Personalized care for patients with EGFR‐mutant nonsmall cell lung cancer: Navigating early to advanced disease management

Abstract

The discovery of activating mutations in the epidermal growth factor receptor (EGFR) gene has revolutionized the management of lung cancer, enabling the development of targeted tyrosine kinase inhibitors (TKIs). These therapies offer improved survival and reduced side effects compared with conventional treatments. Recent advancements have significantly reshaped the treatment paradigm for EGFR‐mutant non‐small cell lung cancer. TKIs are now incorporated into the management of early stage and locally advanced disease, and phase 3 trials have explored combination strategies in metastatic settings. Although these intensified approaches improve progression‐free survival, they come with increased toxicity and higher costs, underscoring the need for precise patient selection to maximize benefit. Emerging data on biomarkers, such as co‐mutations and circulating tumor DNA, show promise for refining treatment decisions. In addition, significant progress in understanding resistance mechanisms to EGFR TKIs has broadened therapeutic options. This review provides a comprehensive overview of the current landscape of EGFR‐mutant nonsmall cell lung cancer, highlighting recent breakthroughs and discussing strategies to optimize treatment based on the latest evidence.

INTRODUCTION

The discovery of activating mutations in the epidermal growth factor receptor (EGFR) gene introduced the concept of actionable alterations in nonsmall cell lung cancer (NSCLC) and paved the way for precision medicine in lung cancer. Mutations in the tyrosine kinase domain of EGFR represent one of the most common druggable alterations in NSCLC. EGFR kinase domain mutations are found almost exclusively in lung adenocarcinomas (LUADs), with a prevalence ranging from 15% to 50% depending on ethnicity. ¹ They occur more commonly in patients with no smoking history, females, and patients of Asian ethnicity. ² , ³ Among patients with NSCLC of East Asian genomic ancestry—determined by a single‐nucleotide variant approach—the prevalence of EGFR mutations is approximately 50.3% versus 34.2%, 26.8%, 14%, and 13% for patients of South Asian, admixed American, African, and European genomic ancestry, respectively. ⁴ , ⁵

EGFR‐mutant NSCLC represents an important health issue worldwide. Although the incidence of smoking‐related lung cancers, such as squamous cell carcinoma or small cell lung cancer, is decreasing, ⁶ LUAD, which affects both individuals with and without a history of smoking, has become the most common histologic subtype of lung cancer worldwide. LUAD is frequently associated with oncogenic driver alterations such as EGFR mutations, particularly in the absence of a smoking history. Exposure to air pollution has been associated with an increased risk of LUADs, notably EGFR‐mutant NSCLC, ⁷ and might contribute to their high incidence. Diagnosis of lung cancer in young and never‐smoking patients, often thought to be at low risk of cancer, can be challenging and contributes to delayed diagnosis. Furthermore, despite improvements in survival with tyrosine kinase inhibitors (TKIs) and combination therapies and a better prognosis than that for nononcogene‐addicted NSCLC, EGFR‐mutant NSCLC remains a challenging disease, particularly in the metastatic setting, in which long‐term disease control remains elusive. Recognizing the heterogeneity of EGFR‐mutant NSCLC and the expanding number of treatment options, risk stratification and personalized strategies are relevant clinical questions.

In this review, we aim to provide a comprehensive clinical overview of the management of patients with EGFR‐mutant NSCLC. Unless explicitly stated otherwise, the studies and results discussed in this report apply to the common L858R and exon 19 deletions (ex19del) EGFR mutations. The treatment of insertion mutations in exon 20 (ex20ins) is not covered in this review because ex20ins represents a distinct disease subtype with unique biology, prognosis, and therapy response.

BIOLOGY OF EGFR‐MUTANT NSCLC

Functional consequences of EGFR kinase domain mutations

EGFR is a transmembrane protein localized at the cellular surface and a member of the ERBB family of tyrosine kinase receptors (Figure 1). EGFR is commonly expressed by various epithelial tissues, such as skin, lung, or gastrointestinal mucosa, in which it plays a role in regulating tissue regeneration and wound healing. In normal conditions, the EGFR protein is activated through binding of its ligand, mainly epidermal growth factor and TGF‐α. Ligand binding induces dimerization of the receptor and triggers activation of the intracellular kinase domain, which then activates intracellular pathways implicated in the survival, growth, and proliferation of the cell, such as the MAP kinase and PI3K/AKT/MTOR pathways (Figure 1). ⁸ EGFR mutations occurring in exons 18 through 21 often lead to constitutive activation of the tyrosine kinase domain.

FIGURE 1.

Structural and functional organization of the EGFR protein. EGFR exists as an inactive monomer embedded in the cell membrane. Upon ligand binding (e.g., epidermal growth factor), the receptor dimerizes, transitioning to an activated state. This activation triggers intracellular signaling cascades, such as MAPK, PI3K‐AKT‐MTOR, which regulate cellular proliferation, survival, and differentiation. Mutations in the kinase domain of EGFR can lead to a constitutive and ligand‐independent activation of the protein and downstream pathways. EGFR indicates epidermal growth factor receptor. Created in BioRender (Maxime Borgeaud, and Timothée Olivier, 2025; https://BioRender.com/s16f601).

EGFR mutation classification

Historically, the L858R point mutation in exon 21 and ex19del, which include the loss of amino acids E746 through A750, have been recognized as the most frequent EGFR mutations, exhibiting high sensitivity to EGFR TKIs. Consequently, most clinical trials primarily include patients with L858R and ex19del mutations. ⁹ , ¹⁰ , ¹¹ Therefore, they are referred to as the common EGFR mutations (Figure 2) and represent around 80% of cases. Ex20ins mutations, which are not the subject of this review, are the most frequent among the uncommon EGFR mutations. In addition to these alterations, many other mutations in the kinase domain are also oncogenic and exhibit diverse sensitivity to EGFR TKIs.

FIGURE 2.

EGFR tyrosine kinase domain mutations. The EGFR gene is located on chromosome 7 (7p11.2). Most frequent mutations within the tyrosine kinase domain in exons 18–21 are shown. Classical‐like mutations include deletions in exon 19, L858R mutations (both are usually referred to as common EGFR mutations), and other rare point mutations. PACC mutations include most other non‐ex20ins and T790M mutations. Insertions in exon 20 represent 8%–9% of EGFR mutations. Most frequent EGFR resistance mutations, acquired after exposure to EGFR TKIs, are shown. T790M represents a resistance mutation to first‐generation or second‐generation TKIs, whereas C797S, G724S, and L718Q represent resistance mutations to third‐generation EGFR TKIs. ex20ins indicates insertion mutation at exon 20; PACC, P‐loop and αC‐helix–compressing; TKIs, tyrosine kinase inhibitors; TM, transmembrane EGFR domain. Created in PowerPoint (Timothée Olivier, 2025).

Various classifications have been proposed regarding EGFR mutations, ¹² , ¹³ often based on sensitivity to EGFR TKIs, which are small molecules designed to inhibit the activated EGFR protein. One of the most comprehensive classifications, The University of Texas MD Anderson Cancer Center classification, ¹⁴ clusters different EGFR mutations by their structural impact on the EGFR kinase domain and sensitivity to different generations of EGFR TKIs into four distinct clusters: classical‐like mutations, P‐loop and αC‐helix–compressing (PACC) mutations, ex20ins mutations, and T790M‐like mutations. These four clusters affect differentially on the adenosine triphosphate (ATP) binding pocket of the kinase domain, which constitutes the binding site of EGFR TKIs. Classical‐like mutations represent mutations distant from the ATP binding pocket and do not affect the binding of EGFR TKIs, such as ex19del, L858R, and L861Q. PACC mutations consist of mutations leading to the compression of the region between the P‐loop and the αC‐helix of the EGFR protein, which results in modification of the ATP binding pocket, such as G719X and S768I. The second‐generation TKI afatinib exhibits higher inhibiting properties against PACC mutations than first‐generation or third‐generation TKIs in preclinical models. ¹⁴ T790M mutation is a well established gatekeeper resistance mutation that hinders the binding of first‐generation and second‐generation EGFR TKIs to the ATP pocket. This structure‐based classification represents a practical approach that may help in the selection of treatment for rare EGFR mutations based on the prediction of their structural impact.

Genomic landscape of EGFR‐mutant NSCLC

EGFR‐mutant and smoking‐related LUAD represent different diseases from a molecular perspective. EGFR‐mutant cancer cells are addicted to EGFR activation of the MAP kinase pathways and are very sensitive to EGFR inhibition. Activating EGFR mutations are usually mutually exclusive with other oncogenic driver alterations, such as KRAS mutations, at initial diagnosis. Concurrent activating driver mutations were found to exhibit synthetic lethality in preclinical models. ¹⁵ Although a rare event, the presence of other oncogenic drivers, such as PIK3CA co‐mutations, seem to be more frequent in patients with a smoking history and are associated with a poorer outcome on EGFR TKIs. ¹⁶ Moreover, patients with a higher tumor mutational burden—a feature that is also associated with smoking history and reflects heterogenous disease biology—derive less benefit from EGFR TKIs. ¹⁷ Conversely, new oncogenic alterations can arise under the selective pressure of EGFR TKIs as a mechanism of resistance to therapy. Unlike oncogenic driver genes, alterations in tumor‐suppressor genes are rather frequent in EGFR‐mutant NSCLC, with TP53 being the most common, occurring in 40%–50% of cases. ¹⁸ TP53 and other tumor‐suppressor genes are thought to favor the acquisition of copy number alterations and genomic instability in cancer cells, which can facilitate resistance to therapy. ¹⁹ Indeed, convincing evidence indicates that TP53 mutation and some other concomitant alterations in tumor‐suppressor genes, ²⁰ , ²¹ such as ARID1A or RB1, seem to confer a further worse prognosis. ¹⁹

