Testing Crizotinib as a Potential Targeted Treatment in Cancers With ALK Genetic Changes (MATCH-Subprotocol F): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Testing Crizotinib as a Potential Targeted Treatment in Cancers With ALK Genetic Changes (MATCH-Subprotocol F) (NCT04439266), a na phase 2 study. The primary outcome is Objective Response Rate (ORR).

2. Clinical Question

3. Trial Design

This is a phase II, single-group treatment study sponsored by the National Cancer Institute. The registry lists enrollment of 5 and describes a study of crizotinib in patients with ALK-rearranged tumors other than adenocarcinoma of lung or ALCL. The listed conditions span advanced and refractory cancers, including lymphoma, malignant solid neoplasms, hematopoietic and lymphoid cell neoplasms, and refractory multiple myeloma. These categories describe the registry’s condition listings; they do not establish how many participants had each tumor type.

The intervention was crizotinib, taken by mouth twice daily on days 1-28. Treatment cycles repeated every 28 days while there was no disease progression or unacceptable toxicity. The study was unmasked, and there was no randomized allocation or parallel control group. Thus, the design can describe outcomes among treated participants but cannot make a direct, randomized comparison with a different treatment.

The registry lists a start date of 2015-08-12, primary completion on 2020-03-20, and status as active, not recruiting. These registry fields describe the study record and do not supply a final efficacy analysis. The primary endpoint was objective response rate among eligible patients who received protocol treatment. Secondary outcomes were the 6-month progression-free survival rate and progression-free survival.

The enrollment count is not necessarily the primary endpoint denominator: the endpoint population is defined by eligibility and treatment, and the record does not report that denominator separately. A sample-size justification, a detailed statistical analysis plan, interim-analysis details, and a comparison-based efficacy analysis 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: To be assigned to a specific MATCH subprotocol, patients needed to submit a tumor biopsy for molecular characterization and those with molecular variants addressed by treatments included in the trial entered corresponding MATCH subprotocol. For the subprotocol F, patients had to have an ALK mutation.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Patients who were eligible and received protocol treatment

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: Eligible and treated patients

6. Primary Endpoint

The registry lists the primary outcome as Objective Response Rate (ORR), measured over Tumor assessments occurred at baseline, then every 2 cycles for the first 26 cycles and every 3 cycles thereafter until disease progression, up to 3 years post registration.

Objective response rate measures the proportion of patients whose best overall response is complete response or partial response. It is a binary endpoint at the patient level: each eligible, treated patient contributes according to whether the response definition was met. A response rate is used to describe how often tumor shrinkage meeting prespecified criteria occurred in the study population. It does not by itself describe how long a response lasted or how long patients lived without progression.

The registry states that tumor assessments began at baseline, then occurred every 2 cycles for the first 26 cycles and every 3 cycles thereafter until disease progression, up to 3 years post registration. Response was assessed using Response Evaluation Criteria in Solid Tumors version 1.1, the Cheson (2014) criteria for lymphoma patients, and Response Assessment in Neuro-Oncology criteria for glioblastoma patients. The record refers readers to the protocol for detailed response definitions.

A 90% two-sided binomial exact confidence interval was specified for the response rate. Such an interval describes uncertainty in the rate estimate under the binomial model; it is not a guarantee that an individual patient will respond. The registry record does not report the observed rate or interval, so neither the frequency of response nor the estimate’s precision can be evaluated from this record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is a proportion. In general, a proportion is formed by counting patients who meet a defined event criterion and relating that count to the relevant analysis population. Here, the event criterion is best overall response of complete response or partial response, and the endpoint population is patients who were eligible and received protocol treatment. The registry record does not report the numerator or denominator, so the response proportion cannot be reconstructed.

What does the confidence interval describe?

The endpoint description specifies a 90% two-sided binomial exact confidence interval. In general, a confidence interval conveys uncertainty around an estimate under the method’s assumptions. A binomial exact procedure is designed for a count of events among a defined number of participants and does not rely on a large-sample normal approximation in the same way as a simple normal-based interval. This can be useful when the sample is small or the observed proportion is near an extreme. The interval’s actual limits are not reported, so its width and the range of values compatible with the observed data cannot be assessed.

A confidence interval is not an individual prediction, and it does not remove uncertainty about study design or generalizability. It describes statistical uncertainty for the defined endpoint and population, conditional on the data and method. It cannot correct for selection into the study, differences among tumor types, missing outcomes, or the absence of a control group. Those issues require design and reporting information beyond an interval calculation.

How should the response criteria be understood?

The record identifies different response criteria for different disease contexts: Response Evaluation Criteria in Solid Tumors version 1.1, the Cheson (2014) criteria for lymphoma patients, and Response Assessment in Neuro-Oncology criteria for glioblastoma patients. Using disease-specific criteria recognizes that response may be assessed differently across tumor types. The registry refers to the protocol for detailed definitions, which are not reproduced in the record. It does not report how many participants were evaluated under each set of criteria or whether any assessments were unavailable.

