A Study of Tepotinib Plus Osimertinib in Osimertinib Relapsed MET Amplified NSCLC (INSIGHT 2): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study of Tepotinib Plus Osimertinib in Osimertinib Relapsed MET Amplified NSCLC (INSIGHT 2) (NCT03940703), a non-randomized phase 2 study. The primary outcome is Combined Therapy (Tepotinib+Osimertinib): Number of Participants Experiencing Dose-Limiting Toxicities (DLTs) According to National Cancer Institute Common Terminology Criteria for Adverse Events Version (NCI-CTCAE v 5.0).

2. Clinical Question

3. Trial Design

INSIGHT 2 is listed as a phase II, two-arm, parallel study with non-randomized allocation and no masking. The registry record lists enrollment of 140 and describes the study as completed. The listed start date is 2019-09-19, and the primary completion date is 2023-05-11. These record details describe the trial’s design and status; they do not establish how many participants were ultimately treated in each arm or included in each analysis.

The study investigated tepotinib, a MET inhibitor, in combination with osimertinib, a third-generation EGFR inhibitor, in participants with advanced or metastatic non-small cell lung cancer and acquired resistance to prior osimertinib. A separate arm received tepotinib monotherapy. Because allocation was non-randomized, the two arms should not be treated as exchangeable groups. The registry record does not report a random assignment procedure or a masking procedure, consistent with the listed design classifications.

The stated primary outcome was the count of participants experiencing dose-limiting toxicities during the early safety run-in period. The time frame was up to Day 21 of Cycle 1, with each cycle described as 21 days. Secondary outcomes included objective response under RECIST version 1.1 assessed by an independent review committee or investigator, depending on the outcome and testing method. The record also lists adverse events, laboratory abnormalities, vital signs, ECOG performance status, and ECG findings among the secondary outcomes.

The record lists 47 outcome measures and zero analyses. It does not report the rationale for the enrollment target, the planned statistical analysis methods, or a protocol-level explanation of how participants entered each arm. These missing details limit what can be concluded about the study’s intended precision and how its outcome measures were to be interpreted.

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: Safety run-in analysis set (SRIAS): all FAS/SAF participants treated during the safety run-in period who received at least 75% of the tepotinib and osimertinib planned dose and completed the DLT period (3 weeks after start of study treatment), or who experienced a DLT during the DLT period regardless of the received amount of each study treatment component.

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: Safety (SAF) analysis set included all participants who were administered any dose of any study treatment.

6. Primary Endpoint

The registry lists the primary outcome as Combined Therapy (Tepotinib+Osimertinib): Number of Participants Experiencing Dose-Limiting Toxicities (DLTs) According to National Cancer Institute Common Terminology Criteria for Adverse Events Version (NCI-CTCAE v 5.0), measured over Up to Day 21 of Cycle 1 (each Cycle is of 21 days).

The primary endpoint counts participants who experienced a dose-limiting toxicity during the safety run-in period. The specified time frame is up to Day 21 of Cycle 1. The registry record defines dose-limiting toxicities using NCI-CTCAE version 5.0 and requires that the investigator and/or sponsor judge the toxicity not attributable to the disease or disease-related processes under investigation.

The listed events include prolonged Grade 4 neutropenia; Grade 3 or higher febrile neutropenia; Grade 4 thrombocytopenia or Grade 3 thrombocytopenia with non-traumatic bleeding; and Grade 3 or higher nausea, vomiting, and/or diarrhea that has not improved within 72 hours despite adequate and optimal treatment. The definition also includes other Grade 3 or higher non-hematological adverse events, with stated exceptions for alopecia and specified gastrointestinal events that improve within 72 hours with optimal treatment.

Counting participants with these events is useful in an early safety run-in because it provides a direct summary of how often protocol-defined toxicities occurred during the specified observation period. It is not a measure of all possible harms, longer-term safety, or tumor response. General statistical interpretation: a count is most informative when its analysis population and observation period are clear, since participants who are not observed for the same period may have different opportunities for an event.