In addition to somatic EGFR mutations, a T790M germline variant in EGFR was first reported in a European descendant family with several members who developed LUAD. ²² Although T790M is the most described germline EGFR variant, several other germline mutations have been identified, including G724S, T725M, R776H, V769M, V834L/I, and K757R. ²³ The prevalence of these germline mutations is probably very low in the general population and among people who have lung cancer, with a frequency of less than 1%. ²⁴ Most patients are female, of European ancestry, without a smoking history, and present with multiple, bilateral, ground‐glass pulmonary nodules. ²⁵ , ²⁶ Oxnard et al. reported that 95% of patients with a germline EGFR mutation have a concomitant somatic EGFR‐activating mutation as a second hit. ²⁵ The risk of developing lung cancer for people carrying a germline EGFR T790M mutation is estimated to exceed 50% before age 60 years, highlighting the need for tailored lung cancer screening in this population. ²⁵

Tumor immune microenvironment in EGFR‐mutant NSCLC

The tumor immune microenvironment (TME) in EGFR‐mutant NSCLC presents distinct characteristics compared with smoking‐related LUAD, with significant implications for immunotherapy strategies. ²⁷ Whereas preclinical models suggest that EGFR mutations could promote programmed death‐ligand 1 (PD‐L1) expression, ²⁸ clinical observations indicate that these mutations are often associated with lower PD‐L1 expression levels. ²⁹ This discrepancy may stem from the unique immunogenic profiles of EGFR‐mutant tumors, typically exhibiting a reduced tumor mutational burden, which is generally correlated with a diminished neoantigen load, leading to a less robust immune response to immune checkpoint inhibitors (ICIs). ³⁰ The TME in EGFR‐mutant NSCLC is generally immunosuppressive, with low CD8‐positive/PD‐L1‐positive tumor‐infiltrating lymphocytes; dysfunctional dendritic cells secondary to the upregulation of IL‐6/STAT3 signaling; higher regulatory T cells, which suppress antitumor immunity; the presence of myeloid‐derived suppressor cells; and predominantly M2 immune‐suppressive tumor‐associated macrophages. ²⁷ Moreover, EGFR‐mutant NSCLC tend to demonstrate an interferon‐γ low immune phenotype, further suggesting an immune‐suppressive TME. ³¹

MANAGEMENT OF PATIENTS WITH EGFR‐MUTANT NSCLC

The relation between EGFR kinase domain mutations and response to EGFR TKIs was first demonstrated in 2004 in two pivotal studies for patients with metastatic NSCLC. ³² , ³³ Three generations of EGFR TKIs are now available in clinical practice for the treatment of patients with EGFR‐mutant NSCLC. The first‐generation TKIs (erlotinib, gefitinib, and icotinib) are reversible EGFR TKIs that compete with ATP binding in the ATP‐binding pocket in the kinase domain. They also have activity on the wild‐type EGFR kinase domain, which is responsible for some of the adverse drug events. Afatinib and dacomitinib represent second‐generation TKIs that bind irreversibly to EGFR and also to some extent inhibit the HER‐2 and HER‐4 kinase domains and the wild‐type EGFR protein. Third‐generation TKIs osimertinib and lazertinib are highly potent, irreversible inhibitors that covalently bind to the C797 residue, ³⁴ , ³⁵ with several advantages over first‐generation and second‐generation drugs, including higher affinity for the mutant EGFR protein and less inhibition of the wild‐type protein, higher intracranial activity, and sustained activity against the T790M gatekeeper resistance mutation. EGFR TKIs have since become the mainstay of treatment of metastatic EGFR‐mutant NSCLC, as discussed below. In the last 5 years, EGFR TKIs have also entered the treatment landscape of nonmetastatic EGFR‐mutant NSCLC.

Management of patients with stage I–III EGFR‐mutant NSCLC

The current management of patients with nonmetastatic NSCLC depends on the resectability of the tumor. For resectable stage I–III tumors, upfront surgery represents the preferred strategy and remains the cornerstone of treatment in otherwise medically fit patients. Definitive chemoradiation is the preferred approach for nonresectable stage III tumors. The following section describes how EGFR TKIs have also become a standard of care in the management of patients with early stage EGFR‐mutant NSCLC as adjuvant therapy after local treatment of the primary tumor and lymph nodes.

Adjuvant tyrosine kinase inhibitors

Gefitinib was the first EGFR TKI evaluated in the adjuvant setting for EGFR‐mutant NSCLC. In the CTONG1104 phase 3 trial (ClinicalTrials.gov identifier NCT01405079), 2 years of adjuvant gefitinib improved median disease‐free survival (DFS) compared with adjuvant cisplatin plus vinorelbine (median DFS, 30.8 vs. 19.8 months; hazard ratio [HR], 0.56; 95% confidence interval [CI], 0.40–0.79). ³⁶ The updated analysis with a follow‐up of 80 months indicates a convergence of the DFS curves, with DFS at 3 and 5 years comparable between the two treatment arms. There was no difference in overall survival (OS) between treatments, likely because most patients in the chemotherapy arm received an EGFR TKI at the time of progression. ³⁷ Thereafter, multiple other randomized trials in the adjuvant setting ³⁸ , ³⁹ , ⁴⁰ involving first‐generation TKIs demonstrated similar findings that a DFS benefit was lost on longer follow‐up after TKI cessation, and no significant OS benefit was observed.

Osimertinib, a third‐generation TKI, is approved for patients who have stage IB–IIIA NSCLC with classical EGFR mutations after surgery, based on results of the phase 3 ADAURA trial (ClinicalTrials.gov identifier NCT02511106). ⁴¹ In this trial, patients with stage IB–IIIA disease according to the American Joint Committee on Cancer's AJCC Cancer Staging Manual, seventh edition (stage IB corresponding to tumor >3 cm without lymph node involvement), were randomly assigned to 3 years of adjuvant osimertinib or placebo, with optional use of adjuvant chemotherapy allowed in both arms. Osimertinib improved DFS in patients who had stage IB–IIIA disease (median DFS, not reached vs. 27.5 months; HR, 0.20; 99% CI, 0.14–0.30) and led to a significant absolute OS benefit of 12% at 5 years for those with stage II–IIIA disease (85% vs. 73%; HR, 0.49; 95% CI, 0.33–0.73). Patients with stage IIIA disease derived the greatest benefit from osimertinib in terms of DFS (HR, 0.12; 95% CI, 0.07–0.20), regardless of adjuvant chemotherapy use. ⁴² Osimertinib also was associated with a lower rate of central nervous system (CNS) relapse (CNS DFS in stage II–IIIA: HR, 0.24; 95% CI, 0.14–0.42). Notably, however, most CNS relapses in the osimertinib group occurred after the end of the treatment. ⁴³

One of the criticisms of ADAURA has been the suboptimal access to standard‐of‐care therapy at progression in the control group, with only 40% of patients receiving osimertinib at progression, thus potentially affecting OS. Because, in the updated analysis of ADAURA, the benefit of osimertinib seems to be reduced after TKI cessation, ⁴¹ a pertinent question is whether adjuvant TKI prevents rather than delays the onset of metastatic disease. However, a counter argument is whether this matters if we improve 5‐year OS rates by 12% with adjuvant osimertinib versus adjuvant chemotherapy, which achieves a less than 5% improvement. ⁴⁴ Although treatment duration was 3 years in ADAURA, it is currently unknown whether a longer duration would result in a greater benefit. The TARGET phase 2 single‐arm study (ClinicalTrials.gov identifier NCT05526755) is currently evaluating 5 years of adjuvant osimertinib. ⁴⁵ To that end, improving risk‐stratification strategies to adapt adjuvant treatment strategies is an important research priority. Whether the benefits of adjuvant treatment extend to people with stage IA lung cancer is the subject of the phase 3 ADAURA2 study (ClinicalTrials.gov identifier NCT05120349), which is currently evaluating 3 years of adjuvant osimertinib versus placebo. ^{46 , 44}

In summary, based on these data, we recommend adjuvant osimertinib for 3 years in patients with stage IB (>3 cm) to IIIA resected NSCLC harboring EGFR L858R or ex19del mutations after adjuvant platinum‐based chemotherapy and adjuvant osimertinib for those who are not eligible for chemotherapy. Although criticism regarding the access to osimertinib for people randomly assigned to placebo in the ADAURA trial is valid, in our opinion, it should not prevent the adoption of this recommendation.

Consolidation osimertinib after chemoradiation

For unresectable stage III NSCLC, concurrent chemoradiation represents the current standard of care. Maintenance durvalumab for 1 year has been the standard of care for stage III NSCLC with nonprogressive disease after definitive chemoradiation, based on the PACIFIC trial (ClinicalTrials.gov identifier NCT02125461). ⁴⁷ However, for the subset of patients with EGFR mutations included in PACIFIC, durvalumab did not appear to improve DFS in a post‐hoc analysis. ⁴⁸

The phase 3 LAURA trial (ClinicalTrials.gov identifier NCT03521154) evaluated consolidation therapy with osimertinib, which was given until patients developed intolerance or progression, for stage III EGFR‐mutant NSCLC after chemoradiation. ⁴⁹ The study showed an improvement in progression‐free survival (PFS) compared with placebo (median PFS, 39.1 vs. 5.6 months; HR, 0.16; 95% CI, 0.10–0.24). A trend toward improved OS was also reported in a subsequent follow‐up analysis. ⁵⁰ In addition, the CNS relapse rate was significantly lower for osimertinib (CNS PFS: HR, 0.17; 95% CI, 0.09–0.32). ⁵¹ People on the control arm did poorly, with only 22% of patients alive and without disease progression at 12 months after the initiation of placebo, highlighting that a large proportion of patients with stage III EGFR‐mutant NSCLC harbor micrometastatic disease. In this trial, nearly 50% of patients did not have baseline positron emission tomography/computed tomography staging, ⁴⁹ raising the possibility that people with occult stage IV disease were inappropriately included. A subgroup analysis was presented at the European Society of Medical Oncology congress subsequently with stratification by whether or not a baseline positron emission tomography/computed tomography scan was performed. In this analysis, there was no difference in PFS with osimertinib versus placebo (HR, 0.23 [95% CI, 0.14–0.40] and 0.24 [95% CI, 0.13–0.38], respectively). ⁴⁸ , ⁵¹ If anything, these data suggest that patients with stage III EGFR‐mutant NSCLC are at a very high risk of relapse. Despite the benefit observed in LAURA, long‐term treatment with osimertinib raises a few issues, such as increased treatment cost and long‐term toxicities, which would be of higher concern if OS were not significantly improved.

