Comprehensive Genomic Profiling of Circulating Cell-Free DNA Distinguishes Focal MET Amplification from Aneuploidy in Diverse Advanced Cancers

Abstract

Amplification (amp) of MET can be observed in cases of focal gene copy number gain, such as MET-driven amp, or with a gain of chromosome 7, such as aneuploidy. Several studies have shown that only high-level focal MET amp (MET/CEP7 ratio ≥5) is oncogenic, with such tumors responding to targeted therapy. However, there are few reports on how to distinguish between focal amplification and aneuploidy using next-generation sequencing (NGS). A total of 1025 patients with advanced solid tumors (typically pre-treated) were tested with a non-invasive comprehensive cfDNA NGS panel (Guardant360) from July 2014 to June 2019. Since bioinformatics upgrades of Guardant360 were undergoing in September 2018, focal MET amp was determined by our independent algorithm using the cohorts tested before September 2018 (291 patients), and validation was performed in the remaining cohort (734 patients). MET alterations (alts) associated with aberrant signaling were found in 110 patients (10.7%) among nine different cancer types, most commonly in non-small cell (12.2%, 62/510) and small cell (33.3%, 3/9) lung cancers, gastroesophageal cancer (19.4%, 7/36), and prostate adenocarcinoma (15.6%; 5/32). Among 291 patients tested before September 2018, 37 (12.7%) had MET alts. Among these, 24 (64.9%) had amps, 5 (13.5%) had exon 14 skipping, and 13 (35.1%) had single nucleotide variants (SNVs). Co-alterations, such as amp + SNVs, were found in four samples (10.8%). Among 24 MET amps, 29.2% (7/24) were focal according to our algorithm. MET copy number was significantly higher with focal amp compared to non-focal amp (mean copy number 3.26 vs. 2.44, respectively, p = 0.00304). In 734 patients tested after September 2018, our definition of focal MET amp was detected in 4.2% (31/734). Overall, focal amplification based on our algorithm was 3.7% (=38/1025). This study describes an approach to distinguish focal and non-focal MET amplification using comprehensive genomic profiling of cfDNA in advanced cancer patients. Focal MET amp accounted for ~30% of all MET amp, which was found in 3.7% of patients with diverse cancers and was associated with a higher plasma copy number. Clinical studies are warranted to assess the clinical utility of targeted therapies for tumors with focal MET amplification detected by NGS of cfDNA.

1. Introduction

MET, also known as c-MET, is a receptor tyrosine kinase that regulates many physiological processes, including proliferation, scattering, morphogenesis, and survival. The binding of its ligand, hepatocyte growth factor (HGF), to the MET receptor leads to the activation of several downstream signaling pathways, including RAS-ERK/MAPK, PI3K-AKT, or PLCgamma-PKC [1]. MET is a proto-oncogene that encodes MET, and activating MET mutations have been reported in diverse carcinomas [2].

MET alterations (alts), such as MET amplification (amp), exon 14 skipping, and single nucleotide variants (SNVs), can be oncogenic drivers. Furthermore, activating MET alts may also be acquired as a mechanism of resistance to tyrosine kinase inhibitors. MET exon 14 skipping has been described as a driver mutation in non-small cell lung cancer (NSCLC) [3], and approximately 3% of NSCLC patients are reported to have MET exon 14 skipping [4]. Recently, MET amp has also been described as oncogenic, and studies using fluorescence in situ hybridization (FISH) have found MET amp in up to ~5% of patients with NSCLC or gastric adenocarcinoma [5]. However, a phase I study reported that patients harboring MET amp did not respond to therapy with a c-MET inhibitor [2].

MET amp can occur from focal gene copy number gain of the MET gene alone (focal amplification), co-amplification with adjacent genes in chromosome 7q such as CDK6 and BRAF (non-focal amplification), or gain of chromosome 7q copy number (e.g., aneuploidy). Several studies have shown that only high-level focal MET amp (MET/CEP7 ratio ≥5), but not non-focal or lower levels of focal MET amp, are oncogenic and respond to targeted therapy [6,7,8]. MET inhibitor-sensitive lung cancers with high-level MET amp have been reported in the absence of other sensitizing MET alts, such as exon 14 skipping, particularly among those with higher MET to chromosome 7 ratios. Therefore, it is critical to distinguish between focal amplification and non-focal amplification. A typical companion diagnostic test interrogates only the target gene but may not include adjacent genes; therefore, in such cases, the distinction between MET focal amplification and non-focal amplification is not feasible. There are few reports on how to distinguish between focal amplification and aneuploidy using next-generation sequencing (NGS).

