Are lazertinib and osimertinib comparable for treating advanced EGFR-mutant non-small cell lung cancer?—insights and limitations from the MARIPOSA study Authors: Edyta Maria Urbanska, Eric Santoni-Rugiu, Morten Grauslund, Jens Benn Sørensen Published: 2026 As the number of treatment options for advanced epidermal growth factor receptor (EGFR)-mutant non-small cell lung cancer (NSCLC) increases, careful comparative evaluation of the efficacy and safety of third-generation EGFR inhibitors becomes increasingly important. In the three-arm, randomized, double-blind, phase III MARIPOSA study for previously untreated patients with advanced NSCLC harboring common EGFR-activating mutations [exon 19 deletion (Ex19del) and the exon 21 L858R substitution], the efficacy of the EGFR-mesenchymal-epithelial transition (MET gene) bispecific antibody, amivantamab, combined with the highly selective, central nervous system (CNS)-penetrant, third-generation EGFR-tyrosine kinase inhibitor (TKI), lazertinib (1,2), was directly compared with lazertinib monotherapy and osimertinib monotherapy (3,4). The amivantamab-lazertinib combination led to significantly prolonged progression-free survival (PFS) and overall survival (OS) as compared with osimertinib monotherapy. The MARIPOSA design uniquely enabled direct comparison of two third-generation EGFR-TKIs, thereby allowing for a clearer assessment of the individual contribution of lazertinib to the combination treatment regimen. In this regard, Lee et al. have recently presented an additional exploratory analysis from the MARIPOSA trial to compare efficacy and safety of lazertinib versus osimertinib (4). Despite the inherent complexity of three-arm trials, the MARIPOSA trial (n=1,074) was sufficiently powered to evaluate treatment efficacy across all randomized arms with formal statistical power primarily allocated to the prespecified primary comparison. The amivantamab-lazertinib arm and the osimertinib arm each included 429 patients, and the lazertinib arm 216 patients. Baseline demographic and disease characteristics were well balanced between the osimertinib and lazertinib monotherapy groups, including gender distribution, ethnicity, performance status (PS), smoking history, and adenocarcinoma histology (97% vs. 98%). In both arms, Ex19del mutations were more frequent than L858R, and 40% of patients had baseline brain metastases. Time-to-event endpoints were assessed by blinded independent central review (BICR) using log-rank testing stratified by mutation type (Ex19del vs. L858R), race (Asian vs. non-Asian), and history of brain metastases (present vs. absent) and Cox proportional hazards modeling. Enrollment in the MARIPOSA study required identification in tumor tissue and/or plasma of EGFR Ex19del or L858R variants using the real-time polymerase chain reaction (PCR)-based cobas® EGFR Mutation Test v2 by Roche Diagnostics. Thus, even if this PCR test can detect the major uncommon EGFR mutations, only participants harboring Ex19dels and L858R were enrolled in the study, thereby impeding a direct head-to-head comparison between the efficacy of lazertinib and osimertinib in NSCLC patients harboring atypical EGFR variants and exon 20 insertions. In this respect, the structure-function-based classification of EGFR variants by Robichaux et al. predicts a similarly reduced responsiveness of the atypical mutations to these two third-generation EGFR-TKIs (5). Moreover, recent data on the use of lazertinib in patients with advanced NSCLC harboring uncommon EGFR mutations have shown efficacy and safety profile similar to that observed with osimertinib in other studies, with both drugs being less efficacious in these patients than in those with Ex19dels or L858R (6-10). The efficacy of the Amivantamab-Lazertinib combination in patients with uncommon EGFR mutations has been explored in the CHRYSALIS-2 Cohort C study displaying an objective response rate (ORR) of 57% in treatment-naïve patients, median PFS (mPFS) of 19.5 months, and median duration of response (mDOR) of 20.7 months, suggesting improved activity relative to that previously reported for EGFR-TKI monotherapy (10,11). Although amivantamab combined with chemotherapy has demonstrated superiority over chemotherapy alone in NSCLC with EGFR exon 20 insertions (Ex20ins) (12), no patients with EGFR Ex20ins were included in the MARIPOSA trial. EGFR protein variants resulting from Ex20ins, except for A763_Y764insFQEA/LQEA, alter the tyrosine-kinase domain conformation and confer resistance to third-generation EGFR-TKIs (5,13). Indeed, osimertinib exhibits severely impaired clinical efficacy in patients with Ex20ins (7,13), and whether lazertinib