In summary, consolidation osimertinib improves PFS in unresectable stage III EGFR‐mutant NSCLC after chemoradiation and is preferred over durvalumab, which is not recommended in this setting. Future trials should explore escalation and de‐escalation treatment strategies.

Role of neoadjuvant EGFR tyrosine kinase inhibitors

The role of neoadjuvant TKI monotherapy for resectable EGFR‐mutant NSCLC has been explored in phase 2 studies. ⁵² Multiple single‐arm trials demonstrated that a neoadjuvant first‐generation EGFR TKI was generally safe, although modest pathologic responses were achieved. ⁵³ , ⁵⁴ The phase 2 EMERGING‐CTONG1103 study (ClinicalTrials.gov identifier NCT01407822) was the only randomized trial that enrolled patients with stage IIIA N2 EGFR‐mutant NSCLC and compared perioperative erlotinib (6 weeks and 12 months before and after surgery, respectively) versus perioperative gemcitabine and cisplatin (two cycles before and after surgery). There was no improvement in the objective response rate (ORR) or OS with erlotinib compared with chemotherapy, and the major pathologic response (MPR) rate was low in both groups (9.7% and 0% for erlotinib and chemotherapy, respectively). However, PFS, which was a secondary end point, was significantly improved with erlotinib (median PFS, 21.5 vs. 11.4 months; HR, 0.36; 95% CI, 0.21–0.61). ⁵⁵ Similarly, several phase 2 trials have evaluated neoadjuvant osimertinib, with rather low MPR and pathologic complete response rates ranging from 10.7% to 24% and from 0% to 3.6%, respectively. ⁵⁶ , ⁵⁷ , ⁵⁸ These findings suggest that EGFR TKIs might have limited capacity for tumor eradication in the neoadjuvant setting.

In the phase 3 NeoADAURA trial (ClinicalTrials.gov identifier NCT04351555), patients with resectable stage II–IIIB NSCLC harboring classical EGFR mutations were randomly assigned either to neoadjuvant osimertinib with or without chemotherapy or to chemotherapy alone. ⁵² , ⁵⁹ Of note, adjuvant use of osimertinib was offered to all trial participants and was received by greater than 80% of the patients in each arm. ⁵² The primary outcomes of MPR was significantly improved with osimertinib monotherapy (MPR, 25%; 95% CI, 17%–34%) and osimertinib chemotherapy (MPR, 26%; 95% CI, 18%–34%) versus chemotherapy alone (MPR, 2%; 95% CI, <1% to 6%). ⁵⁹ The pathologic complete response rate was less than 10% in both osimertinib groups. The selection of MPR as the primary end point in this trial is a subject of debate, given the uncertainty surrounding its validity as a surrogate for OS and long‐term disease control in the context of targeted therapies. ⁶⁰ Event‐free survival (EFS) also trended in favor of the osimertinib arms in the NeoADAURA trial; however, the data are still relatively immature to date (15% maturity). Until full maturation of EFS and OS data, the benefit of adding neoadjuvant osimertinib compared with the current standard of care of surgery followed by adjuvant chemotherapy and osimertinib remains uncertain.

Role of immune checkpoint inhibitors in the management of patients with stage I–III disease

There is currently no evidence supporting the use of neoadjuvant, adjuvant or perioperative ICIs in EGFR‐mutant NSCLC. The IMpower010 and Pearls/KEYNOTE‐091 studies (ClinicalTrials.gov identifiers NCT02486718 and NCT02504372, respectively), which evaluated adjuvant atezolizumab and pembrolizumab, respectively, included a limited number of patients with EGFR‐mutant NSCLC. ⁶¹ , ⁶² Within these small subgroups, a positive EFS trend was seen for adjuvant ICIs, with an HR for DFS of 0.44 (95% CI, 0.23–0.84) in the Pearls/KEYNOTE‐091 study ⁶¹ and 0.57 (95% CI, 0.26–1.24) for patients who had EGFR‐mutant disease with a PD‐L1 level ≥1% in IMpower010. ⁶² The phase 3 trials of ICIs in the perioperative setting either excluded ⁶³ , ⁶⁴ , ⁶⁵ or included only a very limited number of patients with EGFR‐mutant NSCLC. ⁶⁶ , ⁶⁷ In KEYNOTE‐671 (ClinicalTrials.gov identifier NCT03425643), the subgroup with EGFR‐mutant NSCLC seemed to derive significant EFS benefit from perioperative pembrolizumab ⁶⁶ ; whereas, in AEGEAN (ClinicalTrials.gov identifier NCT03800134), no benefit was seen with perioperative durvalumab. ⁶⁷ However, these small sample sizes preclude any conclusion; and, in addition to these results, all available data about immunotherapy in patients with EGFR‐mutant NSCLC do not support this idea. An important consideration about ICIs and EGFR‐mutant NSCLC is the increased risk of toxicities for patients receiving osimertinib after ICIs. Several studies have indeed reported an increased risk of pneumonitis and hepatitis. ⁶⁸ Therefore, we do not recommend the use of neoadjuvant or perioperative ICIs in patients with early stage EGFR‐mutant NSCLC.

Metastatic recurrence during or after adjuvant osimertinib

Recurrence occurring during adjuvant osimertinib will often indicate resistance to osimertinib, whereas recurrence after the end of adjuvant treatment can indicate either resistance or disease flare after cessation of therapy. In all patients, a new biopsy, if feasible, is recommended to evaluate resistance mechanisms to osimertinib. The management of these patients depends on the timing of the recurrence regarding adjuvant treatment, the site and extent of recurrence, and the presence or not of resistance mechanisms. For recurrence occurring after adjuvant osimertinib, without a resistance mechanism on re‐biopsy, we would consider osimertinib rechallenge as a potential option (see Management of patients with metastatic EGFR‐mutant NSCLC, below), although there are no prospective studies supporting this approach. Patients who develop disease recurrence while on adjuvant osimertinib will usually be managed the same as those who have disease progression in the metastatic setting, except in special circumstances, such as oligoprogression amenable to local ablative therapy, in which osimertinib continuation could be considered (see Treatment options at progression after EGFR TKI, below).

Management of patients with metastatic EGFR‐mutant NSCLC

First‐line treatment

EGFR TKI monotherapy

EGFR TKIs have replaced chemotherapy as the first‐line treatment of EGFR‐mutant NSCLC since 2009, when it was demonstrated that gefitinib improved ORR and PFS compared with platinum‐based chemotherapy in the IPASS trial (ClinicalTrials.gov identifier NCT00322452). ¹⁰ Similar improvements in ORR and PFS were observed for other first‐generation EGFR TKIs (Figure 3). ⁶⁹ , ⁷⁰ The second‐generation TKIs afatinib and dacomitinib have both been shown to improve PFS compared with first‐generation TKIs. ¹¹ , ⁷¹ Third generation osimertinib was first approved for the treatment of patients whose disease progressed after a first‐generation TKI treatment, with a T790M EGFR mutation as a resistance mechanism. ⁷² The safety profile of osimertinib was more favorable than that of first‐generation TKIs, with less frequent skin rash and diarrhea. Osimertinib was then tested against first‐generation TKIs in the first‐line metastatic setting in the FLAURA trial (ClinicalTrials.gov identifier NCT02296125), in which it demonstrated an improvement in PFS (median PFS, 18.9 vs. 10.2 months; HR, 0.46; 95% CI, 0.37–0.57) and OS (median OS, 38.6 vs. 31.8 months; HR, 0.80; 95% CI, 0.64–1.00). ⁹ Since then, osimertinib has been the standard of care for the first‐line treatment of metastatic EGFR‐mutant NSCLC. Lazertinib demonstrated a comparable PFS benefit compared with first‐generation TKIs in the LASER301 trial (ClinicalTrials.gov identifier NCT04248829). ⁷³ Aumolertinib, furmonertinib, and befotertinib represent other third‐generation TKIs that were developed in China and have also demonstrated prolonged PFS compared with first‐generation TKIs. ⁷⁴ , ⁷⁵ , ⁷⁶

FIGURE 3.

Key clinical data from trials in the first‐line metastatic setting of EGFR‐mutant NSCLC. *The IPASS trial included 10 patients with uncommon EGFR mutations. **The LASER301 trial only reported treatment‐related grade 3 or greater adverse events. AE indicates adverse events; Chemo, chemotherapy; ex20ins, insertion mutation at exon 20; Mo, months; NR, not reached; NSCLC, nonsmall cell lung cancer. Created in PowerPoint (Timothée Olivier, 2025).

In the FLAURA trial, approximately 40% of patients whose cancer progressed on study treatment did not receive further systemic treatment, and 43% of patients whose cancer progressed on a first‐generation TKI received osimertinib at progression. ⁷⁷ This raises the question of the feasibility of a sequential approach (a first‐generation TKI followed by osimertinib at progression in patients with T790M mutation) and of the advantage of upfront osimertinib. This question was evaluated in the phase 2 APPLE trial (ClinicalTrials.gov identifier NCT02856893). ⁷⁸ No differences in OS was observed between the two groups, but patients in the sequential approach arm had higher rates of brain progression, and no benefit was observed for the sequential approach in terms of toxicities.

Combination strategies

Despite the high efficacy of EGFR TKIs, treatment resistance invariably develops. Thus several trials have evaluated various combination strategies with TKIs in the first‐line setting (Figure 3).