The primary population is described as the safety run-in analysis set: treated participants during the safety run-in who received at least 75% of planned doses and completed the DLT period, or who experienced a DLT during that period. The registry record’s description is truncated after “regardles,” so further population details are not reported in the registry record. No primary endpoint count is reported.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry describes a binary-event outcome expressed as a count of participants: whether each participant experienced a protocol-defined dose-limiting toxicity during the safety run-in window. The endpoint’s unit is participants, and the specified observation period is up to Day 21 of Cycle 1. Although the outcome is a count, interpretation depends on the relevant analysis population and the number of participants observed. The primary endpoint result itself is not reported in the registry record.

What does the endpoint definition contribute?

A prespecified event definition helps make the endpoint more reproducible: the record identifies toxicity grades, event types, timing criteria for certain symptoms, and a judgement about whether the event is attributable to the disease or disease-related processes. General statistical explanation: a consistent definition reduces ambiguity in deciding which observations count as events. It does not eliminate judgement, particularly where attribution is assessed by an investigator and/or sponsor. The registry record does not report additional adjudication procedures.

The DLT analysis set is described as participants treated during the safety run-in who received at least 75% of the planned tepotinib and osimertinib dose and completed the DLT period, or who experienced a DLT during that period. Such eligibility conditions matter because the count is not necessarily a count among everyone enrolled. Including participants who experience a DLT even if they do not complete the period can preserve those events in the safety summary. The population description is truncated in the record, so any further eligibility rules are not reported in the registry record.

How should a count and its precision be read?

General statistical explanation: a count tells how many observed participants experienced an event, while an event proportion relates that count to the number of participants in the defined population. A confidence interval communicates uncertainty around an estimate and is generally wider when information is limited. These concepts help distinguish a raw count from an estimate of risk in a broader population. No primary endpoint proportion or confidence interval is reported in the registry record, so the primary result’s precision cannot be evaluated from the available details.

The broader adverse-event summaries include numbers at risk and numbers affected, which allow the displayed counts to be read in their group context. However, those summaries are not the primary DLT endpoint, and the registry record does not establish that their definitions, time windows, or analysis populations match the DLT population. Serious adverse events and other adverse events should also be kept distinct: they are separate categories in the record, not interchangeable summaries.

What follows from the non-randomized design?

General statistical explanation: random allocation is designed to balance, on average, both measured and unmeasured baseline factors between treatment groups. Without randomization, observed differences between groups can be associated with treatment, participant selection, disease characteristics, or other differences. This does not make descriptive results unusable, but it limits causal interpretation. The record explicitly notes that the arms are not a like-for-like control comparison. No adjustment strategy or covariate model is reported in the registry record.

The parallel-arm structure means that the listed interventions were assigned to distinct groups rather than described as a crossover. The record identifies no masking. These design features are relevant to interpretation, but they do not by themselves specify how outcomes were collected, whether assessors were independent for every outcome, or how missing observations were handled. Those details are not reported in the registry record beyond the named independent review committee and investigator response assessments.

What is the role of the planning illustrations?

The stated planning illustration assumptions are a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. The effect to detect is described as a stated fraction of the observed effect. For time-to-event tables, the stated quantity is total events using the Schoenfeld approach; other tables are described as participants per arm. These are planning assumptions, not observed trial outcomes or observed power. The registry record does not report that these assumptions were the trial’s original design justification, and the actual planning illustration values are not reported in the registry record for this analysis.

General statistical explanation: sample-size planning links an assumed effect, variability or event rate, significance threshold, and desired power to an approximate amount of information needed. If the assumed effect is based on an observed effect, the illustration is conditional on that estimate and may not represent the effect that would be seen in a future study. Such planning values do not show whether the completed study had a particular probability of detecting an effect. The trial’s own sample-size justification, target power, and complete statistical plan are not reported in the registry record.

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("Tepotinib + Osimertinib")
n      <- c(9)
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  Tepotinib + Osimertinib  9  1  11.1  0.3  48.2  2.0  43.5  2.5  44.5

11. Multiplicity, Missing Data and Analysis Populations

The registry lists 47 outcome measures, including the primary DLT count, response outcomes, and multiple safety and clinical-status measures. Testing or interpreting many outcomes can raise multiplicity concerns: general statistical explanation, when many hypotheses are assessed, the chance of at least one apparently notable result can increase unless the analysis plan accounts for that feature. The registry record does not report a multiplicity adjustment, a hierarchy of secondary endpoints, or which outcomes were intended to support confirmatory claims. It lists zero analyses, and no outcome-level inferential results are reported in the registry record.