NGS of cell-free plasma DNA (cfDNA) from patients with advanced cancers is a validated, non-invasive technique that can be used for comprehensive genomic profiling (CGP) of tumor-derived DNA. It can detect gene mutations with a relatively good correlation with tissue sequencing. In this study, we analyzed MET alterations associated with aberrant MET signaling in 1025 patients with advanced solid tumors by using a comprehensive cfDNA NGS panel. We examined the pattern of focal and non-focal MET amps and established a way to investigate the prevalence of this potentially targetable alteration on NGS panels.

2. Materials and Methods

2.1. Patients

We reviewed the results of CGP from plasma of 1025 mostly pre-treated patients with advanced solid tumors. CGP of cfDNA was performed as a routine clinical practice between July 2014 and June 2019 from patients seen at institutes in Asia, Middle East, and Africa. Results from patients participating in prospective clinical trials were excluded. De-identified results were shared by Guardant Health Japan, Corp. Ethics review committee approval was granted by the institutional review board at Tokyo Medical and Dental University (G2020-021).

2.2. Next-Generation Sequencing

All 1025 patients were tested with Guardant360^®, a comprehensive cfDNA NGS panel performed at Guardant Health, Inc. (Redwood City, California), a Clinical Laboratory Improvement Amendment (CLIA)-certified and College of American Pathologists (CAP)-accredited clinical laboratory. This assay interrogates single nucleotide substitutions, indels, amplification, and gene fusions in selected genes. During the study period, several versions of the assay were employed. In each version, MET amplification, MET single nucleotide substitutions, MET exon 14 skipping, and amplifications of BRAF, CDK6, and EGFR were included. Prior to September 2018, Guardant360 reported MET amplification regardless of focality; after bioinformatics upgrades, only focal MET amplification was reported. Therefore, for purposes of the present analysis, unfiltered Guardant360 amplification results for MET, BRAF, CDK6, and EGFR were provided for all samples. Copy number variants were scored based on the baseline copy number pooled in previous CGP data at Guardant Health, Inc. The result of amplification was reported as observed copy number.

2.3. Determination of Focal Amplification

In chromosome 7, four genes were examined in Guardant360: EGFR in 7p11.2, CDK6 in 7q21.2, MET in 7q31.2, and BRAF in 7q34. (Figure 1) Focal amplification was defined as MET amplification without aneuploidy or an increase in the chromosome copy number itself. In this case, the MET gene was considered amplified if there were no co-amplification of adjacent genes, such as CDK6 or BRAF. On the other hand, MET non-focal amplification was defined as MET copy number increase associated with aneuploidy, in which MET copy number was increased together with either CDK6 and/or BRAF. For example, a sample with MET amplification only would be categorized as MET focal amplification. A sample with amplification of MET and EGFR without CDK6 amplification would be categorized as MET focal amplification because CDK6 is located between MET and EGFR, and aneuploidy could not occur without increasing the copy numbers of all three genes together. A sample with co-amplification of MET and EGFR would also be defined as focal amplification when both CDK6 and BRAF genes were not co-amplified, given that EGFR is in a distant location (chromosome 7p). (Figure 2 and Figure S1).

In this study, additional analysis to distinguish focal and non-focal amplification was performed at an independent academic institution (Tokyo Medical and Dental University) using an original algorithm not used in the standard Guardant360 workflow. We divided the whole (1025 patients) into two cohorts; the cohorts tested before September 2018 (291 patients) was used to define focal amplification, and validation was performed in the remaining cohort (734 patients). Using the former cohort, we determined the algorithm to describe focal amplification as follows.

(a)

MET copy number ≥2.2.

(b)

MET gene is amplified without co-amplification of CDK6 and BRAF. Co-amplification status was defined as “increased together” when the copy number of other gene (CDK6 or BRAF) ≥2.2, and the difference with MET amplification is within +/−0.5.

(c)

MET amplification that satisfies both (a) and (b) is defined as focal.

2.4. Statistical and Outcome Analysis

Patient characteristics were summarized using descriptive statistics. Medians and respective 95% confidence intervals and range were calculated whenever possible. T-test or Mann–Whitney U test was used for numerical data, and Fisher’s exact test or chi-square test was used for categorical data. We set the accepted level of significance at 0.05 (p-value). Statistical analysis was performed using EZR software (version 1.5.4 or later) and JMP software (version 14.2). For the validation cohort, positive percent agreement, negative percent agreement, positive predictive value, and negative predictive value were assessed using the updated Guardant360 bioinformatics pipeline results as the gold standard.

