A Study of Axatilimab at 3 Different Doses in Participants With Chronic Graft Versus Host Disease (cGVHD): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study of Axatilimab at 3 Different Doses in Participants With Chronic Graft Versus Host Disease (cGVHD) (NCT04710576), a randomized phase 2 study. The primary outcome is Overall Response Rate (ORR) in the First 6 Cycles as Defined by the 2014 NIH Consensus Development Project on Criteria for Clinical Trials in Chronic Graft-Versus-Host Disease (cGVHD).

2. Clinical Question

The question is about response observed during a defined early treatment period, not about long-term response alone. The primary endpoint covers the first 6 cycles, up to Cycle 7 Day 1; each cycle is 4 weeks. Complete response required resolution of all manifestations in each organ or site. Partial response required improvement in at least 1 organ or site without progression in any other organ or site.

Because the study randomized participants among three axatilimab schedules, the posted results allow dose-group descriptions and comparisons between those groups. They do not provide a comparison with an untreated group or another therapy. The groups also differ in dosing interval: the 3 mg/kg schedule is every 4 weeks, while the other schedules are every 2 weeks. The registry record does not report a separate question isolating the effect of dose amount from the effect of schedule.

3. Trial Design

This is a phase 2, randomized, multicenter, parallel-group study of axatilimab in participants with recurrent or refractory active chronic graft-versus-host disease after at least 2 lines of systemic therapy. It is open-label, with no masking reported. The registry record lists an enrollment of 296 and a Recruiting status. It lists the study start as 2021-03-04 and primary completion as 2023-04-07; those record entries do not, by themselves, establish the status of every follow-up or analysis.

Participants were allocated among three active dose cohorts. One received axatilimab 0.3 mg/kg intravenously every 2 weeks; another received 1 mg/kg intravenously every 2 weeks; and the third received 3 mg/kg intravenously every 4 weeks. Each regimen could continue for up to 2 years. The record does not describe the randomization procedure, allocation concealment, or any stratification factors.

The primary outcome is overall response rate during the first 6 cycles, as defined by the 2014 NIH consensus criteria for chronic graft-versus-host disease clinical trials. The posted arm results use an intent-to-treat analysis set defined separately for each dose level as all participants randomized to that level. The arm-specific analysis counts are 80, 81, and 80, respectively. These are the denominators reported with the primary outcome; the record does not explain why those denominators differ from the stated enrollment.

The registry lists secondary outcomes covering response, symptom scores, response duration and persistence, organ-specific response, corticosteroid dose reduction, and corticosteroid discontinuation. The registry record does not report a control treatment outside the three axatilimab schedules. It also does not report the rationale for the dose schedules, the original sample-size justification, or a detailed statistical analysis plan.

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.

Before assignment: The primary analysis results are reported. The final results will be posted after completion of the study.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: The ITT analysis set for each dose level included all participants randomized to the dose level.

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.

Population: The intent-to-treat (ITT) analysis set for each dose level included all participants randomized to the dose level.

6. Primary Endpoint

The registry lists the primary outcome as Overall Response Rate (ORR) in the First 6 Cycles as Defined by the 2014 NIH Consensus Development Project on Criteria for Clinical Trials in Chronic Graft-Versus-Host Disease (cGVHD), measured over First 6 cycles (up to Cycle 7 Day 1; each cycle = 4 weeks).

The primary endpoint is overall response rate (ORR): the percentage of participants with an objective response, defined as complete response or partial response, during the first 6 cycles. The assessment period extends up to Cycle 7 Day 1, and each cycle is 4 weeks. For this outcome, complete response means resolution of all manifestations in each affected organ or site. Partial response means improvement in at least 1 organ or site, with no progression in another organ or site.

This endpoint summarizes whether participants met a prespecified clinical response definition across affected organs or sites during the assessment window. General statistical context: a binary response outcome is useful for presenting the proportion meeting a defined criterion, and it can be compared across groups as a difference in response probabilities. It does not, by itself, describe how quickly responses began, how long they lasted, or the experience of participants who did not meet the response definition. The registry lists duration of response and sustained response rate separately among secondary outcomes.

