Crizotinib in Lobular Breast, Diffuse Gastric and Triple Negative Lobular Breast Cancer or CDH1-mutated Solid Tumours: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Crizotinib in Lobular Breast, Diffuse Gastric and Triple Negative Lobular Breast Cancer or CDH1-mutated Solid Tumours (NCT03620643), a non-randomized phase 2 study. The primary outcome is Percentage of Breast Cancer Cohort Participants With Confirmed Response Assessed Using RECIST v1.1.

2. Clinical Question

3. Trial Design

This was a phase II, parallel, non-randomized study with single masking. The sponsor was Royal Marsden NHS Foundation Trust. The registry lists the study as completed, with a start date of 2019-05-09 and primary completion date of 2024-08-30. Total enrollment is recorded as 33.

The record describes two treatment cohorts: a basket cohort treated with crizotinib monotherapy and a lobular breast cancer cohort treated with crizotinib plus fulvestrant. The arm descriptions each state n=29 participants. Those arm-description figures do not reconcile with the overall enrollment figure of 33, and the registry record does not explain the discrepancy. They should not be combined or interpreted as a verified allocation count.

The treatment description specifies crizotinib 250 mg twice daily in the monotherapy cohort, on a continuous schedule, with a cycle defined as 28 days. The combination description lists crizotinib 250mg twice daily plus fulvestrant 500 mg per cycle, administered intramuscularly; the posted regimen text is truncated after “q28,” so its full schedule is not reported in the registry record.

The primary endpoint concerns confirmed response in the breast cancer cohort. Secondary outcomes listed are adverse events and serious adverse events, progression-free survival assessed using RECIST v1.1 in both cohorts, and overall survival in each cohort. Although the study is described as single-masked, who was masked and how masking was implemented are not reported in the registry record. Random allocation was not used, so the treatment cohorts do not constitute randomized comparison groups.

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

Analysis population for the primary outcome, as defined in the registry: Patients who receive at least one cycle of the trial medication and had at least one response (RECIST 1.1 scan) post treatment start. Patients who stopped study treatment before the first response evaluation scan has been performed with clinical progression were included in the disease progression category.

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 Percentage of Breast Cancer Cohort Participants With Confirmed Response Assessed Using RECIST v1.1, measured over From Day 1 to Progressive Disease, assessed up to end of study (Approximately 57 months).

The primary endpoint is the percentage of participants in the breast cancer cohort with confirmed response assessed using RECIST v1.1. In general, a response endpoint classifies tumour change according to predefined imaging criteria. Requiring confirmation means that the response assessment is not based solely on a single unconfirmed finding, although the specific confirmation procedures are not reported in the registry record.

The registry describes response assessment from Day 1 until progressive disease, assessed up to the end of the study, with the timeframe stated as approximately 57 months. The endpoint is intended to characterize tumour response in the advanced E-cadherin-negative, ER-positive lobular breast cancer cohort receiving crizotinib and fulvestrant. The record’s population description specifies participants receiving at least one cycle and having at least one post-treatment response scan; participants who stopped before their first response scan because of clinical progression were included in the disease-progression category. The remainder of that population description is incomplete in the registry record.

In general, response rate is useful because it summarizes the proportion meeting a defined tumour-response criterion. It does not by itself describe how long a response lasts, time to progression, overall survival, or comparative benefit against another treatment. The posted primary outcome contains no response count, percentage, confidence interval, or test result. The registry record therefore does not report the observed primary endpoint result or its precision.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The methods can be understood by separating the endpoint definition, the design, and the information available for analysis. The primary endpoint is binary: each participant is classified according to whether a confirmed RECIST v1.1 response was recorded. In general, a binary endpoint summarizes the number or proportion of participants meeting a specified condition. The registry identifies the measure as a percentage of participants but does not post the response numerator, denominator, or analysis result.

What does confirmation add?

In general, confirmation aims to distinguish a response that persists on assessment from one that is not confirmed. The exact scan schedule, confirmation interval, imaging review process, and handling of uncertain assessments are not reported in the registry record. The endpoint description says response is assessed from Day 1 until progressive disease, up to the end of the study, with a timeframe of approximately 57 months. That timeframe does not establish how frequently participants were assessed.

Why does the cohort structure matter?

The study is parallel but non-randomized. In a randomized comparison, assignment can help make treatment groups comparable on average; here, no random allocation was used. The registry describes a basket cohort receiving crizotinib alone and a lobular breast cancer cohort receiving crizotinib with fulvestrant. Because the cohorts differ in their stated clinical populations, their outcomes cannot be interpreted as if treatment were the only difference. A simple difference between their results, even if available, would not establish a treatment effect.

The record reports single masking, but does not identify who was masked or what procedures maintained masking. It also does not report whether response assessments were independently reviewed or blinded to treatment. These details matter in general because knowledge of treatment or clinical status may influence some assessments or decisions. Their absence from the registry record prevents a more specific appraisal.