3. Results

3.1. Patient Demographic Characteristics

In 1025 total samples, the mean age at the time of testing was 62.4 years (CI 95%, 61.6–63.3). The cohort included 482 men (47.0%) and 543 women (53.0%). The most common cancer types were NSCLC (49.8%, n = 510), non-colorectal and non-gastroesophageal gastrointestinal cancers (“other GI cancers”) (15.4%, n = 158), breast cancer (9.0%, n = 92), and colorectal cancer (7.5%, n = 77). There was no statistically significant imbalance of diagnosis between the MET alteration and MET non-alteration groups (

Table 1. Demographic comparison of 1025 patients with or without MET alt.

Characteristics	Total Patients, n = 1025 (100%)	MET Alt Not Detected, n = 915 (89.3%)	MET Alt Detected, n = 110 (10.7%)	p-Values **
Mean age at time of testing, y	62.4	62.4	62.4	0.966
(CI 95%), n = 880 known	(61.6–63.3)	(61.5–63.3)	(59.4–65.3)
Gender				0.00234
Men	482 (47.0%)	415 (86.1%)	67 (13.9%)
Women	543 (53.0%)	500 (92.1%)	43 (7.9%)
Type of cancer				0.0863
NSCLC	510 (49.8%)	448 (87.8%)	62 (12.2%)
SCLC	9 (0.9%)	6 (66.7%)	3 (33.3%)
Colorectal	77 (7.5%)	71 (92.2%)	6 (7.8%)
Gastroesophageal	36 (3.5%)	29 (80.6%)	7 (19.4%)
Other gastrointestinal	158 (15.4%)	147 (94.2%)	9 (5.8%)
Breast	92 (9.0%)	84 (91.3%)	8 (8.7%)
Unknown primary	20 (2.0%)	18 (94.7%)	1 (5.3%)
Gynecologic	46 (4.5%)	43 (93.5%)	3 (6.5%)
Prostate	32 (3.1%)	27 (84.4%)	5 (15.6%)
Other	45 (4.4%)	42 (87.5%)	6 (12.5%) *

* Includes patients with neuroblastoma (n = 2), neuroendocrine tumor (n = 2), melanoma (n = 1), and other (n = 1). ** p-values were calculated using the t-test for linear variables (age at time of testing), Fisher’s exact test (gender), and the chi-square test (type of cancer) for categorical variables.

).

3.2. MET Alterations and Associations with Patient Characteristics

MET alts were defined as MET amplification (amp), exon 14 skipping, and non-synonymous single nucleotide variants (SNVs). In total, MET alts were detected by Guardant360 in 110 of 1025 patients (10.7%), which was similar to a previous report [9].

MET alts were commonly found in SCLC 33.3% (3/9), gastroesophageal cancer 19.4% (7/36), prostate cancer 15.6% (5/32), and NSCLC 12.2% (62/510) (Table 1).

MET alts were found more frequently in men than in women (13.9% vs. 7.9%, p = 0.00234) (Table 1). This may also be due to the fact that MET mutations tended to be more common in prostate cancer (15.6%) than in breast (8.7%) and gynecological cancer (6.5%), although the number of cases of each was small. When analyzed by cancer type, there were significantly more MET alts in males than in females with non-colorectal and other GI cancers (p = 0.00478). However, no gender differences were observed for the frequency of MET alterations in other cancer types.

3.3. MET Alteration Types and Focal vs. Non-Focal MET Amp

Among 291 patients tested before Sep 2018, 37 (12.7%) had MET alts, according to Guardant360. Among these, 24 (64.9%) had amps, 5 (13.5%) had exon 14 skipping, and 13 (35.1%) had SNVs. Co-alterations, such as amplification + SNVs, were found in four samples (10.8%) (Figure 3).

Among the 24 samples with MET amp, we found several patterns of gene amplification on chromosome 7 (

Table 2. Patterns of copy number of four genes on chromosome 7.

Case	EGFR	CDK6	MET	BRAF	Feature	Focal/Non-Focal
1	2.2 *	2.2	2.1	2.3	MET copy number is lower than 2.2	non-focal
2	2.2	2.1	2.1	2.3		non-focal
3	2.2	2.2	2.1	2.0		non-focal
4	3.1	2.9	2.8	3	MET copy number is increased together with either CDK6 and/or BRAF	non-focal
5	2.9	2.8	2.6	2.8		non-focal
6	2.6	2.6	2.6	2.5		non-focal
7	2.2	2.3	2.2	2.2		non-focal
8	2.8	3.6	3.3	3.6		non-focal
9	2.4	2.3	2.6	2.5		non-focal
10	2.0	2.6	2.8	2.6		non-focal
11	3	2.3	2.2	2.2		non-focal
12	3.3	2.3	2.3	2.3		non-focal
13	80	2.6	2.5	2.6		non-focal
14	2.0	2.5	2.3	2.4		non-focal
15	2.0	2.6	2.5	2.5		non-focal
16	2.1	2.0	2.3	2.2		non-focal
17	2.0	2.2	2.2	2.0		non-focal
18	3.2	2.0	3.3	2.0	MET gene is amplified without co-amplification of CDK6 and BRAF	focal
19	2.5	2.0	2.5	2.0		focal
20	2.0	2.0	2.6	2.0		focal
21	2.0	2.5	3.7	2.0		focal
22	2.0	2.0	2.2	2.0		focal
23	2.2	2.3	6.2	N/A **		focal
24	2.0	2.0	2.3	2.0		focal

