Purpose:

MET tyrosine kinase inhibitors (TKIs) can achieve modest clinical outcomes in MET exon 14–altered lung cancers, likely secondary to primary resistance. Mechanisms of primary resistance remain poorly characterized and comprehensive proteomic analyses have not previously been performed.

Experimental Design:

We performed hybrid capture-based DNA sequencing, targeted RNA sequencing, cell-free DNA sequencing, selected reaction monitoring mass spectrometry (SRM-MS), and immunohistochemistry on patient samples of MET exon 14–altered lung cancers treated with a MET TKI. Associations between overall response rate (ORR), progression-free survival (PFS), and putative genomic alterations and MET protein expression were evaluated.

Results:

Seventy-five of 168 MET exon 14–altered lung cancers received a MET TKI. Previously undescribed (zygosity, clonality, whole-genome duplication) and known (copy-number focality, tumor mutational burden, mutation region/type) genomic factors were not associated with ORR/PFS (P > 0.05). In contrast, MET expression was associated with MET TKI benefit. Only cases with detectable MET expression by SRM-MS (N = 15) or immunochemistry (N = 22) responded to MET TKI therapy, and cancers with H-score ≥ 200 had a higher PFS than cancers below this cutoff (10.4 vs. 5.5 months, respectively; HR, 3.87; P = 0.02).

Conclusions:

In MET exon 14–altered cancers treated with a MET TKI, a comprehensive analysis of previously unknown and known genomic factors did not identify a genomic mechanism of primary resistance. Instead, MET expression correlated with benefit, suggesting the potential role of interrogating the proteome in addition to the genome in confirmatory prospective trials.

This article is featured in Highlights of This Issue, p. 657

Translational Relevance

MET inhibitors are active in MET exon 14–altered lung cancers. Crizotinib is listed in the National Cancer Center Guidelines and selective MET inhibitors are approved for this indication. For many of these agents, however, response and progression-free survival are modest compared with targeted therapy for other oncogene-driven lung cancers. We performed genomic and proteomic testing on pretreatment samples from patients treated with a MET tyrosine kinase inhibitor. Genomic factors outside of an activating MET exon 14 alteration by next-generation sequencing did not predict benefit with a MET tyrosine kinase inhibitor. In contrast, we show that MET expression by immunohistochemistry or mass spectrometry may predict benefit. These findings highlight that proteomic factors may modify response to targeted therapy in an oncogene-driven cancer.

MET exon 14 alterations occur in 4% of non–small cell lung cancers (NSCLCs). Most MET exon 14 alterations involve splice acceptor or donor sites flanking exon 14, which lead to exclusion of exon 14 that contains the Y1003 ubiquitin-binding site. Other alterations include MET fusions and Y1003 substitutions that recapitulate the biology of MET exon 14 splice site mutations. This family of MET alterations is presumed to lead to impaired MET protein ubiquitination, degradation, and increased oncogenic signaling (1).

MET tyrosine kinase inhibitors (TKIs) such as the multikinase inhibitor crizotinib are clinically active in MET exon 14–altered NSCLCs. In these tumors, crizotinib achieved an overall response rate (ORR) of 32% and a median progression-free survival (PFS) of 7.3 months (2). The more selective and potent MET inhibitors, tepotinib and capmatinib, are also active (3, 4) and were recently approved by Japanese regulatory authorities and the FDA, respectively, for treatment of MET exon 14–altered NSCLCs (5, 6). Unfortunately, response rates for many MET TKIs are modest relative to the activity of targeted therapy in other oncogene-driven lung cancers (e.g., those with sensitizing EGFR mutations) where ORRs are more consistently greater than 60%. For example, while capmatinib achieved an ORR of 68% in treatment-naïve patients, the ORR was only 41% in pretreated patients (4). The ORR with tepotinib was 48%, and response was similar regardless of prior therapy (3). The response to savolitinib, another selective MET inhibitor, was 48% (7). The fact that most of the reported response rates fall below 50% suggests that primary resistance to MET inhibition may represent a major issue.

We and others have reported that genomic profiling has been unsuccessful at identifying biomarkers of resistance to MET inhibitors in MET exon 14–altered lung cancers (2–4, 8). Variability in genomic factors, such as MET exon 14 splice region or mutation type, have not been associated with lack of clinical benefit. We hypothesized that primary therapeutic resistance is mediated by one or more previously unexplored pretreatment genomic or proteomic factors such as zygosity, clonality, mutational burden, or protein expression.

Clinical samples and study endpoints

All human tissues were obtained with Institutional Review Board approval and patients provided written informed consent. The study was conducted in accordance with the U.S. Common Rule. Patients were retrospectively included if they had advanced lung cancers with MET exon 14 alterations identified by DNA- or RNA-based next-generation sequencing (NGS) diagnosed between 2008 and 2018. Clinicopathologic characteristics, including age, sex, histology, and smoking history were collected. Patients were followed until June 2019. Response to therapy was evaluated using RECIST v1.1.

DNA- and RNA-based NGS

Tumor nucleic-acid testing was performed with targeted NGS of DNA using MSK-IMPACT or Foundation One (Foundation Medicine; ref. 9). Targeted NGS of tumor RNA was performed using an anchored multiplex PCR (MSK Solid Fusion Panel; ref. 10).

Clonality and zygosity analyses

Samples sequenced by MSK-IMPACT were analyzed for zygosity using FACETS (11). FACETS is an open-source integrated software tool designed to quantify several factors, including zygosity, tumor purity, ploidy, and clonal heterogeneity. Mutations called by the MSK-IMPACT pipeline were annotated for cancer cell fraction (CCF) using FACETS-suite (12); variants were classified as clonal if the upper-bound CCF was >85%. MET zygosity was labeled as high amplification if ≥6 total copies. Focal amplifications were called if the MET copy number was ≥6 and ≥3 more than the calculated copies of chromosome 7q. For other genetic alterations, amplification was defined as >2.0-fold change and deletion was defined as <−2.0-fold change.

