Targeted Treatment for ALK Positive Patients Who Have Previously Been Treated for Non-squamous Non-small Cell Lung Cancer: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Targeted Treatment for ALK Positive Patients Who Have Previously Been Treated for Non-squamous Non-small Cell Lung Cancer (NCT03737994), a randomized phase 2 study. The primary outcome is Objective Response Rate (ORR), Per Investigator Assessment Using the Response Evaluation Criteria in Solid Tumors (RECIST) v1.1 Criteria.

2. Clinical Question

3. Trial Design

This was a phase II, randomized, parallel-group trial with no masking, sponsored by the National Cancer Institute. Its stated focus was biomarker-directed treatment for previously treated patients with ALK-positive non-squamous non-small cell lung cancer. Enrollment was 10. The registry lists the start date as 2019-07-25, primary completion as 2021-07-21, and current status as terminated.

The treatment options were assigned according to biomarker groups. For example, the descriptions list crizotinib for ALK L1198F mutation and MET amplification, and different possible ALK inhibitors for several other mutation groups. Patients with no ALK-resistance mutations could receive one of several ALK inhibitors or pemetrexed, with or without cisplatin or carboplatin. The record identifies randomized allocation but does not report how randomization was conducted, how treatment choices were assigned within biomarker groups, or whether allocation was concealed.

The arm descriptions generally specify repeating cycles every 21 days while there was no disease progression or unacceptable toxicity. The description for the no-mutation group is incomplete in the registry record. The recorded intervention list includes alectinib, brigatinib, carboplatin, ceritinib, cisplatin, crizotinib, ensartinib, and lorlatinib.

Secondary outcomes included progression-free survival, duration of overall response, overall survival, intracranial objective response rate, and adverse events. Although the design is labeled randomized, the small enrollment and biomarker-specific treatment choices mean that the trial's intended comparisons should not be read as a large, uniform head-to-head comparison of every listed drug.

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: Tissue and blood submissions were required after first step registration, before second step registration (treatment assignment). Of 16 participants screened, 10 were assigned to treatment.

Only the first 6 of 7 groups are shown.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Eligible participants who started protocol treatment and had post-baseline disease assessment. Treatment arms with no evaluable participants are not shown.

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 participants assigned to treatment (second step registration)

6. Primary Endpoint

The registry lists the primary outcome as Objective Response Rate (ORR), Per Investigator Assessment Using the Response Evaluation Criteria in Solid Tumors (RECIST) v1.1 Criteria, measured over Baseline to 24 weeks.

Objective response rate measures the proportion of evaluable participants whose best overall response is complete response or partial response. In this trial, investigators assessed response using RECIST v1.1 from baseline to 24 weeks. A complete response requires disappearance of target and non-target lesions, no new lesions, and pathological lymph nodes below 10mm. A partial response requires a 30% decrease in target lesions, with non-target lesions not progressed or not evaluated and no new lesions.

As a general statistical concept, ORR is a binary outcome at the participant level: each evaluable participant either meets the response definition or does not. It offers a direct summary of tumor response within the specified assessment period, but does not by itself describe how long a response lasts, whether disease later progresses, or overall survival. Those are distinct outcomes, and the registry lists duration of overall response, progression-free survival, and overall survival as secondary measures.

The denominator matters. The registry defines the primary outcome population as eligible participants who started protocol treatment and had a post-baseline disease assessment. Participants without an evaluable post-baseline assessment are therefore not represented in that defined denominator. The registry does not provide numerical response counts in the record available for this analysis, so no ORR estimate or precision measure can be interpreted.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary measure is a binary response outcome. A participant is classified as having an objective response if the best overall response during the specified assessment window is complete response or partial response. Best overall response is defined from treatment start until progression or new anticancer therapy, or otherwise the last follow-up. This endpoint definition determines both who counts as a responder and which assessments contribute to the classification.

What does the planned Fisher exact test assess?

As a general statistical explanation, Fisher's exact test evaluates whether the distribution of a binary outcome differs between groups, using the observed counts in each outcome category. It can be used when group counts are small because it calculates a test result from the possible arrangements of those counts rather than relying on a large-sample approximation. The registry planned this test for comparisons by mutation status within ALK inhibitor arms and for experimental-arm comparisons with pemetrexed among participants without mutations.

Those tests were not reported as performed. The outcome description explicitly says that early accrual closure resulted in response counts without statistical testing. The registry record therefore does not provide a p-value or a test-based conclusion for the primary outcome. A planned method describes an analysis that was intended; it is not itself evidence that the analysis was completed.

How should effect estimates and uncertainty be understood?