The NEJ009 trial (Japanese University Hospital Medical Information Network [UMIN] Clinical Trials Registry identifier UMIN000006340) and a study by Noronha et al. demonstrated that the addition of from four to six cycles of chemotherapy to gefitinib improved PFS and PFS2 (defined as the time from randomization to progression on both platinum‐based therapy and gefitinib), at the cost of increased toxicity. ⁷⁹ , ⁸⁰ An OS benefit was also seen in both trials at the time of the initial report. However, in an updated analysis of the NEJ009 trial with longer follow‐up, the difference in OS was no longer statistically significant. ⁸¹

FLAURA2 (ClinicalTrials.gov identifier NCT04035486) was a global trial that evaluated a similar strategy of combined TKI plus chemotherapy using osimertinib with or without concomitant carboplatin plus pemetrexed for four cycles followed by maintenance pemetrexed plus osimertinib versus osimertinib monotherapy. ⁸² Compared with osimertinib alone, its combination with chemotherapy led to a statistically significant improvement in PFS (median PFS, 25.5 vs. 16.7 months; HR, 0.62; 95% CI, 0.49–0.79). ⁸² However, the combination was also associated with more grade 3 or greater adverse events (64% vs. 27%, respectively). The increased toxicities mainly consist of the typical side effects of chemotherapy, such as hematologic adverse events. Although data are immature, a trend toward an OS benefit was observed in the secondary analysis. ⁷⁷ Interestingly, the combination led to an improvement in CNS PFS compared with osimertinib alone (HR, 0.58; 95% CI, 0.33–1.01). ⁸³ It is worth mentioning, however, that brain imaging was mandated only for patients with baseline brain metastasis and was performed only at the time of progression for patients without a history of brain involvement. In patients with baseline brain metastasis, the PFS was improved with the addition of chemotherapy (median PFS, 24.9 vs. 13.8 months; HR, 0.47; 95% CI, 0.33–0.66). This finding was unanticipated because chemotherapy supposedly has poor intracranial activity compared with third‐generation EGFR TKIs. Interestingly, although the addition of chemotherapy to osimertinib improved PFS by 8 months in the first‐line setting, in the second line, the median PFS observed with chemotherapy alone is only 4.2–4.4 months. ⁷² , ⁸⁴

The MARIPOSA trial (ClinicalTrials.gov identifier NCT04487080) evaluated the combination of lazertinib plus amivantamab, an EGFR‐MET–bispecific antibody, in the first‐line metastatic setting for NSCLC with common EGFR mutations. ⁸⁵ Patients were randomized in a 2:2:1 ratio to amivantamab plus lazertinib, osimertinib monotherapy, and lazertinib monotherapy. Compared with osimertinib monotherapy, amivantamab plus lazertinib improved PFS (median PFS, 23.7 vs. 16.6 months; HR, 0.70; 95% CI, 0.58–0.85). The PFS for lazertinib was similar to that observed for osimertinib.

In the updated analysis, amivantamab plus lazertinib demonstrated a significant survival benefit (median OS, not reached vs. 36.7 months; HR, 0.75; 95% CI, 0.61–0.92). ⁸⁶ Crossover to amivantamab (e.g., the MARIPOSA‐2 regimen [ClinicalTrials.gov identifier NCT04988295], as discussed below [see Treatment options at progression after EGFR TKI: Agnostic strategies]) at progression was not allowed in the control group. The combination strategy was associated with more toxicities, with grade 3 or greater adverse events in 75% versus 43% of patients treated with osimertinib; paronychia, rash, and infusion reaction were the most common adverse events. Thromboembolic events were also more frequent in the amivantamab plus lazertinib arm (37% vs. 9%). Most thromboembolic events occurred in the first 4 months of treatment, leading to the adoption of prophylactic anticoagulation for the first 4 months in the ongoing trials of the combination, as also stated in the US Food and Drug Administration [FDA] approval of this combination. Interestingly, the combination also seemed to be superior to osimertinib monotherapy in high‐risk subgroups, such as patients with CNS or liver metastasis, TP53 co‐mutation or the persistence of detectable circulating tumor DNA (ctDNA) after two treatment cycles. ⁸⁷ Of note, serial brain magnetic resonance imaging was mandated in MARIPOSA: every 8 weeks for the first 30 months and 12 weeks thereafter for patients with baseline brain metastases, and every 24 weeks for patients without brain metastasis. This imaging schedule might have led to earlier identification of brain metastasis in MARIPOSA compared with other trials that included less frequent brain monitoring. In the phase 3 PALOMA trial (ClinicalTrials.gov identifier NCT05388669), subcutaneous amivantamab, in combination with lazertinib, was noninferior to intravenous administration in terms of pharmacokinetics and ORR and was associated with lower rates of infusion reactions (13% vs. 66%) and thromboembolic events (9% vs. 14%). ⁸⁸ Of note, 81% of patients received prophylactic anticoagulation in this trial. Administration time was significantly shorter with the subcutaneous route: approximately 5 minutes versus 5 hours for intravenous amivantamab. Patients' and health care providers' experience was improved overall with subcutaneous administration. The lower rate of infusion reactions and the convenience of administration potentially could allow the delivery of this treatment at home in the future. Interestingly, median duration of response and OS were superior for subcutaneous administration (HR for OS, 0.62; 95% CI, 0.42–0.92), suggesting a potential effect of the route of administration on treatment efficacy. A potential explanation for this difference is the interaction of amivantamab with subcutaneous immune cells. ⁸⁹

Regarding toxicity of amivantamab plus lazertinib, the phase 2 COCOON study (ClinicalTrials.gov identifier NCT06120140) demonstrated that a prophylactic intensive regimen (consisting of oral doxycycline followed by topical clindamycin, chlorhexidine wash, and skin moisturizers) reduced moderate‐to‐severe dermatologic reactions by almost 50% and increased treatment adherence. ⁹⁰ The use of dexamethasone premedication 8 mg twice daily for 2 days and 8 mg 1 hour before infusion also was shown to reduce infusion‐related reactions. ⁹¹

Several studies have evaluated the role of anti‐angiogenic agents in EGFR‐mutant NSCLC. Two phase 3 studies evaluated the addition of bevacizumab, a humanized monoclonal antibody targeting vascular endothelial growth factor (VEGF), to erlotinib in the first line treatment of advanced or metastatic EGFR‐mutant NSCLC. In both trials, compared with erlotinib alone, the combination of erlotinib and bevacizumab improved PFS (median PFS, 17.9 vs. 11.2 months [HR, 0.55; 95% CI, 0.41–0.73 ⁹² ] and 16.9 vs. 13.3 months [HR, 0.605; 95% CI, 0.417–0.877], respectively). ⁹³ A similar benefit was observed with the combination of erlotinib with ramucirumab, another anti‐VEGF antibody (median PFS, 19.4 vs. 12.4 months for erlotinib monotherapy; HR, 0.59; 95% CI, 0.46–0.76). ⁹⁴ Importantly, none of these trials demonstrated an OS benefit and also showed increased toxicity. The data regarding the combination of anti‐VEGF with third generation TKI in the first‐line metastatic setting are more limited, with three randomized phase 2 studies and no phase 3 trials. The WJOG9717L and OSIRAM trials (UMIN Clinical Trials Registry identifiers UMIN000030206 and UNIM2080224085) demonstrated no benefit in PFS from the addition of bevacizumab ⁹⁵ or ramucirumab ⁹⁶ to osimertinib. Conversely, in the RAMOSE ⁹⁷ trial (ClinicalTrials.gov identifier NCT03909334), the combination of osimertinib and ramucirumab improved PFS (median PFS, 24.8 vs. 15.6 months; HR, 0.55; 95% CI, 0.32–0.93) with no difference in OS and a higher rate of grade 3 or greater treatment‐related adverse events (53% vs. 41%). ⁹⁸ Taken together, these data do not support the addition of anti‐VEGF therapies to first‐line osimertinib, especially in light of FLAURA2 and MARIPOSA.

Clinical and molecular factors that may affect first‐line treatment decisions are discussed further below (see Metastatic disease first line—Risk stratification).

Consolidative local therapies

Local ablative treatments of residual disease in patients responding to first‐line EGFR TKIs represent a conceptually interesting strategy because they target potential drug‐persistent cells thus retarding the acquisition of resistance and progression. Moreover, EGFR‐mutant NSCLC frequently progresses at the site of primary disease. ⁹⁹ Two randomized trials evaluated the addition of radiotherapy to the site of residual disease in the oligometastatic setting. In the phase 2 SINDAS study (ClinicalTrials.gov identifier NCT02893332), patients with oligometastatic disease (five or less locations) were randomized to a first‐generation EGFR TKI with or without stereotactic radiotherapy on all disease sites. ¹⁰⁰ The addition of radiotherapy improved PFS (median PFS, 20.2 vs. 12.5 months; HR, 0.22; 95% CI, 0.17–0.46) and OS (median OS, 25.5 vs. 17.6 months; HR, 0.44; 95% CI, 0.28–0.68). This trial, however, has several limitations, such as the presence of mainly bone metastasis, the exclusion of brain metastasis, and a high screen‐failure rate, suggesting that the study involved a very specific and selected population. In another phase 3 study, treatment‐naive patients with oligometastatic disease were randomized to the first‐generation TKI icotinib with or without concurrent thoracic radiotherapy. ¹⁰¹ Radiotherapy to other metastatic sites was allowed at the physician's discretion. The patients receiving thoracic radiotherapy had an improved median PFS of 17.1 versus 10.6 months in the TKI‐alone group (HR, 0.57; p = .004) and improved OS (median OS, 34.4 vs. 26.2 months; HR, 0.62; p = .029). Thoracic radiotherapy was associated with a higher incidence of treatment‐related adverse events of grade 3 or greater (11.9% vs. 5.1%). Conversely, continuation of osimertinib during thoracic radiation therapy has been associated with higher rates of severe pneumonitis. ¹⁰² Therefore, although we suggest a temporary interruption of osimertinib during thoracic radiation, we also caution against a prolonged interruption given the risk of tumor flares or progression in other disease sites. Indeed, disease flares have been reported in up to 23% of patients upon TKI cessation, with a median time of 8 days. ¹⁰³

We suggest that, in oligometastatic disease, radiotherapy to residual sites of disease can be considered in selected patients who respond to first‐line EGFR TKIs. Currently available evidence does not inform the optimal duration of response or depth of response before consolidative therapy can be offered.