Protein expression evaluation

Protein expression of pre-TKI tumor samples was performed using a targeted selected reaction monitoring-mass spectrometry panel (SRM-MS; NantOmics) as described previously (13). In brief, tissues from sectioned formalin-fixed paraffin-embedded (FFPE) blocks were placed onto DIRECTOR microdissection slides. These were deparaffinized and stained with hematoxylin, then microdissected and solubilized to tryptic peptides. These peptides were analyzed with TSQ Quantiva triple-quadrupole mass spectrometers (Thermo Fisher Scientific). Standardization for quantification and quality control for MET were reported previously. Positive MET expression by SRM-MS was previously defined as >150 amol/μg (13).

Peripheral hepatocyte growth factor Measurement

Peripheral blood was collected from select patients preprogression, on-progression, and postprogression of a MET TKI and plasma was separated and frozen at −80°C until further evaluation. Healthy control human plasma from five subjects was obtained from the New York Blood Center, aliquoted into individual cryovials, and frozen at −80°C until further evaluation. Peripheral hepatocyte growth factor (HGF) levels from each patient with available sample were quantified with an HGF ELISA Kit (SHG00B, R&D Systems; ref. 14). The data were normalized such that the mean level of five healthy control plasma samples were kept equivalent across experiments.

Statistical analysis

Mann–Whitney test was used to compare ORR between groups. For ORR in three or more groups, Kruskal–Wallis test was used. P values for PFS were determined using the log-rank test. P≤0.05 was considered significant. HRs and confidence intervals were generated using the Mantel–Haenszel test. Statistical analyses were performed with Prism 7 (GraphPad Software).

Among 168 patients with MET exon 14–altered NSCLCs, 167 were identified by DNA-based NGS (Supplementary Fig. S1). The remaining case was identified by an RNA-based NGS assay (MSK-Fusion). Of cases identified by DNA with sufficient additional tissue, subsequent RNA testing confirmed exon skipping in 97% (N = 97/100). RNA testing helped identify deep intronic or exonic MET exon 14 alterations that were difficult to interpret from DNA sequencing alone (Fig. 1). For instance, MSK-Fusion confirmed two cases of MET exon 14 skipping that were deep into intron 13 (MET c.2888-32_2888-29delinsG, c.2888-46_2888-22delCATGATAGCCGTCTTTAACAAGCTC) and one deep into exon 14 (c.2913_2962delCGATGCAAGAGTACACACTCCTCATTTGGATAGGCTTGTAAGTGCCCGAA). In contrast, MET exon 14 skipping was not confirmed in three samples, including two mutations deep within exon 14 (c.2993_3008delinsG and c.2967_2976delinsGGCAGTCCAA) and one mutation deep into intron 14 (c.3028+1221G>A). This underscores the complementary diagnostic utility of RNA analysis as a means of determining that select mutations detected by DNA analysis are unexpected to lead to splicing defects.

Figure 1.

Landscape of MET exon 14 alterations in lung cancer and sequencing by RNA-based anchored multiple PCR. For many patients, the likelihood of exon 14 being skipped in their cancer based on DNA-based sequencing was high (light-green). The rest of the cases were nominated for RNA-based confirmation via the MSK-Fusion panel. In patients, with sufficient tissue for MSK-Fusion testing, cases where exon 14 skipping was confirmed are shown in dark green, while those where exon 14 skipping was not observed in RNA are shown in red. Of note, one mutation not confirmed by RNA (c.3028+1221G > A) is deep into intron 14 and is not shown. Patients whose cancers were nominated for MSK-Fusion testing but did not have sufficient tissue for test completion are shown in gray. +, two patients; *, three patients.

Figure 1.

Landscape of MET exon 14 alterations in lung cancer and sequencing by RNA-based anchored multiple PCR. For many patients, the likelihood of exon 14 being skipped in their cancer based on DNA-based sequencing was high (light-green). The rest of the cases were nominated for RNA-based confirmation via the MSK-Fusion panel. In patients, with sufficient tissue for MSK-Fusion testing, cases where exon 14 skipping was confirmed are shown in dark green, while those where exon 14 skipping was not observed in RNA are shown in red. Of note, one mutation not confirmed by RNA (c.3028+1221G > A) is deep into intron 14 and is not shown. Patients whose cancers were nominated for MSK-Fusion testing but did not have sufficient tissue for test completion are shown in gray. +, two patients; *, three patients.

Close modal

Of these patients, 75 received a MET TKI. Demographics are summarized in Table 1. Most had lung adenocarcinoma (79%, N = 59/75) or sarcomatoid carcinoma (9%, N = 7/75). The majority of patients (88%, N = 66) received crizotinib. Most MET exon 14 alterations were detected by MSK-IMPACT (95%, N = 71/75); Foundation One detected the rest (5%, N = 4/75). The first major observation was that MET exon 14 alterations were highly heterogeneous (Fig. 2A and B) beyond previously described factors (e.g., splice site region/mutation type). Specifically, we examined novel genomic characteristics such as zygosity, clonality, and whole-genome duplication, in addition to concurrent MET amplification, focality, and tumor mutational burden (Fig. 2B), the frequencies of which are in Table 2. Interestingly, none of these factors correlated with ORR (P>0.05) or PFS (P>0.05; Table 2; Supplementary Figs. S2 and S3) with MET TKI therapy, although no responses were seen in subclonal tumors versus clonal ones (ORR 0% and 44%, respectively; P = 0.50). Tumor mutational burden ranged from 0.9 to 18 mutations per megabase (Fig. 2C) and did not affect MET TKI ORR (P = 0.44) or PFS (P = 0.80; Table 2; Supplementary Fig. S3B).

