Efficacy and Safety of Selumetinib in Adults With NF1 Who Have Symptomatic, Inoperable Plexiform Neurofibromas: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Efficacy and Safety of Selumetinib in Adults With NF1 Who Have Symptomatic, Inoperable Plexiform Neurofibromas (NCT04924608), a randomized phase 3 study. The primary outcome is Confirmed Partial and Complete Response Rate (ORR) by End of Cycle 16 Using Volumetric MRI Analysis as Determined by ICR (Per REiNS Criteria) in Participants With NF1 Who Have Symptomatic, Inoperable PN..

2. Clinical Question

The clinical question is whether selumetinib is associated with a different rate of confirmed tumor response than placebo among adults with neurofibromatosis type 1 (NF1) who have symptomatic, inoperable plexiform neurofibromas (PNs). The registry’s primary endpoint is the objective response rate by the end of Cycle 16, assessed by volumetric MRI and determined by ICR under REiNS criteria. A response requires a confirmed complete response or partial response.

The randomized comparison is between selumetinib 25 mg/m2 and placebo. The outcome is a proportion of participants, not a direct measure of pain, quality of life, or change in every tumor. Those outcomes are addressed separately among the posted secondary endpoints. The primary outcome’s timeframe runs from the first dose until progression, if it occurs before the end of Cycle 16, or the last evaluable assessment through that cycle. MRI during a prolonged study-intervention interruption is excluded; the registry defines that interruption as at least 28 days.

The registry record does not report a broader clinical decision rule for interpreting the response-rate difference. The numerical results describe this trial’s recorded comparison and should not be treated as a complete account of every outcome relevant to participants.

3. Trial Design

KOMET is a Phase III, multicentre, international, randomized, parallel-group, placebo-controlled study. The registry describes a 2 arm design and quadruple masking. The planned enrollment was 145. The recorded groups were selumetinib 25 mg/m2, with 71 participants, and placebo, with 74 participants. The record labels the trial active, not recruiting.

The study start date is 2021-11-19, and the primary completion date is 2024-08-05. The listed sponsor is AstraZeneca. The trial’s stated aim is to assess selumetinib’s efficacy and safety in adults with NF1 who have symptomatic, inoperable PNs. The registry record does not report which groups of people were masked under the quadruple-masking designation, how masking was maintained, or whether masking was assessed during the trial.

Random allocation is useful because, in expectation, it balances participant characteristics across groups and supports a causal comparison of assigned interventions. It does not ensure that every characteristic is identical in a particular trial. A placebo comparator and masking can also reduce the influence of expectations on care, assessment, and reporting, although the details of implementation are not reported in the registry record.

The primary analysis population is described as the full analysis set: all patients randomized to study intervention. This is a stated population definition, not a description of how missing outcome assessments were handled. The registry record does not report a detailed accounting of participants who discontinued, missed scans, or lacked an evaluable assessment for the primary endpoint.

The primary outcome is a confirmed response determined by ICR using volumetric MRI and REiNS criteria. The posted secondary outcomes include pain intensity at Cycle 12, with primary and supplemental analyses, and PlexiQoL total score at Cycle 12. These outcomes differ in measurement scale and timing from the primary response-rate endpoint. The record does not report the full protocol, detailed allocation procedures, or the design assumptions used to choose the enrollment target.

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:

Analysis population for the primary outcome, as defined in the registry: Full analysis set - All patients who are randomized to study intervention

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 Confirmed Partial and Complete Response Rate (ORR) by End of Cycle 16 Using Volumetric MRI Analysis as Determined by ICR (Per REiNS Criteria) in Participants With NF1 Who Have Symptomatic, Inoperable PN., measured over From first dose up until progression (if it occurs prior to the end of Cycle 16), or the last evaluable assessment up to and including the end of Cycle 16, excluding MRI during prolonged study intervention interruption (defined as interruption >= 28 days).

The primary endpoint is the confirmed objective response rate by the end of Cycle 16. It is the proportion of participants who had a confirmed complete response or confirmed partial response, assessed through volumetric MRI by ICR under REiNS criteria. A complete response is defined as disappearance of the target PN, confirmed by a consecutive scan within 3 to 6 months after the first response. A partial response requires a target PN volume decrease of at least 20% compared with baseline, also confirmed by a consecutive scan within 3 to 6 months after the first response.

This endpoint expresses whether a participant met a prespecified tumor-volume response definition during the stated period. It is not the same as the average change in tumor volume across all participants, nor does it directly measure pain or quality of life. The registry also defines progression as an increase in target PN volume of 20% or more compared with baseline, or compared with the time of best response after a partial response has been documented.