Treatment of patients with NSCLC who have uncommon (non‐ex20ins) kinase domain EGFR mutations

Uncommon EGFR mutations, representing activating mutations of the kinase domain other than L858R, ex19del, and ex20ins, represent a challenge. Taken altogether, uncommon mutations generally exhibit lower response rates and PFS with EGFR TKIs. ¹⁴ They represent a very heterogenous group, with variable sensitivity to EGFR TKIs. Because of their rarity, uncommon EGFR mutations were excluded from the ADAURA and NeoADAURA (ClinicalTrials.gov identifier NCT04351555) trials, resulting in a lack of prospective studies exploring the role of perioperative TKIs for individuals with NSCLC harboring these mutations. In the metastatic setting, few prospective trials have been conducted for uncommon EGFR mutations. A post‐hoc analysis of the LUX‐Lung trials (ClinicalTrials.gov identifiers NCT00525148, NCT00949650, and NCT01121393) with afatinib reported a response rate of 71.0% (95% CI, 54.1%–84.6%) and a PFS of 10.7 months (95% CI, 5.6–14.7 months) for all uncommon mutations altogether. ¹² The ACHILLES study/Thoracic Oncology Research Group study TORG1834 (UMIN Clinical Trials Registry identifier UMIN031180175) is the only randomized trial specifically evaluating a EGFR TKI versus platinum‐pemetrexed chemotherapy in uncommon EGFR mutations. ¹⁰⁴ Patients were randomized to receive either afatinib (n = 73) or chemotherapy (n = 36), and the median PFS was 10.6 months with afatinib versus 5.7 months with chemotherapy (HR, 0.422; 95% CI, 0.256–0.694). The overall response rate was also superior with afatinib, at 61.4% versus 47.1% with chemotherapy. At the time of this writing, OS data had not been reported. Several single‐arm studies have evaluated third‐generation TKIs in this setting, with response rates of 50%–55% and a median PFS of 8.2–9.4 months. ¹⁰⁵ , ¹⁰⁶ The combination of amivantamab plus lazertinib has demonstrated promising activity in uncommon EGFR mutations in the phase 1/1B CHRYSALIS‐2 study (ClinicalTrials.gov identifier NCT04077463), with a 55% response rate (95% CI, 40%–69%) in treatment‐naive patients, with median duration of response not estimable (NE; 95% CI, 9.9 months to NE) and a promising median PFS of 19.5 months (95% CI, 11.0 months to NE). ¹⁰⁷

Although the trials described above aggregated all uncommon mutations into a single group, it is now well known that sensitivity to EGFR TKIs depends on the specific mutations. Here, The University of Texas MD Anderson Cancer Center classification described above, which offers a classification of EGFR mutations based on their structural impact and their sensitivity to different generations of TKI, could offer some guidance. PACC mutations are predicted in preclinical models to be more sensitive to second‐generation TKIs, such as afatinib, and less sensitive to third‐generation osimertinib. ¹⁴ However, in patients, the activity of osimertinib is not negligible even for PACC mutations, with response rate of approximately 45% in mostly retrospective data. ¹⁰⁸ For classical‐like mutations, such as L861Q, in which the structural effect is predicted to be comparable to that of L858R and ex19del mutations, the use of osimertinib seems reasonable. Novel TKIs are also in development for uncommon mutations, such as furmonertinib, a third‐generation EGFR TKI, which led to an ORR of 81.8% (N = 22; 95% CI, 59.7%–94.8%) in patients with NSCLC harboring EGFR PACC mutations in an ongoing phase 1/2 trial. ¹⁰⁹

Currently, afatinib is the only guideline‐recommended and FDA‐European Medicines Agency–approved drug for the treatment of patients with uncommon EGFR mutations. For uncommon EGFR mutations, we suggest the use of afatinib for patients harboring uncommon PACC mutations. For classical‐like mutations, such as L861Q, as well as for PACC mutations in some circumstances, such as patients with brain metastasis, we recommend osimertinib if available because of its CNS activity.

Mechanisms of resistance to EGFR TKI therapy

Understanding these mechanisms of resistance is of paramount importance to identify personalized treatment strategies in subsequent lines (Figure 4A). A repeat tumor tissue biopsy at the time of resistance offers broad possibilities of analysis, such as the characterization of additional somatic tumor genomic changes, tumor histology, and the quantification of protein expression levels on the cell surface. It allows for the detection of resistance mechanisms in approximately 50% of patients who progress on osimertinib. ¹¹⁰ Detection of ctDNA by liquid biopsy in plasma allows for the detection of acquired genomic alterations as mechanisms of resistance to targeted treatment. ¹¹⁰ , ¹¹¹ Liquid biopsy, however, has several limitations, such as nonshedding tumors, inability to detect overexpression of targets like HER2 and MET, and inability to detect histologic transformation. Therefore, the reliability of ctDNA testing should be questioned for cases in which no genomic alteration is retrieved. Modern ctDNA techniques can now determine the proportion of tumor cell DNA in a blood sample (tumor fraction) by using various approaches in combination (the presence of aneuploidy, variant allele frequency, and the presence of canonical alterations). In patients with a high tumor fraction in liquid biopsy, a negative result carries a good negative predictive value, minimizing the need for repeat tissue biopsy to identify an actionable target. ¹¹² Some methods using epigenomic ctDNA profiling might also allow for the detection of small cell transformation from liquid biopsy. ¹¹³

FIGURE 4.

(A, B) This figure illustrates key resistance mechanisms to third‐generation TKIs in EGFR‐mutant NSCLC. On‐target resistance occurs through secondary EGFR mutations, such as C797S and G724S, or through EGFR amplification. Cell cycle alterations, including the loss of CDKN2A/B or amplification of CDK4/6 and CCND1, drive uncontrolled proliferation. Resistance can also arise from bypass pathway activation, such as MET or HER2 amplification, or oncogenic mutations in KRAS, BRAF, or PIK3CA. In addition, histologic transformation enables tumor cells to evade therapy by transforming into small cell or squamous cell carcinoma. ctDNA+ indicates circulating tumor DNA‐positive; NSCLC, nonsmall cell lung cancer; TKIs, tyrosine kinase inhibitors. Created in BioRender (Maxime Borgeaud and Timothée Olivier, 2025; https://BioRender.com/s16f601).

The mechanisms of resistance to targeted therapy often consist in the reactivation of specific pathways arising from on‐target (EGFR secondary alterations) or off‐target alterations. Nongenomic acquisition of resistance may occur through transcriptomic changes arising from epigenetic modifications, such as histologic transformation. The paragraphs below describe these mechanisms of resistance.

Acquired on‐target genomic alterations

Secondary mutations of EGFR represent a common mechanism of resistance to EGFR TKIs. The most common mechanisms of resistance to first‐generation and second‐generation TKIs is the acquisition of the T790M mutations that occur in greater than 50% of patients, which increases the affinity for ATP and reduces the binding of TKIs. ¹¹⁴ Resistance to osimertinib can arise secondary to several point mutations, such as C797 (C797S and C797G), G724S, ¹¹⁵ L718 (L718Q and L718V), ¹¹⁶ , ¹¹⁷ and L792 (L792F and L792H). ¹¹⁸ The C797S mutation is the most common because osimertinib forms a covalent bond precisely with the C797 residue. EGFR amplification of the wide‐type allele is also a well known resistance mechanism, leading to increased EGFR signaling, despite the presence of a TKI.

Acquired off‐target genomic alterations

Amplification of MET is the most common mechanism of resistance to osimertinib monotherapy, representing approximately 15% of cases when considering gene amplification ¹¹⁹ and up to 34% of cases when considering MET overexpression by immunohistochemistry (IHC). ¹²⁰ Various other genomic alterations conferring resistance to osimertinib have been reported with a frequency of less than 5%, including amplifications of HER2, HER3, or KRAS; activating mutations of BRAF (V600E), KRAS, or PIK3CA; and ALK, RET, or BRAF fusions. ¹¹⁰ , ¹²⁰ In the detection of MET amplification, the gold standard is fluorescence in situ hybridization (FISH) on tissue biopsy. Next‐generation sequencing (NGS) on liquid biopsy can also detect MET amplification, especially in cases of high focal amplification, as illustrated in the INSIGHT‐2 trial (ClinicalTrials.gov identifier NCT03940703), ¹²¹ in which 38% of patients who had MET amplification by FISH on tissue biopsy had detectable MET amplification on liquid biopsy NGS. Patients who had MET amplification detected on liquid biopsy NGS had mostly focal MET amplification and high gene copy numbers.

Resistance mechanisms may vary with the use of combination strategies. Based on ctDNA analysis, the acquisition of secondary EGFR mutations and MET amplifications were less frequent with amivantamab plus lazertinib than with osimertinib. ¹²² In FLAURA‐2, no major difference in the pattern of resistance mechanism was observed between osimertinib and osimertinib plus chemotherapy, but acquired genomic alterations were detected less frequently with disease progression after osimertinib plus chemotherapy. ¹²³

Histologic transformation

Histologic transformation is identified in 10%–15% of patients with EGFR‐mutant NSCLC who progress on EGFR TKIs, mostly among those small cell carcinoma morphology. ¹¹⁰ , ¹²⁴ The occurrence of small cell transformation has been suspected to be more prevalent with the use of third‐generation TKIs. ¹²⁴ Evidence suggests that specific genomic alterations, such as RB1 or TP53, could be associated with small cell transformation. ¹²⁵ , ¹²⁶ Beyond genomic alterations, multi‐omics analyses indicate that small cell transformation is primarily driven by transcriptional changes through epigenetic reprogramming, ¹²⁷ , ¹²⁸ which may be favored in particular genomic contexts, including the loss of the 3p chromosome and possibly TP53 and RB1 silencing. Squamous cell transformation represents another rare histologic transformation and also is associated with significant genomic changes, such as acquired PIK3CA mutations, chromosome 3q amplification, or FGF amplification. ¹²⁴ Evidence regarding histologic transformation after combination strategies is scarce because the reports on resistance mechanisms in the FLAURA2 and MARIPOSA trials are based on ctDNA analysis. Interestingly, TP53 and RB1 mutations, which sometimes are associated with histologic transformation, were also slightly less frequent after amivantamab plus lazertinib. ¹²²

Drug‐tolerant persister cells

Recent studies have highlighted the presence of drug‐tolerant persister cells in patients with EGFR‐mutant NSCLC, ¹²⁹ , ¹³⁰ characterized by reduced cell cycle activity and division and the ability to withstand treatment lethality. These can survive targeted therapies, contributing to disease relapse and progression. ¹³¹ , ¹³² , ¹³³ Evidence suggests that epigenetic mechanisms play a crucial role in the adaptation of these cells, enabling them to withstand therapeutic pressures. ¹³⁴ , ¹³⁵ Moreover, it also has been demonstrated that the TME plays a role in the emergence and promotion of drug persister cells, such as tumor‐associated fibroblasts. ¹³⁰

Understanding the interactions between epigenetic mechanisms and the TME is crucial for developing novel therapeutic strategies.