Table 1.

Clinicopathologic features of patients with advanced MET exon 14–altered lung cancers.

Patients treated with MET TKI (N = 75)
Age, median (range) 73 years (44–91 years) 
Sex, % (N 
 Female 52% (39) 
Cigarette smoking, % (N 
 Never 44% (33)  
 Former or current 56% (42) 
Histology, % (N 
 Adenocarcinoma 79% (59)  
 Sarcomatoid 9% (7)  
 Other 12% (9) 
Number of MET TKIs, % (N 
 1 75% (56)  
 2 or more 25% (19) 
First MET TKI received, % (N 
 Crizotinib 88% (66)  
 Cabozantinib 1% (1)  
 Tepotinib 11% (8) 
DNA-based NGS, % (N 
 MSK-IMPACT 95% (71)  
 Foundation One 5% (4) 
MET splice site region, % (N 
 Intron 13 acceptor 33% (25)  
 Intron 14 donor 60% (45)  
 Fusion 3% (2)  
 Other 3% (2)  
 Not detected by DNA-based NGS 1% (1) 
MET mutation type, % (N 
 Base substitution 48% (36)  
 Insertion/deletion 35% (26)  
 Large deletion (>35 bp) 8% (6)  
 Fusion 3% (2)  
 Other 3% (2)  
 Not detected by DNA-based NGS 4% (3) 
Patients treated with MET TKI (N = 75)
Age, median (range) 73 years (44–91 years) 
Sex, % (N 
 Female 52% (39) 
Cigarette smoking, % (N 
 Never 44% (33)  
 Former or current 56% (42) 
Histology, % (N 
 Adenocarcinoma 79% (59)  
 Sarcomatoid 9% (7)  
 Other 12% (9) 
Number of MET TKIs, % (N 
 1 75% (56)  
 2 or more 25% (19) 
First MET TKI received, % (N 
 Crizotinib 88% (66)  
 Cabozantinib 1% (1)  
 Tepotinib 11% (8) 
DNA-based NGS, % (N 
 MSK-IMPACT 95% (71)  
 Foundation One 5% (4) 
MET splice site region, % (N 
 Intron 13 acceptor 33% (25)  
 Intron 14 donor 60% (45)  
 Fusion 3% (2)  
 Other 3% (2)  
 Not detected by DNA-based NGS 1% (1) 
MET mutation type, % (N 
 Base substitution 48% (36)  
 Insertion/deletion 35% (26)  
 Large deletion (>35 bp) 8% (6)  
 Fusion 3% (2)  
 Other 3% (2)  
 Not detected by DNA-based NGS 4% (3) 

Note: Percentages in select demographic groups do not add up to 100% due to rounding.

Figure 2.

Pretreatment genomic features of MET exon 14–altered lung cancers and MET inhibitor activity. A, PFS and best objective response with MET inhibition. Each column is an individual patient/biopsy (N = 75). B, Splice site region, zygosity, whole-genome duplication, copy-number changes, focality, and clonality (clonal if > 80%). C, Tumor mutational burden (TMB). D, Sample origin, previous exposure to chemotherapy, and concurrent genomic alterations.

Figure 2.

Pretreatment genomic features of MET exon 14–altered lung cancers and MET inhibitor activity. A, PFS and best objective response with MET inhibition. Each column is an individual patient/biopsy (N = 75). B, Splice site region, zygosity, whole-genome duplication, copy-number changes, focality, and clonality (clonal if > 80%). C, Tumor mutational burden (TMB). D, Sample origin, previous exposure to chemotherapy, and concurrent genomic alterations.

Close modal
Table 2.

Objective response and PFS by MET genomic aberrations.

FactorsN (%)Responses (ORR %)P95% CIPFS (months)PHR95% CI
MET zygosity 
 WT copies 23 (31%) 7 (30%)  15%–51% 7.9    
 CN LOH 24 (32%) 9 (38%) 0.88 21%–57% 7.2 0.84   
 Heterozygous loss 5 (7%) 1 (20%)  2%–64%    
 Amplification 9 (12%) 3 (33%)  12%–65% 6.5    
 NE 14 (19%)  
Whole-genome duplication 
 Yes 17 (23%) 9 (53%) 0.14 31%–74% 7.3 0.66 1.2 0.6–2.3 
 No 50 (67%) 15 (30%)  19%–44% 8.8    
 NE 8 (11%)  
MET clonality 
 Subclonal 2 (3%) 0 (0%) 0.5 0%–71% 4.5 0.1 2.0 0.5–8.2 
 Clonal 43 (57%) 19 (44%)  30%–59% 8.9    
 NE 30 (40%)  
MET focality 
 Broad 2 (3%) 1 (50%) 1.00 9%–91% 11.0 0.51 0.6 0.1–3.3 
 Focal 7 (9%) 2 (29%)  8%–65% 5.5    
MET nonamplifieda 65 (87%)        
 NE 1 (10%)        
Tumor mutational burden 
 ≥ 4.5 mut/Mb 33 (44%) 14 (42%) 0.44 27%–59% 7.4 0.80 0.9 0.5–1.7 
 < 4.5 mut/Mb 33 (44%) 10 (30%)  17%–47% 7.2    
 NE 9 (12%)  
FactorsN (%)Responses (ORR %)P95% CIPFS (months)PHR95% CI
MET zygosity 
 WT copies 23 (31%) 7 (30%)  15%–51% 7.9    
 CN LOH 24 (32%) 9 (38%) 0.88 21%–57% 7.2 0.84   
 Heterozygous loss 5 (7%) 1 (20%)  2%–64%    
 Amplification 9 (12%) 3 (33%)  12%–65% 6.5    
 NE 14 (19%)  
Whole-genome duplication 
 Yes 17 (23%) 9 (53%) 0.14 31%–74% 7.3 0.66 1.2 0.6–2.3 
 No 50 (67%) 15 (30%)  19%–44% 8.8    
 NE 8 (11%)  
MET clonality 
 Subclonal 2 (3%) 0 (0%) 0.5 0%–71% 4.5 0.1 2.0 0.5–8.2 
 Clonal 43 (57%) 19 (44%)  30%–59% 8.9    
 NE 30 (40%)  
MET focality 
 Broad 2 (3%) 1 (50%) 1.00 9%–91% 11.0 0.51 0.6 0.1–3.3 
 Focal 7 (9%) 2 (29%)  8%–65% 5.5    
MET nonamplifieda 65 (87%)        
 NE 1 (10%)        
Tumor mutational burden 
 ≥ 4.5 mut/Mb 33 (44%) 14 (42%) 0.44 27%–59% 7.4 0.80 0.9 0.5–1.7 
 < 4.5 mut/Mb 33 (44%) 10 (30%)  17%–47% 7.2    
 NE 9 (12%)  