Missing-data handling is also not reported in the registry record. The endpoint’s population definition requires completion of the DLT period or occurrence of a DLT, with a dose-exposure condition also described. The record does not explain how participants who discontinued for reasons other than a DLT were handled, how incomplete follow-up affected the denominator, or whether any sensitivity analyses were planned. Those details are important because participants with limited observation may have fewer opportunities for a toxicity to be detected.

The registry record does not report a complete analysis-population plan or the number assessed for the primary endpoint. It does name a safety run-in analysis set and provides adverse-event counts at risk in the two arms, but these figures should not be assumed to represent the same population as the primary endpoint. The record also does not report how protocol deviations, missing assessments, or differing follow-up were incorporated into analyses. Without those methods and endpoint results, readers cannot assess the robustness of the findings or the effect of missing information.

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: general physical health deterioration (8.3%), cancer pain (8.3%), disease progression (3.9%), covid-19 (3.9%), pneumonia (3.9%).

For the tepotinib-and-osimertinib group, the registry lists 128 participants at risk for serious adverse events and 49 affected; it lists 123 of 128 as affected by other adverse events and 59 deaths. For tepotinib monotherapy, it lists 12 participants at risk for serious adverse events and 2 affected; it lists 12 of 12 as affected by other adverse events and 7 deaths. These counts describe the groups as recorded, but the registry record does not report the follow-up duration associated with these summaries in the registry record for this analysis.

Serious adverse events are not the same category as dose-limiting toxicities. The primary endpoint applies a specific event definition and an early time window, while the adverse-event summaries describe broader categories. The record does not state that all serious adverse events were treatment-related. It separately lists treatment-emergent and treatment-related adverse events among secondary outcomes, but their numerical results are not reported in the registry record.

The listed serious-event terms include general physical health deterioration and cancer pain at 8.3% each; disease progression, COVID-19, pneumonia, and dyspnoea are listed at 3.9% each. The record does not give the denominator or calculation details for those term percentages. The death counts should likewise not be interpreted as treatment-attributed deaths: causality is not reported for these figures. Because treatment allocation was non-randomized and the arm sizes differ, the group counts do not constitute a randomized safety comparison.

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

The registry record establishes that INSIGHT 2 investigated tepotinib plus osimertinib in a phase II, non-randomized study, with a separate tepotinib monotherapy arm. It also establishes that the primary endpoint was a safety run-in count of participants with protocol-defined dose-limiting toxicities through Day 21 of Cycle 1. The record describes the toxicity criteria and names several secondary response and safety outcomes.

The registry entries available do not include the primary DLT count, numerical response results, or inferential estimates for the secondary outcomes. They therefore do not support a conclusion about the frequency of primary-endpoint toxicities, the magnitude or precision of a response outcome, or whether an efficacy difference was observed. The listed adverse-event counts provide descriptive information, but they do not replace the missing primary endpoint results or establish treatment attribution.

Nor does this design establish a causal comparison between the combination and monotherapy groups. Allocation was non-randomized, and the registry record explicitly cautions that the arms are not a like-for-like control comparison. Differences in their recorded counts could reflect differences between the groups or their observation, as well as any treatment-related factors. No adjustment methods are reported that would address such differences.

General statistical explanation: a trial result is most informative when the endpoint estimate is read alongside its analysis population, observation period, uncertainty, and analysis plan. Here, the primary endpoint definition and time frame are described, but its result and precision are not. This means the available information supports an account of the study question and design, not a definitive statistical assessment of the primary outcome or comparative benefit-risk profile.

The record’s status as completed does not itself supply missing outcome estimates. Likewise, a listed enrollment of 140 should not be treated as the number analyzed for every outcome; the adverse-event groups list different numbers at risk, and the primary analysis population has separate eligibility criteria. The amount of information contributing to each outcome cannot be fully assessed from the record details available.

