MET-Aberrant non-small cell lung cancer: from kinase dependence to cell-surface targetability—mechanistic basis and biomarker framework for bispecific antibodies and antibody-drug conjugates

Highlights

• •MET-aberrant NSCLC requires platform-aligned biomarker interpretation.
• •Kinase dependence and surface targetability frame therapeutic positioning.
• •MET-TKIs are most evidence-supported in METex14 and selected high-level MET amplification.
• •EGFR × MET BsAbs show activity in broad post-osimertinib EGFR-mutant NSCLC.
• •MET-ADCs support drug-specific IHC selection for high c-MET protein expression.

Abstract

MET-aberrant non-small cell lung cancer (NSCLC) is not a uniform therapeutic entity. Its biology, diagnostic pathways, and treatment sensitivity differ across MET exon 14 skipping alteration (METex14), MET amplification, and MET overexpression. This heterogeneity cannot be fully explained by conventional event-based classification and is reflected in the distinct clinical activity of MET tyrosine kinase inhibitors (MET-TKIs), bispecific antibodies (BsAbs), and antibody-drug conjugates (ADCs). With the emergence of antibody-based therapies, MET has evolved from a signaling driver to a cell-surface target for receptor modulation and payload delivery. We therefore propose a clinically anchored two-dimensional framework for interpreting therapeutic relevance in MET-aberrant NSCLC: kinase dependence and cell-surface targetability. Neither dimension should be regarded as a directly measurable binary variable. Kinase dependence is inferred from genomic and treatment-contextual proxies, most strongly METex14 and, more conditionally, high-level focal MET amplification. Cell-surface targetability is approximated by drug-specific IHC assessment of assay-defined c-MET protein expression; however, receptor internalization, intracellular trafficking, and payload delivery capacity remain incompletely measurable in routine clinical practice. Within this framework, MET-TKIs have the most evidence-supported established role in tumors with evidence of MET-driven kinase dependence. EGFR × MET BsAbs have demonstrated clinical activity in broad post-osimertinib EGFR-mutant NSCLC, while EGFR/MET co-dependence or MET-mediated bypass activation provides a mechanistic rationale for their use; MET-defined preferential benefit remains to be prospectively established. MET-directed antibody-drug conjugates (MET-ADCs) are supported in drug- and assay-defined populations with high c-MET protein overexpression, although the predictive relevance of delivery-related factors remains hypothesis-generating. Accordingly, MET testing should shift from single-event detection to platform-oriented stratification: next-generation sequencing (NGS) for driver alterations and resistance profiles, fluorescence in situ hybridization (FISH) for high-level focal amplification, and immunohistochemistry (IHC) for surface expression relevant to antibody-based therapies. This framework is intended to organize current biological and clinical evidence rather than to replace drug-specific companion diagnostics, regulatory indications, or prospectively validated treatment-selection algorithms. Precision treatment of MET-aberrant NSCLC is thus moving from event-based drug selection toward mechanism-based therapeutic matching. Future priorities include standardizing biomarkers, defining optimal target populations, and aligning biological subtypes, diagnostic strategies, and therapeutic platforms.

Graphical abstract

Introduction

The MET gene, located on chromosome 7, encodes the transmembrane receptor tyrosine kinase c-MET. This receptor comprises an extracellular region containing the semaphorin (SEMA), plexin-semaphorin-integrin (PSI), and immunoglobulin-plexin-transcription factor (IPT) domains, and an intracellular region containing the juxtamembrane (JM) region and the kinase domain [1,2]. Under physiological conditions, hepatocyte growth factor (HGF) binding induces c-MET dimerization and downstream signaling, thereby regulating tissue repair and regeneration [3,4]. In cancer, however, the MET axis can be persistently activated through HGF-dependent or HGF-independent mechanisms, including gene amplification, protein overexpression, and pathogenic mutations, thereby promoting proliferation, invasion, and metastasis. From a therapeutic perspective, c-MET is both accessible on the cell surface and capable of receptor-mediated internalization. It is therefore targetable not only by small molecules that inhibit intracellular kinase activity, but also by antibody-based agents that induce receptor downregulation, immune effector function, or cytotoxic payload delivery.

Over the past decade, systemic treatment for advanced non-small cell lung cancer (NSCLC) has continued to evolve, yet long-term survival remains limited by tumor heterogeneity and acquired resistance [5,6]. In epidermal growth factor receptor (EGFR)-mutant NSCLC, for example, the third-generation EGFR tyrosine kinase inhibitor osimertinib has extended median overall survival (OS) to 38.6 months, but disease progression remains almost inevitable [6]. Within the complex resistance network, aberrant activation of the MET signaling axis is one of the most important bypass mechanisms and among the most clinically actionable resistance events. MET signaling can compensate for EGFR blockade by reactivating downstream pathways such as phosphoinositide 3-kinase (PI3K)-AKT and RAS-mitogen-activated protein kinase (MAPK), thereby reducing tumor dependence on EGFR signaling [7,8]. Given the marked heterogeneity of osimertinib resistance, dynamic reassessment of MET status at progression by tissue or liquid biopsy may inform subsequent-line treatment selection; however, the interpretation of MET findings should remain treatment-context- and platform-specific [9].

Clinically, MET aberration is not a single entity but includes MET exon 14 skipping alteration (METex14), MET amplification, and MET overexpression. These subtypes differ fundamentally in both biology and therapeutic relevance. METex14 is a canonical driver event and provides the most evidence-supported genomic proxy for MET kinase dependence. MET amplification may function either as a primary driver or, after EGFR tyrosine kinase inhibitor (EGFR-TKI) resistance, may indicate MET-mediated bypass activation or EGFR/MET co-dependence; its biological and therapeutic relevance depends on amplification magnitude, focality, co-occurring alterations, and treatment context. In contrast, MET overexpression more often represents a phenotypic alteration with unstable oncogenic dependence and substantial spatial heterogeneity, and high expression alone should not be interpreted as evidence of stable MET kinase dependence. Importantly, biomarker requirements are also shifting across drug platforms. Small-molecule TKIs primarily rely on evidence of tumor kinase dependence. EGFR × MET bispecific antibodies (BsAbs) require extracellular target accessibility but are primarily positioned in EGFR/MET co-driven or MET bypass-activated states, whereas ADCs rely more directly on assay-defined c-MET protein expression, target accessibility, and receptor-mediated delivery. Accordingly, MET companion diagnostics should be interpreted through an integrated model combining next-generation sequencing (NGS) and fluorescence in situ hybridization (FISH) for genomic alterations with immunohistochemistry (IHC) for protein expression [10,11]. IHC provides an indirect proxy for assay-defined c-MET protein expression, but the relevant staining pattern, scoring system, and cutoff vary by antibody clone and drug context; IHC does not directly measure receptor internalization, intracellular trafficking, or payload release. Substantial heterogeneity remains across clinical trials and real-world practice. Differences in testing platforms, antibody clones, sample sources, and positivity thresholds can all affect MET classification and patient enrichment, limiting both cross-trial comparability and reproducibility of patient selection [12,13].

Although MET-targeted TKIs have long dominated the treatment of MET-aberrant NSCLC, their limitations are increasingly apparent. Clinical benefit is most consistently observed in subgroups with clear kinase dependence, whereas activity is limited in MET overexpression and low- to intermediate-level amplification, where dependence is weaker. In addition, in the complex resistance setting after EGFR-TKI treatment, kinase inhibition alone may be insufficient in some patients to suppress multi-pathway escape [14,15]. As a result, MET-directed therapy is expanding from the intracellular kinase domain to the extracellular target space. Antibody-based platforms, particularly BsAbs and ADCs, have expanded the therapeutic repertoire by enabling coordinated receptor blockade, immune-mediated activity, or payload delivery [16]. Notably, the potential bystander effect of ADCs may partially mitigate treatment escape driven by spatial heterogeneity, although its clinical contribution in MET-directed ADC therapy remains to be established [17]. Recent studies of MET-directed antibody-drug conjugates (MET-ADCs), exemplified by telisotuzumab vedotin, suggest that clinical benefit can be further stratified by MET protein expression, with more enriched objective responses observed in the high-expression subgroup [18]. These findings support the clinical relevance of assay-defined high c-MET protein expression in the telisotuzumab vedotin setting, but do not establish surface targetability as a universal predictive biomarker across antibody-based platforms. Taken together, treatment stratification in MET-aberrant NSCLC can no longer be defined by a single genomic event alone. Beyond conventional driver-centered classification, two interpretive dimensions may be considered simultaneously: the inferred degree of tumor dependence on MET signaling output and the accessibility of MET as a cell-surface antigen for therapeutic engagement. Both dimensions should be interpreted within drug-specific evidence boundaries. On this basis, this review summarizes the biological spectrum of MET aberrations, the mechanisms of action of BsAbs and ADCs, the key clinical evidence, and current testing strategies, and proposes an aligned framework linking mechanism, diagnostics, and treatment selection to support biomarker interpretation, study design, and cross-trial interpretation. This framework is intended as a clinically anchored conceptual model rather than a prospectively validated treatment-selection algorithm.

Fig. 1 provides a schematic overview of the distinct mechanisms of action of MET tyrosine kinase inhibitors, BsAbs, and ADCs. Together, these platforms provide the mechanistic basis for platform-specific therapeutic rationale in MET-aberrant NSCLC.

Fig. 1.

Mechanistic overview of major MET-targeted therapeutic platforms in MET-aberrant NSCLC. (A) MET-TKIs suppress intracellular MET kinase signaling and downstream pathways, such as PI3K–AKT and ERK. Their activity is most closely aligned with kinase-dependent states, including METex14 and selected high-level MET-amplified tumors, although acquired kinase-domain mutations can mediate resistance. (B) EGFR × MET BsAbs simultaneously engage EGFR and MET, promoting receptor blockade, receptor internalization and degradation, and immune-mediated effector functions such as ADCC and ADCP. (C) MET-ADCs bind surface MET, undergo internalization and lysosomal processing, and release cytotoxic payloads such as MMAE, resulting in microtubule disruption and potential bystander effects.

Abbreviations: ADC, antibody-drug conjugate; ADCC, antibody-dependent cellular cytotoxicity; ADCP, antibody-dependent cellular phagocytosis; BsAb, bispecific antibody; METex14, MET exon 14 skipping alteration; MMAE, monomethyl auristatin E; NSCLC, non-small cell lung cancer; TKI, tyrosine kinase inhibitor.

The MET spectrum: from oncogenic dependence to surface targetability

With the evolution of anticancer drug platforms, clinical stratification of MET aberrations has expanded from the traditional single-axis framework centered on oncogenic dependence to a two-dimensional model that also incorporates surface targetability [10,19]. The therapeutic premise of conventional tyrosine kinase inhibitors (TKIs) is that tumor cell survival and proliferation remain highly dependent on sustained activation of a single aberrant signaling pathway, a state referred to as oncogenic addiction [20]. By contrast, newer platforms represented by ADCs and BsAbs do not rely solely on persistent MET kinase activity. For ADCs, therapeutic activity is more directly related to cell-surface target abundance and receptor-mediated payload delivery, whereas BsAbs require extracellular target accessibility but are primarily positioned in EGFR/MET co-driven or MET bypass-activated states [21]. In MET-aberrant NSCLC, surface targetability therefore represents an additional biological dimension for interpreting the therapeutic relevance of MET across distinct drug platforms, rather than a direct substitute for kinase dependence.