Note: Percentages may not add up to 100% due to rounding.

Abbreviations: CI, confidence interval; CN LOH, copy neutral loss of heterozygosity; HR, hazard ratio; Mb, megabase; mut, mutations; NE, not evaluable (no evaluable biomarker and/or not RECIST-evaluable); ORR, objective response rate; PFS, progression-free survival; WT, wild-type.

aMET focality can only be evaluated in MET-amplified cases.

Pretreatment concurrent genomic alterations were then assessed. These commonly involved TP53 (41%), MDM2 (29%), CDK4 (21%), TERT (19%), and CDKN2A (17%, Fig. 2D). Most alterations did not impact ORR or PFS (Supplementary Table S1; Supplementary Fig. S3C–S3G). Previous studies showed that acquired RAS activation was a mechanism of MET inhibitor resistance (15, 16). However, RAS or NF1 alterations did not affect response (20% vs. 39% in mutated vs. wild-type, respectively; P = 0.64; Supplementary Table S1; Figs. S3H and S4A) or PFS (7.5 vs. 7.3 months in mutated and wild-type, respectively; P = 0.36; Supplementary Table S1; Supplementary Fig. S3H). In contrast, the ORR was numerically lower in tumors with PI3KCA or PTEN comutations (13% vs. 41% in mutated vs. wild-type, respectively; P = 0.24; Supplementary Table S1; Supplementary Fig. S4B).

Conceptually, MET exon 14 alterations are thought to mediate oncogenesis by increasing MET expression. However, the degree of MET protein expression by targeted SRM-MS proteomics or immunohistochemistry (IHC) (SP44 antibody, Ventana) was heterogeneous. Surprisingly, only 67% (N = 10/15) of cases had detectable MET expression by SRM-MS in pretreatment biopsies (Fig. 3A) despite the high sensitivity of this assay (13). MET IHC H-scores (% cells multiplied by 1/2/3+ MET staining intensity) were: 0 (N = 1), 1–149 (N = 3), 150–199 (N = 5), and ≥200 (N = 13; Fig. 3B); MET expression by SRM-MS correlated with IHC H-scores (Spearman rho = 0.77; P = 0.008, Fig. 3C). High MET IHC expression (H-score ≥ 200) clustered with higher MET copy number (>3 copies; Supplementary Fig. S5A).

Figure 3.

MET expression in MET exon 14–altered lung cancers and acquired resistance. A, MET protein expression by mass spectrometry (SRM-MS). B, MET expression by IHC. C, Correlation of MET protein expression between SRM-MS and IHC. D,MET exon 14 skipping detected by RNA-based anchored multiplex PCR (blue, skipping present; open circle, insufficient tissue). E, Best response to MET inhibition by IHC expression. F, Best response to MET inhibition by protein expression (SRM-MS). G, PFS in cancers stratified by H-score.

Figure 3.

MET expression in MET exon 14–altered lung cancers and acquired resistance. A, MET protein expression by mass spectrometry (SRM-MS). B, MET expression by IHC. C, Correlation of MET protein expression between SRM-MS and IHC. D,MET exon 14 skipping detected by RNA-based anchored multiplex PCR (blue, skipping present; open circle, insufficient tissue). E, Best response to MET inhibition by IHC expression. F, Best response to MET inhibition by protein expression (SRM-MS). G, PFS in cancers stratified by H-score.

Close modal

Factors responsible for undetectable MET expression were explored. DNA-based NGS, MSK-Fusion, and SRM-MS were performed to determine whether DNA-level exon 14 alteration resulted in RNA-level loss of exon 14 and decreased MET protein expression. MET expression was undetectable in two tumors with DNA- and RNA-confirmed exon 14 skipping (Fig. 3D).

Response to MET inhibition was not observed in cancers with undetectable MET expression by SRM-MS. The ORR was 60% (N = 6 of 10) versus 0% (N = 0 of 5) in cases with detectable versus undetectable MET (P = 0.04; Fig. 3E). ORRs were higher (62%, N = 8 of 13) in cancers with an H-score ≥ 200 than with an H-score of 150–199 (25%, N = 1 of 4) or 1–149 (33%, N = 1 of 3; P = 0.39; Fig. 3F). No response was seen in the one case without MET expression by IHC. Depth of response was not associated with the degree of MET protein expression by SRM-MS or IHC. The median PFS was longer in tumors with an H-score ≥ 200 than an H-score <200 (10.4 versus 5.5 months, respectively; HR, 3.87; P = 0.02; Fig. 3G). The median PFS with crizotinib alone in tumors with an H-score ≥ 200 compared with an H-score < 200 was 6.7 vs. 5.4 months, respectively; HR, 2.65; P = 0.10), recognizing that the latter group included tumors that expressed MET, albeit at lower levels. No significant differences in clinicopathologic characteristics, including treatment with crizotinib or tepotinib, were seen in H-score ≥ 200 compared with H-score < 200 groups (Supplementary Table S2).