performs differently in this setting or in other uncommon EGFR mutations remains uncertain. While MARIPOSA trial focused exclusively on common EGFR mutations (Ex19del and L858R), the mechanism of action of Amivantamab is not restricted to specific EGFR kinase domain mutations and has demonstrated activity across a broader spectrum of EGFR alterations, including Ex20ins (12). Therefore, the exclusion of exon 20 insertions from MARIPOSA limits generalizability of the monotherapy comparison but underscores the potential rationale for combination strategies incorporating agents with broader mutational coverage. Furthermore, the MARIPOSA trial, apart from the possible presence of TP53 co-mutations, provided no information about genetic co-alterations in the tumors of the included patients. EGFR-mutant NSCLC with driver co-alterations may represent a high-risk molecular subgroup and may negatively impact the response to treatment with osimertinib (14). Plasma circulating tumor DNA (ctDNA) was evaluated using the U.S. Food and Drug Administration (FDA)-approved Guardant360® CDx in 636 patients (amivantamab-lazertinib, n=320; osimertinib, n=316) at baseline and at cycle 3, day 1 (C3D1) (3,15). Among patients with TP53 co-mutations, mPFS was 18.2 months in the combination arm vs. 12.9 months in the osimertinib arm. Among patients with detectable baseline ctDNA, treatment with amivantamab-lazertinib significantly prolonged PFS (20.3 vs. 14.8 months) (15). Similar findings were observed in patients without ctDNA clearance at C3D1 (16.5 vs. 9.1 months) (15). Importantly, this biomarker-driven analysis compared the amivantamab-lazertinib combination with osimertinib, and corresponding ctDNA-based efficacy data for Lazertinib monotherapy have not been reported, limiting conclusions regarding its independent molecular activity in these high-risk subsets. Considering the key efficacy endpoints of Lee et al.’s exploratory analysis (4) (Table 1), at the median follow-up of 22 months, comparable outcome as assessed by BICR was observed, with mPFS 18.5 months for lazertinib and 16.6 months for osimertinib, but without statistical significance [hazard ratio (HR) 0.98; 95% confidence interval (CI): 0.79–1.22, P=0.86]. Similarly, in the two groups the percentages of patients alive and progression-free at 12, 18 and 24 months as well as the mPFS among prespecified subgroups (age category; gender; race; weight category; PS score; history of smoking; EGFR Ex19del vs. L858R) were highly comparable across subgroups. Table 1Comparative efficacy and safety of lazertinib and osimertinib in the exploratory analysis of the MARIPOSA trialParameterLazertinibOsimertinibInterpretationEfficacy Median PFS (months)18.516.6No significant difference Hazard ratio (95% CI)0.98 (0.79–1.22)–Confidence interval crossed unity Nominal P value0.86–No multiplicity adjustments Objective response rate (%)8385Comparable Median duration of response (months)16.816.6Comparable 24-month PFS rate (%)7169Similar CNS mPFS (baseline brain metastases, months)16.413.0Exploratory; limited sample sizeSafety QT interval prolongation >450 ms (%)917Lower with lazertinib LVEF decline (%)14Lower with lazertinib Diarrhea (%)3244Higher with osimertinib Rash (%)4531Higher with lazertinib Statistical frameworkExploratory comparisonExploratory comparisonNot powered for superiority testingThe comparison was not powered for definitive superiority testing and relied on nominal P values without multiplicity adjustment. CI, confidence interval; CNS, central nervous system; LVEF, left ventricular ejection fraction; mPFS, median progression-free survival; PFS, progression-free survival. Comparative efficacy and safety of lazertinib and osimertinib in the exploratory analysis of the MARIPOSA trial The comparison was not powered for definitive superiority testing and relied on nominal P values without multiplicity adjustment. CI, confidence interval; CNS, central nervous system; LVEF, left ventricular ejection fraction; mPFS, median progression-free survival; PFS, progression-free survival. The mPFS was also consistent among patients with high-risk features, defined by presence of brain metastases, detectable ctDNA at baseline, and TP53 co-mutations. Likewise, the ORR was equivalent in the lazertinib arm and osimertinib arm (83% vs. 85%) and so were the percentages of patients with complete (4% vs. 4%) and partial response (81% vs. 79%), the mDOR (16.8 vs. 16.6 months), and interim OS, which was not estimable in either arm due to immature data. The percentages of participants who were alive at 18 and 24 months were 78% and 71% in the lazertinib arm and 79% and 69% in the osimertinib arm, respectively. Although