Assessment timing also matters for a best-overall-response endpoint. The record says assessments occurred at baseline, then every 2 cycles for the first 26 cycles and every 3 cycles thereafter until disease progression, up to 3 years post registration. In general, the schedule defines when opportunities to observe response occur. The registry does not report participant-level assessment histories, the timing of responses, or whether follow-up differed among participants.

What does a single-group design permit?

In a single-group study, the response rate describes outcomes in the treated study population. There is no concurrent randomized comparison to estimate what would have occurred to similar participants under another treatment. General statistical explanations of causal effects rely on a comparison that helps distinguish treatment effects from other influences; this study’s listed design does not provide that randomized contrast. The result, if reported, would therefore need to be interpreted as a description of this group rather than a direct estimate of comparative treatment effect.

The study is unmasked, meaning masking is listed as none. In general, knowledge of treatment assignment can influence some parts of clinical assessment or reporting, although the registry information alone does not establish that such influence occurred. The use of named response criteria provides a defined assessment framework, but the record does not report central review, assessor masking, or additional safeguards. Those details are not reported in the registry record.

Why does over-dispersion matter?

For a binary outcome, the binomial model represents variation in event counts under a common event probability for the population being analyzed. In general, over-dispersion means that observed variation is greater than that model expects. It can arise when participants differ substantially in their underlying event probabilities or when the data structure is more heterogeneous than the model assumes. A simple binomial interval may then describe uncertainty too narrowly if that extra variation is ignored. The registry record does not report an over-dispersion assessment or a model allowing for it. Given the range of listed cancers, differences in disease context are relevant to interpretation, but the record does not permit quantification of their influence.

What can planning calculations establish?

Planning calculations are hypothetical illustrations, not observed outcomes and not a substitute for the study’s design rationale. The assumptions described for the planning illustration are a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation, with the effect to detect expressed as a stated fraction of the observed effect. Time-to-event illustrations are described as using a total number of events; other illustrations are described in participants per arm. These assumptions do not match the registered single-group design’s lack of a comparator. They also do not establish the study’s original sample-size justification. The actual planning values and a protocol-based justification are not reported in the registry record. Such calculations should not be interpreted as observed power.

Finally, the registry lists no analyses. That means the record does not provide inferential results that could be examined for model choice, uncertainty, or adjustment. It does not license reconstruction of missing estimates from enrollment or adverse-event summaries. Statistical interpretation must remain limited to the endpoint definitions, design information, and safety counts actually recorded.

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("Treatment (Crizotinib)")
n      <- c(4)
events <- c(2)
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  Treatment (Crizotinib)  4  2  50.0  6.8  93.2  15.0  85.0  14.7  85.3

11. Multiplicity, Missing Data and Analysis Populations

The registry lists objective response rate as the primary outcome and progression-free survival and the 6-month progression-free survival rate as secondary outcomes. Considering multiple outcomes can raise a multiplicity question: in general, evaluating several hypotheses can increase the chance of at least one apparently noteworthy finding if each is judged separately. Whether and how to address that issue depends on the prespecified analysis plan and the intended claims. A multiplicity strategy, such as a hierarchy or adjustment, is not reported in the registry record. No efficacy analyses are reported, so there are no findings from these outcomes to assess jointly.

The record specifies the population for the primary endpoint as eligible patients who received protocol treatment. It does not describe analysis populations for the secondary outcomes, nor does it report an intention-to-treat analysis population. Enrollment is 5, but the record does not say whether every enrolled participant was eligible, treated, and evaluable for each endpoint. Those populations should not be assumed to be identical.

Missing-data handling is also not reported in the registry record. The outcome description gives an assessment schedule, but it does not explain how missing scans, incomplete follow-up, withdrawal, or unevaluable response assessments were handled. For progression-free survival outcomes, the record does not specify censoring rules or how follow-up was summarized. These omissions matter because the analysis population and handling of unavailable observations can affect an estimate’s interpretation.

The registry record does not report interim analyses, stopping rules, or a detailed statistical analysis plan. No assumptions about these procedures can be made from the listed design and outcome descriptions alone.

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: aspartate aminotransferase increased (40.0%), alanine aminotransferase increased (20.0%), blood bilirubin increased (20.0%), hyponatremia (20.0%), hypophosphatemia (20.0%).

The registry lists a treatment group with 5 participants at risk for serious adverse events and 2 affected. For other adverse events, it lists 5 at risk and 5 affected. These are separate summaries and should not be combined. The record also lists 4 deaths. It does not provide enough detail in these figures to determine the timing, circumstances, or cause of each death, or whether a death was considered related to crizotinib.

The serious adverse-event terms with the largest listed percentages are aspartate aminotransferase increased, 40.0%; alanine aminotransferase increased, 20.0%; blood bilirubin increased, 20.0%; hyponatremia, 20.0%; and hypophosphatemia, 20.0%. These percentages describe the registry’s event summaries; they do not establish causation. The record does not specify the event-specific denominators or provide additional context about severity, duration, recurrence, or attribution in this summary.