Cases with concomitant MET amplification had increased MET expression. Almost all cases with greater than neutral MET copy numbers by FACETS (copy number >2) had H-scores of ≥250 (Supplementary Fig. S5A). Because MET expression increased with MET copy-number gain in these cancers, we evaluated whether utilizing both MET copy number and H-score could further select for therapeutic benefit. MET exon 14–altered cases that either had low MET expression (H-score <200) with neutral MET copy numbers (copy number = 2) or lost a mutated or wild-type MET allele (copy number <2) were less likely to respond to MET TKI therapy (0%, N = 0/6; Supplementary Fig. S5B). Those that had gained MET alleles (copy number >2) or had an H-score ≥200 were more likely to respond (64%, 9 of 14; P = 0.0141). The median PFS was also significantly longer in cases with MET copy number >2 or MET IHC ≥200 (13.8 months, N = 14) compared with cases with MET copy number ≤2 and IHC <200 (4.6 months; HR, 10.5; 95% CI, 2.3–47.8; P = 0.003; Supplementary Fig. S5C).

We then evaluated whether low or high levels of MET expression by IHC were associated with the presence of coalterations. Tumors that expressed MET at low levels were more likely to have a concomitant mutation in the RAS or PIK3CA/PTEN pathways (44%; N = 4/9; Supplementary Table S3) than tumors that expressed MET at higher levels (0%; N = 0/13, P = 0.017). These results suggest that low MET-expressing tumors may be reliant on more than one oncogenic pathway and can bypass MET inhibition.

Although HGF is the ligand for MET (1), peripheral HGF was not associated with benefit from a MET inhibitor. HGF levels were obtained from 28 patients with MET exon14–altered lung cancers pretreatment (N = 9), on therapy (N = 15), and postprogression (N = 15; Supplementary Fig. S5D). These plasma levels were then pooled at each timepoint of therapy for statistical analysis. Compared with normal healthy controls (N = 5), HGF was elevated in patients treated with MET exon 14–altered lung cancers regardless of treatment with a MET TKI (P = 0.009). HGF levels did not differ across treatment phase (pre-TKI, on-TKI, postprogression; P = 0.91). While plasma HGF levels were higher in patients with MET exon 14–altered NSCLCs than healthy controls (P = 0.018), the degree of HGF elevation did not correlate with MET inhibitor response (P = 1.0; Supplementary Table S1).

Having explored primary resistance to MET inhibitor therapy, factors mediating acquired resistance were analyzed. Paired pre- and post-MET TKI biopsies were obtained. On-target acquired resistance (METD1228N or HGF amplification) was observed in 20% (3/15) of cases (Fig. 4A). Off-target resistance was found in 33% (5/15) of cases (Fig. 4A). No resistance mechanism was identified in the remaining cases (47%, 7/15). Plasma cell-free DNA collected postresistance (Fig. 4B) identified no on-target resistance; however, a hotspot KRAS mutation was found (9%, 1/11).

Figure 4.

A, On-/off-target mechanisms of acquired resistance in paired tumor biopsies. B, Resistance detected in postprogression circulating tumor DNA.

Figure 4.

A, On-/off-target mechanisms of acquired resistance in paired tumor biopsies. B, Resistance detected in postprogression circulating tumor DNA.

Close modal

Clonal heterogeneity was also seen in one acquired resistance sample (Supplementary Fig. S6). One clone (clone No. 1) had two MET exon 14 splice site mutations (c.3028G>C and c.2887+1G > A) prior to therapy and acquired a MET D1228N on-target mutation at progression. A second clone (clone No. 2) detected at progression was found to have a unique MET c.2888-22_2888-10delinsT exon 14 splice site mutation. The genomic profiles of each clone were distinct on NGS. These clones also had different histologies: clone No. 1 displayed predominantly a solid pattern whereas clone No. 2 displayed an acinar/lepidic pattern.

To date, trials of MET inhibitors in MET exon 14–altered NSCLCs have only performed genomic profiling to determine mechanisms of sensitivity or resistance, and these factors (MET exon 14 mutation region and type) did not predict benefit (2, 8). We evaluated additional genomic factors (zygosity, clonality, whole-genome duplication, and tumor mutational burden) and did not find a correlation with response or survival. High MET expression level by IHC or SRM-MS, in contrast, was associated with benefit from MET inhibition.

In identifying patients for this study, we noted the significant heterogeneity of MET exon 14 alterations. Some MET alterations were deep into the intron 13 or exon 14 and thus pathogenicity was difficult to identify from DNA sequencing alone. Of these six cases, RNA-based testing with MSK-Fusion confirmed MET exon skipping in only half of the cases. In addition, RNA-based testing uncovered MET exon 14 skipping in one case that was not found on DNA sequencing. This patient benefitted from crizotinib for 6.7 months. Previous studies have demonstrated the utility of using RNA sequencing in confirming MET exon 14 skipping or identifying these MET alterations in NSCLCs that were thought to be driver negative (10, 17). These results reflect that DNA-based approaches are limited in their ability to capture the breadth of MET exon 14 alterations that complementary RNA-based approaches can detect (17, 18).

