This page separates reported trial results from statistical interpretation. Numerical results presented here are restricted to the ClinicalTrials.gov trial data posted on ClinicalTrials.gov for MARIPOSA-2.
1. Trial at a Glance
MARIPOSA-2 is a randomized, open-label, parallel-group phase 3 oncology trial evaluating lazertinib and amivantamab in combination with platinum-based chemotherapy against platinum-based chemotherapy in patients with EGFR-mutated locally advanced or metastatic non-small cell lung cancer after osimertinib failure.
| Feature | MARIPOSA-2 |
|---|---|
| Phase | Phase 3 |
| Condition | Carcinoma, Non-Small-Cell Lung |
| Population | Patients with EGFR-mutated locally advanced or metastatic non-small cell lung cancer after osimertinib failure |
| Design | Randomized, parallel-group |
| Masking | None |
| Allocation | Randomized |
| Primary purpose | Treatment |
| Enrollment | 776 |
| Arms | 5 |
| Primary endpoint types | Binary; Time-to-event |
| Results posted | Yes |
| ClinicalTrials.gov | NCT04988295 |
2. Clinical Question
The principal efficacy question represented in the posted statistical analyses is whether treatment combinations containing amivantamab and chemotherapy, with or without lazertinib, improve progression-free survival compared with carboplatin plus pemetrexed in patients with EGFR-mutated locally advanced or metastatic non-small cell lung cancer after osimertinib failure.
Population
Patients with epidermal growth factor receptor (EGFR)-mutated locally advanced or metastatic non-small cell lung cancer after osimertinib failure.
Interventions
Lazertinib, amivantamab, pemetrexed, and carboplatin are the interventions listed in the registry data. The primary comparisons identify lazertinib + amivantamab + carboplatin + pemetrexed and amivantamab + carboplatin + pemetrexed as the active combinations.
Comparator
Carboplatin + pemetrexed (CP).
Primary question
Do the specified combination regimens reduce the hazard of progression or death compared with carboplatin plus pemetrexed?
3. Trial Design
The registry profile identifies five study arms, but the ClinicalTrials.gov record does not provide complete descriptions or sample sizes for all five arms. The formal efficacy analyses identify three relevant treatment groups by name: lazertinib + amivantamab + carboplatin + pemetrexed, carboplatin + pemetrexed, and amivantamab + carboplatin + pemetrexed. This page does not infer the composition of the remaining arms.
Lazertinib + Amivantamab + Chemotherapy
- Lazertinib
- Amivantamab
- Carboplatin
- Pemetrexed
- Serious adverse events: 137 affected / 263 at risk
Carboplatin + Pemetrexed
- Carboplatin
- Pemetrexed
- Serious adverse events: 49 affected / 243 at risk
Amivantamab + Chemotherapy
- Amivantamab
- Carboplatin
- Pemetrexed
- Serious adverse events: 42 affected / 130 at risk
Not Fully Described in registry Data
- The registry profile reports 5 arms.
- The ClinicalTrials.gov record does not provide the complete treatment descriptions or denominators for the other arms.
- No additional arm characteristics are inferred here.
4. Clinical Trial Timeline
Study initiated
The MARIPOSA-2 study began on 2021-11-17.
Primary completion reached
The registered primary completion date was 2023-07-10.
Active, not recruiting
The ClinicalTrials.gov profile lists the study status as ACTIVE_NOT_RECRUITING.
5. Primary Endpoints
| Endpoint | Registered definition / time frame | Type |
|---|---|---|
| Main Study: Progression-free Survival (PFS) According to RECIST Version 1.1 as Assessed by BICR | From randomization to either disease progression or death, whichever occurred first (up to 1 year 7 months). PFS is defined as the time from randomization until the date of objective disease progression or death, whichever came first, as assessed by BICR according to RECIST version 1.1. Progressed disease: Sum of diameters increased by greater than or equal to 20% and greater than or equal to 5 mm from nadir. | Time-to-event |
| Main Study + Extension Cohort: Number of Participants With Treatment-emergent Adverse Events (TEAEs) and Treatment-emergent Serious Adverse Events (TESAEs) | Up to 4 years 10 months. Follow-up for the extension cohort is still ongoing and results from the analysis that includes the extension cohort will be reported upon study completion. | Binary |
| Main Study + Extension Cohort: Progression-free Survival (PFS) as Assessed by Blinded Independent Central Review | Up to 4 years 10 months. | Time-to-event |
The registry profile lists three primary endpoints, while the registry-reported formal statistical analyses consist of two analyses of the main-study PFS endpoint. The two posted formal analyses compare the carboplatin + pemetrexed group with two different active treatment combinations.