Importantly, surface targetability is not equivalent to the static protein expression level captured by a single IHC assessment. For ADCs, antitumor activity is further constrained by dynamic delivery-related parameters, including surface receptor density, internalization kinetics, and lysosomal payload release efficiency [22]. Prior mechanistic studies and ADC-focused reviews suggest that these variables directly influence intracellular payload exposure and antitumor efficacy [23,24]. Accordingly, even in the absence of canonical driver events such as primary MET amplification or METex14, a therapeutically actionable window may still exist if surface antigen expression and delivery conditions exceed a certain threshold [18]. However, this threshold is likely drug-specific and influenced by payload class, linker stability, and the bystander effect; it therefore cannot be inferred from IHC alone. Accordingly, IHC should be regarded as a clinically feasible proxy for c-MET expression, rather than a direct measure of receptor internalization, intracellular trafficking, payload release, or overall delivery competence.

In the setting of clonal evolution after epidermal growth factor receptor (EGFR)-TKI resistance, MET aberrations also display substantial biological plasticity [25]. When MET functions as an acquired bypass driver, particularly in the context of high-level amplification, tumors often retain concurrent dependence on both EGFR and MET signaling, resulting in dual-pathway co-dependence. By contrast, when MET is manifested primarily as high membrane expression without clear evidence of amplification or driver status, kinase dependence is usually weaker. Available evidence supports MET amplification as a major mechanism of acquired resistance in EGFR-mutant NSCLC, and the main therapeutic strategies in this setting include dual EGFR/MET blockade, BsAbs, and MET-ADCs [10,26]. However, the available evidence does not support assigning a uniform therapeutic role to all MET-amplified tumors, because the clinical relevance of MET amplification depends on amplification level, focality, co-occurring alterations, and treatment context. Subsequent-line treatment selection should therefore not be based solely on the presence of a MET aberration, but should instead reflect its underlying biological properties. Tumors with strong kinase dependence have a clearer mechanistic rationale for combined pathway inhibition or dual receptor blockade, whereas those with weaker kinase dependence but high surface targetability may provide a rationale for evaluating antibody-based approaches such as ADCs.

Notably, the LUMINOSITY study showed that c-MET overexpression can be used for patient selection, with more enriched objective responses observed in the high-expression cohort [18]. This finding suggests that surface targetability is not merely a mechanistic concept, but may also serve as a clinically relevant stratification variable. More specifically, these data support high c-MET expression as a drug- and assay-specific selection marker for telisotuzumab vedotin in a defined clinical population; they do not establish surface targetability as a universal predictive biomarker for all ADCs or BsAbs. Taken together, stratification of MET-aberrant NSCLC should move beyond a purely genomic classification and toward a dual-axis framework integrating both kinase dependence and surface targetability. In this review, these dimensions are used as clinically anchored interpretive constructs that must be inferred from available genomic, protein-based, and treatment-contextual proxies. On the basis of these biological distinctions, MET-aberrant NSCLC can be conceptualized within a dual-axis framework integrating kinase dependence and surface targetability, as summarized in Fig. 2.

Fig. 2.

Evidence-layered dual-axis framework for interpreting MET-aberrant NSCLC. The horizontal axis represents the inferred degree of kinase dependence, whereas the vertical axis represents drug-specific surface targetability. Neither axis is directly measurable as a binary clinical variable. METex14 is positioned in the high-kinase-dependence region and is most strongly aligned with MET-TKIs based on current evidence. High-level focal MET amplification occupies an intermediate-to-high region of both axes and represents a context-dependent state that may support MET inhibition or EGFR/MET-targeted strategies, particularly in bypass-driven or co-dependent settings. Intermediate-/low-level MET amplification is more heterogeneous and is shown as a context-dependent category. MET overexpression is positioned in the high-surface-targetability, lower-kinase-dependence region and is therefore more closely aligned with ADCs. High c-MET expression is an evidence-supported, drug-specific selection marker for telisotuzumab vedotin in its defined clinical setting. EGFR-mutant NSCLC after osimertinib progression with MET bypass activation is positioned in the EGFR/MET co-dependence or bypass-activation context. The lower panels summarize the corresponding diagnostic tools, therapeutic alignment, and representative clinical studies. This framework is intended to organize current evidence and does not replace drug-specific companion diagnostics, regulatory indications, or prospectively validated treatment-selection algorithms.

MET exon 14 skipping alteration

METex14 is one of the most representative oncogenic driver events in MET-aberrant NSCLC. It occurs in approximately 3%–4% of lung adenocarcinomas and 1%–2% of other NSCLC subtypes, and is enriched in specific histologies such as pulmonary sarcomatoid carcinoma [27,28]. A systematic review further reported a median prevalence of about 2% in unselected NSCLC, supporting METex14 as a routinely actionable alteration [29].

Mechanistically, exon 14 encodes the intracellular juxtamembrane (JM) region containing the Y1003 site. Under physiological conditions, phosphorylation at this site recruits the E3 ubiquitin ligase c-CBL, thereby promoting receptor degradation [30]. METex14 deletes the c-CBL binding site, impairing receptor ubiquitination and subsequent lysosomal clearance [[31], [32], [33]]. The resulting prolongation of receptor half-life not only sustains downstream signaling but also leads to accumulation of MET on the cell surface. METex14 therefore represents a prototypical alteration with both strong kinase dependence and preserved surface targetability within the dual-axis framework proposed here. METex14 therefore represents the most evidence-supported genomic proxy for MET kinase dependence. Its association with increased cell-surface MET provides a biological rationale for antibody accessibility, whereas receptor internalization and delivery-related competence require further clinical characterization.

Consistent with this biology, METex14-positive tumors are generally sensitive to selective type I MET-TKIs, as supported by prospective studies of capmatinib, tepotinib, and crizotinib in this molecular subgroup [[34], [35], [36]]. However, prolonged exposure to type I MET-TKIs may lead to resistance through secondary kinase-domain mutations, such as alterations at D1228 or Y1230, which reduce drug binding, [37] or through bypass pathway activation involving EGFR signaling [[37], [38], [39]]. Available evidence also suggests that this subtype derives only limited overall benefit from conventional chemotherapy and immune checkpoint inhibitors (ICIs) [34,40,41]. Importantly, resistance to TKIs does not necessarily eliminate the therapeutic relevance of MET. In some patients, although inhibition of the intracellular kinase domain becomes ineffective, the extracellular domain remains intact and surface expression persists. This pattern provides a biological rationale for evaluating antibody-based strategies after MET-TKI resistance; the association between retained surface expression and clinical benefit from BsAbs or ADCs requires prospective clarification. METex14 can therefore be viewed as a subtype with a well-established kinase-dependence signal and potentially retained surface accessibility. MET-TKIs remain the evidence-supported initial targeted treatment approach; BsAbs have early clinical activity in selected post-TKI settings [42], whereas the role of MET-ADCs in METex14-positive disease remains investigational.

MET amplification

MET amplification refers to an abnormal increase in MET gene copy number within tumor cells. It has dual clinical significance, serving either as a primary driver event or as an acquired resistance mechanism after treatment with EGFR or anaplastic lymphoma kinase (ALK) tyrosine kinase inhibitors. As a primary event, MET amplification is relatively uncommon in NSCLC, occurring in approximately 1%–5% of cases, and has been associated with smoking history and poorer prognosis [40,41,43]. Its role is more prominent in acquired resistance, where it is detected in about 5%–22% of tumors after first- or second-generation EGFR-TKI failure and becomes even more frequent after osimertinib, making it one of the most important bypass resistance mechanisms [25,[44], [45], [46]].

Mechanistically, MET amplification increases MET protein expression through a gene-dosage effect and can promote ligand-independent receptor dimerization and constitutive phosphorylation, thereby sustaining downstream oncogenic signaling. Some studies have also linked MET amplification to higher histologic grade and greater metastatic potential [47]. Unlike METex14, however, MET amplification is not a discrete binary driver event, but rather a continuous molecular variable. Its therapeutic relevance depends not only on copy number itself, but also on amplification level, amplification pattern (focal amplification versus polysomy), coexisting driver alterations, and treatment context. Accordingly, high-level focal amplification is more likely to indicate biologically relevant MET pathway dependence than low-level copy-number gain or polysomy, although its clinical interpretation remains assay- and context-dependent [35,48,49].

This heterogeneity directly shapes drug sensitivity. Available data indicate that the efficacy of MET-TKIs in MET-amplified tumors is strongly correlated with gene copy number: high-level amplification usually reflects stronger oncogenic dependence and is therefore more likely to confer sensitivity to MET inhibition, whereas low- or intermediate-level amplification often lacks stable and sufficient kinase dependence, resulting in limited activity of TKIs [35,50]. Accordingly, stratification based solely on the presence or absence of MET amplification is inadequate to explain the marked variation in treatment outcomes across patients.

Within the dual-axis framework, MET amplification is best viewed as a continuum between kinase dependence and surface targetability. High-level amplification is closer to a classical kinase-dependent state and, particularly in the setting of EGFR-TKI resistance, may function as a clear bypass driver. In this context, combined EGFR/MET inhibition or BsAbs has a strong mechanistic basis. By contrast, in low- to intermediate-level amplification, the degree of amplification may be insufficient to sustain stable driver dependence, while the associated increase in MET protein expression may still provide a biological rationale for evaluating antibody-based strategies when high c-MET expression is independently confirmed. The predictive value of copy-number gain for ADC activity, however, remains to be established. MET amplification should therefore not be treated as a uniform therapeutic entity, but should be further stratified according to both kinase dependence and surface expression.

MET overexpression

MET overexpression refers broadly to increased c-MET protein abundance detected by immunohistochemistry (IHC) or other protein-based assays, rather than to a specific genomic alteration. In this review, MET overexpression is used as a biological category when discussing increased c-MET protein expression in general. When discussed in the context of antibody-based patient selection, particularly MET-ADCs, it should be interpreted as drug- and assay-defined c-MET protein positivity based on the relevant antibody clone, staining pattern, scoring system, and cutoff. Therefore, MET overexpression should not be equated with unqualified MET positivity or stable MET kinase dependence without specifying the assay and therapeutic context. Unlike METex14 or MET amplification, MET overexpression arises through more complex mechanisms. It may result from increased gene dosage, but can also be driven by transcriptional upregulation, a hypoxic microenvironment, or crosstalk with other pathways such as epidermal growth factor receptor (EGFR) [51,52]. Epidemiologically, MET overexpression is one of the most common forms of MET abnormality in NSCLC, but its reported prevalence varies widely according to antibody selection, scoring system, and positivity threshold. A universally applicable frequency is therefore difficult to define outside a specific methodological context. Although MET overexpression may coexist with MET amplification or METex14, it also frequently occurs in the absence of a clear genomic alteration [43,49]. Most studies associate MET overexpression with poor clinical outcomes, and some suggest a relationship with programmed death-ligand 1 (PD-L1) expression, although its predictive value for anti-programmed cell death protein 1 (PD-1)/PD-L1 therapy remains inconsistent [53].

With the evolution of treatment strategies, the clinical relevance of MET overexpression has shifted from a possible surrogate of MET pathway activation to a phenotypic marker for stratification with antibody-based therapies. In MET-tyrosine kinase inhibitor (TKI)-based approaches, MET overexpression alone is usually insufficient to predict benefit from monotherapy, because increased protein expression does not necessarily indicate stable driver dependence [[54], [55], [56]]. For example, studies such as TATTON and INSIGHT suggest that patients are more likely to benefit from combined EGFR-TKI and MET inhibition only when IHC shows high MET expression [57,58]. This indicates that, for kinase inhibition strategies, MET overexpression becomes therapeutically meaningful only when it aligns with clear functional dependence.