We then evaluated whether pretreatment genomic coalterations could identify primary resistance to MET inhibition. While previous research showed that acquired RAS activation is a mechanism of MET inhibitor resistance, in this cohort, concomitant RAS/NF1 alterations did not affect response or survival (15, 16). Common coalterations, including TP53, MDM2, CDK4, TERT, and CDKN2A also did not influence these outcomes.

Beyond genomic heterogeneity, MET exon 14–altered NSCLCs were variable at the level of the proteome. Although MET exon 14 skipping leads to MET overexpression in preclinical models (19, 20), one study demonstrated that MET protein expression was heterogeneous in early-stage MET exon 14–altered lung cancers (21). Here, we show that MET protein is also heterogeneously expressed in metastatic MET exon 14–altered lung cancers. RNA transcripts were still measurable in cases with absent protein expression, implying the role of other regulatory factors (i.e., posttranslational modification) in mediating low MET protein levels.

Interestingly, pretreatment MET protein expression by IHC or SRM-MS identified primary resistance to MET inhibition. Lung cancers with undetectable MET expression did not respond to MET TKI therapy. Pending further exploration, these findings are likely generalizable to more selective TKIs such as capmatinib or tepotinib that, like crizotinib, bind the MET kinase domain in a type I fashion; the lack of MET expression on the cell surface is likely to represent a shared liability for these drugs (19). In addition, high MET expression by IHC (or detectable MET protein levels by SRM-MS) correlated with an improved response and longer survival. The addition of MET copy number to IHC appeared to further isolate poor responders. In tumors with H-scores < 200, responses were still seen. However, no responses were seen in those with both low MET expression (H-score < 200) and either a neutral MET copy number or loss of a mutated or wild-type MET allele. H-score appeared to increase with MET copy number in this context.

MET protein expression is likely a surrogate marker of MET dependency in MET exon 14–altered lung cancers. In addition to correlating with response, tumors with low MET protein expression were more likely to have concomitant alterations in the RAS or PI3KCA/PTEN pathways than tumors with high MET protein expression. These results are in line with recent preclinical work showing primary resistance to MET inhibition in MET exon 14–altered cell lines that also had KRAS or PI3KCA/PTEN coalterations (15, 16, 22). Thus, these low MET-expressing tumors may rely on other oncogenic pathways, bypassing the effect of MET inhibition alone. In tumors that are less reliant on MET, other regulatory factors (e.g., posttranslational modification) may contribute to the apparent absence of MET expression despite the presence of DNA- and RNA-level MET exon 14 skipping.

In the acquired resistance setting, recent studies have demonstrated that on-target mechanisms of acquired resistance to MET inhibition are uncommon in MET exon 14–altered NSCLCs (16, 23). In our cohort, 20% of paired tumor tissue cases acquired de novo METD1228N or HGF amplification at resistance. Off-target mechanisms of resistance are more common and appear to be predominantly mediated by the RAS pathway. Here, 33% of paired biopsy cases developed new EGFR, KRAS, or RASA1 alterations at acquired resistance.

Our data suggest that even a thorough analysis of both known and previously undescribed factors fails to strongly nominate a genomic factor that mediates MET TKI benefit beyond the singular identification of a MET exon 14 alteration. In contrast, this series exposes a potential need to perform prospective proteomic profiling as a supplement to genomic testing to identify patients with MET-expressing cancers that may be more poised to benefit from MET TKI therapy, recognizing that this was a single-center experience that did not feature a large proportion of patients diagnosed with widely used assays in the community. In clinical trials, pretreatment tumor biopsies should be tested for MET expression by both IHC, a practical assay that can be run in clinical laboratories, and a proteomic assay such as SRM-MS, with potentially increased reproducibility. In terms of limitations, it is important to recognize that this was a single-center experience that did not feature a large proportion of patients diagnosed with widely used assays in the community. Notably, the small sample size of this study is insufficient to support routine MET protein analysis as a standard-of-care test to exclude patients with MET exon 14–altered and non-MET-expressing cancers from MET TKI therapy, although it is reasonable to closely monitor patients whose cancers fit this phenotype for primary progression.

In summary, pretreatment genomic heterogeneity, including zygosity, clonality, or tumor mutational burden did not correlate with MET inhibitor benefit in MET exon 14–altered NSCLCs. Only undetectable MET protein expression resulted in decreased benefit from MET inhibition, a finding that should be validated in ongoing and future trials.