With only the reported counts and terms, the safety findings are descriptive. They do not support a comparison with another treatment because there is no control group, and they do not establish that every event was caused by treatment. The absence of further detail should not be read as evidence that other events did not occur.

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

The registry record describes a phase II study of crizotinib in a single group of patients whose tumors had ALK rearrangements, with lung adenocarcinoma and ALCL excluded in the official title. It defines objective response rate as the primary endpoint and specifies how response was assessed. It also lists secondary progression-free survival outcomes and adverse-event summaries.

However, the primary response result is not reported in the registry record. There is no reported response count, response-rate estimate, or confidence interval. The absence of those values means that the posted record cannot establish how often complete or partial responses occurred or how precisely a response rate was estimated. It is not evidence that the response rate was zero, and it is not evidence that a response occurred.

The registry’s single-group design also limits what any response estimate could establish. A rate among treated participants would describe that group under the study conditions, but without a randomized comparison it would not show how outcomes differed from another treatment or no study treatment. That limitation is separate from the missing efficacy result: even a reported response proportion would not provide a direct comparative treatment effect.

The adverse-event summaries document counts and serious event terms, and the record lists deaths. These summaries warrant careful description but do not, by themselves, establish cause or treatment attribution. No clinical recommendation follows from these statistical summaries. The most defensible conclusion is that the registry defines the study question and some safety information, while leaving the primary efficacy result unavailable.

14. Why This Trial Matters Statistically

This study illustrates why a clearly named endpoint is not the same as an interpretable result. The registry gives a specific primary outcome—objective response rate—and defines the eligible, treated population and response criteria. Yet without the event count, denominator, estimate, and confidence interval, readers cannot evaluate either the observed response frequency or its statistical precision. Reporting the endpoint definition without its result leaves the main efficacy question unresolved.

The single-group design is also important statistically. It can describe outcomes in the treated participants, but it does not include a randomized comparator that would help distinguish treatment-associated outcomes from other explanations. This distinction remains important even if a response estimate is eventually reported: a descriptive response rate and a comparative treatment effect answer different questions.

The registry’s condition list spans several cancer categories, while response criteria vary according to disease context. This makes the analysis population and its composition relevant to interpretation. The record does not report the number of participants in each tumor category, so it is not possible to assess how much the overall result would reflect any one disease group. This is a question of applicability as well as statistical precision.

Finally, the safety summaries and the efficacy endpoint answer different questions. Counts of serious adverse events, other adverse events, and deaths do not supply a missing response estimate, and event listings alone do not establish causation. Clear reporting of analysis populations, endpoint results, uncertainty, missing-data handling, and event attribution is necessary for readers to judge what the record can support.

15. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

The primary limitation is the absence of reported efficacy results. The registry record does not give an objective response count, response-rate estimate, or 90% confidence interval. It lists 5 enrolled participants, but the primary endpoint population is defined by eligibility and receipt of protocol treatment, and the record does not report that population’s size separately. The enrollment count therefore cannot stand in for the endpoint denominator.

The study has no comparator and is unmasked. The single-group design prevents a direct randomized comparison, while the lack of masking may be relevant to outcome assessment. The record names response criteria but does not report whether assessments were centrally reviewed or whether assessors were masked. It also does not report how many patients had each tumor type, despite the range of conditions and the use of different response criteria across disease contexts.

Secondary outcome results are not reported. The registry does not specify the analysis populations or missing-data procedures for progression-free survival outcomes, and it does not provide censoring details. A detailed statistical analysis plan, sample-size justification, multiplicity approach, and interim-analysis information are not reported in the registry record. These gaps limit assessment of prespecification and the handling of incomplete observations.

The safety information is also limited to event counts, selected serious adverse-event terms, percentages, and deaths. The record does not provide enough detail to establish event timing or treatment attribution. These limitations are specific to the information recorded and should not be treated as proof that the unreported procedures or results had a particular value.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Testing Crizotinib as a Potential Targeted Treatment in Cancers With ALK Genetic Changes (MATCH-Subprotocol F) — NCT04439266.
• Registry record last updated 2026-05-07; status last verified 2026-05.
• Results first posted 2022-09-22.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

19. Record Summary

The registry record describes NCT04439266, a phase II, single-group study of crizotinib in patients with ALK-rearranged tumors other than adenocarcinoma of lung or ALCL. Crizotinib was taken by mouth twice daily on days 1-28, with 28-day cycles continuing in the absence of progression or unacceptable toxicity. Enrollment is listed as 5; the study is unmasked and has no comparator.

The primary endpoint is objective response rate among eligible patients who received protocol treatment, with response assessed using disease-specific criteria and a 90% two-sided binomial exact confidence interval specified. The response result and interval are not reported in the registry record. The record lists progression-free survival outcomes, safety counts, selected serious adverse-event terms, and 4 deaths, but does not provide a comparative efficacy analysis. The listed status is active, not recruiting.