6. Statistical Methodology
Full analysis set and intention-to-treat principle
The registry-reported analysis text states that the full analysis set (FAS) included all randomized participants, classified according to their assigned treatment arm regardless of the actual treatment received. The analysis text also identifies intention-to-treat analysis as a concept used in the statistical analysis.
This approach keeps the primary randomized comparison tied to treatment assignment rather than allowing treatment receipt, discontinuation, or other post-randomization events to redefine the principal efficacy groups.
Kaplan-Meier estimation
PFS is a time-to-event endpoint. For this type of endpoint, Kaplan-Meier estimation is used to describe the distribution of time to progression or death while appropriately accounting for participants who are censored before experiencing the event.
Here, di is the number of events at time ti and ni is the number at risk immediately before that time.
Log-rank test
The formal registry method for both posted PFS analyses is the log-rank test. The log-rank test compares time-to-event experience between groups across the follow-up period rather than comparing only one fixed time point.
Hazard ratio
The effect measure reported for both formal analyses is the hazard ratio. A hazard ratio compares the estimated instantaneous event rates between the groups under the time-to-event analysis framework.
The direction of interpretation depends on the ordering of the groups. A hazard ratio of 0.44 for LACP/ACP-L versus CP therefore indicates a lower estimated hazard of progression or death in the LACP/ACP-L group relative to CP.
Superiority framework
Both posted formal PFS analyses are identified as superiority analyses. The registry therefore frames these comparisons around whether the active regimen demonstrates a statistically detectable reduction in the hazard of progression or death relative to the comparator, rather than around a non-inferiority margin.
7. Results
The ClinicalTrials.gov record contains two formal statistical analyses, both for the main-study PFS endpoint assessed according to RECIST version 1.1 by blinded independent central review. Both analyses use the full analysis set, log-rank testing, hazard ratios, two-sided 95% confidence intervals, and a superiority hypothesis.
Primary PFS Analysis: LACP/ACP-L vs CP
Hazard ratio for progression or death
95% CI: 0.35–0.56 · P < 0.0001
Comparison: Lazertinib + Amivantamab + Carboplatin + Pemetrexed vs Carboplatin + Pemetrexed
An HR of 0.44 means that, within the time-to-event model used for this analysis, the estimated instantaneous hazard of progression or death for the LACP/ACP-L group was 0.44 times the corresponding hazard in the CP group. Expressed as a simple relative interpretation, this corresponds to an estimated 56% lower hazard in the first-listed group.
The HR does not mean that 56% of patients avoided progression, that every patient experienced a 56% reduction in risk, or that the probability of progression or death was reduced by exactly 56% at every point in time.
The 95% CI of 0.35–0.56 describes statistical uncertainty around the estimated hazard ratio. It does not describe the range of individual patient outcomes. Because the entire interval is below 1, the interval is consistent with a lower estimated hazard for the first-listed treatment group under this analysis.
The P < 0.0001 value addresses the statistical evidence against the null hypothesis used for the comparison. A p-value is not a measure of effect size, clinical importance, or the probability that the treatment works.
Interpretation also depends on the time-to-event framework, including censoring and the assumptions underlying a hazard-ratio summary. The ClinicalTrials.gov record identifies the analysis as a log-rank comparison with a hazard ratio but does not provide a separate assessment of proportional hazards.
Primary PFS Analysis: ACP vs CP
Hazard ratio for progression or death
95% CI: 0.36–0.64 · P < 0.0001
Comparison: Amivantamab + Carboplatin + Pemetrexed vs Carboplatin + Pemetrexed
An HR of 0.48 means that, under the reported time-to-event analysis, the estimated instantaneous hazard of progression or death for the ACP group was 0.48 times the hazard in the CP group. As a simple relative interpretation, this corresponds to an estimated 52% lower hazard for the first-listed group.