By contrast, antibody-based platforms, particularly ADCs, have redefined the clinical role of MET overexpression. Their activity depends less on blockade of MET signaling than on the use of surface MET as an entry site for internalization and selective payload delivery into tumor cells. Accordingly, high c-MET expression may serve as a clinically practical, drug-specific marker of target accessibility for selected ADCs [13,18]. The LUMINOSITY study showed that telisotuzumab vedotin (TelisoV) has clinical activity in c-MET-overexpressing, non-squamous, EGFR wild-type NSCLC, with more enriched objective responses observed in the high-expression cohort. These findings support the clinical utility of MET overexpression as a biomarker for antibody-based patient selection, while also highlighting its strong dependence on assay methodology and threshold definition. In this setting, IHC primarily reflects surface antigen expression; receptor internalization, intracellular trafficking, and payload release remain additional delivery-related determinants that are not directly captured by routine IHC assessment.

Taken together, MET overexpression should no longer be viewed simply as a secondary correlate of genomic alteration. Compared with METex14 and high-level MET amplification, it defines a broader but more biologically heterogeneous population. Within the dual-axis framework, its main value lies not in identifying a canonical driver event, but in recognizing tumors that may meet drug-specific expression criteria for antibody-based treatment selection. Its predictive value should therefore be interpreted within the context of the relevant agent, assay, and cutoff, rather than generalized across all ADCs or BsAbs. In advanced NSCLC, MET overexpression is therefore shifting from a surrogate marker of pathway activation to a phenotypic biomarker for the antibody era.

Clinical operationalization and evidentiary boundaries of the dual-axis framework

The dual-axis framework conceptualizes kinase dependence and surface targetability as clinically interpretable dimensions of MET biology. Their assessment integrates genomic, protein-based, and treatment-contextual evidence, thereby providing a structured basis for biomarker interpretation, therapeutic positioning, and prospective study design. The clinical utility of this framework will be further refined through prospective validation across therapeutic platforms and disease settings.

Kinase dependence denotes the extent to which tumor growth and survival are sustained by MET kinase signaling. In routine clinical practice, this biological state is inferred through a hierarchy of genomic and clinical proxies. METex14 currently represents the most robust genomic indicator because it disrupts receptor degradation, sustains MET signaling, and is consistently associated with clinical activity of selective MET-TKIs [[31], [32], [33], [34], [35], [36]]. High-level focal MET amplification serves as a conditional indicator of kinase dependence, particularly when confirmed by FISH or rigorously validated tissue NGS. Its interpretation integrates amplification magnitude, focality, assay platform, co-occurring genomic alterations, and treatment context [35,48,49]. In EGFR-mutant NSCLC following EGFR-TKI progression, acquired MET amplification commonly delineates a MET-mediated bypass or EGFR/MET co-dependence state [15,44,46]. MET protein overexpression gains interpretive value as an indicator of kinase dependence when supported by concordant genomic and clinical evidence [13,43,49].

Surface targetability denotes the presence of membranous MET accessible to antibody-based therapeutic platforms. For ADCs, this dimension is further shaped by antigen abundance, receptor-mediated uptake, and intracellular payload processing. c-MET expression assessed by IHC currently provides the most practical clinical surrogate for target accessibility [13,18]. In the LUMINOSITY study, high c-MET expression defined by a prespecified SP44-based assay and threshold served as a patient-selection marker for telisotuzumab vedotin in a defined clinical population [18]. Whereas IHC characterizes surface antigen expression, receptor internalization kinetics, lysosomal trafficking, linker cleavage, payload release, extracellular-domain integrity, and intratumoral heterogeneity remain research-level determinants of ADC activity [22,23].

The evidentiary basis for the two dimensions differs across therapeutic platforms. METex14 serves as an evidence-supported anchor for MET-TKI-based treatment [[34], [35], [36]]. High-level focal MET amplification represents a context-dependent indicator that may inform MET inhibition or dual EGFR/MET targeting in selected clinical settings [15,35,46,48,49]. High c-MET expression functions as a drug- and assay-specific selection marker for telisotuzumab vedotin within its defined clinical setting [18]. For EGFR × MET BsAbs, the predictive contribution of MET IHC, receptor internalization capacity, and integrated gene–protein models remains an area of active investigation [59,60]. Therapeutic positioning therefore integrates drug-specific companion diagnostics, regulatory indications, prior therapy, co-occurring molecular alterations, and prospective trial eligibility criteria. The clinical proxies, evidentiary status, and boundaries of interpretation for these two dimensions across therapeutic platforms are summarized in Table 1.

Table 1.

Clinical proxies and evidentiary boundaries for kinase dependence and surface targetability in MET-aberrant NSCLC.

Biological dimension	Clinical proxy and assay	Evidence status	Current clinical interpretation	Boundary of interpretation
Kinase dependence	METex14 detected by DNA- and/or RNA-based NGS	Evidence-supported	Provides the most robust genomic indicator of MET kinase dependence and supports MET-TKI-based treatment	Does not independently define sensitivity to BsAbs or ADCs
Kinase dependence	High-level focal MET amplification assessed by FISH or validated tissue NGS	Context-dependent	May support MET inhibition or dual EGFR/MET targeting, particularly in EGFR-TKI-resistant disease	Requires interpretation of amplification magnitude, focality, co-occurring alterations, and treatment context; low- or intermediate-level copy-number gain does not establish stable kinase dependence
EGFR/MET co-dependence or bypass activation	Acquired MET amplification after EGFR-TKI progression, integrated with resistance lineage and clinical context	Context-dependent	May provide a rationale for EGFR/MET co-targeting or EGFR × MET BsAb-based strategies	Does not establish autonomous MET dependence or predict preferential benefit from BsAbs in an individual patient
Surface targetability	High c-MET expression assessed using a drug-specific IHC assay	Evidence-supported in a defined drug setting	Supports patient selection for telisotuzumab vedotin when the validated assay, cutoff, and clinical population are met	Does not establish kinase dependence or provide a universal predictor of response across all ADCs or BsAbs
Delivery-related biology	Receptor internalization, lysosomal trafficking, linker cleavage, payload release, and intratumoral heterogeneity	Research-level determinants	May contribute to heterogeneous ADC activity and inform translational studies	Cannot be directly quantified by routine IHC
BsAb responsiveness	EGFR/MET network state, resistance lineage, co-occurring alterations, and extracellular target accessibility	Active area of investigation	May inform biomarker-enriched clinical-trial design and therapeutic positioning	MET IHC alone has not been prospectively established as a predictive biomarker for EGFR × MET BsAbs

Abbreviations: ADC, antibody-drug conjugate; BsAb, bispecific antibody; EGFR, epidermal growth factor receptor; FISH, fluorescence in situ hybridization; IHC, immunohistochemistry; METex14, MET exon 14 skipping alteration; NGS, next-generation sequencing; TKI, tyrosine kinase inhibitor. 

Detection of MET aberrations and companion diagnostics

As BsAbs and ADCs become increasingly important therapeutic options for MET-aberrant NSCLC, the clinical role of MET testing has expanded beyond qualitative identification of a single driver alteration to the integrated interpretation of clinical proxies related to kinase dependence and surface targetability [5]. MET aberrations span multiple molecular layers, including splice alterations, copy number changes, and protein overexpression, and different drug platforms require fundamentally different types of biomarker information [41]. Accordingly, the central goal of modern molecular pathology is no longer simply to detect a MET abnormality, but to align each diagnostic assay with the relevant therapeutic context—for example, using DNA/RNA NGS to identify METex14 and co-occurring resistance lineages, FISH or validated tissue NGS to characterize high-level focal amplification, and drug-specific IHC to assess c-MET expression—thereby enabling mechanism-based stratification [6,11,61].

Detection strategies for METex14

METex14 is fundamentally a heterogeneous group of splice-related variants that converge on a common transcript-level consequence: exon 14 skipping [62]. The main diagnostic challenge therefore lies not in identifying a genomic variant per se, but in determining whether that variant truly produces an abnormal splicing event with therapeutic relevance [33]. In practice, detection of METex14 requires integrated interpretation based on the coverage and functional resolution of different assay platforms [63].

DNA NGS

DNA-based NGS is currently the most widely used screening approach in clinical practice. It infers METex14 indirectly by identifying genomic alterations that disrupt normal pre-mRNA splicing, such as variants affecting canonical splice sites flanking exon 14 [33,64]. Its main advantages are high throughput, compatibility with formalin-fixed paraffin-embedded (FFPE) samples, and simultaneous characterization of co-occurring genomic alterations [33]. However, as an indirect method, DNA NGS has two major limitations. First, targeted panels often provide insufficient coverage of deep intronic regions, which may lead to missed noncanonical splice-site variants [64]. Second, not all detected splice-related variants necessarily result in functionally relevant exon 14 skipping at the transcript level [65]. Without RNA-based confirmation, DNA findings may therefore be misinterpreted, including false-negative results or positives of uncertain functional significance [66]. In clinically suspected cases with borderline or atypical DNA NGS findings, complementary testing is often required.

RNA NGS

RNA-based NGS is the key method for direct confirmation of the functional consequence of aberrant splicing [67]. By analyzing mature transcripts, it can directly detect the abnormal junction between exon 13 and exon 15 and thereby confirm exon 14 loss without inferring the underlying genomic event [68]. Compared with DNA NGS, RNA NGS can overcome intronic coverage gaps and more comprehensively identify skipping events driven by complex structural alterations or deep intronic variants [66,69]. Its limitations are mainly preanalytical: performance depends heavily on sample quality, particularly RNA integrity in FFPE specimens, and sensitivity may be affected by transcript abundance [66]. Overall, RNA NGS serves as the principal functional validation tool for METex14 detection.

From a clinical testing perspective, DNA- and RNA-based assays should be viewed as complementary rather than interchangeable for METex14 detection. DNA NGS is appropriate as an initial broad profiling approach because it can simultaneously assess METex14-associated splice-region alterations, co-occurring oncogenic drivers, copy-number changes, and resistance lineages. However, DNA-based panels may miss functionally relevant METex14 events when intronic coverage is limited or when the detected variant has uncertain splicing consequence. RNA NGS provides direct evidence of exon 14 skipping at the transcript level and is particularly useful when DNA results are negative but clinical suspicion remains, when a splice-region variant of uncertain significance is detected, or when the DNA panel has limited intronic coverage. In practice, the most actionable approach is to use broad DNA/RNA NGS when available, or to perform reflex RNA testing for equivocal or high-suspicion cases identified by DNA testing [66].

Other methods: RT-qPCR

Real-time quantitative reverse-transcription polymerase chain reaction (RT-qPCR) is a highly targeted assay that rapidly screens for the exon 13–exon 15 junction [70]. Its strengths include short turnaround time and relatively high sensitivity in samples with low tumor cellularity [71]. However, by design, RT-qPCR mainly detects canonical METex14 transcripts and has limited ability to capture rare splice isoforms or provide broader genomic context. It is therefore best used as a rapid adjunct in selected clinical scenarios rather than as a substitute for NGS-based molecular profiling [72]. It is therefore best used as a rapid adjunct in selected clinical scenarios rather than as a substitute for NGS-based molecular profiling [73].

