A Phase 1 Study of AMG 330 in Subjects With Myeloid Malignancies: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Phase 1 Study of AMG 330 in Subjects With Myeloid Malignancies (NCT02520427), a non-randomized phase 1 study. The primary outcome is Number of Participants Who Experienced a Dose-limiting Toxicity (DLT).

2. Clinical Question

3. Trial Design

NCT02520427 was a Phase 1, first-in-human study of AMG 330 in subjects with myeloid malignancies. Its official title describes evaluation of safety, tolerability, pharmacokinetics, pharmacodynamics, and efficacy for AMG 330 administered as a continuous intravenous infusion. The registry lists Amgen as sponsor and an enrollment of 95. The study started on 2015-08-31, had a primary completion date of 2022-01-09, and is marked terminated. The reason for termination is not reported in the registry record.

The design was nonrandomized, unmasked, and sequential. The registry summary says that the study tested increasing doses at multiple sites. It lists groups for R/R AML, MRD-positive AML, and MDS, as well as R/R AML groups using alternative pretreatment or dose schedules. The group descriptions are not a randomized comparison: cohort membership and dose differed, and there was no like-for-like control arm. Between-group differences therefore cannot isolate an effect of AMG 330 from differences in dose, disease group, pretreatment, or other cohort features.

The stated aims were to assess safety and tolerability and determine a maximum tolerated dose and/or a biologically active dose. Safety monitoring included vital signs, electrocardiograms, physical examinations, and laboratory tests. Secondary outcomes included antibody formation, response rates by disease group, duration and time to response, event-free survival, and overall survival. The registry lists 18 outcome measures, but detailed results for these secondary outcomes are not included in the record information. The protocol’s dose-escalation rules, stopping criteria, and rationale for the enrollment target 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.

Before assignment: 95 participants were enrolled and received treatment.

Only the first 6 of 26 groups are shown.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: DLT-evaluable Analysis Set: All DLT-evaluable participants.

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.

Only the first 6 of 27 groups are shown.

6. Primary Endpoint

The registry lists the primary outcome as Number of Participants Who Experienced a Dose-limiting Toxicity (DLT), measured over Day 1 to Day 14.

The primary endpoint is the number of participants who experienced a dose-limiting toxicity (DLT), measured from Day 1 to Day 14. It is a participant count, not a count of all toxicity episodes. The registry specifies that the endpoint applies to the DLT-evaluable Analysis Set. Its description excludes participants who did not meet treatment or observation requirements for reasons other than a study-drug-related adverse event. These requirements include completing the DLT window and receiving investigational product for the specified period at the target dose. Additional rules address participants who could not complete repeat cycles after drug interruptions for reasons other than DLT.

In general, a DLT endpoint is used in early-phase dose-finding research to identify toxicity that may constrain dose escalation. It focuses attention on a prespecified observation period and an evaluable population, helping distinguish assessed toxicity from cases where the required assessment was not completed. Here, the registry’s definition makes evaluability dependent on treatment exposure and completion criteria as well as the observation window.

The registry record does not report the number of DLT-evaluable participants or the number who experienced a DLT. It therefore does not support a numerical estimate of the DLT frequency or a conclusion about the maximum tolerated dose. The trial’s stated aim also included identifying a biologically active dose, but the corresponding result is not reported in the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does the sequential design imply?

In general, a sequential dose-escalation design evaluates cohorts in an ordered progression, using accumulating safety information to guide which dose or schedule is studied next. It differs from random allocation: participants are not assigned by chance among contemporaneous treatment arms. This makes such designs suitable for describing early safety experience, but it limits causal comparisons between cohorts. Dose levels, time in the study, disease characteristics, and treatment schedules may all differ. The registry describes this study as sequential and says it tested increasing doses; the specific escalation algorithm and decision rules are not reported in the registry record.

Why define an evaluable population?