Again, the HR is not an absolute risk reduction and does not mean that 52% of participants were protected from progression or death. It is a relative time-to-event measure derived from the statistical analysis.
The 95% CI of 0.36–0.64 quantifies uncertainty around the estimated HR. Because the interval remains below 1, the registry-reported analysis is consistent with a lower estimated hazard for ACP than for CP.
The P < 0.0001 value indicates strong statistical evidence against the null hypothesis specified for the superiority comparison. It should not be interpreted as an effect-size measure or as the probability that the observed treatment effect is due to chance.
As with the first analysis, the interpretation depends on censoring and the time-to-event model. The ClinicalTrials.gov record does not provide a separate proportional-hazards diagnostic, so the HR should be understood as the reported summary measure rather than as proof that hazards were proportional throughout follow-up.
| Primary PFS comparison | Hazard ratio | 95% CI | P-value | Hypothesis |
|---|---|---|---|---|
| LACP/ACP-L vs CP | 0.44 | 0.35–0.56 | <0.0001 | Superiority |
| ACP vs CP | 0.48 | 0.36–0.64 | <0.0001 | Superiority |
8. Other Registered Primary Endpoints
Treatment-Emergent Adverse Events and Serious Adverse Events
The registry lists the number of participants with treatment-emergent adverse events and treatment-emergent serious adverse events as a primary endpoint for the main study plus extension cohort, with a time frame of up to 4 years 10 months. The ClinicalTrials.gov record states that extension-cohort follow-up is still ongoing and that results from the analysis including the extension cohort will be reported upon study completion.
The ClinicalTrials.gov record does provide serious adverse event counts for three main-study arms. It does not provide complete TEAE counts or a formal statistical comparison for the registered binary primary endpoint.
Main Study + Extension Cohort PFS
The registry also lists PFS assessed by blinded independent central review as a primary endpoint for the main study plus extension cohort, with a time frame of up to 4 years 10 months. The registry-reported statistical-analyses data does not provide a formal estimate, confidence interval, or p-value for this extension-cohort endpoint.
9. Safety Results
The ClinicalTrials.gov record reports treatment-emergent serious adverse events by arm as affected participants divided by participants at risk. These figures are descriptive counts; the ClinicalTrials.gov record does not provide a p-value, confidence interval, risk ratio, odds ratio, or other formal between-arm safety comparison.
| Main-study arm | Serious adverse events | At risk |
|---|---|---|
| Arm A: Lazertinib + Amivanta | 137 | 263 |
| Arm B: Carboplatin + Pemetre | 49 | 243 |
| Arm C: Amivantamab + Carbopl | 42 | 130 |
The registry's abbreviated arm labels are retained in the safety table where they are reported. The formal efficacy analyses provide the corresponding full regimen names for the primary comparisons.
10. Statistical Methods Explained
Why was a log-rank test used for PFS?
PFS measures the time from randomization until progression or death, so participants can have different follow-up durations and can be censored. The log-rank test is designed for comparing such time-to-event distributions across randomized groups rather than treating the endpoint as a simple binary outcome.
What does an HR of 0.44 mean?
An HR of 0.44 means that the estimated instantaneous hazard in the first-listed treatment group was 44% of that in the comparator under the reported analysis. A convenient relative description is a 56% lower estimated hazard. It is not a statement that 56% of patients benefited or that the absolute probability of progression or death fell by 56%.
Why is the confidence interval important?
The 95% confidence interval communicates uncertainty around the estimated HR. For the LACP/ACP-L comparison, the interval is 0.35–0.56; for ACP versus CP, it is 0.36–0.64. Neither interval crosses 1, which is the null value for a hazard ratio. The width of the interval also provides information about precision.
Why does the p-value not measure effect size?
A p-value addresses evidence against a specified null hypothesis under the statistical model. It depends on both the magnitude of the observed effect and the information available for estimating it. Consequently, a very small p-value does not mean that the treatment effect is necessarily large, clinically important, or uniform across patients.