* Copy numbers 2.2 and above are shown in red. ** N/A represents data not available. Algorithm to determine focal amp: (a) MET copy number ≥ 2.2. (b) MET gene is amplified without co-amplification of CDK6 and BRAF. Co-amplification status was defined as “increased together” when the copy number of other genes (CDK6 or BRAF) ≥ 2.2, and the difference with MET amplification is within +/−0.5. (c) MET amplification that satisfies both (a) and (b) is defined as focal. Cases 18 and 19 are considered focal because co-amplification of MET occurred only with the EGFR gene but not with CDK6 and BRAF genes. Cases 21 and 23 were defined as focal amplification because the difference in amplification magnitude between MET and CDK6 was more than 0.5.

). Focal amplification was defined as MET amplification without aneuploidy or an increase in the chromosome copy number itself, and using this cohort as a test set, we established an algorithm for determining focal MET amplification, as described in the Methods. Using this algorithm, we found that seven cases (29.2%) were focal, and 17 (70.8%) were non-focal (Table 2, Figure 3). In the cases of focal amp, the majority (71.4%) had only MET amplification without amplifications of any of the other three genes from chromosome 7 included in the assay, and the minority of patients (28.6%) had co-amplification of MET and EGFR. In non-focal MET amplification, 82.4% of cases had co-amplification of MET, CDK6, and BRAF, suggesting aneuploidy. Other cases had co-amplifications of MET with either CDK6 or BRAF.

We examined the features of focal and non-focal MET amplifications. MET copy number was significantly higher with focal amp compared to non-focal amp (mean copy number 3.26 vs. 2.44, respectively, p = 0.00304). (Figure 4) In the case of focal MET amplification, CDK6 and BRAF copy numbers were lower than the non-focal MET amplification group (CDK6: 2.11 copies vs. 2.48 copies, respectively, p = 0.0278; BRAF: 2.0 copies vs. 2.47 copies, respectively, p = 0.0086). EGFR copy number was not different between the focal MET amp and non-focal MET amp groups (2.27 copies vs. 2.44 copies, respectively; p = 0.421). This illustrated that a higher copy number in MET and low copy numbers in CDK6 and BRAF, but not EGFR, were associated with focal MET amplification.

Next, we investigated the proportions of focal vs. non-focal amplifications among different cancer types (

Table 3. Focal vs. non-focal MET amplification by cancer type in training set (n = 291).

Cancer Type (n Patients, %w/AMP)	Patients w/Focal MET Amp in ≥1 Sample	Patients w/Only Non-Focal MET Amp	Proportion Focal	p-Value
NSCLC (144, 6.9%)	4 (2.8%)	6 (4.2%)	4/10 (40.0%)	0.140
Gastrointestinal (69, 5.8%)	1 (1.4%)	3 (4.3%)	1/4 (25.0%)
Breast (26, 15.4%)	2 (7.7%)	2 (7.7%)	2/4 (50.0%)
Other (52, 11.5%)	0 (0.0%)	6 (11.5%)	0/6 (0%)
Overall (291, 8.2%)	7 (2.4%)	17 (5.8%)	7/24 (29.2%)

). In the NSCLC cohort, 10 out of 144 patients (6.9%) had MET amplification. Among 10 patients with MET amplification, four had focal amplification, and six had non-focal amplification. The proportion of focal amplification in total MET amplification in the NSCLC cohort was 40%. The proportions of MET focal amplifications in gastrointestinal, breast, and other cancers were 25%, 50%, and 0%, respectively. Focal amplification was found in 7 out of 291 samples (2.4%), and the proportion of focal amplification among MET amp was 7 out of 24 samples (29.2%). However, the number of patients who had MET focal amplification was too small to elucidate any meaningful statistical conclusion. Further investigation is warranted in a larger patient cohort.