numerical trends favored lazertinib in biologically and clinically aggressive subgroups, confidence intervals crossed unity, and no statistically significant differences were observed (4). These exploratory findings warrant cautious interpretation. Among these high-risk features, CNS involvement warrants special clinical attention. EGFR-mutant NSCLC is associated with a significant propensity for the development of brain metastases with approximately 25–40% of patients presenting with CNS disease at diagnosis and up to 50–70% developing intracranial progression during the disease course (16-18). Effective and durable CNS control is therefore a key therapeutic objective. In this context, both lazertinib and osimertinib have demonstrated substantial intracranial activity compared with first-generation EGFR-TKIs. In the LASER301 study, lazertinib significantly prolonged intracranial PFS vs. gefitinib (HR 0.42) (2) and achieved high intracranial response rates, whereas in the FLAURA CNS analysis, osimertinib significantly improved CNS PFS versus standard EGFR-TKIs (HR 0.48) and reduced the risk of CNS progression (16). In the MARIPOSA trial mPFS among the patient subgroup with brain metastases were 16.4 versus 13.0 months with lazertinib and osimertinib, respectively. However, this comparison is based on only 86 versus 172 patients, respectively, and further exploration on which TKI to use in case of brain metastases is warranted. Collectively, the results indicate that lazertinib and osimertinib possess highly comparable efficacy profiles (Table 1). Safety findings in MARIPOSA were comparable between arms, with most adverse events (AEs) classified as grade 1–2 and consistent with earlier studies (2,16). Yet, patients in the osimertinib arm experienced higher rates of diarrhea (44% vs. 32% in the lazertinib arm) and thrombocytopenia (20% vs. 9%). Patients receiving lazertinib experienced higher rates of rash (45% vs. 31% in the osimertinib arm), muscle spasms (23% vs. 7%), alanin aminotransferase/aspartate aminotransferase (ALT/AST) elevation (23%/21% vs. 13%/14%) and paresthesia (15% vs. 6%). Episodes of venous thrombo-embolism (VTE) were mostly observed after the first 4 months of treatment in both arms, however with an incidence much lower than that observed in the amivantamab-lazertinib arm (3). Recently, VTEs have also been reported with the combination of amivantamab and osimertinib (19), suggesting a potential association between amivantamab-containing regimens and increased VTE risk; however, a definitive causal relationship cannot be established based on available data. Notably, the cardiac toxicity was different with lower rates of QT interval prolongation (>450 ms) with lazertinib (9%) than with osimertinib (17%). Furthermore, rates of left ventricular ejection fraction (LVEF) decline were lower with lazertinib (1%) as compared with osimertinib (4%). While these differences were modest in absolute terms, they may be clinically relevant in selected patients with baseline cardiovascular risk factors. A potential biological explanation relates to differential off-target inhibition of human epidermal growth factor receptor (HER)-family receptors. HER2 signaling plays an important role in cardiomyocyte survival and stress adaptation through neuregulin-1-mediated HER2/HER4 activation pathways (20). Osimertinib has been shown to exhibit broader HER-family inhibitory activity (21), which could theoretically contribute to the observed differences in cardiac events. However, this mechanistic link remains hypothetical and was not directly evaluated in MARIPOSA. A structured overview of key efficacy and safety parameters is provided in Table 1. In summary, this exploratory head-to-head analysis from MARIPOSA demonstrates that lazertinib achieves efficacy comparable to osimertinib in first-line treatment of advanced NSCLC harboring common EGFR mutations. Although the lazertinib versus osimertinib comparison was not powered for definitive superiority testing and relied on nominal P values without multiplicity adjustments, the consistency of efficacy across endpoints (PFS, ORR, and subgroup analyses) strengthens the robustness of the findings. Lazertinib exhibited a slightly distinct safety profile, including lower rates of QT prolongation and LVEF decline, supporting a potentially more favorable cardiac safety profile. Despite OS immaturity, the results indicate that Lazertinib appears as a clinically reasonable alternative to Osimertinib and a rational backbone for combination strategies in EGFR-mutant NSCLC. Supplementary The article’s supplementary files as Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-1-0485/prf