M. Offin reports other from PharmaMar, Novartis, and Targeted Oncology outside the submitted work. T. Hembrough reports other from NantOmics during the conduct of the study and other from AstraZeneca outside the submitted work; in addition, T. Hembrough has a patent for Multiple pending and issued to NantOmics. B.T. Li reports grants from Amgen, Genentech Roche, Lilly, AstraZeneca, Daiichi Sankyo, Ilumina, GRAIL, BioMedValley Discoveries, Boehringer Ingelheim, and Bolt Therapeutics; personal fees from Thermo Fisher Scientific and Mersana Therapeutics; grants and personal fees from Guardant Health and Hengrui Therapeutics; grants and nonfinancial support from MORE Health; nonfinancial support from Resolution Bioscience and Jiangsu Hengrui Medicine outside the submitted work; in addition, B.T. Li has a patent for US62/685,057 issued and a patent for US62/514,661 issued. C.M. Rudin reports personal fees from AbbVie, Amgen, AstraZeneca, Bicycle, Celgene, Genentech/Roche, Ipsen, Jansen, Jazz, Lilly/Loxo, Pfizer, PharmaMar, Syros, Vavotek, Bridge Medicines, Earli, and Harpoon outside the submitted work. M.G. Kris reports personal fees from AstraZeneca, Daiichi-Sankyo, Pfizer, Regeneron, and Sanofi-Genzyme and nonfinancial support from Genentech outside the submitted work. M.E. Arcila reports personal fees from Invivoscribe and Biocartis outside the submitted work. P.K. Paik reports personal fees and other from EMD Serono, Calithera, Takeda, AstraZeneca, Celgene, and Glaxo Smith Kline outside the submitted work. A. Zehir reports personal fees from Illumina outside the submitted work. A. Drilon reports other from Ignyta/Genentech/Roche, Loxo/Bayer/Lilly, Takeda/Ariad/Millenium, TP Therapeutics, AstraZeneca, Pfizer, Blueprint Medicines, Helsinn, Beigene, BergenBio, Hengrui Therapeutics, Exelixis, Tyra Biosciences, Verastem, MORE Health, Abbvie, 14ner/Elevation Oncology, Remedica Ltd., ArcherDX, Monopteros, Novartis, EMD Serono, and Melendi during the conduct of the study; other from GlaxoSmithKlein, Teva, Taiho, and PharmaMar outside the submitted work; other (food/beverage) from Merck, Puma, Merus, Boehringer Ingelheim; royalties from Wolters Kluwer; CME honoraria from Medscape, OncLive, PeerVoice, Physicians Education Resources, Targeted Oncology, Research to Practice, Axis, Peerview Institute, Paradigm Medical Communications, WebMD, and MJH Life Sciences. No disclosures were reported by the other authors.

R. Guo: Conceptualization, formal analysis, methodology, writing-original draft, writing-review and editing. M. Offin: Formal analysis, writing-original draft. A.R. Brannon: Conceptualization, formal analysis, writing-original draft, writing-review and editing. J. Chang: Investigation, methodology, writing-original draft, writing-review and editing. A. Chow: Writing-original draft, writing-review and editing. L. Delasos: Writing-original draft, writing-review and editing. J. Girshman: Writing-original draft, writing-review and editing. O. Wilkins: Writing-original draft, writing-review and editing. C.G. McCarthy: Writing-original draft, writing-review and editing. A. Makhnin: Writing-original draft, writing-review and editing. C. Falcon: Writing-original draft, writing-review and editing. K. Scott: Methodology, writing-original draft, writing-review and editing. Y. Tian: Methodology, writing-original draft, writing-review and editing. F. Cecchi: Methodology, writing-original draft, writing-review and editing. T. Hembrough: Methodology, writing-original draft, writing-review and editing. D. Alex: Investigation, methodology, writing-original draft, writing-review and editing. R. Shen: Data curation, formal analysis, writing-original draft, writing-review and editing. R. Benayed: Investigation, methodology, writing-original draft, writing-review and editing. B.T. Li: Writing-original draft, writing-review and editing. C.M. Rudin: Writing-original draft, writing-review and editing. M.G. Kris: Writing-original draft, writing-review and editing. M.E. Arcila: Investigation, methodology, writing-original draft, writing-review and editing. N. Rekhtman: Investigation, methodology, writing-original draft, writing-review and editing. P. Paik: Writing-original draft, writing-review and editing. A. Zehir: Writing-original draft, writing-review and editing. A. Drilon: Conceptualization, data curation, formal analysis, supervision, investigation, methodology, writing-original draft, writing-review and editing.

We thank Marc Ladanyi for his comments and critical reading of the article and Clare Wilhelm for his editorial support. We thank Emiliano Cocco and Maurizio Scaltriti for their scholarship and input in the manuscript.