The registry record identifies the 2014 NIH consensus criteria but does not report additional details about assessment schedules, independent review, or how unavailable assessments were handled. Those details matter when interpreting how response status was determined, so they should not be inferred from the endpoint label alone.

7. Results Posted in the Registry

• Axatilimab 1 mg/kg Q2W: 66.7% (54/81) versus 73.8% (59/80) on control; risk difference -7.1 percentage points (95% CI -21.2 to 7.0), risk ratio 0.90 (0.74–1.11); unadjusted two-proportion z-test p = 0.326.
• Axatilimab 3 mg/kg Q4W: 50.0% (40/80) versus 73.8% (59/80) on control; risk difference -23.8 percentage points (95% CI -38.3 to -9.2), risk ratio 0.68 (0.53–0.88); unadjusted two-proportion z-test p = 0.002.

8. Statistical Interpretation

9. Statistical Methods Explained

What does the response percentage summarize?

For each dose group, the primary result is a percentage of participants who met the complete- or partial-response definition during the first 6 cycles. The posted analysis set includes all participants randomized to that dose level. The arm results show 73.8% for 0.3 mg/kg every 2 weeks, 66.7% for 1 mg/kg every 2 weeks, and 50.0% for 3 mg/kg every 4 weeks. The corresponding group sizes are 80, 81, and 80. The response counts shown in the comparison details are 59, 54, and 40, respectively.

General statistical context: an observed percentage is an estimate of the response probability in a population like the trial participants. It is not an exact description of the underlying probability, because it is based on a finite group. That sampling variability is why the posted results include confidence intervals.

How should the confidence intervals be read?

A 95% confidence interval describes uncertainty around an estimate using a procedure that, under its assumptions, would cover the target quantity in 95% of repeated samples. It is not a statement that a particular fixed parameter has a 95% probability of lying in the interval. The arm-level intervals show uncertainty in each response proportion. The pairwise intervals show uncertainty in the risk differences between groups.

General statistical context: when a confidence interval for a risk difference includes zero, the data are compatible with no difference as well as with values on either side of zero under that interval procedure. When it excludes zero, the interval does not include a null risk difference. This is a statement about statistical compatibility, not a judgment of importance or a guarantee that the result will reproduce.

What is a risk difference?

A risk difference subtracts one group's response probability from the other group's response probability. Here, the comparison details use the 0.3 mg/kg every 2 weeks group as the reference. For 1 mg/kg every 2 weeks, the recalculated risk difference is -0.071, with a confidence interval from -0.212 to 0.07. For 3 mg/kg every 4 weeks, it is -0.237, with a confidence interval from -0.383 to -0.092. Negative estimates indicate lower observed response probability in the comparison group than in the reference group. These are simple unadjusted recalculations from the posted summary statistics; the registry record does not report a discrepancy with the sponsor's analysis.

The comparison details also list risk ratios and odds ratios, but the interpretation focuses on the risk difference because it directly expresses the absolute difference in response probability. General statistical context: different effect measures answer related but not identical questions, and an odds ratio is not the same quantity as a risk difference. A measure should be read with its scale and reference group in mind.

What do the p-values say?

The unadjusted p-value is 0.326 for the comparison of 1 mg/kg every 2 weeks with 0.3 mg/kg every 2 weeks, and 0.002 for the comparison of 3 mg/kg every 4 weeks with 0.3 mg/kg every 2 weeks. These values are reported for the posted comparisons. They do not quantify the size of a difference, and they do not address whether a difference is clinically important.

General statistical context: a p-value is calculated under a null model and depends on the data and the test assumptions. An unadjusted p-value does not account for testing multiple comparisons unless an adjustment is explicitly incorporated. The registry record does not report a multiplicity procedure for these comparisons, so the values should be treated as unadjusted rather than as confirmed family-wise error-controlled results.

What can be inferred about the analysis?

The registry reports an intent-to-treat analysis set for each dose, including all participants randomized to that level. That definition identifies who belongs in the analysis set; it does not specify how missing endpoint measurements were treated. The method for handling missing response assessments, the statistical test used for the pairwise comparisons, and any model covariates are not reported in the registry record.