Detection of MET amplification and definition of thresholds

MET amplification may act either as a primary driver event or as an acquired resistance mechanism, particularly after epidermal growth factor receptor (EGFR) tyrosine kinase inhibitor treatment [15]. Unlike METex14, MET amplification represents a continuous copy number variation (CNV) rather than a discrete molecular event. The central diagnostic challenge is to distinguish focal amplification from chromosome 7 polysomy: the former is more likely to be associated with MET pathway dependence, whereas the latter more often indicates broader chromosomal instability and is less likely to predict benefit from targeted therapy [48,74]. The key objective of MET amplification testing is therefore to identify and quantify the subgroup with true high-level focal amplification that carries potential clinical stratification value in the appropriate therapeutic context [41,75].

Fluorescence in situ hybridization

FISH remains a key reference method for assessing MET amplification. By calculating the ratio of MET signals to chromosome 7 centromere (CEP7) signals, FISH can more directly distinguish focal amplification from polysomy, as in the Cappuzzo criteria [41,76,77]. A major advantage of FISH is preservation of spatial heterogeneity within tumor tissue, allowing copy number abnormalities to be assessed at the single-cell level [41]. Clinically, FISH is particularly valuable for confirming high-level focal amplification and for identifying patients who may benefit from dual epidermal growth factor receptor (EGFR)/MET blockade [78]. Its limitations include high technical complexity, low throughput, and the inability to capture other bypass resistance alterations in parallel [79].

NGS

Compared with FISH, the main value of NGS in MET amplification testing lies in its broader molecular coverage [80]. In the setting of EGFR tyrosine kinase inhibitor (EGFR-TKI) resistance, NGS can simultaneously detect MET amplification, METex14, and other bypass alterations such as human epidermal growth factor receptor 2 (HER2) and phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit alpha (PIK3CA), thereby helping define resistance evolution [64]. However, both tissue NGS and circulating tumor DNA (ctDNA) assays are often limited in their ability to separate focal amplification from polysomy algorithmically, which may lead to overestimation of true focal amplification. Accordingly, when NGS indicates increased MET copy number and kinase inhibitor-based treatment is being considered, orthogonal confirmation by FISH is generally recommended [81]. When the goal is to evaluate potential target abundance for ADCs, complementary protein-level assessment by IHC is also important [82,83].

From a clinical implementation perspective, FISH and NGS provide complementary information for MET amplification assessment. FISH remains the most direct method for distinguishing focal MET amplification from chromosome 7 polysomy because it preserves single-cell and spatial information through the MET/CEP7 ratio. This distinction is clinically relevant because focal amplification is more closely associated with MET pathway dependence and response to MET inhibition than polysomy. Tissue NGS provides broader molecular context by simultaneously detecting co-occurring resistance alterations and other oncogenic drivers, but copy-number calling is influenced by tumor purity, ploidy, panel design, bioinformatic algorithms, and the genomic breadth of the assay. Plasma NGS is useful when tissue is unavailable and may capture spatially heterogeneous resistance clones, but low tumor fraction can reduce sensitivity for copy-number gain and may underestimate focal amplification. In practice, tissue NGS can serve as an efficient screening tool in broad molecular profiling, whereas FISH confirmation should be considered when MET amplification is central to treatment selection, particularly for kinase inhibitor-based strategies or clinical-trial eligibility [49,84].

IHC assessment of MET overexpression

With the development of MET-ADCs such as telisotuzumab vedotin (TelisoV), the clinical role of IHC-based assessment of MET overexpression has changed substantially. Rather than serving only as an indirect indicator of MET pathway activation, IHC is increasingly used as a patient-selection tool in specific drug settings [85]. In this context, its main value is the direct assessment of targetable MET antigen abundance on the tumor cell surface.

Technical and scoring considerations for IHC

IHC is the principal method for evaluating surface MET abundance [86]. However, major methodological heterogeneity remains. Different antibody clones, including SP44, differ in membrane versus cytoplasmic staining characteristics, and scoring systems have not been fully standardized [87]. Common approaches include the H-score as a continuous variable from 0 to 300 and categorical methods based on the percentage of tumor cells with defined staining intensity, such as ≥2+ or 3+ staining [88,89].

The clinical meaning of a given IHC threshold also varies across therapeutic platforms. In studies of BsAbs or combination strategies, such as SAVANNAH, IHC is typically interpreted together with FISH or NGS as an auxiliary marker of bypass activation [90,91]. By contrast, in ADC studies such as LUMINOSITY, strong positivity by SP44—for example, ≥50% of tumor cells showing 3+ membrane staining—has served as a key enrollment criterion for telisotuzumab vedotin, and treatment activity was enriched in the high-expression subgroup [18]. These findings indicate that the clinical definition of “MET-high” is highly drug-specific and platform-dependent.

For clinically actionable SP44-based testing, both staining intensity and the percentage of stained tumor cells should be reported. In the LUMINOSITY program, c-MET overexpression was defined by the proportion of tumor cells showing 3+ staining, with c-MET-high defined as ≥50% and c-MET-intermediate as ≥25% to <50% of tumor cells with 3+ staining. In the telisotuzumab vedotin setting, high c-MET protein overexpression is defined by an FDA-approved test as ≥50% of tumor cells with strong (3+) staining. The VENTANA MET (SP44) RxDx Assay further specifies strong membrane and/or cytoplasmic staining as part of the clinical scoring framework. These thresholds should therefore be interpreted as drug- and assay-specific criteria for patient selection, rather than as universal cutoffs for all MET-targeted therapies [13,18,92].

Several technical and biological variables can affect MET IHC interpretation. Pre-analytical factors include specimen type, fixation quality, tumor cellularity, necrosis, decalcification, tissue age, and whether the specimen is archival or recently obtained. Small biopsies may be clinically practical but can underrepresent spatial heterogeneity, particularly when MET expression varies between the invasive front, primary tumor, and metastatic sites. Prior systemic therapy may also influence MET expression status in selected settings. For these reasons, MET IHC should ideally be performed on adequately preserved tumor tissue with sufficient viable tumor content, and results should be reported together with the antibody clone, scoring method, cutoff, specimen source, and clinical context. When IHC is used to guide ADC selection, it should be regarded as a measure of surface antigen abundance rather than a direct readout of receptor internalization, trafficking, or payload delivery competence [13].

As MET-targeted BsAbs and ADCs continue to advance, standardized IHC workflows, optimized antibody clones, and reproducible cross-cohort cutoff values will be essential for more precise measurement of surface targetability [93,94]. In this setting, IHC serves as a core, platform-specific component of biomarker stratification for MET-ADCs and selected antibody-based strategies. It captures surface antigen abundance, whereas receptor internalization, intracellular trafficking, and payload processing require separate functional characterization. Its clinical interpretation should therefore be integrated with the relevant agent, assay, cutoff, genomic context, and treatment setting [22,23].

In clinical practice, MET testing should be guided by the therapeutic question and sample context. Broad DNA-based NGS is suitable for initial molecular profiling because it can assess METex14-associated splice-region alterations, MET copy-number changes, co-occurring drivers, and resistance lineages in parallel; reflex RNA testing is useful when METex14 is suspected but DNA findings are negative, equivocal, or limited by insufficient intronic coverage [66]. For MET amplification, tissue NGS provides broad molecular context, whereas FISH remains valuable for confirming clinically decisive amplification and distinguishing focal amplification from chromosome 7 polysomy [49]. At progression, particularly after EGFR-TKI therapy, tissue re-biopsy enables histology, FISH, and IHC assessment in the same disease context, while plasma ctDNA can provide rapid resistance profiling when tissue is unavailable; however, negative or noninformative plasma results should be interpreted cautiously when MET-driven resistance remains clinically plausible [95,96]. For SP44-based IHC, reports should specify specimen source, fixation adequacy, staining intensity, percentage of positive tumor cells, cutoff, and treatment timing, because pre-analytical variables and spatial heterogeneity may affect interpretation [13].

MET-TKI therapy as a benchmark for platform comparison

MET-TKIs provide the principal clinical benchmark for interpreting the therapeutic relevance of MET across the biological contexts considered in this review. Their activity is most reproducible in METex14-positive NSCLC, where selective MET inhibition directly addresses a well-established kinase-dependent driver state. In de novo MET amplification, clinical benefit is more variable and is concentrated in tumors with high-level focal amplification, whereas low- or intermediate-level copy-number gain provides a less consistent signal of MET dependence. In EGFR-mutant NSCLC with acquired MET amplification after EGFR-TKI progression, continued EGFR inhibition combined with MET blockade provides a biologically and clinically supported strategy. By contrast, MET protein overexpression alone has not been validated as a stand-alone selection biomarker for MET-TKI monotherapy. Table 2 summarizes selected key clinical datasets across these distinct MET contexts, providing a benchmark for comparison with antibody-based therapeutic platforms.

Table 2.

Key clinical evidence for MET-TKI-based therapy across biologically distinct MET aberrations in non-small cell lung cancer.

MET biological context	Biomarker definition / clinical setting	Representative MET-TKI–based regimen and study	Population	Evidence level	Regulatory status	Key efficacy outcomes	Reference
METex14	MET exon 14 skipping alteration	Capmatinib; final GEOMETRY mono-1 analysis	Treatment-naive: n = 60; previously treated: n = 100	Prospective phase II; regulatory-supported biomarker-defined evidence	FDA-approved for metastatic NSCLC with METex14 skipping alteration	Treatment-naive: ORR 68%, median DoR 16.6 months, median PFS 12.5 months; previously treated: ORR 44%, median DoR 9.7 months, median PFS 5.5 months	[97]
METex14	MET exon 14 skipping alteration	Tepotinib; VISION long-term analysis	Pooled Cohorts A and C: n = 313	Prospective phase II; regulatory-supported biomarker-defined evidence	FDA-approved for metastatic NSCLC with METex14 skipping alteration	ORR 51.4%, median DoR 18.0 months, median PFS 11.2 months	[98]
De novo MET amplification	Tumor-tissue MET gene copy number ≥10	Capmatinib; GEOMETRY mono-1 high-level amplification cohorts	Treatment-naive: n = 15; previously treated: n = 69	Prospective phase II subgroup evidence	Not approved for MET amplification-defined NSCLC	Treatment-naive: ORR 40%, median PFS 4.2 months; previously treated: ORR 29%, median PFS 4.1 months	[35]
Acquired MET amplification after first-line osimertinib	Central tissue FISH-confirmed MET amplification: MET GCN ≥5 or MET/CEP7 ratio ≥2	Tepotinib + osimertinib; INSIGHT 2 phase II	Primary activity population: n = 98	Prospective phase II single-arm evidence	Investigational for acquired MET amplification after osimertinib	Confirmed ORR 50.0%, median DoR 8.5 months, median PFS 5.6 months, median OS 17.8 months	[99]
Acquired MET amplification after EGFR-TKI progression	Study-defined acquired MET amplification	Savolitinib + osimertinib versus platinum-based chemotherapy; SACHI phase III	ITT population: n = 211; savolitinib + osimertinib, n = 106; chemotherapy, n = 105	Randomized phase III evidence	NMPA-approved in China for EGFR-mutant, MET-amplified NSCLC after EGFR-TKI progression; not FDA-approved	ORR 58% versus 34%; median PFS 8.2 versus 4.5 months; HR 0.34; median DoR 8.4 versus 3.2 months	[100]
Acquired MET overexpression and/or amplification after osimertinib	MET IHC 3+ in ≥90% of tumor cells and/or FISH 10+	Savolitinib + osimertinib; SAVANNAH phase II primary efficacy population	n = 80	Prospective phase II biomarker-enriched evidence	Investigational / biomarker-enriched post-osimertinib setting	Investigator-assessed ORR 56.3%, median DoR 7.1 months, median PFS 7.4 months	[101]

Abbreviations: DoR, duration of response; EGFR, epidermal growth factor receptor; FISH, fluorescence in situ hybridization; GCN, gene copy number; HR, hazard ratio; IHC, immunohistochemistry; ITT, intention-to-treat; METex14, MET exon 14 skipping alteration; NMPA, National Medical Products Administration; NSCLC, non-small cell lung cancer; ORR, objective response rate; OS, overall survival; PFS, progression-free survival; TKI, tyrosine kinase inhibitor.