This work was supported by the NCI at the NIH (grant numbers P30-CA-008748, P01-CA-129243, T32-CA-009207 to A. Chow).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Drilon
A
,
Cappuzzo
F
,
Ou
SI
,
Camidge
DR
. 
Targeting MET in lung cancer: will expectations finally be MET?
J Thorac Oncol
2017
;
12
:
15
26
.
2.
Drilon
A
,
Clark
JW
,
Weiss
J
,
Ou
SI
,
Camidge
DR
,
Solomon
BJ
, et al
Antitumor activity of crizotinib in lung cancers harboring a MET exon 14 alteration
.
Nat Med
2020
;
26
:
47
51
.
3.
Paik
PK
,
Felip
E
,
Veillon
R
,
Sakai
H
,
Cortot
AB
,
Garassino
MC
, et al
Tepotinib in non-small-cell lung cancer with MET exon 14 skipping mutations
.
N Engl J Med
2020
;
383
:
931
43
.
4.
Wolf
J
,
Seto
T
,
Han
JY
,
Reguart
N
,
Garon
EB
,
Groen
HJM
, et al
Capmatinib in MET exon 14-mutated or MET-amplified non-small-cell lung cancer
.
N Engl J Med
2020
;
383
:
944
57
.
5.
Merck KGaA
. 
TEPMETKO (Tepotinib) Approved in Japan for Advanced NSCLC with METex14 Skipping Alterations
; 
2020
.
Available from
: https://www.emdgroup.com/en/news/tepotinib-25-03-2020.html.
6.
U.S. Food and Drug Administration
. 
FDA Approves First Targeted Therapy to Treat Aggressive Form of Lung Cancer.
; 
2020
.
Available from
: https://www.fda.gov/news-events/press-announcements/fda-approves-first-targeted-therapy-treat-aggressive-form-lung-cancer.
7.
Lu
S
,
Fang
J
,
Li
X
,
Cao
L
,
Zhou
J
,
Guo
Q
, et al
Phase II study of savolitinib in patients (pts) with pulmonary sarcomatoid carcinoma (PSC) and other types of non-small cell lung cancer (NSCLC) harboring MET exon 14 skipping mutations (METex14+)
.
J Clin Oncol
38
:
15s
, 
2020
(
suppl; abstr 9519
).
8.
Lu
S
,
Fang
J
,
Cao
L
,
Li
X
,
Guo
Q
,
Zhou
J
, et al
Preliminary efficacy and safety results of savolitinib treating patients with pulmonary sarcomatoid carcinoma (PSC) and other types of non-small cell lung cancer (NSCLC) harboring MET exon 14 skipping mutations [abstract]
.
In
:
Proceedings of the American Association for Cancer Research Annual Meeting 2019
; 
2019 Mar 29–Apr 3
;
Atlanta, GA. Philadelphia (PA)
:
AACR
;
Cancer Res 2019;79 (13 Suppl): Abstract nr CT031
.
9.
Cheng
DT
,
Mitchell
TN
,
Zehir
A
,
Shah
RH
,
Benayed
R
,
Syed
A
, et al
Memorial sloan kettering-integrated mutation profiling of actionable cancer targets (MSK-IMPACT): a hybridization capture-based next-generation sequencing clinical assay for solid tumor molecular oncology
.
J Mol Diagn
2015
;
17
:
251
64
.
10.
Benayed
R
,
Offin
M
,
Mullaney
K
,
Sukhadia
P
,
Rios
K
,
Desmeules
P
, et al
High yield of RNA sequencing for targetable kinase fusions in lung adenocarcinomas with no mitogenic driver alteration detected by DNA sequencing and low tumor mutation burden
.
Clin Cancer Res
2019
;
25
:
4712
22
.
11.
Shen
R
,
Seshan
VE
. 
FACETS: allele-specific copy number and clonal heterogeneity analysis tool for high-throughput DNA sequencing
.
Nucleic Acids Res
2016
;
44
:
e131
.
12.
Bielski
CM
,
Donoghue
MTA
,
Gadiya
M
,
Hanrahan
AJ
,
Won
HH
,
Chang
MT
, et al
Widespread selection for oncogenic mutant allele imbalance in cancer
.
Cancer Cell
2018
;
34
:
852
62.e4
.
13.
Catenacci
DV
,
Liao
WL
,
Thyparambil
S
,
Henderson
L
,
Xu
P
,
Zhao
L
, et al
Absolute quantitation of Met using mass spectrometry for clinical application: assay precision, stability, and correlation with MET gene amplification in FFPE tumor tissue
.
PLoS One
2014
;
9
:
e100586
.
14.
Llovet
JM
,
Pena
CE
,
Lathia
CD
,
Shan
M
,
Meinhardt
G
,
Bruix
J
, et al
Plasma biomarkers as predictors of outcome in patients with advanced hepatocellular carcinoma
.
Clin Cancer Res
2012
;
18
:
2290
300
.
15.
Suzawa
K
,
Offin
M
,
Lu
D
,
Kurzatkowski
C
,
Vojnic
M
,
Smith
RS
, et al
Activation of KRAS mediates resistance to targeted therapy in MET exon 14-mutant non-small cell lung cancer
.
Clin Cancer Res
2019
;
25
:
1248
60
.
16.
Rotow
JK
,
Gui
P
,
Wu
W
,
Raymond
VM
,
Lanman
RB
,
Kaye
FJ
, et al
Co-occurring alterations in the RAS-MAPK pathway limit response to MET inhibitor treatment in MET exon 14 skipping mutation-positive lung cancer
.
Clin Cancer Res
2020
;
26
:
439
49
.
17.
Davies
KD
,
Lomboy
A
,
Lawrence
CA
,
Yourshaw
M
,
Bocsi
GT
,
Camidge
DR
, et al
DNA-based versus RNA-based detection of MET exon 14 skipping events in lung cancer
.
J Thorac Oncol
2019
;
14
:
737
41
.
18.
Poirot
B
,
Doucet
L
,
Benhenda
S
,
Champ
J
,
Meignin
V
,
Lehmann-Che
J
. 
MET exon 14 alterations and new resistance mutations to tyrosine kinase inhibitors: risk of inadequate detection with current amplicon-based NGS panels
.
J Thorac Oncol
2017
;
12
:
1582
7
.
19.
Kong-Beltran
M
,
Seshagiri
S
,
Zha
J
,
Zhu
W
,
Bhawe
K
,
Mendoza
N
, et al
Somatic mutations lead to an oncogenic deletion of met in lung cancer
.
Cancer Res
2006
;
66
:
283
9
.
20.
Abella
JV
,
Peschard
P
,
Naujokas
MA
,
Lin
T
,
Saucier
C
,
Urbe
S
, et al
Met/Hepatocyte growth factor receptor ubiquitination suppresses transformation and is required for Hrs phosphorylation
.
Mol Cell Biol
2005
;
25
:
9632
45
.
21.
Awad
MM
,
Oxnard
GR
,
Jackman
DM
,
Savukoski
DO
,
Hall
D
,
Shivdasani
P
, et al
MET exon 14 mutations in non-small-cell lung cancer are associated with advanced age and stage-dependent MET genomic amplification and c-met overexpression
.
J Clin Oncol
2016
;
34
:
721
30
.
22.
Jamme
P
,
Fernandes
M
,
Copin
MC
,
Descarpentries
C
,
Escande
F
,
Morabito
A
, et al
Alterations in the PI3K pathway drive resistance to MET inhibitors in NSCLC harboring MET exon 14 skipping mutations
.
J Thorac Oncol
2020
;
15
:
741
51
.
23.
Recondo
G
,
Bahcall
M
,
Spurr
LF
,
Che
J
,
Ricciuti
B
,
Leonardi
GC
, et al
Molecular mechanisms of acquired resistance to MET tyrosine kinase inhibitors in patients with MET exon 14 mutant NSCLC
.
Clin Cancer Res
2020
;
26
:
2615
25
.