The assessment window begins with the first dose and ends at progression if progression occurs before the end of Cycle 16, or at the last evaluable assessment through the end of that cycle. MRI during prolonged study-intervention interruption, defined in the record as at least 28 days, is excluded. A response-rate endpoint is useful for describing the share of participants who meet a clinically specified response criterion; its interpretation depends on the criteria, confirmation process, assessment schedule, and completeness of outcome assessment. Details beyond those recorded are not reported in the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is binary at the participant level: a participant either met the confirmed response definition within the assessment window or did not. The registry expresses each group’s result as a percentage and gives a 95% confidence interval for each rate. A confidence interval describes the statistical uncertainty associated with an estimate under the method used. It is not a range that guarantees the true rate lies within it, and it is not a direct interval for the difference between the two groups.

How to read the primary comparison

The reported primary comparison uses Fisher Exact. This test evaluates the evidence for an association between randomized group and the response count in a two-group setting. It is suited to categorical count data, including settings where some observed counts are small. Its p-value summarizes how compatible the observed count pattern is with the test’s null hypothesis, under the test assumptions. The registry gives p = 0.011. That value does not quantify the response-rate difference, and the posted results do not include a confidence interval for that difference.

Fisher Exact differs from the continuous-outcome methods used for the secondary endpoints. For a continuous outcome, the reported analyses use mixed models and least-squares means. Least-squares means are model-based group estimates; a reported least-squares mean difference compares the modeled change between groups. The model specification, covariates, and assumptions are not reported in the registry record, so the fitted analysis cannot be fully reconstructed from the posted summaries alone.

Reported and recalculated secondary comparisons

For the continuous secondary endpoints, the posted group summaries include least-squares mean changes and standard errors. The registry also reports a mixed-model difference, a 95% confidence interval, and a p-value for each analysis. Recalculated figures are simple unadjusted comparisons derived from the posted summary statistics, rather than re-fits of the sponsor’s model. They are useful as a transparent cross-check, but do not replace the registry’s model-based analysis.

The unadjusted recalculations use approximate standard deviations inferred from the reported standard errors, with a pooled standard deviation for standardized effect-size calculations. That approximation cannot reproduce all the information used by a mixed model, such as participant-level data, covariance structure, or any model adjustments. As a result, recalculated confidence intervals and p-values may differ from the reported model-based results. The registry record does not report enough model detail to resolve every possible source of difference.

For the first key secondary outcome’s primary analysis, the recalculated unadjusted mean difference is -0.7, with a 95% interval from -1.518 to 0.118 and p = 0.093. The registry’s mixed-model result is -0.8, with a 95% confidence interval from -1.6 to 0.1 and p = 0.070. For its supplemental analysis, the recalculated difference is -0.7, with a 95% interval from -1.296 to -0.104 and p = 0.021; the registry reports -0.7, with a 95% confidence interval from -1.3 to -0.1 and p = 0.024. For PlexiQoL, the recalculated unadjusted difference is -0.1, with a 95% interval from -1.334 to 1.134 and p = 0.874; the registry reports -0.1, with a 95% confidence interval from -1.2 to 1.1 and p = 0.918.

Planning illustrations are not trial design evidence

The sample-size illustrations are calculated from observed effects, not from the trial’s own design assumptions. They assume a two-sided significance level of 0.05, equal allocation between arms, and a normal approximation. For each illustration, the target effect is a stated fraction of the observed effect. Such calculations show how sample-size requirements change under those assumptions; they are not observed power and do not establish the power the trial was designed to achieve. The registry record does not report the trial’s sample-size justification or whether its original planning used these assumptions.

Planning figures based on observed effects are especially sensitive to the estimated effect being used. An observed effect can be imprecise, and a planning illustration that treats it as fixed does not account for that uncertainty. These values should therefore be understood as conditional examples, not as evidence that a larger or smaller study would necessarily produce a particular result.

What the statistics do not establish

A p-value is not a measure of clinical importance, and a confidence interval does not account for every source of uncertainty. Incomplete outcome data, measurement choices, model assumptions, and selective reporting can affect interpretation. Randomization supports a comparative interpretation of assigned groups, but the registry summary alone does not establish how missing scans were addressed or how well the analysis followed the planned protocol. Those details are not reported in the registry record.