Bispecific antibodies in MET-Aberrant NSCLC

Mechanisms of action and biological positioning

In MET-aberrant NSCLC, BsAbs act primarily through coordinated extracellular blockade and receptor regulation, and are therefore biologically distinct from MET-TKIs, which mainly suppress intracellular kinase activity. At the signaling level, BsAbs can simultaneously engage MET and functionally related receptors, most notably epidermal growth factor receptor (EGFR), thereby co-inhibiting the EGFR/MET axis under both ligand-dependent and ligand-independent conditions and attenuating key downstream pathways, including phosphoinositide 3-kinase (PI3K)-AKT, mitogen-activated protein kinase (MAPK), and signal transducer and activator of transcription (STAT) signaling. The prototypical EGFR × MET BsAb amivantamab has been shown to bind both EGFR and MET and to exert antitumor activity through receptor blockade, receptor downregulation, and immune-mediated effects. Its therapeutic rationale therefore lies not only in dual-target engagement, but also in its mechanistic fit for tumors driven by EGFR/MET co-dependence [[102], [103], [104]].

At the receptor level, simultaneous binding of MET and a cooperating receptor can promote receptor clustering, internalization, and degradation, thereby reducing the availability of surface MET on tumor cells. This mechanism may be particularly relevant in tumors that retain an intact extracellular domain and persistent receptor activation, including some cases of METex14 and high-level MET amplification. In addition, BsAbs with an intact fragment crystallizable (Fc) domain may induce antibody-dependent cellular cytotoxicity (ADCC) and antibody-dependent cellular phagocytosis (ADCP), further enhancing innate immune-mediated antitumor activity. BsAbs therefore function not only as signal inhibitors, but also as immune effector molecules, providing a mechanistic basis for combination with other systemic therapies [102,105,106].

Unlike MET-TKIs, whose activity depends mainly on intracellular kinase inhibition, the main advantage of BsAbs lies in extracellular dual blockade and receptor regulation. Unlike ADCs, whose efficacy centers on cytotoxic payload delivery, the core value of BsAbs is not target-mediated delivery but simultaneous interference with cross-pathway signaling networks such as EGFR/MET. Biologically, BsAbs have a mechanistic rationale in disease states characterized by MET-mediated bypass activation in an EGFR-dependent setting, EGFR/MET co-driven signaling, or retention of extracellular receptor targetability after TKI resistance. Retention of extracellular receptor accessibility after TKI resistance may support continued antibody engagement; its relationship with clinical benefit, however, requires further prospective validation. Importantly, the most mature clinical evidence for this platform has been generated largely in patients with EGFR-mutant disease after progression on osimertinib, rather than in prospectively enriched cohorts defined solely by MET positivity. This distinction is central to the interpretation of CHRYSALIS-2 and MARIPOSA-2: their findings support clinical activity of amivantamab-based regimens in broad post-osimertinib EGFR-mutant NSCLC, but do not quantify the component of benefit specifically attributable to MET-driven biology. The optimal population for BsAbs should therefore be determined by integrating resistance context, molecular lineage, and protein expression features [59,60].

Key clinical studies and efficacy patterns

Table 3 summarizes the key clinical evidence for MET-related bispecific antibodies, including biomarker definition, evidence level, regulatory status, major efficacy outcomes, and the boundaries of MET-specific attribution.

Table 3.

Key clinical evidence for MET-related bispecific antibodies in NSCLC.

Study / regimen	Population / setting	Biomarker definition	Evidence level	Regulatory status	Key efficacy outcomes	Interpretation / evidence boundary
CHRYSALIS; amivantamab	Advanced METex14 NSCLC; treatment-naive or previously treated; some patients had prior MET-targeted therapy	METex14	Prospective phase I MET-defined cohort	Not approved for METex14-defined NSCLC	N = 97; ORR 32%; median DoR 11.2 months; median PFS 5.3 months; median OS 15.8 months	Provides MET-defined evidence that EGFR × MET BsAb therapy has measurable activity in METex14 NSCLC; optimal sequencing relative to MET-TKIs remains undefined
CHRYSALIS-2 cohort A; amivantamab + lazertinib	EGFR exon 19del/L858R-mutant NSCLC after osimertinib and platinum chemotherapy	Not prospectively restricted to MET-positive disease; exploratory ctDNA resistance analyses	Phase I/Ib single-arm evidence in broad post-osimertinib EGFR-mutant disease	Not approved for this post-osimertinib setting; not a MET-selected strategy	N = 162; investigator-assessed ORR 28%; BICR-assessed ORR 35%; median DoR 8.3 months; median PFS 4.5 months; median OS 14.8 months	Demonstrates activity in a molecularly heterogeneous post-osimertinib population; does not establish MET aberrations as predictive biomarkers
MARIPOSA-2; amivantamab + chemotherapy	EGFR-mutant advanced NSCLC after progression on osimertinib	EGFR exon 19del/L858R; not restricted by MET positivity	Randomized phase III evidence in broad post-osimertinib EGFR-mutant disease	FDA-approved with carboplatin and pemetrexed for EGFR exon 19del/L858R NSCLC after EGFR-TKI progression; not MET-selected	N = 131; ORR 64% versus 36% with chemotherapy; median PFS 6.3 versus 4.2 months; PFS HR 0.48; median intracranial PFS 12.5 versus 8.3 months	Establishes a regulatory-supported post-EGFR-TKI regimen in EGFR-mutant NSCLC; benefit should not be attributed specifically to MET-positive biology
MARIPOSA-2; amivantamab + lazertinib + chemotherapy	EGFR-mutant advanced NSCLC after progression on osimertinib	EGFR exon 19del/L858R; not restricted by MET positivity	Randomized phase III evidence in broad post-osimertinib EGFR-mutant disease	Not approved for post-osimertinib or MET-defined use	N = 263; ORR 63% versus 36% with chemotherapy; median PFS 8.3 versus 4.2 months; PFS HR 0.44; median intracranial PFS 12.8 versus 8.3 months	Shows randomized efficacy in a broad EGFR-mutant population, but MET-defined preferential benefit remains unestablished
MCLA-129	Advanced NSCLC / solid tumors in early-phase development	EGFR × c-MET targeting; biomarker-enriched subsets variably reported	Early-phase investigational evidence	Investigational	Mature NSCLC efficacy dataset not yet established	Supports platform expansion, but current evidence is insufficient to define a clinical role in MET-aberrant NSCLC
EMB-01	Early solid-tumor development including NSCLC-related programs	EGFR × MET targeting; biomarker strategy not yet clinically established	Early-phase investigational evidence	Investigational	Mature NSCLC efficacy data not reported	Mechanistically relevant, but no established clinical role in MET-aberrant NSCLC
REGN5093 / davutamig	Early MET-targeted development	Biparatopic MET-directed approach	Early-phase investigational evidence	Investigational	Mature NSCLC efficacy data remain limited	Supports feasibility of MET-directed antibody targeting; clinical utility remains investigational

Abbreviations: BICR, blinded independent central review; BsAb, bispecific antibody; ctDNA, circulating tumor DNA; DoR, duration of response; EGFR, epidermal growth factor receptor; HR, hazard ratio; METex14, MET exon 14 skipping alteration; NSCLC, non-small cell lung cancer; ORR, objective response rate; OS, overall survival; PFS, progression-free survival; TKI, tyrosine kinase inhibitor. 5.2.1 Early Clinical Evidence for Bispecific Antibodies in METex14 NSCLC. 

Clinical evidence for BsAbs in METex14-positive NSCLC remains largely exploratory. The CHRYSALIS study provided initial clinical support for the use of amivantamab in MET-driven tumors. According to the final results published in 2025, 97 patients with advanced METex14 NSCLC were enrolled, including 16 treatment-naive patients, 28 previously treated but MET tyrosine kinase inhibitor (MET-TKI)-naive patients, and 53 previously treated patients who had received MET-targeted therapy. The overall objective response rate (ORR) was 32%, with ORRs of 50%, 46%, and 19% in these three subgroups, respectively. The overall clinical benefit rate (CBR) was 69%, with corresponding rates of 88%, 64%, and 66%. Median duration of response (DoR) was 11.2 months, and 61% of responders maintained response for at least 6 months. Median progression-free survival (PFS) was 5.3 months, and median overall survival (OS) was 15.8 months [42].

These findings indicate that amivantamab has measurable antitumor activity in advanced METex14 NSCLC, with greater benefit in treatment-naive and MET-TKI-naive patients, while still providing some clinical benefit after prior MET inhibitor failure. However, the central biological feature of METex14 remains oncogenic kinase dependence. Current evidence for BsAbs in this setting is therefore better viewed as support for post-MET-TKI development or sequential strategies, rather than as a replacement for MET-TKIs. Biologically, although METex14 tumors are primarily kinase-dependent, they often retain an intact extracellular domain and persistent surface expression after TKI resistance, providing a rationale for continued antibody-based targeting. The METex14 cohort in CHRYSALIS thus suggests that BsAbs warrant further development in MET-aberrant NSCLC beyond EGFR-related settings [107]. Their optimal timing and sequencing relative to MET-TKIs, however, remain to be defined in prospectively stratified studies.

Clinical evidence for EGFR × MET bispecific antibodies after osimertinib progression: broad clinical activity and the limits of MET-Defined attribution

In epidermal growth factor receptor (EGFR)-mutant NSCLC, acquired resistance to osimertinib is often accompanied by bypass pathway reprogramming, among which MET amplification or MET pathway activation is one of the most important mechanisms. In this setting, tumor biology is not defined by MET aberration alone, but rather by persistent EGFR dependence combined with compensatory MET activation. EGFR × MET BsAbs are therefore biologically well positioned when MET-mediated bypass activation or EGFR/MET co-dependence is present. Clinical efficacy evidence, however, should be interpreted separately from this mechanistic rationale. In CHRYSALIS-2 cohort A and MARIPOSA-2, eligibility was defined by sensitizing EGFR mutations and disease progression on osimertinib; MET positivity was neither a prospective enrollment criterion nor a stratification factor. These studies therefore establish clinical activity of amivantamab-based regimens in broad post-osimertinib EGFR-mutant populations, rather than in prospectively defined MET-positive disease [59,60].

In CHRYSALIS-2 cohort A, amivantamab plus lazertinib was evaluated in patients with EGFR exon 19 deletion or L858R-mutant NSCLC who had progressed on osimertinib and platinum chemotherapy. A total of 162 patients were enrolled. The investigator-assessed ORR was 28%, and the blinded independent central review-assessed ORR was 35%. Median DoR was 8.3 months, CBR was 58%, median PFS was 4.5 months, and median OS was 14.8 months. Exploratory ctDNA analyses suggested that responses were observed both in patients with and without EGFR/MET-dependent resistance features [10]. Accordingly, this cohort demonstrates antitumor activity in a molecularly heterogeneous later-line population, while its design does not define a MET-selected subgroup with preferential benefit [59].