Why does intention-to-treat analysis matter?
The registry-reported analysis population is the full analysis set containing all randomized participants, classified according to assigned treatment regardless of actual treatment received. This preserves the comparison generated by randomization and reduces the risk that post-randomization treatment decisions determine the primary efficacy groups.
What does RECIST progression mean in this analysis?
The registered PFS definition uses objective disease progression or death, whichever occurs first, assessed by blinded independent central review according to RECIST version 1.1. The registry-reported definition specifies progressive disease as a sum of diameters increase of at least 20% and at least 5 mm from nadir, including baseline if it was the smallest sum.
Why is the extension-cohort endpoint not given a hazard ratio here?
The ClinicalTrials.gov record identifies the extension-cohort PFS endpoint but does not provide a formal statistical analysis with an estimate, confidence interval, and p-value. Reporting a hazard ratio would therefore require information not contained in the ClinicalTrials.gov record.
11. Understanding the Two PFS Comparisons
MARIPOSA-2 is statistically interesting because the ClinicalTrials.gov record contains two separate superiority comparisons against the same CP comparator. The first evaluates the combination containing both lazertinib and amivantamab; the second evaluates amivantamab with carboplatin and pemetrexed.
LACP/ACP-L vs CP
HR 0.44, with a two-sided 95% CI of 0.35–0.56 and P < 0.0001. The estimate is below 1, indicating a lower estimated hazard of progression or death for the first-listed regimen.
ACP vs CP
HR 0.48, with a two-sided 95% CI of 0.36–0.64 and P < 0.0001. The estimate is also below 1, indicating a lower estimated hazard of progression or death for the first-listed regimen.
The two hazard ratios should not be interpreted as though they came from a direct head-to-head comparison of LACP/ACP-L versus ACP. The analyses posted on ClinicalTrials.gov compare each active regimen with CP. A smaller HR in one comparison does not by itself establish that one active regimen is superior to the other.
12. Censoring and Time-to-Event Interpretation
Progression-free survival is not simply a count of patients who progressed. It incorporates the timing of progression or death. Participants who have not experienced the event by their last evaluable assessment can contribute information up to the time they are censored.
A valid time-to-event analysis uses each participant's available follow-up rather than requiring every participant to be observed until progression or death.
The ClinicalTrials.gov record does not provide the detailed censoring rules or censoring distributions needed for a more granular assessment. Accordingly, this page explains the statistical role of censoring without attributing additional censoring assumptions to MARIPOSA-2.
13. Hazard Ratios: Relative Effects vs Absolute Outcomes
The estimated hazard of progression or death was lower for LACP/ACP-L than for CP under the reported model. The numerical estimate was 0.44.
It does not state the percentage of patients who avoided progression, the median PFS, the probability of being progression-free at a particular date, or the absolute difference in those probabilities.
The ACP versus CP analysis produced an estimated hazard ratio of 0.48, indicating a lower estimated hazard for ACP under the reported analysis.
Absolute survival probabilities, median time-to-event estimates, and numbers of events can make treatment effects easier to interpret clinically. Those measures are not reported in the ClinicalTrials.gov record, so they are not added here.
14. Multiplicity and Multiple Comparisons
The ClinicalTrials.gov record identifies two formal primary PFS analyses, both using superiority hypotheses and both reporting P values below 0.0001. It does not provide the prespecified multiplicity-adjustment procedure, alpha allocation, hierarchical testing strategy, or information about interim-analysis boundaries.
| Feature | What the ClinicalTrials.gov record establishes |
|---|---|
| Formal PFS comparisons | Two statistical analyses are posted. |
| Hypothesis type | Superiority for both analyses. |
| P-value direction | Both are reported as <0.0001. |
| Confidence intervals | Both are two-sided 95% confidence intervals. |
| Multiplicity adjustment | Not specified in the ClinicalTrials.gov record. |
| Interim-analysis procedure | Not specified in the ClinicalTrials.gov record. |
15. Stratification and Randomization
the ClinicalTrials.gov record establishes randomization and a parallel-group design. Randomization is central to the causal interpretation of the efficacy comparisons because treatment assignment, rather than post-randomization treatment receipt, defines the full analysis set.