14. Why This Trial Matters Statistically

INSIGHT 2 raises a recognizable statistical challenge: evaluating a combination regimen in a population described by prior treatment resistance and MET amplification, while also including a separate monotherapy arm. The design combines early safety assessment with response and broader tolerability outcomes. Clear distinctions among those endpoints matter because a safety run-in count, a RECIST response assessment, and an adverse-event summary answer different questions.

The record’s primary endpoint focuses on dose-limiting toxicities during a defined early treatment period. This emphasizes how the study operationalized early safety, but it does not by itself summarize longer-term adverse events or efficacy. The secondary outcomes extend the scope to response assessment by independent review committee or investigator, adverse events, laboratory values, vital signs, performance status, and ECG findings. Interpreting a collection of outcomes requires knowing which were primary, which were exploratory, and how multiplicity was handled; those statistical-plan details are not reported in the registry record.

The non-randomized allocation is especially important for readers considering the two arms side by side. A separate monotherapy group may provide descriptive context, but without randomization the groups cannot be assumed to be comparable. A raw difference would not isolate the effect of adding osimertinib. The record itself warns that the arms are not a like-for-like control comparison.

Finally, the difference between a listed outcome and a reported result is consequential. Although the registry describes 47 outcome measures and the study status is completed, the primary endpoint result and inferential analyses are not reported in the registry record. This makes transparent outcome reporting, population definitions, and uncertainty estimates central to any statistical appraisal. The record supplies enough detail to understand the intended early safety question, but not enough to judge its numerical answer.

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 endpoint is narrowly defined by its safety run-in period, event criteria, and analysis-set rules. It should not be generalized to all adverse events across the full treatment course. The registry record does not report the observed DLT count, the number included in the primary endpoint analysis, or a precision estimate. Its population description is truncated, leaving additional eligibility detail unreported in the registry record.

The arms were not randomized, and the registry explicitly states that they are not a like-for-like control comparison. The record does not report baseline characteristics or an adjustment plan that would permit assessment of group comparability. Therefore, descriptive contrasts between combination therapy and tepotinib monotherapy cannot be interpreted as isolated treatment effects.

The record lists serious adverse-event and other adverse-event counts, but does not report the follow-up duration for those summaries in the registry record for this analysis. It also does not state whether the listed deaths were treatment-related. The serious-event term percentages are listed without denominator or calculation details. The outcome definitions, populations, and observation windows for these summaries should not be assumed to match those of the primary DLT endpoint.

Further limitations arise from unreported analysis details. The record does not describe missing-data handling, multiplicity procedures, a complete statistical plan, or the trial’s sample-size justification. It lists 47 outcome measures and zero analyses, while numerical primary and secondary outcome results are not reported in the registry record. These omissions limit assessment of precision, robustness, and the role of secondary findings. The stated sample-size assumptions describe planning illustrations, not necessarily the original trial design, and do not establish observed power.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: A Study of Tepotinib Plus Osimertinib in Osimertinib Relapsed MET Amplified NSCLC (INSIGHT 2) — NCT03940703.
• Registry record last updated 2026-08-21; status last verified 2026-07.
• Results first posted 2024-06-04.
• 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

INSIGHT 2 (NCT03940703) is listed as a completed phase II, two-arm, parallel, non-randomized study with no masking. It investigated tepotinib plus osimertinib in advanced or metastatic non-small cell lung cancer with activating EGFR mutations, acquired resistance to prior osimertinib, and MET amplification. A separate arm received tepotinib monotherapy. The registry lists enrollment of 140, a start date of 2019-09-19, and a primary completion date of 2023-05-11.

The primary endpoint was the number of participants experiencing protocol-defined dose-limiting toxicities through Day 21 of Cycle 1 in the safety run-in analysis set. Secondary outcomes covered response assessments and several safety and clinical measures. The record lists 47 outcome measures and zero analyses. It reports adverse-event counts for both arms, but the primary endpoint count, response results, inferential estimates, missing-data methods, multiplicity plan, and complete statistical analysis plan are not reported in the registry record. The arms are not a like-for-like control comparison.