For a DLT endpoint, interpretation depends on whether participants received enough treatment and completed the required observation period. A participant who leaves early for a reason unrelated to an adverse event, for example, may not have had the opportunity for a complete DLT assessment. The study’s definition therefore establishes conditions for inclusion in the DLT-evaluable Analysis Set. It also addresses treatment interruptions and completion of repeat cycles. In general, excluding participants who do not meet evaluability rules can make the endpoint denominator smaller than the enrollment figure. The registry does not report the endpoint denominator or the number excluded, so the scale of any such difference cannot be assessed.

How should a participant count be read?

A count of participants with an event answers how many people met the endpoint definition; it is not the same as a total count of event episodes. A frequency or proportion would also require a denominator. The primary endpoint is specified as a participant count, but the result and denominator are not reported in the registry record. No rate should therefore be inferred from the enrollment figure. In general, count outcomes can be compared or modeled in different ways, but a valid analysis requires the analysis population, event definition, and statistical method to be stated.

What do confidence intervals and significance tests add?

In general, a confidence interval expresses the uncertainty around an estimated effect or event frequency under a specified statistical model. A narrow interval indicates greater precision than a wide interval, while neither alone establishes that an effect is clinically important. A significance test evaluates compatibility between data and a specified null hypothesis under its assumptions; a p-value is not the probability that the null hypothesis is true. The registry information lists no statistical analyses, and no confidence interval, p-value, or model-based estimate for the primary endpoint is reported. There is consequently no inferential result to interpret for that endpoint.

What is the role of the planning assumptions?

The stated planning illustration assumptions are a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. The illustration is described as using a fraction of the observed effect. In general, such calculations explore the sample size that might be needed under chosen assumptions; they do not describe the sample size actually analyzed and are not observed power. They also depend on a meaningful observed effect and a comparison that can be specified. This trial is nonrandomized, and its arms are not a like-for-like control comparison. The registry’s own design justification and sample-size rationale are not reported in the registry record. No planning result should be used as evidence that the trial had adequate power to answer a treatment-effect question.

Why does over-dispersion matter in general?

For count data, over-dispersion means that variability is greater than a simple statistical model expects. If ignored, uncertainty can be understated, making intervals too narrow and significance tests too optimistic. Appropriate modeling depends on how counts were generated, exposure time, repeated events, and the study design. The primary endpoint is a participant count, and the registry does not report a statistical model or analysis. There is no basis in the record for assessing over-dispersion or selecting a particular count model.

What can the safety summaries establish?

Descriptive counts show events recorded in the listed cohorts, but they do not on their own provide a controlled estimate of excess risk. General comparisons of safety require attention to denominators, duration of exposure, event definitions, attribution, and cohort composition. Those features are not fully described in the registry information used for this interpretation. The adverse-event figures should therefore be treated as descriptive registry summaries rather than as adjusted or comparative estimates.

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("Group 1 - Cohort 1 AMG 330 0.5 µg/Day", "Group 1 - Cohort 2 AMG 330 1.5 µg/Day", "Group 1 - Cohort 3 AMG 330 4 µg/Day", "Group 1 - Cohort 4 AMG 330 10 µg/Day", "Group 1 - Cohort 5 AMG 330 30 µg/Day", "Group 1 - Cohort 6 AMG 330 10/30 µg/Day")
n      <- c(1, 1, 3, 3, 2, 3)
events <- c(0, 0, 0, 0, 1, 0)
for (i in seq_along(n)) {
  ci <- binom.test(events[i], n[i])$conf.int          # Clopper-Pearson exact interval
  cr <- qbeta(c(0.025, 0.975), 1 + events[i], 1 + n[i] - events[i])   # Bayesian, uniform Beta(1,1) prior
  wi <- (function(x, n) { p <- x / n; z <- qnorm(0.975); d <- 1 + z^2 / n
    c((p + z^2 / (2 * n) - z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d, (p + z^2 / (2 * n) + z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d) })(events[i], n[i])
  cat(sprintf("ARM|%s|%d|%d|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f\n", arm[i], n[i], events[i], 100 * events[i] / n[i],
              100 * ci[1], 100 * ci[2], 100 * wi[1], 100 * wi[2], 100 * cr[1], 100 * cr[2]))
}