The ClinicalTrials.gov record does not identify specific randomization strata. Therefore, no stratification factors are attributed to MARIPOSA-2 on this page.
Randomization
Creates the basis for comparing assigned treatment groups while reducing systematic differences in measured and unmeasured baseline factors in expectation.
FAS analysis
Keeps participants classified according to their assigned treatment, regardless of the actual treatment received.
16. Non-Inferiority, Crossover, and Bayesian Methods
The ClinicalTrials.gov record identifies the primary PFS analyses as superiority analyses. No non-inferiority margin is reported, so no non-inferiority interpretation is appropriate for the posted comparisons.
The ClinicalTrials.gov record does not identify crossover as a design feature, does not provide a crossover analysis, and does not report a Bayesian statistical method. These topics are therefore not treated as trial-specific methods in this analysis.
| Design topic | MARIPOSA-2 evidence in the ClinicalTrials.gov record |
|---|---|
| Superiority | Yes; both formal PFS analyses are identified as superiority analyses. |
| Non-inferiority margin | Not reported. |
| Crossover | Not reported. |
| Factorial design | Not reported; the registered design model is parallel. |
| Bayesian analysis | Not reported. |
| Interim analysis | Not reported in the ClinicalTrials.gov record. |
| Missing-data / imputation method | Not reported in the ClinicalTrials.gov record. |
17. What the PFS Endpoint Measures
The registered PFS endpoint begins at randomization and ends at the first of two events: objective disease progression or death. This construction is important because death is included as an event rather than treated as ordinary loss to follow-up.
Starting point
Randomization.
Event 1
Objective disease progression assessed by blinded independent central review according to RECIST version 1.1.
Event 2
Death, whichever occurred first.
Progression criterion reported
Sum of diameters increased by at least 20% and at least 5 mm from nadir, including baseline if it was the smallest sum.
This endpoint combines radiographic disease progression and death into one time-to-event outcome. Consequently, the PFS hazard ratio is not a pure measure of tumor shrinkage; it summarizes the relative rate of the defined composite event over time.
18. Interpretation of the Confidence Intervals
| Comparison | Estimate | Two-sided 95% CI | Interpretive point |
|---|---|---|---|
| LACP/ACP-L vs CP | HR 0.44 | 0.35–0.56 | The interval remains below the null HR of 1. |
| ACP vs CP | HR 0.48 | 0.36–0.64 | The interval remains below the null HR of 1. |
The confidence intervals provide more information than the point estimates alone. An HR of 0.44 is compatible with values from 0.35 to 0.56 within the reported 95% confidence framework, while the HR of 0.48 has a corresponding interval of 0.36–0.64. Neither interval crosses 1, the value representing equal hazard.
These intervals should not be read as saying that 95% of individual patients experience hazards inside those ranges. Confidence intervals describe uncertainty in the estimated treatment effect, not individual-level variation in outcomes.
19. Why the P-Values Should Be Read Carefully
Both formal PFS comparisons have P < 0.0001. This is strong statistical evidence against the corresponding null hypothesis under the reported testing framework.
Confidence interval → How precisely is it estimated?
P-value → How much evidence is there against the null hypothesis?
These are complementary quantities. None of them alone describes the entire clinical meaning of a treatment comparison.
The ClinicalTrials.gov record does not state whether an overall type I error allocation was used across the two formal comparisons or across other endpoints. Consequently, the reported p-values are reproduced exactly but no additional multiplicity-adjusted interpretation is imposed.
20. Analysis Population and Causal Interpretation
The FAS definition posted on ClinicalTrials.gov for both formal analyses is particularly important: all randomized participants are included and classified according to assigned treatment regardless of actual treatment received.
Why assignment matters
Randomization establishes the treatment groups before post-randomization events occur. Maintaining assignment in the efficacy analysis preserves that randomized comparison.
What FAS does not solve
FAS analysis does not automatically eliminate every source of bias. Censoring, missing assessments, treatment discontinuation, subsequent treatment, and departures from model assumptions can still affect interpretation.