How should the safety counts be read?

The adverse-event record gives at-risk and affected counts for serious events and other adverse events. These are descriptive counts, not comparative estimates adjusted for differences between cohorts. The serious-event count is 4 among 6 at risk in the monotherapy group and 5 among 27 at risk in the combination group. For other adverse events, the counts are 6 among 6 and 27 among 27, respectively. The record also lists deaths counts of 6 and 20. It does not state event attribution, timing, exposure-adjusted rates, or whether the deaths were considered treatment-related. The selected serious-event term percentages are listed without an explicit denominator, so they cannot be translated into counts from the posted information.

In general, an at-risk denominator defines which participants could contribute to a reported event measure. A difference between the at-risk denominator and the enrollment or arm-description count may reflect distinct analysis populations or reporting conventions, but the reason cannot be assumed. Here, enrollment is 33, each arm description states n=29, and safety at-risk counts are 6 and 27. The registry record does not explain how those figures relate.

What can planning illustrations tell us?

The record describes sample-size planning illustrations using a two-sided significance level of 0.05, equal allocation, a normal approximation, and effects defined as fractions of an observed effect. Such calculations generally estimate the number of participants per arm needed to detect a specified effect under stated assumptions. They are planning quantities, not observed power and not evidence about the results of the completed study. No numerical planning values or underlying observed effect are reported in the registry record. The record also does not report a trial-specific sample-size justification or the statistical plan used for the primary analysis.

For a binary response endpoint, analysis choices can include how to define the response denominator, how to handle participants without an evaluable scan, and whether to report an interval around the response proportion. The registry’s population description gives some inclusion details, including treatment exposure and a post-treatment scan, and says that some early progressions were included in the disease-progression category. However, the full description and detailed missing-data rules are not reported. Without the response count and denominator, neither an estimate nor its precision can be reconstructed.

In general, confidence intervals express the uncertainty associated with an estimated effect or proportion. A narrow interval indicates greater precision than a wide interval, while neither interval width nor a point estimate alone establishes causality. No confidence interval is reported for the primary endpoint. No p-value or model-based analysis result is reported either. Since the record lists zero analyses, interpretation should remain descriptive and should not imply an unreported statistical test.

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("Crizotinib Oral Capsule [Xalkori] Plus Fulvestrant")
n      <- c(27)
events <- c(1)
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  Crizotinib Oral Capsule [Xalkori] Plus Fulvestrant  27  1  3.7  0.1  19.0  0.7  18.3  0.9  18.3

11. Multiplicity, Missing Data and Analysis Populations

The registry lists five outcome measures and identifies the primary response endpoint, alongside secondary outcomes covering adverse events, progression-free survival, and overall survival. A study evaluating multiple outcomes can face multiplicity: the more hypotheses assessed, the greater the chance of a nominally notable finding arising by chance if each is tested separately. In general, a prespecified testing strategy or adjustment can address that issue. A multiplicity strategy, hierarchy, or adjustment is not reported in the registry record, so it is not possible to determine how multiple outcomes were handled.

The record lists no analyses and does not report a primary endpoint result. It also does not describe an analysis population beyond the partial primary endpoint population description. That description refers to participants who received at least one cycle and had at least one post-treatment response scan, while participants stopping before the first scan with clinical progression were included as disease progression. The full rule and the treatment of other missing or unevaluable assessments are not reported in the registry record.

In general, missing outcome data can affect an estimate when the probability of being observed is related to response or clinical status. Appropriate interpretation depends on knowing how many observations were missing, why they were missing, and how the analysis handled them. Those details are not reported for the primary outcome. The registry also does not report whether all outcomes were analyzed in an intention-to-treat population, a treated population, or another prespecified group. No interim analysis plan, stopping rule, or detailed statistical analysis plan is reported in the registry record.

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: fracture (16.7%), gastric ulcer (16.7%), abdominal pain (16.7%), pleural effusion (16.7%), ascites (7.4%).

The registry reports serious adverse events among 4 of 6 at-risk participants in the crizotinib monotherapy group and 5 of 27 at-risk participants in the crizotinib-plus-fulvestrant group. For other adverse events, it reports 6 of 6 and 27 of 27, respectively. These are descriptive participant counts. The adverse-event denominators do not match the arm-description figures, and the registry record does not explain the difference.

The registry lists serious-event terms with percentages: Fracture, Gastric Ulcer, Abdominal Pain, and Pleural Effusion at 16.7% each; Ascites and Vomiting at 7.4% each. The denominator and whether these percentages refer to participants or events are not reported in the registry record. The percentages should not be converted into counts or treated as directly comparable rates between cohorts.

Deaths counts are listed as 6 for monotherapy and 20 for combination treatment. The record does not report timing, cause, relationship to treatment, or a defined analysis period for these counts. Cohort differences and the non-randomized design further limit direct comparisons. These figures describe what the registry records; they do not establish that treatment caused or prevented any event.