The study is open-label and uses clinical response assessments. General statistical context: knowledge of assigned treatment can affect behavior or assessment in some settings, so masking and assessment procedures are relevant when evaluating a clinical endpoint. The registry record does not describe blinded outcome review. It is therefore appropriate to report the observed estimates and their uncertainty without assuming additional protections or procedures.

The comparison calculations are unadjusted recalculations from posted summary statistics. They are useful for explaining the arithmetic scale and uncertainty shown in the posted results, but they do not replace a complete analysis plan or the sponsor's primary analysis. The registry record does not report a detailed plan specifying tests, covariate adjustment, missing-data sensitivity analyses, or multiplicity control.

10. Confidence Intervals and Precision

The posted 95% confidence intervals are important because the arm percentages are estimates rather than exact population response rates. For 0.3 mg/kg every 2 weeks, the response estimate is 73.8%, with an interval from 62.7% to 83.0%. For 1 mg/kg every 2 weeks, it is 66.7%, with an interval from 55.3% to 76.8%. For 3 mg/kg every 4 weeks, it is 50.0%, with an interval from 38.6% to 61.4%. The intervals convey that the underlying response proportions remain uncertain, even though the observed percentages are reported to one decimal place.

For the comparison with 1 mg/kg every 2 weeks, the risk-difference interval extends from -0.212 to 0.07 and includes zero. The interval for 3 mg/kg every 4 weeks extends from -0.383 to -0.092 and lies below zero. These intervals give a more informative account of uncertainty than the point estimates alone: one comparison is compatible with no difference, while the other interval is entirely on the negative side of the reference contrast.

General statistical context: interval width reflects uncertainty under the method used, and interpretation depends on the measure being estimated and its assumptions. Overlap or non-overlap of separate arm-level intervals is not a substitute for examining the interval for the direct comparison. Here, the comparison intervals are posted directly. Precision is also distinct from clinical importance: a relatively precise estimate can still be of uncertain relevance, and an imprecise estimate can encompass effects in more than one direction.

The registry record does not report further details about the confidence-interval calculation method or any adjustment for multiple comparisons. The intervals should therefore be read as the posted estimates, without assuming a particular modeling approach beyond what is reported.

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 table and curve are planning illustrations based on the observed effect in each posted comparison. The entries are participants per arm for 80% power and participants per arm for 90% power; they are not totals across both arms. The listed comparisons are between 0.3 mg/kg every 2 weeks and each of the other dose schedules. These figures are not observed power.

The stated planning assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect to detect is set to the stated fraction of the observed effect. The table labels planning for 100%, 75%, or 50% of the observed effect. The 80% power column represents a planning target for the chance of detecting an effect of the specified size under the stated assumptions if that effect is present. The 90% column uses a higher target chance of detection under those assumptions.

For the comparison of 1 mg/kg every 2 weeks with 0.3 mg/kg every 2 weeks, The table shows 654 participants per arm for 80% power and 874 for 90% power when planning for 100% of the observed effect. For the comparison of 3 mg/kg every 4 weeks with 0.3 mg/kg every 2 weeks, the corresponding entries are 65 and 86 participants per arm. The remaining rows show how the planning counts change when the target effect is set to 75% or 50% of the observed effect; those values are shown in the table and curve.

General statistical context: smaller effects are harder to distinguish from random variation, so detecting a smaller target effect generally requires more participants under otherwise similar assumptions. The registry record does not report the trial's original sample-size justification or original design assumptions. These illustrations are based on the observed comparisons and should not be read as the design calculation that determined trial enrollment.

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 = 0.738; treatment proportion p₁ = 0.667; average p̄ = 0.702.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.6468 = 1.2677.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.6448 = 0.5427.
4. n = (1.2677 + 0.5427)² / (-0.0708)² = 653.2, rounded up to 654 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 <- 0.7375                          # control proportion
p1 <- 0.666666666666667                          # 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(4710576)
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 = 654 (80% power), 874 (90% power)
 75% of observed effect: n per arm = 1,142 (80% power), 1,529 (90% power)
 50% of observed effect: n per arm = 2,523 (80% power), 3,378 (90% power)
power.prop.test: 653.2 
Simulated power at n = 654 per arm: 0.802 (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 primary outcome is overall response rate during the first 6 cycles. The registry also lists secondary outcomes for response on study, clinically significant improvement in normalized modified Lee Symptom Scale score, duration and persistence of response, organ-specific response, joints and fascia response, corticosteroid dose reductions, and corticosteroid discontinuation. The record lists 18 outcome measures and posts 17 secondary outcomes. The registry record does not explain the status of any outcome measure without a posted secondary result in this record.