10. Analyses Reported by the Sponsor

For the primary response-rate outcome, the registry reports the sponsor’s Fisher Exact comparison, p = 0.011, and arm-specific rates with 95% confidence intervals. A recalculated two-arm contrast is not available for this count outcome in the posted summary, so there is no independent unadjusted contrast to set alongside the sponsor’s test.

For the continuous secondary outcomes, the registry reports mixed-model estimates and the recalculated summaries use unadjusted calculations from posted group statistics. The recalculated figures are not a substitute for the model-based analyses. In the first key secondary outcome’s primary analysis, the registry’s estimate and p-value are close to the recalculation, while the confidence intervals differ slightly. In the supplemental analysis, the reported and recalculated mean differences agree at the displayed precision, with modest differences in interval limits and p-values. For PlexiQoL, the reported and recalculated estimates are also close, and both intervals include zero.

These comparisons are descriptive: the calculations use different methods and different amounts of information. The registry record does not provide enough detail to reproduce the mixed models. There is no recorded discrepancy flag, but where the sponsor’s own analysis is available, it remains the primary reported analysis. The recalculations are best read as checks on the broad numerical pattern, not as corrected or preferred results.

11. Multiplicity, Missing Data and Analysis Populations

The registry record lists a primary endpoint and several secondary outcomes, including two analyses of the pain endpoint and a quality-of-life endpoint. It does not report a multiplicity-adjustment procedure, a gatekeeping strategy, or how the secondary analyses were ordered for confirmatory inference. Without that information, the interpretation of multiple p-values requires caution. A nominal p-value does not by itself show that a finding remains persuasive after accounting for the number and organization of analyses.

The primary comparison is distinguishable from the secondary analyses because it is identified as the primary endpoint in the registry. However, the record does not report whether the primary analysis was part of a broader testing hierarchy or whether any adjustment was planned for the full family of outcomes. The absence of those details means that the posted p-values should not be treated as evidence of a particular multiplicity plan.

The primary population is defined as all randomized participants. The registry record does not report how participants without an evaluable scan were classified, how missing MRI assessments were handled, or whether sensitivity analyses examined alternative assumptions about missing outcomes. It also does not provide a detailed participant flow for discontinuations or incomplete follow-up. These omissions limit assessment of whether the observed response rates could be influenced by missing data.

For the secondary outcomes, the record lists the analysis populations in the posted group summaries, but does not explain why those populations differ from the randomized totals or how missing measurements were addressed. Nor does the registry record report the full mixed-model specification, including its assumptions about incomplete repeated measurements. Such details matter because a model-based analysis can depend on assumptions about the data that are not verifiable from group summaries alone.

These are reporting limitations, not proof that the analyses were conducted improperly. The available registry record supports describing the analyses and their estimates, but does not support a definitive account of multiplicity control, missing-data methods, or sensitivity to alternative analysis choices.

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: neurofibrosarcoma (4.1%), cellulitis (2.8%), pneumonia (1.4%), joint dislocation (1.4%), neurofibrosarcoma recurrent (1.4%).

The posted safety counts show serious adverse events affecting 10 of 71 participants in the selumetinib group and 9 of 74 in the placebo group. Other adverse events affected 70 of 71 and 54 of 74, respectively. The registry reports zero deaths in each group. These are participant counts for the recorded categories; they do not, by themselves, establish that an event was caused by the assigned intervention or describe event severity beyond the serious-event classification.

The listed serious-event terms and percentages are neurofibrosarcoma, 4.1%; cellulitis, 2.8%; pneumonia, 1.4%; joint dislocation, 1.4%; neurofibrosarcoma recurrent, 1.4%; and headache, 1.4%. The record does not report event-level timing, duration, recurrence, attribution, or a statistical comparison of safety outcomes. The terms should be read as registry listings rather than as a complete assessment of comparative risk.

Safety counts also depend on the observation period and reporting definitions. The registry summary does not provide enough information to compare exposure time or to determine how every event was classified. The figures describe what was posted and do not establish the full safety profile.

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

The primary result describes a randomized comparison of confirmed response rates in adults with NF1 and symptomatic, inoperable PNs. The posted response-rate estimates are higher in the selumetinib group than in the placebo group, and the registry reports a Fisher Exact p-value of 0.011. The arm-specific confidence intervals show that each rate remains uncertain; the registry does not provide a confidence interval for the difference between rates.

This result does not mean that every participant responded, that an individual response is certain, or that the outcome applies to people outside the trial population. It does not establish an effect on pain, quality of life, or every other outcome. Those are separate endpoints and need their own interpretation. Likewise, a statistical test does not determine whether an observed difference is important to participants, and the registry record does not provide a threshold for judging the clinical importance of the response-rate contrast.