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

The study addresses a clinically specific response question: confirmed RECIST v1.1 response in an advanced lobular breast cancer cohort receiving crizotinib with fulvestrant. The record also describes crizotinib monotherapy in a basket cohort and lists safety, progression-free survival, and overall survival as additional outcomes. These features define what the study set out to examine, but the posted record does not include the primary response result or analyses.

Accordingly, no conclusion about the frequency or precision of confirmed response can be drawn from the primary outcome information. Nor can the posted record establish a causal comparison between the combination and monotherapy cohorts. The study was non-randomized, and the cohorts differ in their stated populations; the registry explicitly cautions that they are not a like-for-like control comparison.

The safety counts and serious-event terms provide descriptive information, but they do not identify treatment causality. Their denominators are not aligned with the arm-description counts, and the record does not provide context on event timing, exposure, attribution, or the denominator for the listed serious-event percentages. The deaths counts likewise lack the context needed for a treatment-related interpretation.

In general, an endpoint result should be interpreted together with its denominator, uncertainty, analysis population, missing-data handling, and study design. Those elements are incomplete or absent for the primary endpoint. The registry record also does not report a sample-size justification or a detailed statistical plan. The most defensible reading is therefore that the record describes a completed phase II study and its intended questions, while not reporting enough efficacy analysis to determine the observed primary response outcome.

14. Why This Trial Matters Statistically

This study is statistically notable for the gap between its clearly described clinical question and the limited outcome analysis posted in the registry. The primary endpoint is a confirmed response measure with a named assessment framework, RECIST v1.1, and a defined breast cancer cohort. Yet the response count, percentage, uncertainty interval, and analysis result are not reported. A well-specified endpoint is necessary for interpretation, but it cannot substitute for an observed estimate.

The non-randomized structure is also central to interpreting any cross-cohort comparison. The monotherapy arm is described as a basket cohort, while the combination cohort is a lobular breast cancer cohort. Because treatment assignment was not randomized and the clinical groups are not like-for-like, treatment and cohort characteristics cannot be separated using a simple arm contrast. The record itself flags this limitation.

Discrepancies among enrollment, arm descriptions, and safety at-risk counts illustrate why denominator provenance matters. Enrollment is recorded as 33; both arm descriptions state n=29; and the safety groups have at-risk counts of 6 and 27. These values may represent different populations or reporting contexts, but the registry does not explain their relationship. Statistical interpretation depends on understanding which participants contribute to each estimate.

Finally, the record lists multiple secondary outcomes and planning assumptions but does not provide analyses, a multiplicity strategy, or a detailed analysis plan. This makes it important to distinguish intended measurements from reported results, and planning calculations from observed evidence. The available information supports careful description of design and safety reporting, but not an efficacy estimate or comparative treatment conclusion.

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.

Several limitations specific to this trial record affect interpretation. First, the primary endpoint is named, but its result is absent: no response count, percentage, confidence interval, or analysis is reported. The registry lists no analyses. Consequently, the central efficacy question cannot be answered from the posted outcome information.

Second, cohort allocation was non-randomized. The monotherapy group is described as a basket cohort and the combination group as a lobular breast cancer cohort. These are not like-for-like comparison groups, so any direct contrast would be vulnerable to differences in disease population and other factors. The record does not provide adjusted comparative analyses.

Third, the participant counts do not reconcile in the registry record. Overall enrollment is 33, each arm description states n=29, and safety at-risk counts are 6 and 27. The record does not explain the relationship among these figures or identify which denominator applies to each outcome. The primary population description is also truncated, and rules for other missing or unevaluable response assessments are not reported.

Fourth, the safety information lacks key context. The serious-event term percentages have no reported denominator, and the record does not specify event attribution, timing, exposure duration, or the meaning of the deaths counts. The single-masking description does not identify who was masked. A trial-specific sample-size justification, detailed statistical analysis plan, multiplicity approach, and interim-analysis information are not reported in the registry record.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Crizotinib in Lobular Breast, Diffuse Gastric and Triple Negative Lobular Breast Cancer or CDH1-mutated Solid Tumours — NCT03620643.
• Registry record last updated 2025-10-30; status last verified 2025-10.
• Results first posted 2025-10-30.
• 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 describes a completed, non-randomized phase II study sponsored by Royal Marsden NHS Foundation Trust. It began on 2019-05-09 and has a primary completion date of 2024-08-30. The study examined crizotinib monotherapy in a basket cohort and crizotinib plus fulvestrant in a lobular breast cancer cohort.

The primary endpoint was confirmed response assessed using RECIST v1.1 in the breast cancer cohort. Listed secondary outcomes include adverse events, serious adverse events, progression-free survival, and overall survival. The record reports selected safety counts and serious-event terms but does not report the primary response result or analyses. It also does not explain differences among enrollment, arm-description, and safety denominators. Because treatment allocation was non-randomized and the cohorts are not like-for-like, the arms should not be interpreted as a direct treatment comparison.