In general, an effect estimate describes the observed difference or relationship in a sample, while a confidence interval describes a range of values compatible with the data under the method's assumptions. With sparse data, estimates can be imprecise, and small changes in the observed counts may substantially change the estimated effect. No primary response estimate or confidence interval is reported in the registry record, so the magnitude and precision of response cannot be assessed from it.

The registry's sample-size assumptions describe a planning illustration based on an observed effect, with a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. These are not reported as the trial's own design assumptions. The illustration is not observed power and does not establish that the trial had a particular probability of detecting an effect. No numeric sample-size illustration is included in the record.

Why are the mutation groups important?

The treatment options depended on mutation status, so the study did not pose one identical treatment comparison for every participant. Some mutation groups had several listed drug options, while others had a narrower set. The no-mutation group also included chemotherapy options. This structure makes the definition of each comparison important: an overall count across all participants would combine distinct biomarker groups and treatment choices, and would not answer each planned comparison separately.

As a general statistical concept, small groups produce limited information about response rates and treatment differences. If few participants contribute to a particular group, one participant's outcome can have substantial influence on the observed group summary. Randomized allocation can help balance factors in a comparison, but it cannot make sparse data precise. The registry does not report randomization procedures or the number assigned to each primary-outcome comparison.

What does the safety count format convey?

The safety record presents counts of participants at risk and affected for serious adverse events and for other adverse events, and separately reports deaths by listed group. These are descriptive counts. A count of affected participants is not a measure of event severity for every participant, and the number of participants affected does not indicate how many episodes each person had. General interpretation also requires attention to the group denominator, exposure, follow-up, and event definitions. The record gives small group counts, but does not supply a comparative statistical analysis of safety.

More generally, over-dispersion matters when observed variation exceeds what a simple statistical model expects. It can make uncertainty estimates too narrow if the analysis assumes less variation than the data contain. The registry does not report a model for the primary binary outcome, and the planned Fisher exact test is a count-based comparison rather than a model that estimates over-dispersion. No additional modeling details are reported.

10. Multiplicity, Missing Data and Analysis Populations

The primary outcome description identifies multiple planned comparisons: mutation versus no mutation within each ALK inhibitor treatment arm, and each experimental arm versus pemetrexed among participants with no mutation. When several hypotheses are tested, the chance of at least one apparently unusual result can increase. As a general statistical concept, multiplicity procedures can control that overall error rate, but the registry does not report a multiplicity adjustment or state how results across the planned comparisons would be interpreted together.

In this case, the record says that no statistical testing was conducted because accrual closed early. Therefore, there are no reported primary-outcome p-values to interpret, adjusted or unadjusted. The record also does not report interim analyses, a statistical analysis plan, or a rationale for the planned sample size. These details are not reported in the registry record and should not be inferred from the listed test.

The defined primary population includes eligible participants who started treatment and had a post-baseline disease assessment. The record does not explain how participants without such an assessment were handled, whether any data were imputed, or how deviations from the planned assessment schedule were addressed. It also does not report the count of evaluable participants for each comparison. The population definition therefore clarifies the intended denominator but does not resolve the extent or handling of missing response data.

Analysis populations for the secondary outcomes are not described in detail in the available record. The presence of a listed outcome does not establish that all participants contributed data to it. With a small enrolled population and biomarker-based groups, transparent participant flow and missing-data handling would be important for assessing how representative each outcome summary is; those details are not reported in the registry record.

11. 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: abdominal pain (100.0%), adult respiratory distress syndrome (100.0%), dyspnea (100.0%), pericardial effusion (50.0%), acute kidney injury (50.0%).

The safety entries list participants at risk, participants affected, and deaths for six groups. For example, the Lorlatinib (Mutations) group lists 1 participant at risk for serious adverse events and 0 affected; the Lorlatinib (no Mutations) group lists 2 at risk and 2 affected. The Ensartinib group lists 3 at risk and 1 affected for serious adverse events, with 2 deaths. These are small descriptive counts, not comparative estimates.

The registry also lists serious-event terms with percentages: Abdominal pain, Adult respiratory distress syndrome, and Dyspnea are each listed as 100.0%; Pericardial effusion, Acute kidney injury, and Pleural effusion are each listed as 50.0%. The denominator and participant-level overlap for each term are not reported in the registry record. The percentages should not be combined or interpreted as distinct people without that information.

Serious adverse events and other adverse events are separate count categories in the record. A participant can be represented in more than one event category, so counts should not be added to infer a total number of affected participants. The registry does not report a statistical comparison of safety between treatments.

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

The registry record describes a randomized phase II trial of biomarker-directed treatment in previously treated ALK-positive non-squamous non-small cell lung cancer. It records enrollment of 10 and states that the trial was terminated. The primary endpoint definition and intended statistical test are described, but numerical response counts, an ORR estimate, and a formal statistical test result are not reported in the registry record.