By contrast, MARIPOSA-2 provided higher-level randomized evidence. This phase III trial randomly assigned 657 patients with EGFR-mutant, locally advanced or metastatic NSCLC after progression on osimertinib in a 2:2:1 ratio to amivantamab–lazertinib–chemotherapy, chemotherapy, or amivantamab–chemotherapy. Both amivantamab-containing regimens significantly improved PFS versus chemotherapy alone, with hazard ratios of 0.48 and 0.44, respectively. Median PFS was 6.3 months and 8.3 months, compared with 4.2 months in the chemotherapy arm. ORR was 64% and 63% with the two amivantamab regimens, respectively, versus 36% with chemotherapy. Intracranial PFS was also improved, with medians of 12.5 months and 12.8 months, compared with 8.3 months for chemotherapy [60,108].

A subsequent exploratory baseline ctDNA analysis of the amivantamab–chemotherapy and chemotherapy arms reported PFS benefit across EGFR/MET-dependent, EGFR/MET-independent, and resistance-mechanism-unknown subgroups. In the ctDNA-defined MET-amplification subgroup, the estimated PFS hazard ratio was 0.51 (95% CI, 0.24–1.11; nominal P = 0.078), based on 12 and 30 patients in the amivantamab–chemotherapy and chemotherapy arms, respectively. Thus, they support broad activity of amivantamab–chemotherapy after osimertinib progression, while the predictive contribution of MET amplification remains unresolved [109].

Evidence linking MET protein expression with preferential benefit from EGFR × MET BsAbs has also evolved. An initial exploratory signal suggesting higher activity in MET IHC-positive tumors was not validated in subsequent CHRYSALIS-2 cohorts. In a 2025 WCLC validation analysis, MET IHC positivity did not meet prespecified criteria as a biomarker strategy for response to amivantamab plus lazertinib or amivantamab monotherapy. Current evidence therefore does not support MET IHC as a validated patient-selection biomarker for EGFR × MET BsAb treatment [110].

Taken together, the available evidence supports three related but distinct conclusions. First, confirmed MET-mediated bypass activation or EGFR/MET co-dependence provides a compelling biological rationale for EGFR × MET co-targeting. Second, CHRYSALIS-2 cohort A and MARIPOSA-2 establish clinical activity of amivantamab-based regimens in broad post-osimertinib EGFR-mutant NSCLC populations. Third, MET amplification, integrated resistance-lineage features, and other molecular variables remain candidate enrichment factors that require prospective validation before they can guide preferential BsAb selection. MET IHC should currently be interpreted as an investigational biomarker in this therapeutic setting. A key priority for future research is to establish prospectively defined biomarker-enriched cohorts and to align molecular stratification with the clinical development of BsAbs.

Early development of other bispecific antibodies and the limits of current evidence

Beyond amivantamab, several other BsAbs targeting MET-related pathways have entered early clinical development, but evidence in NSCLC remains limited and is not yet practice-changing. The most notable agents include MCLA-129, EMB-01, and REGN5093 (davutamig). Most of these agents are designed either for coordinated EGFR/MET blockade or for dual-epitope targeting of MET, indicating continued expansion of the BsAb platform, although the maturity of the evidence remains clearly below that of amivantamab [111].

Among them, MCLA-129 is one of the more notable early-stage candidates. This EGFR × c-MET BsAb has entered phase I/II testing in advanced NSCLC and other solid tumors. Early public data suggest preliminary antitumor activity in NSCLC, including encouraging response signals in some patients with METex14. However, available results are still largely derived from conference abstracts or early reports, sample sizes remain small, and patients with different molecular subtypes have often been enrolled together. Its optimal role in MET-aberrant NSCLC therefore remains unclear [112].

By contrast, although EMB-01 has also entered phase I/II development and includes NSCLC-focused studies, currently available information is still limited mainly to trial registration data and preclinical mechanistic evidence, with no mature efficacy results reported. At present, its main significance lies in demonstrating the structural and early safety feasibility of EGFR × MET BsAbs rather than providing clear evidence of clinical benefit [113].

The development path of REGN5093 (davutamig) differs somewhat. This agent is a biparatopic antibody targeting two distinct MET epitopes, designed to enhance receptor internalization and degradation while inhibiting both HGF-dependent and HGF-independent MET signaling. Although phase I/II studies are ongoing, published clinical efficacy data in NSCLC remain sparse and currently serve more as proof of mechanism and feasibility than as mature therapeutic evidence [114].

Overall, with the exception of amivantamab, MET-related BsAbs in NSCLC remain at an early stage of development. Their main evidentiary limitations include small sample sizes, substantial population heterogeneity, nonuniform biomarker stratification, and the absence of randomized comparative trials. At present, these agents are better regarded as evidence of platform expansion and mechanistic validation than as the basis for a clearly defined clinical role in MET-aberrant NSCLC.

Antibody-Drug conjugates in MET-Aberrant NSCLC

Mechanisms of action and biological positioning

In MET-aberrant NSCLC, MET-ADCs extend the therapeutic paradigm from pathway inhibition to surface antigen-mediated delivery. Unlike TKIs, whose efficacy depends primarily on suppression of MET kinase activity, the antitumor activity of MET-ADCs relies more on antibody binding to surface MET, receptor-mediated internalization, and intracellular payload release. In this context, MET functions not only as a signaling molecule, but also as a therapeutic entry site.

At the cellular level, MET-ADCs bind surface MET, undergo receptor-mediated internalization, and traffic to lysosomes, where linker cleavage releases a potent cytotoxic payload that induces cell death. For ADCs with cleavable linkers, payload released from MET-high cells may also diffuse into adjacent tumor cells, generating a bystander effect that could partly mitigate spatial heterogeneity of MET expression. Telisotuzumab vedotin (TelisoV) exemplifies this design: it consists of the anti-MET monoclonal antibody telisotuzumab conjugated to the microtubule inhibitor monomethyl auristatin E (MMAE) through a cleavable valine-citrulline linker. The contribution of receptor internalization, intracellular trafficking, payload release, and the bystander effect to clinical response in individual patients remains an active area of investigation.

Importantly, the potential relevance of MET-ADCs is not restricted to tumors with MET overexpression alone. In NSCLC harboring METex14 or MET amplification, ADC activity may also be biologically plausible if these events increase receptor stability and promote surface accumulation. These biological features provide a rationale for evaluating ADCs in selected METex14- or MET-amplified tumors; the extent to which genomic MET alterations independently enrich for MET-ADC benefit requires prospective characterization. MET-ADCs therefore expand the role of MET from a signaling driver to a delivery portal, providing a rationale for activity even in tumors without fully established kinase dependence. This is precisely where the biological positioning of ADCs differs from that of TKIs.

Table 4 summarizes the key clinical evidence for MET-ADCs, including biomarker definition, evidence level, regulatory status, major efficacy outcomes, and the boundaries of interpretation for c-MET expression-based patient selection.

Table 4.

Key clinical evidence for MET-directed antibody-drug conjugates in NSCLC.

Study / regimen	Population / setting	Biomarker definition	Evidence level	Regulatory status	Key efficacy outcomes	Interpretation / evidence boundary
First-in-human telisotuzumab vedotin study	Advanced solid tumors; NSCLC expansion subset	c-MET-aberrant NSCLC subset	Exploratory phase I evidence	Investigational at the time of study	NSCLC subset: n = 16; ORR 18.8%; median DoR 4.8 months; median PFS 5.7 months	Provided early proof-of-concept for MET-ADCs therapy, but sample size was small and biomarker criteria were not aligned with the later approved companion diagnostic
Phase I study of 2- or 3-week dosing of TelisoV	Advanced NSCLC; schedule-optimization study	c-MET-aberrant NSCLC by IHC	Phase I dose- and schedule-optimization evidence	Investigational dose-development evidence	c-MET-positive efficacy-evaluable NSCLC: n = 40; ORR 23%; median DoR 8.7 months; median PFS 5.2 months	Supported the every-2-week dosing strategy and further clinical development, but was not a registrational efficacy dataset
LUMINOSITY; TelisoV monotherapy	Previously treated, non-squamous, EGFR wild-type advanced NSCLC	SP44 IHC; c-MET-high: ≥50% tumor cells with 3+ staining; c-MET-intermediate: ≥25% to <50% tumor cells with 3+ staining	Prospective phase II; regulatory-supported high-expression subgroup	FDA accelerated approval for previously treated, non-squamous NSCLC with high c-Met protein overexpression; c-MET-intermediate subgroup not approved	172 treated; 161 efficacy-evaluable. Overall ORR 28.6%, median DoR 8.3 months, median PFS 5.7 months; c-MET-high: ORR 34.6%, median DoR 9.0 months; c-MET-intermediate: ORR 22.9%, median DoR 7.2 months	Defines the principal monotherapy dataset for TelisoV; regulatory approval is confined to the high-expression population defined by the approved SP44 companion diagnostic and should not be generalized to all c-MET-overexpressing or MET-aberrant NSCLC
TelisoV + erlotinib	EGFR-mutant, c-MET-aberrant NSCLC after prior EGFR-TKI therapy	c-MET-positive by IHC; c-MET-high commonly defined as H-score ≥225	Phase Ib combination evidence	Investigational combination strategy	42 treated; 36 efficacy-evaluable. EGFR-mutant subgroup ORR 32.1%; c-MET-high EGFR-mutant subgroup ORR 52.6%; median PFS 5.9 months in the efficacy-evaluable population	Suggests potential activity with EGFR inhibition in selected EGFR-mutant/c-MET-positive disease, but predates current post-osimertinib standards and is not a regulatory-supported strategy
TelisoV + osimertinib	EGFR-mutated, non-squamous NSCLC with c-MET protein overexpression after progression on prior osimertinib	c-MET protein overexpression by IHC	Phase I/Ib post-osimertinib combination evidence	Investigational combination strategy	n = 38; ICR-assessed ORR 50.0%; median DoR not reached; median PFS 7.4 months	Provides a clinically relevant combination signal in post-osimertinib EGFR-mutant NSCLC, but remains early-phase and requires prospective confirmation
TelisoV + nivolumab	Advanced c-MET-aberrant NSCLC	c-MET-positive by IHC	Phase Ib combination evidence	Investigational combination strategy	37 treated; 27 efficacy-evaluable. ORR 7.4%; median PFS 7.2 months	Tolerable but showed limited antitumor activity; current evidence does not support immunotherapy combination as a leading MET-ADC development direction

Abbreviations: ADC, antibody-drug conjugate; DoR, duration of response; EGFR, epidermal growth factor receptor; FDA, U.S. Food and Drug Administration; ICR, independent central review; IHC, immunohistochemistry; NSCLC, non-small cell lung cancer; ORR, objective response rate; PFS, progression-free survival; TelisoV, telisotuzumab vedotin. 6.2 Clinical Development of Teliso. V

Clinical evidence for monotherapy

Among MET-ADCs, TelisoV currently has the most mature clinical evidence in c-MET protein-overexpressing NSCLC. Its clinical development has focused primarily on this phenotypic biomarker rather than on MET genomic alterations alone. In the first-in-human phase I study, TelisoV showed a manageable safety profile and preliminary antitumor activity in advanced solid tumors. A subsequent phase I study further supported 1.9 mg/kg every 2 weeks as the recommended schedule and showed sustained activity in c-MET-aberrant NSCLC.

