detail. Protocol-level definitions, analysis plans, interim monitoring, and multiplicity procedures are not reported in the registry record, so they cannot be inferred from the outcome numbers.

14. 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: respiratory failure (8.6%), infection (6.0%), mental status changes (5.7%), renal failure (5.7%), cmv gastritis (2.9%).

The serious adverse event figures are reported by safety group. Arm A lists 6 participants with a serious adverse event among 35 at risk. Arm B lists 9 among 67 at risk. The other, non-serious adverse event figures are 28 among 35 at risk in Arm A and 51 among 67 at risk in Arm B. The registry also lists 15 deaths in Arm A and 20 in Arm B. These are descriptive counts with their listed safety denominators; the record does not provide a comparative statistical analysis of these figures.

The safety denominators differ from the primary outcome denominators, which are 35 for Arm A and 65 for Arm B. The registry record does not explain the difference for Arm B or describe the rules used to define the safety groups. The adverse event figures should therefore not be substituted for the primary endpoint data or treated as if they had identical follow-up and analysis populations.

The registry lists respiratory failure at 8.6%, infection at 6.0%, mental status changes at 5.7%, renal failure at 5.7%, CMV Gastritis at 2.9%, and loss of vision at 2.9% among the top serious adverse event terms. These terms are presented as registry-listed summaries; the record does not establish from this information that each term applies to a particular arm. Counts and terms alone do not provide a complete assessment of event timing, severity, attribution, or comparative risk.

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

The primary result describes the proportion of participants who did not have progression, defined as relapse, by 6 months after transplant. More participants in Arm B met this endpoint in the posted data: 48 of 65, compared with 20 of 35 in Arm A. The estimated risk difference is 0.167 for Arm B compared with Arm A. These are observations from the trial’s recorded outcome groups, and the comparison is accompanied by appreciable uncertainty.

The confidence interval for the risk difference includes zero, and the unadjusted p-value is 0.088. The relative-measure intervals also include their null value. These features mean that the posted result does not establish a statistically clear difference between the regimens. They also do not prove that the regimens have the same effect. A non-significant result is not evidence of equivalence unless an equivalence design and its criteria are specified; such criteria are not reported in the registry record.

The endpoint is limited to relapse status at the 6-month time point. It does not by itself summarize survival, chronic or acute graft-versus-host disease, non-relapse mortality, longer-term relapse, or quality of life. The registry lists several secondary outcomes, but their results are not part of the primary comparison. The primary result cannot answer every question about transplant outcomes or the relative safety of the conditioning strategies.

Randomization is an important feature of the trial design, but detailed information about allocation, analysis populations, missing data, and the statistical plan is not reported in the registry record. The primary outcome denominators differ by arm, and the record does not explain why. These limitations make it important to avoid treating the point estimate as a complete causal or clinical conclusion.

In statistical terms, the posted figures are compatible with a range of possible underlying differences. The data show a higher observed proportion without recorded progression in Arm B, but the uncertainty does not support a definitive claim about the size or direction of the underlying contrast. No treatment recommendation follows from these statistics alone.

16. Why This Trial Matters Statistically

This trial poses a clear statistical comparison: two conditioning strategies are evaluated in a randomized, parallel-group phase II design, and the primary outcome is defined at a fixed time after transplant. The endpoint is a binary participant-level status, which allows the observed counts and proportions to be compared directly. The registry also reports absolute and relative measures, giving more than one view of the contrast.

The central teaching point is the distinction between the observed point estimate and its precision. The risk difference of 0.167 is positive for Arm B, but the confidence interval crosses zero and the unadjusted p-value is 0.088. Reporting only the point estimate could make the result seem more certain than it is. Reading the interval alongside it shows that the plausible range includes a null contrast and values in either direction.

The record also illustrates why endpoint denominators matter. Enrollment is listed as 102, while the primary outcome is reported for 35 participants in Arm A and 65 in Arm B. Safety counts use a different denominator in Arm B. Without participant flow and missing-data information, a reader cannot fully determine how those denominators arose. Transparency about analysis populations is important for understanding both the estimate and its generalizability.

Finally, a planning illustration based on an observed effect should not be confused with the original design calculation or with observed power. The registry does not report the trial’s original sample-size justification, and the illustrative requirements change when the effect being targeted changes. Together, these points show why statistical interpretation requires attention to design, outcome definition, denominators, interval estimates, and what the registry does not report, not just a single p-value.

17. 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 registry record gives a concise account of the design and posted outcome but does not report several details needed for a full evaluation. Eligibility criteria, participant characteristics, the randomization procedure, allocation concealment, and the intended analysis population are not described in the information accompanying the primary result. Although the study is identified as randomized and unmasked, those labels do not answer all questions about implementation or potential bias.

The overall enrollment is 102, whereas the primary outcome denominators are 35 in Arm A and 65 in Arm B. The record does not explain the difference or provide a participant flow diagram in the information. It also does not report how missing follow-up, withdrawals, or other endpoint-status uncertainty were addressed. These gaps limit evaluation of whether the analyzed groups fully represent the randomized groups.

The primary endpoint is recorded at 6 months post-transplant, and progression is defined as relapse. More detail about outcome ascertainment and the treatment of death or other events before the endpoint assessment is not reported in the registry record. The endpoint therefore should not be generalized to longer follow-up or interpreted as a complete measure of transplant outcomes.

The posted statistical comparison is unadjusted, while the detailed statistical analysis plan, original sample-size justification, multiplicity strategy, and interim-analysis plans are not reported. The registry also does not give a full analysis description for the listed secondary outcomes. The safety groups have different denominators from the primary outcome groups, and the reason is not explained. These limitations do not change the reported counts or estimates, but they constrain how confidently the result can be interpreted beyond its stated outcome and analysis.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Treo/Flu/TBI With Donor Stem Cell Transplant for Patients With Myelodysplastic Syndrome or Acute Myeloid Leukemia — NCT01894477.
• Registry record last updated 2021-06-03; status last verified 2021-05.
• Results first posted 2018-05-21.
• 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.

21. Record Summary

NCT01894477 is a randomized, open-label phase II parallel-group study of treosulfan and fludarabine conditioning, with or without low-dose total-body irradiation, before allogeneic hematopoietic cell transplantation for participants with myelodysplastic syndrome or acute myeloid leukemia. Enrollment is listed as 102. The primary outcome is the number of participants who did not progress within 6 months after transplant, with progression defined as relapse.

The primary outcome lists 20 of 35 participants in Arm A and 48 of 65 in Arm B as not progressing by the endpoint. The reported risk difference for Arm B compared with Arm A is 0.167, with a confidence interval that includes zero, and the unadjusted p-value is 0.088. Secondary outcomes and safety counts are also listed, but detailed analysis plans, missing-data handling, and the original sample-size justification are not reported in the registry record. Current study status is listed as unknown.