The result is also not a complete evaluation of harms and benefits. The safety summary reports adverse-event counts, but does not provide all details needed to assess timing, attribution, or exposure. The registry record does not report enough information to assess missing outcome data, sensitivity analyses, or the full statistical plan. These limits affect how precisely and broadly the results can be interpreted.

Randomization and masking are design features that support a fair comparison, but they do not remove every source of uncertainty. The findings should be understood as the recorded results for this trial, under its stated endpoint definitions and analysis population. They do not, on their own, settle how the outcome compares across other populations, settings, or measures that are not described in the registry record.

14. Why This Trial Matters Statistically

This trial is statistically informative because it compares a defined response endpoint between randomized groups, rather than describing outcomes in a single group alone. The placebo-controlled parallel design gives a direct framework for assessing whether response counts differ by assigned group. The primary endpoint also uses a specific confirmation rule and volumetric MRI assessment, making clear what counts as a response for this analysis.

The results illustrate why an estimate and its uncertainty should be considered together. Each arm has a response-rate estimate and a 95% confidence interval, while the between-group test supplies a separate measure of evidence against equal response rates. Because no confidence interval for the primary contrast is posted, the strength of evidence and the precision of the contrast are not conveyed by one statistic alone. Keeping those questions distinct is central to interpreting clinical trial results.

The secondary outcomes also demonstrate that different endpoints can produce different statistical patterns. Pain intensity, quality of life, and MRI-defined response are not interchangeable outcomes. Their differing populations, measures, and analyses make multiplicity and missing-data reporting important parts of interpretation. Recalculated summaries can help readers understand the posted group statistics, but they cannot recover details of the sponsor’s fitted models.

Finally, the registry record highlights the distinction between a trial’s observed estimates and its original design rationale. Sample-size illustrations based on observed effects can be useful for planning discussions, but they do not reveal the study’s own justification or establish observed power. The registry record does not report the original sample-size assumptions or full statistical plan, so transparent reporting of those details would be necessary for a fuller statistical appraisal.

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 comparison is reported with arm-specific response rates and confidence intervals, but no between-group effect estimate or confidence interval. This limits direct assessment of the contrast’s magnitude and precision. The Fisher Exact p-value provides evidence about the count comparison, but does not replace an effect estimate with uncertainty.

The registry record does not report how missing or unevaluable MRI assessments were handled, whether sensitivity analyses were conducted, or how participants who discontinued were treated in the endpoint analysis. Although the primary population is defined as all randomized participants, that definition alone does not explain how unavailable outcome data were incorporated. This is especially relevant for an endpoint requiring evaluable and confirmatory imaging.

Details of the mixed models used for secondary outcomes are not reported, including model covariates and assumptions. The recalculated unadjusted comparisons rely on posted summary statistics and approximate standard deviations, so they cannot reproduce participant-level analyses or all model adjustments. They should not be treated as replacements for the reported analyses.

The record also does not report a multiplicity strategy for the several secondary analyses, or the trial’s sample-size justification. The planning illustrations use observed effects and stated assumptions, not the original trial design. Finally, the safety summary does not provide detailed event timing, attribution, exposure-adjusted comparisons, or a complete account of event management. These reporting limits constrain the conclusions that can be drawn from the registry record; they do not establish that unreported analyses or procedures were absent.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Efficacy and Safety of Selumetinib in Adults With NF1 Who Have Symptomatic, Inoperable Plexiform Neurofibromas — NCT04924608.
• Registry record last updated 2026-03-23; status last verified 2026-03.
• Results first posted 2025-10-20.
• 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

KOMET is a Phase III randomized, parallel-group, placebo-controlled study in adults with NF1 and symptomatic, inoperable PNs. It lists quadruple masking, planned enrollment of 145, and 71 participants in the selumetinib 25 mg/m2 group and 74 in the placebo group. The registry status is active, not recruiting.

The primary endpoint is confirmed objective response by the end of Cycle 16, determined by ICR using volumetric MRI and REiNS criteria. The registry reports response rates of 19.7% and 5.4%, respectively, with arm-specific 95% confidence intervals and a Fisher Exact p-value of 0.011. The posted secondary outcomes cover pain intensity and PlexiQoL at Cycle 12.

The registry record does not report the full statistical plan, multiplicity handling, missing-data methods, or the trial’s sample-size justification. Those omissions limit reconstruction and interpretation of the analyses beyond the posted results.