Treatment options at progression after EGFR TKI

Local therapies for oligoprogression

In patients who present with disease progression isolated to limited metastatic sites (usually from less than three to five sites), local ablative therapy offers a conceptually interesting approach. Local therapy may help to prolong the duration of current therapy and delay the need for a switch in systemic treatment by eradicating the resistant clones. Retrospective data support the use of radiotherapy in this setting, directed to oligo‐progressive sites followed by continuation of a EGFR TKI, with an improvement in PFS. ¹³⁶ In a single‐arm phase 2 trial, patients who had oligo‐progression on erlotinib were treated with radiotherapy and continuation of erlotinib, with a median PFS of 6 months (95% CI, 2.5–11.6 months) and a median OS of 29 months (95% CI, 21.7–36.3 months). ¹³⁷

In the setting of oligo‐progression, we suggest local ablative therapy, although no prospective data are available comparing different methods of local ablative therapy.

Biomarker‐selected treatment approaches

Among patients who present with secondary EGFR resistance mutations, treatment strategy differs according to the type of mutation. For patients treated with first‐generation or second‐generation TKIs, the acquisition of a T790M mutation should be treated with osimertinib. ⁷² Some data suggest that patients who present with secondary G724S mutations or L718Q/V mutations could respond to the second‐generation TKI afatinib, ¹³⁸ , ¹³⁹ whereas secondary C797S mutations in trans could be sensitive to first‐generation TKIs such as gefitinib. ¹⁴⁰ Fourth‐generation TKIs designed to block the C797S mutation are currently being assessed in early phase trials ¹⁴¹ but are not currently available.

In patients who present with MET amplification as a mechanism of resistance, the combinations osimertinib plus MET inhibitors have been tested in several phase 2 trials. INSIGHT‐2 evaluated the combination of osimertinib plus the MET inhibitor tepotinib, ¹²¹ whereas the SAVANNAH and ORCHARD trials (ClinicalTrials.gov identifiers NCT03778229 and NCT03944772) combined osimertinib and savolitinib. In the INSIGHT‐2 trial, among patients who had a MET amplification defined by FISH on tissue (gene copy number five or greater or an MET‐to‐CEP7 ratio ≥2.0) the ORR was 50% (95% CI, 39.7%–60.3%), with a median PFS of 5.6 months (95% CI, 4.2–8.1 months). In SAVANNAH and ORCHARD, the ORRs were 32% (95% CI, 26%–39%) and 41% (80% CI, 25%–59%), respectively. ¹²⁰ , ¹⁴² In the SAVANNAH trial, the response rate correlated with MET amplification or expression, with an ORR of 56% (95% CI, 39%–59%) in patients with high levels of expression (an IHC score of 3+ in ≥90% of tumor cells) or amplification (gene copy numbers ≥10), compared with 9% (95% CI, 4%–18%) in those with lower expression and amplification levels. ¹²⁰ , ¹⁴³ Conversely, in INSIGHT‐2, patients who had lower levels of MET amplification experienced only a slightly lower ORR compared with those who had higher levels of amplification (ORR 42.2%; 95% CI, 27.7%–57.8%) versus 56.6% (95% CI, 42.3%–70.2%) for patients with MET gene copy numbers <10 versus ≥10, respectively. Importantly, in INSIGHT‐2 and SAVANNAH, ¹⁴⁴ patients receiving tepotinib or savolitinib monotherapy exhibited a much lower ORR (tepotinib: ORR 8.3% [95% CI, 0.2%–38.5%]; savolitinib: ORR, 16% [95% CI, 5%–36%], respectively) highlighting the importance of maintaining EGFR blockade. More recently, the combination of savolitinib plus osimertinib also produced an improvement in PFS compared with chemotherapy among patients with EGFR‐mutant NSCLC who had MET amplification after a previous EGFR TKI, in the phase 3 SACHI trial (ClinicalTrials.gov identifier NCT05015608). ¹⁴⁵

Interestingly, the few patients who harbored concurrent bypass mechanisms of resistance in the INSIGHT‐2 trial (BRAF, ALK, RAS) exhibited no response to the combination, highlighting the importance of exhaustive and comprehensive molecular evaluation at the time of progression. The most common mechanisms of resistance to osimertinib plus tepotinib in INSIGHT‐2 consisted of an on‐target alteration of MET or EGFR (loss of MET amplification and secondary resistance mutations of the kinase domain of MET or EGFR C797 mutations). A phase 1 study of combined osimertinib plus telisotuzumab vedotin (a cMet‐targeting antibody–drug conjugate) was recently reported in patients with c‐MET–expressing tumors after progression on prior osimertinib, ¹⁴⁶ with an ORR of 58%.

Regarding the concomitant use of inhibitors of bypass resistance mechanisms and osimertinib, the evidence is scarce; therefore, the inclusion in biomarker‐selected clinical trials, ¹⁴⁷ , ¹⁴⁸ wherever available, should be the priority. Selpercatinib and pralsetinib have been combined with osimertinib in several cases reports and cases series for patients presenting an RET fusion as a mechanism of resistance. ¹⁴⁹ , ¹⁵⁰ , ¹⁵¹ The combination of osimertinib with alectinib or with BRAF/MEK inhibitors have been reported in the context of acquired ALK fusion or BRAF mutations, respectively. ¹⁵⁰ , ¹⁵² , ¹⁵³ , ¹⁵⁴

For amplification of the EGFR wild‐type allele, a potential approach could consist of the use of monoclonal EGFR antibodies or an earlier generation of TKIs that also portend activity on the wild‐type EGFR protein. However, because amivantamab also seems to be efficient in the case of EGFR amplification, we would suggest not using another strategy outside of a standard‐of‐care amivantamab plus chemotherapy regimen.

Preclinical data and a few case reports seem to suggest that capivasertib or CDK4/6 inhibitors, in combination with osimertinib, may overcome resistance driven by PI3K pathway or cell cycle gene alterations, respectively. ¹⁵⁵ , ¹⁵⁶

In HER‐2 amplification as an acquired resistance mechanism, the phase 1/2 TRAEMOS trial (ClinicalTrials.gov identifier NCT03784599) studying osimertinib plus trastuzumab emtansine revealed limited efficacy. ¹⁵⁷

Although none of the combination strategies discussed here are FDA approved, up to 45% of patients (Figure 4B) who have disease progression on osimertinib harbor alterations that are potentially actionable with FDA‐approved drugs.

Agnostic strategies

In the second‐line setting, platinum plus pemetrexed chemotherapy has long represented the standard of care. With median a PFS and OS after TKI therapy of 4–5 months ⁷² , ⁸⁴ and 14–15 months, respectively, ¹⁵⁸ , ¹⁵⁹ chemotherapy seems to demonstrate better efficacy in EGFR‐mutant compared versus nononcogene‐addicted NSCLC. The question of pursuing EGFR TKIs beyond progression in addition to chemotherapy—to target both TKI‐resistant and TKI‐sensitive clones and to maintain CNS protection—remains controversial. In the phase 3 IMPRESS trial (ClinicalTrials.gov identifier NCT01544179), continuing gefitinib along with platinum‐based doublet chemotherapy after disease progression on first‐line gefitinib did not improve PFS. ¹⁶⁰ The COMPEL trial (ClinicalTrials.gov identifier NCT04765059) is currently evaluating the efficacy of pursuing osimertinib along with chemotherapy in patients with EGFR‐mutant NSCLC who exhibit disease progression after first‐line osimertinib. ¹⁶¹

Several other agnostic strategies have been tested and compared with chemotherapy in the second‐line setting, independent of the mechanisms of resistance. The combination of amivantamab plus chemotherapy with or without lazertinib was evaluated in the MARIPOSA‐2 trial among patients with progressive disease after osimertinib. ⁸⁴ The primary end point of PFS was improved by the combination of amivantamab plus chemotherapy (median PFS, 6.3 vs. 4.2 months for chemotherapy alone; HR, 0.48; 95% CI, 0.36–0.64) and amivantamab plus chemotherapy plus lazertinib (median PFS, 8.3 vs. 4.2 months for chemotherapy alone; HR, 0.44; 95% CI, 0.35–0.56). The CNS PFS was also improved with amivantamab plus chemotherapy (median CNS PFS, 12.5 vs. 8.3 months for chemotherapy; HR, 0.55; 95% CI, 0.38–0.79) and amivantamab plus chemotherapy plus lazertinib (median CNS PFS, 12.8 vs. 8.3 months for chemotherapy; HR, 0.58; 95% CI, 0.44–0.78). In a biomarker analysis of the CHRYSALIS‐2 trial, which evaluated the efficacy of amivantamab plus lazertinib in the post‐osimertinib and chemotherapy‐naive setting, MET overexpression, defined as an IHC score of 3+ (staining of ≥25% cancer cells), was predictive of a response to the combination, with an ORR of 61% (95% CI, 41%–78%) for MET‐positive tumors versus 12% (95% CI, 5%–25%) for MET‐negative tumors. ¹⁶² In this analysis, only one patient exhibited MET amplification on ctDNA and also had MET overexpression on tissue IHC. An analysis of outcomes in MARIPOSA‐2 based on osimertinib‐resistance mechanisms indicated that amivantamab plus chemotherapy improved median PFS regardless of the resistance mechanism. ¹⁶³