ARM  Group 1 - Cohort 1 AMG 330 0.5 µg/Day  1  0  0.0  0.0  97.5  0.0  79.3  1.3  84.2
ARM  Group 1 - Cohort 2 AMG 330 1.5 µg/Day  1  0  0.0  0.0  97.5  0.0  79.3  1.3  84.2
ARM  Group 1 - Cohort 3 AMG 330 4 µg/Day  3  0  0.0  0.0  70.8  0.0  56.1  0.6  60.2
ARM  Group 1 - Cohort 4 AMG 330 10 µg/Day  3  0  0.0  0.0  70.8  0.0  56.1  0.6  60.2
ARM  Group 1 - Cohort 5 AMG 330 30 µg/Day  2  1  50.0  1.3  98.7  9.5  90.5  9.4  90.6
ARM  Group 1 - Cohort 6 AMG 330 10/30 µg/Day  3  0  0.0  0.0  70.8  0.0  56.1  0.6  60.2

11. Multiplicity, Missing Data and Analysis Populations

The study lists multiple secondary outcomes, including anti-AMG 330 antibody formation, response rates for R/R AML, MRD-positive AML, and MDS, duration and time to response, event-free survival, and overall survival. Testing or estimating many outcomes can raise a multiplicity issue: in general, the chance of at least one apparently notable finding can increase when many hypotheses are examined. A prespecified hierarchy or adjustment can address some multiplicity concerns, but the registry record does not report a multiplicity plan, a testing hierarchy, or adjustment methods. It also lists no statistical analyses.

Missing data are especially relevant to the primary endpoint because DLT evaluability depends on treatment exposure and completion of the observation requirements. The registry definition identifies circumstances in which participants would not be DLT-evaluable, but it does not report how many participants met those circumstances or the primary endpoint denominator. The handling of missing secondary outcome data, censoring rules for time-to-event outcomes, and sensitivity analyses are not reported in the registry record.

The record identifies a DLT-evaluable Analysis Set for the primary endpoint, but analysis populations for the secondary outcomes are not described in the information available. It is not possible to determine whether analyses used intention-to-treat, per-protocol, safety, or other populations, or how participants were assigned to those populations. The absence of these details limits assessment of potential selection effects and the completeness of follow-up. No conclusions about the direction or impact of missing data can be drawn without the relevant counts and analysis methods.

12. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 5% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: febrile neutropenia (100.0%), cytokine release syndrome (100.0%), arthritis infective (100.0%), sepsis (100.0%), acute myeloid leukaemia (100.0%).

The serious adverse-event data are organized by dose cohort and other study groups. Each entry gives counts for participants at risk and participants affected, alongside other adverse-event counts and deaths. These are descriptive figures, not randomized comparisons. Cohort sizes are small in the listed entries, and the cohorts differ in dose, disease group, or treatment schedule; the counts should not be interpreted as a dose-response estimate or a comparison of treatment strategies.

Serious-event terms listed include febrile neutropenia, cytokine release syndrome, arthritis infective, sepsis, acute myeloid leukaemia, and pneumonia. The record attaches percentages to these terms, including 100.0% and 66.7%. The denominator for those percentages is not reported in the registry record, so they should not be read as percentages of all enrolled participants. The record information also does not establish whether events were judged related to AMG 330. A listed death is not, by itself, evidence that treatment caused the death.

In general, safety interpretation should consider event definitions, exposure time, attribution, and the population used as the denominator. Those details are not fully. These counts document events recorded in the listed groups; they do not establish comparative risk or a causal relationship.

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

The registry record describes a first-in-human Phase 1 study that explored continuous intravenous AMG 330 in people with myeloid malignancies. Its sequential dose design and DLT endpoint were directed toward early safety assessment and dose finding. The record also lists several disease groups and secondary outcomes covering antibody formation, response, and survival. This documents the scope of the study, but a list of outcomes is not evidence that an outcome was achieved.

The primary endpoint result is not reported in the registry record. As a result, the available information does not show how many DLT-evaluable participants had a DLT, what the endpoint frequency was, or what dose was considered the maximum tolerated or biologically active dose. The listed enrollment cannot supply the missing endpoint denominator. No statistical analysis results are listed.