Two pairwise primary-outcome comparisons are displayed against the 0.3 mg/kg every 2 weeks group. Their p-values are explicitly unadjusted. The registry record does not report a testing hierarchy, family-wise error procedure, or other multiplicity adjustment. Because multiple dose comparisons and secondary outcomes are considered, the unadjusted values should not be interpreted as though the record had documented control of the chance of false-positive findings across all analyses.

The primary analysis population is described as the intent-to-treat set for each dose level, including all participants randomized to that dose. This is a clear statement of group membership, but it does not answer how missing endpoint assessments were handled. The method for assigning response status when an assessment was absent, the amount and pattern of missing data, and any sensitivity analyses are not reported in the registry record.

Likewise, the record does not report subgroup analyses, covariate-adjusted analyses, or a detailed analysis plan. This absence does not establish that such analyses were not performed; it means their methods are not documented in the registry record. The posted estimates and comparisons can be described as shown, while broader claims about prespecification, missing-data robustness, or multiplicity control remain unsupported by this record.

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 5% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: pneumonia (9.9%), covid-19 (6.3%), cellulitis (3.8%), respiratory failure (3.8%), hypoxia (3.7%).

The safety figures summarize participants recorded as having serious adverse events, other adverse events, and deaths in each dose group. For serious adverse events, the record reports 30 affected participants among 79 at risk in the 0.3 mg/kg every 2 weeks group, 33 among 81 in the 1 mg/kg every 2 weeks group, and 38 among 79 in the 3 mg/kg every 4 weeks group. The other-adverse-event counts are 74 of 79, 78 of 81, and 78 of 79, respectively. The posted death counts are 2, 7, and 12.

The listed serious adverse-event terms include pneumonia at 9.9%, COVID-19 at 6.3%, cellulitis at 3.8%, respiratory failure at 3.8%, hypoxia at 3.7%, and generalised oedema at 2.5%. The registry record does not show these terms by dose group in the listed summary.

These are descriptive counts, not evidence that axatilimab caused each event or death. The record does not provide event-level timing, attribution, exposure-adjusted rates, or comparative statistical analyses for these figures. The at-risk denominators for safety also differ from some primary-outcome denominators; the reasons are not reported. The figures should be read as posted safety summaries rather than as a complete account of individual clinical circumstances.

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

The primary results show the proportion of participants in each randomized axatilimab schedule who met the response definition during the first 6 cycles. The observed percentages were 73.8% for 0.3 mg/kg every 2 weeks, 66.7% for 1 mg/kg every 2 weeks, and 50.0% for 3 mg/kg every 4 weeks. The confidence intervals show uncertainty around each estimate, and the direct comparisons indicate different degrees of compatibility with a zero risk difference.

These results do not establish that one dose schedule is universally preferable, that any schedule causes a response, or that the observed ordering will be the same in another population. The registry record does not list a placebo or non-axatilimab arm, so these comparisons do not estimate the difference between axatilimab and no axatilimab or another treatment. Since dose amount and dosing interval vary together across schedules, the results also do not isolate the effect of dose amount alone.

The endpoint is limited to an objective response assessment in a defined early period. It does not itself establish how long responses lasted, whether symptoms improved for every participant, or how participants fared after the assessment window. The record lists separate secondary outcomes relevant to symptom scores, response duration, organ-specific response, and corticosteroid use, but their results are not reported in this record.

General statistical context: a confidence interval and a p-value address uncertainty under statistical assumptions; neither is a direct measure of treatment benefit for an individual. The unadjusted p-values should not be treated as a complete accounting of all comparisons in the study. Safety counts likewise describe events recorded in the groups, but do not establish causality. Interpretation should remain tied to the design, endpoint, and information actually documented in the registry record.