Toxicity was increased by the addition of amivantamab to chemotherapy and increased further by the addition of lazertinib, with grade 3 or greater treatment‐emergent adverse events in 72% of patients who received amivantamab plus chemotherapy, 92% in those who received amivantamab plus lazertinib plus chemotherapy, and 48% in those who received chemotherapy. ⁸⁴ Discontinuation of one of the study agents because of toxicities occurred in 18% and 34% of patients who received amivantamab plus chemotherapy and amivantamab plus chemotherapy plus lazertinib, respectively, compared with 4% of those who received chemotherapy alone. Hematologic toxicities were predominant and notably increased by the addition of amivantamab and lazertinib. Given the higher toxicity observed with the addition of lazertinib to amivantamab plus chemotherapy and fairly comparable efficacy parameters between amivantamab plus chemotherapy and lazertinib plus amivantamab plus chemotherapy, the FDA approved amivantamab plus chemotherapy for patients who progress after treatment with osimertinib. ¹⁶⁴

Numerous trials have evaluated the role of ICIs in combination with chemotherapy ¹⁵⁹ , ¹⁶⁰ or with chemotherapy and an anti‐VEGF ¹⁶⁵ , ¹⁶⁶ , ¹⁶⁷ for patients with EGFR‐mutant NSCLC who have progressive disease while receiving a TKI. The CheckMate 722 and KEYNOTE‐789 trials (ClinicalTrials.gov identifiers NCT02864251 and NCT03515837, respectively), which evaluated chemotherapy with nivolumab and pembrolizumab, respectively, were negative. ¹⁵⁸ , ¹⁵⁹ Several trials evaluating the combination of an ICI, chemotherapy, and an anti‐VEGF reported a modest improvement in PFS for ICI plus anti‐VEGF plus chemotherapy combinations and no improvement in OS. ¹⁶⁵ , ¹⁶⁶ In the IMpower150 phase 3 trial (ClinicalTrials.gov identifier NCT02366143), the combination of platinum‐based chemotherapy, bevacizumab and atezolizumab demonstrated an improvement PFS and OS compared with chemotherapy plus bevacizumab in an exploratory analysis of the small subgroup of patients with EGFR mutations. ¹⁶⁷ These results were not reproduced in EGFR‐mutant NSCLC in the phase 3 IMpower151 trial (ClinicalTrials.gov identifier NCT04194203) evaluating the same strategy. ¹⁶⁸ Several meta‐analyses confirm a modest gain in PFS of ICI plus chemotherapy and ICI plus anti‐VEGF plus chemotherapy versus chemotherapy alone, with higher toxicity rates and no definitive benefit in terms of OS. ¹⁶⁹ , ¹⁷⁰

Ivonescimab, a programmed cell death protein 1 (PD‐1)/VEGF–bispecific antibody, was evaluated in combination with chemotherapy in the phase 3 HARMONi‐A trial (ClinicalTrials.gov identifier NCT05184712) for patients with EGFR‐mutant NSCLC who progressed on a previous TKI. ¹⁷¹ Ivonescimab plus chemotherapy improved PFS compared with placebo plus chemotherapy (median PFS, 7.1 vs. 4.8 months; HR, 0.46; 95% CI, 0.34–0.62). OS data are not yet mature.

In the trials that included an anti‐VEGF agent, the relative contribution of the anti–PD‐1/PD‐L1 component in the small PFS gain remains uncertain because all of these trials (except IMpower150) lacked an anti‐VEGF plus chemotherapy arm.

Because several randomized phase 3 trials have reported no benefit in terms of OS and only a modest benefit in terms of PFS, mostly for ICI plus anti‐VEGF plus chemotherapy combinations, at the moment, we do not suggest the use of ICI‐based combinations for EGFR‐mutant NSCLC.

Several antibody–drug conjugates are also being assessed in EGFR‐mutant NSCLC. Datopotamab‐deruxtecan (dato‐DXd), an antitrophoblast cell surface antigen 2 (TROP2) with a topoisomerase I inhibitor payload, was evaluated in the TROPION‐Lung05 phase 2 trial (ClinicalTrials.gov identifier NCT04484142) in 137 previously treated patients with actionable genomic alterations. ¹⁷² Dato‐DXd resulted in an ORR of 43.6% (95% CI, 32.4%–55.3%) in 78 patients with EGFR mutations. The median PFS was 5.8 months, and grade 3 or greater adverse events occurred in 47% of patients. Moreover, the phase 3 TROPION‐Lung01 trial (ClinicalTrials.gov identifier NCT04656652), which compared dato‐DXd versus docetaxel in the second line, included a subgroup of patients with actionable genomic alterations. ¹⁷³ These patients, who harbored mainly EGFR mutations, achieved better PFS on dato‐DXd, with an HR of 0.38 (95% CI, 0.22–0.65). ¹⁷³ The combination of dato‐DXd plus osimertinib in patients with disease progression on osimertinib showed promising results in the ORCHARD phase 2 trial. ¹⁷⁴

Patritumab deruxtecan, an anti‐HER3 with a topoisomerase I inhibitor payload, demonstrated an interesting ORR in an early phase trial ¹⁷⁵ but achieved only a modest improvement in PFS compared with chemotherapy in the second‐line setting after an EGFR TKI in the phase 3 HERTHENA‐Lung02 trial (ClinicalTrials.gov identifier NCT04619004; median PFS, 5.8 vs. 5.4 months; HR, 0.77; 95% CI, 0.63–0.94). ¹⁷⁶ Other agents under investigation include sacituzumab tirumotecan, an anti‐TROP2 agent with a topoisomerase I inhibitor payload. ¹⁷⁷ In the OptiTROP‐Lung03 study (ClinicalTrials.gov identifier NCT05631262), sacituzumab tirumotecan demonstrated an improvement in the ORR and in PFS compared with docetaxel in the third‐line setting after a third‐generation TKI and platinum‐based chemotherapy (ORR, 45.1% vs. 15.6%; median PFS, 6.9 vs. 2.8 months; HR, 0.30; 95% CI, 0.20–0.46). ¹⁷⁸

Small cell carcinoma transformation

Small cell carcinoma transformation is associated with a change in disease phenotype, with a more aggressive course and a poor prognosis. ¹²⁵ , ¹⁷⁹ The standard treatment consists of platinum plus etoposide treatment, but the median OS is usually less than 1 year in this situation. ¹⁷⁹ Although it is routine to continue EGFR inhibition in cases of other acquired resistance mechanisms, translational studies suggest that, in the context of small cell transformation, even if EGFR mutation is conserved, EGFR expression tends to be downregulated, calling into question the utility of continued EGFR blockade. ¹⁸⁰ No prospective trial has specifically tested this strategy in small cell transformation, but some retrospective data suggest that continuing an EGFR TKI during chemotherapy could be associated with better PFS in this setting, potentially treating persisting, EGFR‐driven clones. ¹⁸¹ , ¹⁸² Another question is the potential benefit of ICIs in combination chemotherapy, as they lead to a small but significant increment of OS in de‐novo small cell lung cancer ¹⁸³ , ¹⁸⁴ but have limited efficacy in EGFR‐mutant NSCLC, as discussed above. Retrospective studies suggest that an ICI plus chemotherapy combination could improve PFS or OS compared with chemotherapy alone. ¹⁸¹ , ¹⁸⁵ Such small retrospective studies carry a high risk of bias, and larger cohorts or prospective trials are needed to validate this strategy. Caution should be taken when ICI therapy is considered concomitantly with osimertinib because severe grade 3–5 pulmonary adverse events have been reported with combining these agents. ¹⁸⁶ Therefore, this is not a strategy we recommend for patients who have small cell carcinoma transformation.

Future directions and current challenges in personalizing EGFR‐directed therapies

With the advent of these numerous intensification strategies, which improve PFS or DFS, but at the cost of increased toxicity, it is critical to develop risk‐stratified approaches from early to advanced stages to avoid overtreating patients with a good prognosis while improving outcomes for poor‐risk patients.

Early stage and molecular residual disease

In the adjuvant setting, it is imperative to identify patients at higher risk of relapse who potentially could benefit more from prolonged adjuvant treatment with an EGFR TKI. Several trials have shown promise for ctDNA in NSCLC in the adjuvant setting, ¹⁷⁴ including EGFR‐mutant tumors. ¹⁸⁷ The half‐life of tumor ctDNA in the blood is relatively short, typically less than 2.5 hours, making it a dynamic biomarker for the monitoring of treatment response. ¹⁸⁸ In an analysis of the ADAURA trial, a tumor‐informed ctDNA panel identified molecular residual disease (MRD) with a median lead time of 4.7 months ahead of DFS events in both arms.¹⁹⁰ However, in the treatment arm, the majority of MRD events occurred after the end of osimertinib treatment. This underlies the finding that, in the adjuvant setting, efficient targeted therapy can either suppress or decrease ctDNA levels below the detection threshold of current techniques. Although this suggests that the use of MRD detection potentially could be interesting before the introduction of adjuvant targeted treatment, in the study by Herbst et al., only a very small proportion of patients who further relapsed were MRD‐positive at that time. ¹⁸⁹ Furthermore, approximately 30% of patients who relapsed were not previously positive for MRD. Better technology with improved sensitivity is needed before MRD‐informed approaches can routinely be used for clinical decision making. ¹⁹⁰

Metastatic disease first line—Risk stratification

In the metastatic setting, combination strategies have demonstrated advantages in terms of PFS, and the MARIPOSA regimen, a combination of amivantamab plus lazertinib, led to a statistically significant and clinically meaningful reduction in mortality compared with osimertinib (HR, 0.75; 95% CI, 0.61–0.92). At 3 years, 60% of participants were alive in the experimental arm compared with 51% in the osimertinib arm. ⁸⁶ After these OS results, the MARIPOSA regimen thus represents a standard first‐line option. However, EGFR‐mutant NSCLC can sometimes follow a long course with prolonged disease control on a single TKI. Given the increased toxicities of combination approaches, we believe that some patients might still benefit from a sequential approach with osimertinib in the first‐line setting. The sequential approach offers the convenience of oral treatment for a prolonged period and minimizes toxicities. The key challenge lies in distinguishing patients who could be offered TKI monotherapy from those who would require a treatment‐intensification approach. Several clinical or molecular poor prognostic factors may help to guide management in this scenario (Figure 5). ⁸⁷ , ¹⁹¹ , ¹⁹² , ¹⁹³

FIGURE 5.