The ClinicalTrials.gov record does not provide enough detail to quantify these issues beyond the stated FAS definition. The appropriate interpretation is therefore that the formal efficacy estimates are randomized-group comparisons analyzed according to assigned treatment within the registry-reported FAS framework.
21. Limitations
- Incomplete arm information: the ClinicalTrials.gov record reports 5 arms, but the ClinicalTrials.gov record does not fully describe all five arms. Only the arms appearing in the formal comparisons and safety data are described in detail here.
- Limited reported endpoint statistics: formal estimates, confidence intervals, and p-values are posted on ClinicalTrials.gov for the main-study PFS analyses, but not for every registered primary endpoint.
- Extension-cohort uncertainty: the registry states that extension-cohort follow-up is still ongoing for the TEAE/TESAE endpoint and that the corresponding analysis will be reported upon study completion.
- No median PFS reported: the ClinicalTrials.gov record provides hazard ratios and confidence intervals but does not provide median PFS estimates or time-specific PFS probabilities.
- No subgroup analyses reported: no subgroup-specific hazard ratios or interaction tests are included in the ClinicalTrials.gov record.
- No multiplicity procedure reported: the data identifies two formal superiority comparisons but does not specify the error-control strategy across comparisons.
- No proportional-hazards assessment reported: the HR is reported, but the ClinicalTrials.gov record does not provide a diagnostic assessment of whether the proportional-hazards assumption is appropriate.
- No missing-data method reported: the ClinicalTrials.gov record does not specify imputation or missing-assessment procedures for the primary PFS analysis.
- Safety is descriptive in the ClinicalTrials.gov record: serious adverse event counts are available for three arms, but no formal comparative safety analysis is reported.
22. Why This Trial Matters Statistically
MARIPOSA-2 is a useful teaching example because its the ClinicalTrials.gov record combines randomized treatment assignment, a clearly defined time-to-event endpoint, blinded independent central review, log-rank testing, hazard-ratio estimation, two-sided confidence intervals, and a full-analysis-set framework.
| Statistical concept | How it appears in MARIPOSA-2 |
|---|---|
| Randomization | The trial uses randomized allocation in a parallel-group phase 3 design. |
| Intention-to-treat principle | The FAS includes randomized participants and classifies them according to assigned treatment. |
| Time-to-event endpoint | PFS measures time from randomization to progression or death. |
| Blinded independent review | The main-study PFS endpoint is assessed by BICR according to RECIST version 1.1. |
| Log-rank test | The reported formal method for both primary PFS analyses. |
| Hazard ratio | The reported effect measure for both formal PFS comparisons. |
| Confidence interval | Both analyses report two-sided 95% CIs around the HR. |
| Superiority testing | Both formal analyses are identified as superiority hypotheses. |
| Multiple comparisons | Two formal PFS comparisons are reported against the same CP comparator; the ClinicalTrials.gov record does not specify the multiplicity procedure. |
| Safety analysis | Serious adverse event counts are posted on ClinicalTrials.gov for three main-study arms. |
23. Related Tutorials
Learn more about the methods used in this trial:
24. Related Calculators
25. Sources
- ClinicalTrials.gov: MARIPOSA-2, NCT04988295.
Continue with Clinical Biostats statistical methods
Explore tutorials and statistical calculators covering the time-to-event methods, treatment-effect measures, and clinical-trial concepts illustrated by MARIPOSA-2.
26. Record Summary
MARIPOSA-2 provides a focused example of randomized time-to-event analysis. The ClinicalTrials.gov record reports a full-analysis-set approach, blinded independent central review of PFS according to RECIST version 1.1, log-rank testing, hazard-ratio estimation, two-sided 95% confidence intervals, and superiority hypotheses. The two formal PFS comparisons produced HRs of 0.44 and 0.48, respectively, with both P values reported as <0.0001. The most important statistical distinction is between what the hazard ratios establish as relative time-to-event effects and what the ClinicalTrials.gov record does not provide, such as median PFS, absolute PFS probabilities, subgroup effects, multiplicity procedures, or extension-cohort formal analyses.