16. Why This Trial Matters Statistically

This study is statistically informative because it places three active axatilimab schedules in randomized parallel groups and reports a common response endpoint for each. The arm-level estimates and their confidence intervals make the observed response proportions and uncertainty visible, while the risk-difference comparisons provide a direct scale for describing differences between the dose groups.

The pairwise results illustrate why effect estimates, intervals, and p-values should be read together. The 1 mg/kg comparison has an interval that includes zero and an unadjusted p-value of 0.326, while the 3 mg/kg comparison has an interval below zero and an unadjusted p-value of 0.002. These are not interchangeable with judgments about clinical importance, and the record does not report multiplicity control for the comparisons.

The design also raises questions that matter for interpretation: it is open-label, uses clinical response criteria, and changes both dose amount and schedule across cohorts. The available record gives the endpoint definition and group results, but not the assessment-blinding details, missing-data method, or full statistical plan. Recognizing what is documented and what is not prevents a narrow endpoint result from being overextended into claims about durability, comparative benefit against other treatments, or causation of safety events.

The planning illustration is another useful statistical feature, provided it is kept separate from the trial's own design. It shows how hypothetical sample-size requirements depend on the observed effect and the target effect size under stated assumptions. It does not recover the original enrollment rationale or establish power for the completed analyses.

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.
• Participant counts were reconstructed from posted percentages and arm sizes, so small rounding differences are possible.

The primary endpoint is assessed over the first 6 cycles, so the posted primary result does not answer how durable a response was or describe outcomes outside that window. Although the registry lists response duration and sustained response rate as secondary outcomes, results for those outcomes are not posted in this record.

The study has no listed placebo or non-axatilimab comparison group. It compares three axatilimab schedules, and the 3 mg/kg group also has a different dosing interval from the other groups. Consequently, the comparisons do not isolate dose amount from schedule and do not provide a direct estimate against another treatment or no treatment.

The open-label design is relevant because the endpoint uses clinical assessments. The registry record does not report whether assessors were blinded, whether central review was used, or what procedures supported consistency across centers. It also does not report the randomization procedure, allocation concealment, missing-response handling, or sensitivity analyses. These details cannot be inferred from the posted arm estimates.

The record states an enrollment of 296, while the primary outcome denominators are 80, 81, and 80 and the safety denominators are 79, 81, and 79. The reasons for the differences among these denominators are not reported in the registry record. The record also does not report the original sample-size justification, a complete statistical analysis plan, or a multiplicity adjustment. The pairwise p-values are unadjusted, and the planning illustrations are based on observed effects rather than the original trial design assumptions.

Finally, the safety summary gives counts and selected serious adverse-event terms but does not provide event-level timing, attribution, or exposure-adjusted rates in the information. These limitations restrict how far the posted efficacy and safety summaries can be generalized or interpreted.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: A Study of Axatilimab at 3 Different Doses in Participants With Chronic Graft Versus Host Disease (cGVHD) — NCT04710576.
• Registry record last updated 2026-06-16; status last verified 2026-05.
• Results first posted 2024-10-10.
• 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

The registry record describes a randomized, open-label, multicenter phase 2 study of axatilimab in participants with recurrent or refractory active chronic graft-versus-host disease after at least 2 lines of systemic therapy. It compares 0.3 mg/kg every 2 weeks, 1 mg/kg every 2 weeks, and 3 mg/kg every 4 weeks, with treatment for up to 2 years.

The primary outcome is overall response rate during the first 6 cycles, assessed using the 2014 NIH consensus criteria. The posted estimates are 73.8%, 66.7%, and 50.0% across the three groups, respectively, with arm-level 95% confidence intervals. The direct comparison with 1 mg/kg every 2 weeks has an unadjusted p-value of 0.326; the comparison with 3 mg/kg every 4 weeks has an unadjusted p-value of 0.002.

The registry lists secondary outcomes and safety summaries, but it does not report a full statistical plan, missing-data methods, multiplicity procedures, or the original sample-size justification. The posted results support a description of the observed response proportions and their uncertainty; broader interpretations should remain within the limits of the record.