Clinical and molecular poor prognostic factors. PFS outcomes are shown for different treatments groups from the MARIPOSA and FLAURA2 trials. This figure combines subgroup and exploratory analyses from two separate clinical studies to provide a comprehensive overview; however, caution is warranted when performing cross‐trial comparisons. Ami‐laz indicates amivantamab plus lazertinib; HR, hazard ratio; NE, not estimable; Osi, osimertinib; Osi‐CT, osimertinib plus chemotherapy; PFS, progression‐free survival. Created in BioRender (Maxime Borgeaud, 2025; https://BioRender.com/s16f601).

The type of EGFR mutation is a well known prognostic factor. Patients harboring an L858R mutation in exon 21, for instance, have a worse prognosis on osimertinib monotherapy, with a median PFS of 12.4 months in the FLAURA trial compared with 21.4 months for patients with ex19del. ⁹ Although patients with L858R mutations also had a poorer prognosis than patients with ex19del in the FLAURA2 and MARIPOSA trials, ⁸² , ⁸⁵ both combination regimens improved PFS compared with osimertinib alone in the L858R subgroup. In FLAURA2, osimertinib plus chemotherapy extended PFS to 24.7 months versus 13.9 months for osimertinib alone in patients with L858R mutations. The HR for PFS was comparable between the L858R and ex19del subgroups (ex19del: HR, 0.6; L858R: HR, 0.63). In MARIPOSA, amivantamab plus lazertinib extended PFS to 18.4 months versus 14.8 months for osimertinib in patients with L858R mutations. The PFS benefit appeared greater for ex19del (HR, 0.65) than for L858R (HR, 0.78), and the same trend was observed for OS. ⁸⁶ Hence, the relative benefit of both combinations did not seem to be higher in patients with L858R mutations versus those with ex19del, even if the absolute PFS was longer with the combination, underpinning a prognostic rather than predictive role for the type of mutation with regard to combination strategies.

Patients with CNS metastasis represent another poor‐prognosis group. Both combination regimes in the FLAURA2 and MARIPOSA trials demonstrated an improvement in CNS PFS in patients with baseline brain metastasis. ⁸³ , ⁸⁶ The magnitude of benefit from combination therapy seemed greater in patients who had baseline CNS metastases in the FLAURA2 trial, but this trend was not observed in the MARIPOSA trial.

The combination of amivantamab plus lazertinib also improved PFS compared with osimertinib monotherapy in other high‐risk subgroups who are usually associated with a worse prognosis on osimertinib monotherapy. ⁸⁷ The combination improved PFS for patients harboring TP53 mutations (HR, 0.65; 95% CI, 0.48–0.87), patients without clearance of ctDNA after two cycles (HR, 0.64; 95% CI, 0.48–0.87), and patients with baseline liver metastasis (HR, 0.58; 95% CI, 0.37–0.91).

In an exploratory analysis of FLAURA2, the detection of baseline ctDNA was prognostic and appeared to be predictive of clinical benefit with the combination of osimertinib plus chemotherapy. ¹⁹³ Among those who had no detectable ctDNA at baseline, compared with single‐agent osimertinib, PFS was prolonged with the combination, although it did not reach statistical significance (median PFS, 33.3 vs. 30.3 months with osimertinib alone; HR, 0.93; 95% CI, 0.51–1.72). For patients who had detectable ctDNA at baseline, PFS was significantly prolonged with the combination (median PFS, 24.8 vs. 13.9 months; HR, 0.60; 95% CI, 0.45–0.80). The clearance of ctDNA at 3 weeks was associated with improved PFS compared with no clearance. However, the relative benefit of the combination versus osimertinib alone was similar regardless of whether there was clearance of ctDNA at 3 weeks, suggesting that ctDNA clearance might be prognostic, but not predictive, in this setting. Thus several trials are evaluating intensification strategies based on ctDNA guidance (ClinicalTrials.gov identifiers NCT04410796 and NCT06020989). For example, the PACE‐LUNG phase 2 trial is currently evaluating osimertinib followed by the addition of chemotherapy in patients who are positive for ctDNA at 3 weeks (EudraCT [European Union Drug Regulating Authorities Clinical Database] registration number 2019‐004757‐88).

These observations in high‐risk subgroups emerge purely from exploratory analyses. Nevertheless, they may inform patient discussions and contribute to shared decision making, particularly when assessing the potential benefits and risks of treatment‐intensification strategies. As technology improves, dynamic ctDNA monitoring may help to individualize treatment decisions in the future. Combinations of third‐generation EGFR TKIs with chemotherapy or with amivantamab are currently approved options.

CONCLUSION

In this review, we have discussed the 2‐decade‐long journey of discovery, identification, and targeting of EGFR mutations in lung cancer. Although these advances have improved survival and quality of life, patients with early stage disease still face a high risk of recurrence, and long‐term disease control remains a challenge in the metastatic setting. Progress is undeniable, but so is the work that remains. We are now entering a new era of precision oncology, driven by improvements in real‐time disease monitoring, ultrasensitive disease‐detection thresholds, and the development of more nuanced treatment strategies based on co‐alterations, tumor burden, and patient characteristics. We predict that novel therapeutics possibly allowing the complete inactivation of EGFR pathways and the ability to bypass mechanisms of resistance will make inroads in the EGFR realm, further improving outcomes in this disease.

Finally, even more for instances in which there are conflicting data or uncertain benefits, we encourage treating clinicians to consider patient comorbidities, financial challenges, preferences, and personal values to facilitate shared decision making and to ensure informed decisions.

CONFLICT OF INTEREST STATEMENT

Maxime Borgeaud reports financial support to attend scientific meetings for registration, travel, and accommodations from AstraZeneca, Bayer, Janssen Pharmaceuticals, Merck Sharp and Dohme (MSD), and PharmaMar outside the submitted work. Jair Bar reports institutional research funding from AbbVie, AstraZeneca/MedImmune, Immunai, MSD Novartis, Oncohost, Pfizer, Roche, and Takeda Pharmaceutical Company; personal/consulting fees from AstraZeneca, Bayer, BeiGene, Bristol Myers Squibb Company, Immunity Bio, Jazz Pharmaceuticals, Merck Serono, MSD Novartis, Regeneron, Rigel Pharmaceuticals, Roche, and Takeda Pharmaceutical Company; and travel support from Rigel Pharmaceuticals outside the submitted work. Stephanie Pei Li Saw reports grants from AstraZeneca; personal/consulting fees from AstraZeneca, Bayer, Boehringer Ingelheim, Daiichi Sankyo Company, Johnson & Johnson, and Pfizer; honoraria from AstraZeneca, Bristol‐Myers Squibb, Daiichi Sankyo Company, MSD, Pfizer, Roche, and Takeda Oncology;, meeting support from AstraZeneca, MSD, and Pfizer; and advisory board participation for AstraZeneca, Boehringer Ingelheim, Daiichi Sankyo Company, and Pfizer outside the submitted work. Kaushal Parikh reports payment made to his institution from AstraZeneca, Dava Oncology, Guardant Health, Intellisphere, ImmunityBio, Medscape, Regeneron, and Rigel Pharmaceuticals; personal/consulting fees from AstraZeneca, Immunity Bio, Jazz Pharmaceuticals, Regeneron, and Rigel Pharmaceuticals; and travel support from Rigel Pharmaceuticals outside the submitted work. Giuseppe Luigi Banna reports personal/consulting or advisory fees from Accord, Amgen, AstraZeneca, and Merck; and fees for serving on speakers' bureaus from Amgen, Astellas, AstraZeneca, Bayer, Merck, and Pfizer outside the submitted work. Jill Feldman reports personal/consulting fees from AstraZeneca, Blueprint Medicine, Janssen, Johnson & Johnson, and EQRx outside the submitted work. Xiuning Le reports institutional research funding from ArriVent, Boehringer Ingelheim, Dizal, Eli Lilly and Company, EMD Serono, Janssen, Regeneron, Takeda, Teligene, and ThermoFisher; personal/consulting fees from AbbVie, Abion, ArriVent, AstraZeneca, Bayer, BlossomHill, Blueprint Medicines, Boehringer Ingelheim, Daiichi Sankyo Company, Eli Lilly and Company, EMD Serono (Merck KGaA), Hengrui Therapeutics, Janssen Biotech, Novartis, Regeneron Pharmaceuticals, Sensei Biotherapeutics, Spectrum Pharmaceuticals, SystImmune, Taiho, and Teligene; travel support from EMD Serono, Janssen, and Spectrum Pharmaceutics; and owns stock options in BlossomHill outside the submitted work. Alfredo Addeo reports grants from AstraZeneca; personal/consulting or advisory fees from Amgen, Astellas, AstraZeneca, Bristol Myers Squibb Company, Eli Lilly and Company, Merck, MSD, Novartis, Pfizer, and Roche; and fees for serving on speakers' bureaus from Amgen, AstraZeneca, and Eli Lilly and Company outside the submitted work. Timothée Olivier and Claudio De Vito disclosed no conflicts of interest.

Borgeaud M, Olivier T, Bar J, et al. Personalized care for patients with EGFR‐mutant nonsmall cell lung cancer: navigating early to advanced disease management. CA Cancer J Clin. 2025;75(5):387‐409. doi: 10.3322/caac.70024