The safety counts describe events in cohorts that were not randomized and were not designed as like-for-like control groups. They do not establish whether an event was caused by AMG 330, nor do differences between cohorts establish a dose effect. Likewise, the record does not report results for the secondary efficacy outcomes in the information. No conclusion about comparative efficacy or survival follows from the trial description.

In general, early-phase data can inform subsequent research by describing exposure, safety observations, and dose choices. The strength of such information depends on complete outcome reporting, clearly defined populations, and appropriate uncertainty estimates. The record’s design and safety summaries can be understood descriptively, but without primary and secondary results and their analysis details, the treatment effect and its precision cannot be determined.

14. Why This Trial Matters Statistically

This trial is statistically instructive because its central question is not the same as a conventional randomized efficacy comparison. The sequential, nonrandomized design tests increasing doses and focuses the primary endpoint on dose-limiting toxicity during a defined assessment window. That structure highlights how early-phase research uses exposure and short-term safety assessment to inform dose decisions. It also makes clear why cohort counts should not be treated as if they came from randomized, comparable arms.

The definition of the DLT-evaluable population matters as much as the endpoint label. Required treatment exposure and completion of observation determine who can contribute to the primary assessment. If the population and denominator are not reported, even a simple participant count cannot be converted into an interpretable event frequency. This illustrates the importance of reporting the flow from enrollment to endpoint evaluation alongside outcome counts.

The listed adverse events also demonstrate the distinction between a safety signal and a causal estimate. Counts across cohorts can describe what was recorded, but differing doses, diagnoses, and schedules complicate comparisons. Exposure duration, attribution, and uncertainty are needed to interpret those observations responsibly.

Finally, statistical transparency requires the results and methods to accompany the design. The registry information lists many outcomes but no analysis results. Without estimates, uncertainty measures, missing-data handling, or multiplicity methods, the record cannot support a quantitative efficacy conclusion. The study’s statistical value therefore lies in its early-phase design and clearly stated endpoint concept, while the incomplete outcome information constrains what can be learned from the posted record.

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 registry record marks the study as terminated but does not report the reason. It does not provide a primary endpoint result, the number of DLT-evaluable participants, or the number excluded from that population. Although the enrollment is listed as 95, that figure cannot be treated as the denominator for the DLT endpoint.

The record describes a sequential, nonrandomized design without masking or a like-for-like control comparison. The listed groups differ in dose, disease category, and, in some cases, pretreatment or schedule. Cohort comparisons therefore have substantial design limitations and cannot establish causal dose or treatment effects.

The adverse-event data are presented across groups, but the information does not fully describe event attribution, exposure duration, or the denominator used for the listed serious-event term percentages. The counts and deaths cannot be interpreted as comparative risks or as evidence of treatment causation. The registry’s primary endpoint description sets out evaluability rules, but does not state how many participants failed them or how missing data were addressed.

Although secondary outcomes are listed, their results are not reported in the registry information described. Statistical analysis methods, multiplicity handling, missing-data procedures, time-to-event censoring rules, and the sample-size rationale are also not reported in the registry record. These gaps prevent assessment of the precision and robustness of outcome estimates and limit interpretation to the study design and descriptive safety information.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: A Phase 1 Study of AMG 330 in Subjects With Myeloid Malignancies — NCT02520427.
• Registry record last updated 2024-07-31; status last verified 2024-02.
• Results first posted 2024-07-31.
• 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

NCT02520427 is a terminated, nonrandomized, unmasked, sequential Phase 1 study of continuous intravenous AMG 330 in myeloid malignancies. The registry lists an enrollment of 95 and groups covering R/R AML, MRD-positive AML, and MDS, including alternative pretreatment or dosing schedules. The primary endpoint was the number of DLT-evaluable participants who experienced a DLT from Day 1 to Day 14. The registry record does not report that endpoint’s result or statistical analysis results. It lists secondary response, antibody, and survival outcomes and provides cohort-level adverse-event information. Because the cohorts were not randomized comparisons and the endpoint denominator and analysis methods are not reported, the posted information supports descriptive interpretation, not a comparative efficacy or causal safety conclusion.