The phase II LUMINOSITY trial provides the key evidence for TelisoV monotherapy. LUMINOSITY was a two-stage, multicohort study designed to identify the c-MET protein-overexpressing NSCLC population most suitable for TelisoV monotherapy and to expand the selected cohort for further efficacy evaluation. The primary analysis focused on previously treated patients with advanced, non-squamous, epidermal growth factor receptor (EGFR) wild-type NSCLC with c-MET protein overexpression defined by the SP44 assay. c-MET-high was defined as ≥50% of tumor cells with 3+ staining, and c-MET-intermediate as ≥25% to <50%. At the primary data cutoff, by independent central review, the objective response rate (ORR) was 34.6% in the c-MET-high group, 22.9% in the c-MET-intermediate group, and 28.6% overall. Median DoR was 9.0, 7.2, and 8.3 months, respectively; median PFS was 5.5, 6.0, and 5.7 months; and median OS was 14.6, 14.2, and 14.5 months. A subsequent peer-reviewed analysis with approximately 6 months of additional follow-up included 168 efficacy-evaluable patients, with 84 patients each in the c-MET-high and c-MET-intermediate subgroups. The updated ORR was 29.2% overall, 34.5% in the c-MET-high subgroup, and 23.8% in the c-MET-intermediate subgroup; median DoR was 7.2 months overall. These results confirmed a consistent response pattern, with numerically higher response rates in the c-MET-high subgroup [115].

These are trial-reported outcomes for the overall LUMINOSITY analysis population and its prespecified c-MET-high and c-MET-intermediate subgroups. The updated analysis of LUMINOSITY confirmed a numerically higher response rate in the c-MET-high subgroup, while the absolute efficacy estimates reflected a later data cutoff [116].

The regulatory evidence base should be interpreted separately from the full trial population. The U.S. Food and Drug Administration (FDA) accelerated approval of telisotuzumab vedotin-tllv was supported by an efficacy population of 84 previously treated patients with non-squamous, EGFR wild-type NSCLC and high c-MET protein overexpression. In this regulatory analysis, the blinded independent central review-assessed ORR was 35%, and the median DoR was 7.2 months. Thus, the FDA-supported efficacy dataset was restricted to the high-expression population and should not be conflated with the trial-level outcomes reported for c-MET-intermediate disease or for the overall LUMINOSITY cohort.

LUMINOSITY also showed that the non-squamous EGFR-mutant cohort and the squamous cohort did not generate efficacy signals sufficient to support further expansion. These trial-reported findings define the monotherapy activity of TelisoV within the selected non-squamous, EGFR wild-type, c-MET-overexpressing LUMINOSITY population, with c-MET-high and c-MET-intermediate disease analyzed as prespecified biomarker subgroups. The regulatory interpretation of the high-expression subgroup is considered separately in Section 6.2.3. Within this trial context, c-MET expression should be interpreted as a drug-specific stratification variable, rather than extrapolated across all MET-aberrant subtypes.

Progress in combination strategies

Combination studies suggest that TelisoV may be particularly relevant in selected molecular settings. In a phase Ib trial, TelisoV plus erlotinib showed encouraging antitumor activity and acceptable toxicity in previously EGFR-TKI-treated patients with EGFR-mutant, c-MET-aberrant NSCLC.

This rationale appears even stronger in the more clinically relevant setting of osimertinib resistance. A phase Ib study published in 2025 showed that TelisoV plus osimertinib produced promising antitumor activity with a manageable safety profile in patients with c-MET-overexpressing, EGFR-mutant, non-squamous NSCLC after progression on prior osimertinib. Biologically, this combination is better aligned with a state of persistent EGFR dependence plus a retained c-MET delivery window than with enrollment based on MET genomic status alone.

By contrast, evidence for combination with immunotherapy remains limited. In a phase Ib study, TelisoV plus nivolumab was generally tolerable but showed limited antitumor activity. Current data therefore do not support immunotherapy combinations as a leading development direction for MET-ADCs. Compared with EGFR-TKI combinations, this strategy has not shown comparable therapeutic potential.

Overall, available combination data support further development of TelisoV in the setting of acquired EGFR-TKI resistance, particularly in patients with progression on osimertinib who retain c-MET protein overexpression. In this context, combination with EGFR-TKIs appears more promising than immunotherapy-based approaches.

Efficacy patterns, evidence boundaries, and clinical positioning

Taken together, current studies indicate that TelisoV has consistent antitumor activity in c-MET protein-overexpressing NSCLC. Monotherapy activity is more pronounced in the high-expression subgroup, whereas combination with EGFR-TKIs appears better matched to selected resistance settings. Even so, important challenges remain, including heterogeneity in IHC scoring, uncertainty regarding the optimal combination partner, and the lack of clearly defined sequencing strategies after resistance.

A major milestone was reached on May 14, 2025, when the U.S. Food and Drug Administration (FDA) granted accelerated approval to telisotuzumab vedotin-tllv for adults with previously treated, locally advanced or metastatic, non-squamous NSCLC with high c-MET protein overexpression, defined as ≥50% of tumor cells with strong (3+) staining, as determined by an FDA-approved test. On the same day, the VENTANA MET (SP44) RxDx Assay was approved as the companion diagnostic. The FDA-labeled indication and its supporting efficacy analysis should be distinguished from the broader LUMINOSITY trial dataset. LUMINOSITY reported outcomes in a selected cohort of previously treated, non-squamous, EGFR wild-type NSCLC with c-MET overexpression, including prespecified c-MET-high and c-MET-intermediate subgroups. By contrast, the FDA accelerated approval was supported by the 84-patient high c-MET expression cohort, in which the blinded independent central review-assessed objective response rate was 35% and median duration of response was 7.2 months. High c-MET expression, assessed using the approved companion diagnostic, therefore functions as a drug-specific selection biomarker for the approved monotherapy setting. Although EGFR wild-type status characterized the LUMINOSITY cohort and the regulatory efficacy population, it is not specified as a restriction in the FDA-labeled indication. Continued approval remains contingent upon verification of clinical benefit in confirmatory trial(s) [117].

Clinically, TelisoV has established a later-line monotherapy role within the FDA-labeled population defined by non-squamous histology, high c-MET protein overexpression, and prior systemic therapy. The broader LUMINOSITY dataset remains informative for understanding the association between expression level and treatment activity, whereas the approved indication is confined to the high-expression population. Its expansion into post-EGFR-TKI combination therapy and other MET-aberrant subgroups remains investigational.

More broadly, the development of TelisoV and other MET-ADCs in NSCLC should not be discussed independently of the two key biological dimensions proposed in this review: kinase dependence and surface receptor abundance. Further randomized studies are needed to define their optimal role across the broader spectrum of MET-aberrant NSCLC.

Discussion

MET-aberrant NSCLC is not a single therapeutic entity, and the populations most likely to benefit differ across drug platforms. Traditional classification based on METex14, MET amplification, and MET overexpression is no longer sufficient to explain the marked efficacy heterogeneity observed with MET-TKIs, BsAbs, and ADCs. In addition, acquired resistance to MET-TKIs is highly diverse, involving on-target alterations in the juxtamembrane region or tyrosine kinase domain (TKD), downstream effector alterations, and compensatory bypass activation [39,118]. This complexity indicates that the therapeutic significance of MET aberrations should not be defined by a single molecular event alone, but by the biological state it represents.

A more informative framework should therefore integrate two dimensions: kinase dependence and surface targetability. These dimensions provide clinically anchored interpretive constructs: kinase dependence is inferred from genomic alterations and resistance context, whereas surface targetability is approximated by drug-specific c-MET expression. Within this framework, MET-TKIs are principally aligned with tumors showing evidence of MET-driven kinase dependence, most clearly METex14. MET-ADCs are positioned in drug- and assay-defined populations with high c-MET expression, in which surface antigen availability provides the basis for receptor-mediated payload delivery. EGFR × MET BsAbs occupy a distinct position: their biological rationale is strongest in EGFR/MET co-dependence or MET-mediated bypass activation, whereas their most mature clinical evidence has been generated in broad post-osimertinib EGFR-mutant NSCLC populations. Current clinical studies illustrate these distinctions. The final CHRYSALIS results demonstrated that amivantamab has clinically meaningful activity in METex14 NSCLC, supporting continued development of antibody-based strategies in this molecular subgroup [42]; LUMINOSITY established high c-MET expression, assessed using a defined assay and cutoff, as a drug-specific selection marker for telisotuzumab vedotin [18]. In contrast, CHRYSALIS-2 and MARIPOSA-2 demonstrated clinical activity of amivantamab-based regimens in broad post-osimertinib EGFR-mutant cohorts that were not prospectively enriched for MET-positive disease [59,60]. These studies support regimen activity in this treatment setting, while the extent to which MET-defined biology identifies preferential beneficiaries remains to be prospectively established. Accordingly, the current value of the dual-axis framework lies in structuring biomarker interpretation and therapeutic positioning across distinct drug platforms. Its role in standardized treatment selection will require prospective validation in biomarker-defined clinical cohorts.

On this basis, we propose an evidence-layered framework for interpreting MET aberrations across therapeutic platforms. In METex14-positive disease, the dominant biology is strong oncogenic dependence, and first-line treatment should remain centered on MET-TKIs. After TKI resistance, however, retained extracellular MET accessibility may provide a rationale for evaluating BsAbs or ADCs; the optimal sequencing of these platforms and their predictive biomarkers require prospective definition. In high-level focal MET amplification, particularly after EGFR-TKI resistance, tumors may exhibit EGFR/MET co-dependence, making dual EGFR/MET blockade or BsAbs mechanistically attractive. By contrast, MET overexpression alone does not necessarily indicate stable kinase dependence, but in the antibody era—especially for ADCs—it is increasingly becoming a phenotypic biomarker of target accessibility for selected ADCs. In other words, future treatment stratification should move beyond the binary question of whether a target is present and integrate the degree of signal dependence with drug-specific surface-expression requirements. This framework complements genomic classification by linking mechanism, phenotype, and drug-specific biomarker requirements.

One major reason why current studies are difficult to compare directly is the marked heterogeneity in how MET positivity is defined. METex14, MET amplification, and MET overexpression are biologically distinct events that reflect different levels of driver dependence and different therapeutic windows. In addition, companion diagnostic platforms and positivity thresholds vary substantially across studies, making definitions such as “MET-high” or “MET-positive” highly platform-dependent. For example, based on LUMINOSITY, TelisoV was approved for previously treated, EGFR wild-type, non-squamous NSCLC using a threshold of ≥50% of tumor cells with IHC 3+ staining [18]. In contrast, trials such as SAVANNAH, which explored savolitinib plus osimertinib in EGFR-mutant resistant disease, applied a more stringent MET-positive definition, requiring IHC 3+ in ≥90% of tumor cells and/or FISH gene copy number ≥10 [101]. These differences indicate that the optimal threshold for MET overexpression is not absolute, but drug-specific. Future translational studies will need to determine whether the surface-expression threshold required for effective ADC delivery is fundamentally different from the pathway-dependence threshold required for MET-TKIs or BsAbs.