That absence means the posted record cannot quantify how often participants had a complete or partial response, how different response was across planned comparisons, or how precisely any difference was estimated. It also cannot support an inference about equivalence or a lack of difference. No conclusion about treatment benefit or lack of benefit follows from the absence of a reported test.

The design included several mutation-specific treatment options and a no-mutation group with both targeted therapy and chemotherapy options. The trial therefore addressed multiple biomarker-defined treatment questions rather than a single comparison of one treatment against one comparator across a uniform population. Its small enrollment and early closure further limit the amount of information available for those questions.

The safety counts offer a record of affected participants and deaths in the displayed groups, but the counts are too sparse to provide a reliable comparative assessment. The record does not report a reason for termination, detailed missing-data handling, or the procedures used for randomization. Those omissions constrain interpretation; they do not imply that any particular process or outcome occurred.

Overall, the registry provides useful information about the intended population, treatment structure, outcome definition, and reasons the primary analysis was limited. It does not provide enough reported outcome data to estimate response or compare the listed treatments statistically. The appropriate reading is descriptive and cautious, without extending these limited data to broader conclusions.

13. Why This Trial Matters Statistically

This trial raises a central design issue in biomarker-directed research: the treatment question may differ across molecular subgroups. Here, the registry describes distinct treatment options for mutation groups and a separate set of choices for participants without ALK-resistance mutations. Analysis must therefore preserve the meaning of each subgroup comparison rather than treating all participants as if they faced the same alternatives.

The record also illustrates the distinction between randomization as a design label and the information available to evaluate the resulting comparisons. The registry classifies allocation as randomized, but enrollment was 10, the trial was terminated, and details of randomization procedures are not reported. In small studies, randomization does not guarantee that the observed group estimates will be precise, especially when participants are divided among biomarker-defined treatment groups.

Another statistical lesson is that a well-defined endpoint and a planned test do not substitute for reported results. The primary outcome uses a specified response definition and planned Fisher exact testing, but the record says that early accrual closure led to response counts without statistical testing. Because the numerical counts are not reported in the registry record, neither effect size nor uncertainty can be assessed.

Finally, registry status and reporting completeness are part of statistical interpretation. The terminated status, small enrollment, and absence of inferential results set clear boundaries on what can be concluded. The record is informative about the intended research question and endpoint, while demonstrating why subgroup structure, accrual, analysis plans, and complete outcome reporting all matter when interpreting clinical trial evidence.

14. 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.

Enrollment of 10 and early accrual closure sharply limit the information available for the biomarker-specific comparisons. The registry says that few subjects led to response counts without statistical testing, but the numerical response counts are not reported in the registry record. As a result, neither the observed response pattern nor its precision can be assessed.

The trial's treatment structure is complex: options varied by mutation group, and the no-mutation group included chemotherapy as well as ALK inhibitors. The record does not provide participant counts for each primary-outcome comparison, detailed randomization procedures, or the reason for termination. It also does not report interim analyses or a statistical analysis plan.

The primary population requires a post-baseline disease assessment, but the record does not report how many eligible participants lacked one or how missing assessments were handled. The description of the no-mutation treatment arm is incomplete. These gaps make it difficult to reconstruct participant flow, treatment exposure, and the exact analysis population.

Safety groups are small, and the record does not give sufficient detail about event-level denominators or overlap among serious-event terms. The listed counts therefore should not be treated as comparative safety estimates. Collectively, these limitations constrain what can be learned from the posted registry information; they do not establish what the unreported outcomes or procedures would have been.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Targeted Treatment for ALK Positive Patients Who Have Previously Been Treated for Non-squamous Non-small Cell Lung Cancer — NCT03737994.
• Registry record last updated 2026-04-23; status last verified 2026-03.
• Results first posted 2023-11-21.
• 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.

18. Record Summary

The record describes a randomized, unmasked, parallel-group phase II trial sponsored by the National Cancer Institute. It studied biomarker-directed treatment for previously treated patients with stage IV ALK-positive non-squamous non-small cell lung cancer. Enrollment was 10, the status is terminated, and the registry lists 2019-07-25 as the start date and 2021-07-21 as the primary completion date.

The primary outcome was investigator-assessed ORR by RECIST v1.1 from baseline to 24 weeks. The registry describes intended Fisher exact comparisons but states that early accrual closure led to response counts without statistical testing. Numerical response counts and analyses are not reported in the registry record. Safety entries give small group counts for serious and other adverse events and deaths. The reason for termination, detailed missing-data handling, and randomization procedures are not reported.