Different testing platforms capture complementary biological dimensions. DNA- or RNA-based NGS is particularly informative for identifying functional splice alterations and co-occurring molecular lineages; FISH, owing to its spatial resolution, is valuable for confirming high-level focal amplification; and IHC provides clinically practical assessment of surface c-MET expression for selected antibody-based platforms, particularly ADCs [119]. Threshold heterogeneity further complicates cross-trial interpretation. MET amplification may be defined by gene copy number (GCN), MET/CEP7 ratio, or other criteria, whereas IHC results vary according to antibody clone, scoring system, and cut-off. As a result, patients labeled as “MET-positive” across different studies do not necessarily represent the same biological population. Without further standardization of MET definitions and multi-omic testing approaches, cross-trial comparison and real-world generalizability will remain limited.

Clinically, we propose that MET-aberrant NSCLC should be interpreted across the disease course according to biological context rather than a single molecular label. In tumors with clear driver dependence, such as METex14 and selected high-level focal MET-amplified tumors, MET-TKIs remain the most evidence-supported targeted approach, whereas BsAbs or MET-ADCs may represent investigational or context-dependent strategies after phenotypic evolution under treatment pressure. In EGFR-mutant NSCLC after osimertinib progression, confirmed MET-mediated bypass activation provides a strong biological rationale for BsAbs or combined EGFR/MET inhibition. Randomized evidence from MARIPOSA-2 supports amivantamab-based regimens in the broader post-osimertinib EGFR-mutant population; whether MET-defined bypass activation identifies preferential beneficiaries remains to be prospectively established [59,60]. In patients meeting drug- and assay-defined criteria for high c-MET protein overexpression, MET-ADCs—most notably telisotuzumab vedotin in its approved setting—provide an evidence-supported treatment option. This approach is supported by LUMINOSITY and further reflected in the U.S. Food and Drug Administration (FDA) accelerated approval of telisotuzumab vedotin-tllv in May 2025 for previously treated, locally advanced or metastatic, non-squamous NSCLC with high c-MET protein overexpression confirmed by an approved companion diagnostic [18]. The key question, therefore, is not how broadly to define MET positivity, but how to align a given MET biological state with the most appropriate diagnostic and therapeutic platform within current evidence boundaries.

In the era of antibody-based therapy, durable clinical benefit depends not only on initial antitumor activity but also on the ability to maintain treatment with acceptable toxicity. For BsAbs, especially amivantamab, adverse events are strongly platform-specific and include infusion-related reactions (IRRs), skin toxicity, paronychia, and hypoalbuminemia. IRRs occur mainly during the first infusion and remain a major barrier to treatment convenience and early adherence. The recently reported SKIPPirr study was the first to systematically evaluate multiple prospective premedication strategies and showed that prophylactic oral dexamethasone (8 mg) significantly reduced IRRs with intravenous amivantamab [120]. This finding suggests that optimized premedication can substantially improve clinical feasibility. At the same time, management of the common dermatologic toxicities relies on early recognition and grading within a multidisciplinary framework, including intensive moisturization, topical corticosteroids, and antibiotics when needed, to minimize dose interruption or discontinuation [121]. In practice, toxicity management for BsAbs is therefore defined by prevention and early intervention.

By contrast, the toxicity profile of MET-ADCs is more clearly payload-driven. In LUMINOSITY, the most common treatment-related adverse events with TelisoV monotherapy were peripheral sensory neuropathy, peripheral edema, and fatigue, with peripheral sensory neuropathy representing the main grade ≥3 toxicity. In the phase Ib study of TelisoV plus osimertinib, peripheral sensory neuropathy, peripheral edema, and nausea were also common, while serious events such as anemia and pulmonary embolism required attention [18]. These data indicate that safety management for ADCs is centered not on infusion reactions, but on cumulative neurotoxicity, fluid retention, and hematologic risk in combination settings. Clinically, this requires baseline and serial neurologic assessment, timely dose delay or reduction, and close monitoring for edema and thrombotic risk. If the management focus of BsAbs is first-dose risk, that of MET-ADCs is cumulative toxicity and long-term tolerability.

Future safety management should also move upstream into trial design and formulation development. New routes of administration may fundamentally reshape platform toxicity. In PALOMA-3, the rate of IRRs with subcutaneous amivantamab was 13%, compared with 66% with intravenous administration (P < 0.001), and the rate of grade 3 IRRs fell from 4% to 0.5% [122]. These data show that formulation optimization itself can improve tolerability. Likewise, future ADC studies should incorporate patient-reported outcomes (PROs) and long-term functional toxicity more systematically to define the safety boundaries required for sustainable treatment rather than initial response alone. As BsAbs and ADCs move from early development to clinical implementation, toxicity management is no longer merely supportive care; it is becoming a determinant of treatment continuity and real-world benefit.

From a clinical-management perspective, assessment of MET aberrations should also shift from single-time-point testing to a dynamic, multilayered, and mechanism-aligned companion diagnostic strategy. At diagnosis, NGS should be used whenever possible to identify METex14 and co-occurring driver alterations [119]. At progression on EGFR-TKIs, repeat tissue biopsy or ctDNA analysis should be considered to track MET amplification and other bypass mechanisms [123]. When antibody-based therapy is being considered, standardized IHC should be incorporated to determine whether surface expression is sufficient to support macromolecular targeting and delivery [101]. In the ADC era, IHC is no longer merely a surrogate for genomic alteration, but a key tool for treatment-sequence selection. The ideal future pathway may therefore move beyond the linear model of “test first, then choose treatment” toward a mechanism-aligned model in which the intended drug platform determines the diagnostic requirement: genomic assessment for driver dependence, resistance-lineage profiling for signaling-network co-dependence, and protein quantification for delivery-based strategies [124].

Beyond current MET-directed agents, trispecific antibodies and MET-mediated immune-cell engagers are also showing early translational promise in advanced solid tumors. ABBV-303, for example, is a novel engager designed to redirect natural killer (NK) cells and CD8-positive T cells against MET-expressing tumors through a MET-binding arm, an NKG2D-binding arm, and an Fc region engaging CD16a [125]. It is currently being evaluated in a phase I study as monotherapy or in combination with the anti-programmed cell death protein 1 (PD-1) antibody budigalimab. BG-TIN187, an EGFR × MET × MET trispecific antibody, is likewise being tested in phase I studies across solid tumors as monotherapy and in combination regimens [126]. These emerging platforms suggest that future MET-directed therapy may extend beyond inhibition or payload delivery toward integrated receptor control and immune activation.

Limitations and future directions

Despite rapid progress, several important limitations should be acknowledged. First, many antibody-based strategies in MET-aberrant NSCLC remain at an early stage of clinical development. Amivantamab-based regimens have generated the most mature evidence among EGFR × MET BsAbs, particularly in broad post-osimertinib EGFR-mutant NSCLC [59,60], and telisotuzumab vedotin has achieved regulatory approval in an assay-defined, high c-MET protein-overexpressing population [18]. However, most other MET-related BsAbs, MET-ADCs, trispecific antibodies, and immune-cell engagers are still supported mainly by early-phase trials, small cohorts, preclinical studies, or preliminary reports [[127], [128], [129]]. Their optimal target populations, sequencing strategies, comparative roles relative to MET-TKIs, and long-term safety profiles remain insufficiently defined.

Second, predictive biomarkers for BsAbs and MET-ADCs are not fully validated. For EGFR × MET BsAbs, MET amplification, MET IHC, resistance-lineage features, and EGFR/MET co-dependence remain candidate enrichment factors rather than prospectively validated predictive markers. Importantly, exploratory biomarker signals should not be interpreted as established selection criteria unless validated in prospectively defined cohorts. For MET-ADCs, drug-specific IHC assessment of high membranous or membrane-enriched c-MET expression is clinically actionable in defined settings, but routine IHC does not directly measure receptor internalization, lysosomal trafficking, linker cleavage, intracellular payload release, bystander effect, or intratumoral heterogeneity. These delivery-related variables may influence ADC activity but remain difficult to quantify in routine clinical practice.

Third, MET testing remains limited by assay heterogeneity and disease evolution. Biomarker definitions differ across clinical trials, including thresholds for MET amplification, FISH criteria, NGS-based copy-number calling, IHC antibody clones, staining intensity, and positivity cutoffs. Spatial and temporal heterogeneity further limits the ability of a single tissue sample or plasma assay to capture the full MET landscape across the disease course. Therefore, the framework proposed in this review should be regarded as an evidence-layered conceptual model rather than a prospectively validated treatment-selection algorithm. Future studies should standardize biomarker definitions, validate platform-specific predictive markers, integrate tissue and liquid biopsy at progression, and define rational sequencing strategies for MET-TKIs, BsAbs, and MET-ADCs.

Summary

The key clinical implication of this review is that MET testing should be interpreted according to the intended therapeutic platform rather than as a single positive-or-negative result. For kinase-dependent disease, particularly METex14 and selected high-level focal MET amplification, MET-TKIs remain the most evidence-supported targeted approach. For EGFR-mutant NSCLC after osimertinib progression, EGFR × MET BsAbs have established clinical activity in broad post-osimertinib populations, but whether MET amplification, MET IHC, or integrated resistance-lineage features can identify preferential beneficiaries remains unresolved. For MET-ADCs, high c-MET protein expression defined by a drug-specific IHC assay currently provides the most clinically actionable selection strategy, whereas receptor internalization, intracellular trafficking, payload release, bystander effect, and intratumoral heterogeneity remain incompletely captured by routine testing.

These distinctions support a platform-aligned diagnostic pathway: DNA/RNA NGS for METex14 and resistance-lineage profiling, FISH or validated tissue NGS for high-level focal MET amplification, and drug-specific IHC for c-MET protein expression relevant to antibody-based therapy. The major unresolved questions are how to harmonize biomarker thresholds across assays, how to integrate tissue and plasma testing at progression, how to sequence MET-TKIs, BsAbs, and MET-ADCs after resistance, and how to prospectively validate predictive biomarkers for antibody-based strategies. Thus, precision therapy for MET-aberrant NSCLC is moving from event-based classification toward evidence-layered therapeutic matching, but prospective validation remains essential before this framework can function as a treatment-selection algorithm.

Availability of data and materials

Not applicable.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

All figures and tables in this manuscript are original and were generated by the authors from the underlying data. No material has been reproduced or adapted from published or unpublished sources, and the content has not appeared elsewhere.

Funding

This work was supported by grants from the Noncommunicable Chronic Diseases-National Science and Technology Major Project (No.2024ZD0521103), Tianjin Public Health Science and Technology Major Youth Project (No.24ZXGQSY00090), Science and Technology Project of Haihe Laboratory of Modern Chinese Medicine (No.GZY-KJS-2025–056), National Clinical Research Center for Chinese Medicine Acupuncture and Moxibustion Open Funding Project (No.NCRCOP2023007), Tianjin Key Research Projects in Traditional Chinese Medicine (No.2025011), Hebei Provincial Administration of Traditional Chinese Medicine Research Project (No.T2025083 & No.T2025059), Pilot Demonstration Project for the Inheritance and Innovative Development of Traditional Chinese Medicine in Nankai District, Tianjin (No.20240204019), Tianjin Municipal Key Disciplines and Key Specialties Construction Program in Medicine (TJYXZDXK-010A) and Special Fund for Clinical Research of Wu Jieping Medical Foundation (No.320.6750.2025–6–87).

CRediT authorship contribution statement

Haoyu Lu: Writing – original draft, Data curation, Conceptualization. Nana Zhao: Writing – original draft. Na Wang: Resources, Conceptualization. Yuan Meng: Formal analysis, Conceptualization. Yaning Luo: Project administration, Conceptualization. Xinyue Zhang: Supervision, Data curation. Xin Guan: Resources. Fanming Kong: Writing – review & editing.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.