Next, we examined whether there were differences in the coexistence of EGFR driver mutations between focal and non-focal MET amps in NSCLC patients. Among 17 NSCLC patients with MET alts, 10 had MET amp. Among them, four had focal MET amp, and six had non-focal. Co-occurrence of EGFR driver mutations was found in five MET amp patients, two with a focal amp. EGFR T790M was also present in both patients with focal MET amp and one patient with a non-focal amp. This is suggestive of acquired resistance to third-generation EGFR TKI [10], although the treatment histories of these patients are unknown.

Lastly, we compared our focal amplification definition to the amplification determined by a CLIA (Clinical Laboratory Improvement Amendments) certified laboratory at Guardant Health, Inc using an independent cohort. (original data; Table S1) In 734 patients tested after Sep/2018, our definition of focal amplification was detected in 31 out of 734 patients (4.2%). Guardant360 reported MET amplification in 30 patients (4.1%). The positive percent agreement (PPA; focal in both TMDU and G360/focal in G360), negative percent agreement (NPA; non-focal in both TMDU and G360/non-focal in G360), positive predictive value (PPV; focal in both TMDU and G360/focal in TMDU), and negative predictive value (NPV; non-focal in both TMDU and G360/non-focal in TMDU) were 83.3% (25/30), 98.7% (471/477), 80.6% (25/31), and 98.9% (471/476), respectively. Using our algorithm in the complete study cohort (test and validation sets combined, n = 1025), focal amplification of MET was found in 38 patients (3.7%).

4. Discussion

We describe a method for determining focal MET amplification using unfiltered data from comprehensive genomic profiling of cfDNA in advanced cancer patients. Focal MET amp was found in 3.7% (38/1025) of patients with diverse cancers and accounted for only ~30% of all MET amps. This distinction has clinical importance. Focal MET amplification is likely a driver alteration and, therefore, a therapeutic target. Non-focal amplification or aneuploidy is unlikely to be a driver alteration and, therefore, not an ideal therapeutic target.

In previous studies using tumor tissue, MET amp has been defined using the ratio of MET to CEP7 by FISH, and MET amps were found in up to ~5% of patients with NSCLC or gastric cancer [5]. In a phase I study, in which the definition of MET amp was as follows: “the MET: CEP7 signal ratio was ≥2.0 or when this ratio was <2.0, but there were >20 copies of MET signals in more than 10% of the tumor nuclei counted,” MET amp was detected in 2.5% of patients with advanced solid tumors in different cancer types. However, there was no significant difference in sensitivity to MET inhibitors between the MET amp group and the not amplified group [2]. Therefore, a higher focal MET amp was proposed as a threshold, and a correlation with the therapeutic effect was seen in patients whose MET/CEP was 5 or more [6,7,8]. This suggests that it is necessary to find a group with MET-only amplification, without aneuploidy of chromosome 7, to enhance the therapeutic effect of MET inhibitors. Our study showed that a higher MET copy number in plasma was more likely to be due to focal amplification.

Furthermore, this result indicates that it is necessary to interrogate not only MET but also other adjacent genes on the same chromosome when designing companion diagnostic tests for MET amplification. Currently, a typical biomarker for amplification investigates only the target gene and does not include adjacent genes. However, this study illustrated that focal MET amplification was observed in only approximately 30% of “MET amplification,” and the rest were likely aneuploidy of chromosome 7. This distinction is critical for selecting a patient population in which MET amp drives cancer and may therefore be a target population for treatment with MET inhibitors. The idea of considering adjacent genes to identify focal gene amplification might be critical not only for MET but also for FGFR1 and FGFR2, for which there are emerging targeted therapeutics.

However, the algorithm needs to be optimized. We found several patterns of co-amplification of three genes on chromosome 7 (BRAF, CDK6, and EGFR) in this study and defined aneuploidy as having CDK6 and/or BRAF amplified to the same degree. We defined the copy number of CDK6 or BRAF “increasing together” as within ±0.5 of the MET copy number, but the condition of less than 0.5 may not have been necessary because even if the copy number of CDK6 and BRAF is much higher (>0.5) than that of MET, it is still considered aneuploidy when all of them are high. For example, if our algorithm had excluded this rule, the NPA, PPV, and NPV would have been improved to 99.4% (from 98.7%), 89.3% (from 80.6%), and 99.0% (from 98.9%), respectively (PPA would have been same, 83.3%). Another point is whether or not to include EGFR in the definition. Unlike MET, CDK6, and BRAF genes on 7q, EGFR on 7p often behave independently, with apparently different values of copy number. Therefore, we removed EGFR from the definition of aneuploidy in this study. Further research is needed to validate this definition and its correlation with the clinical utility of MET targeting agents. Some of the MET tyrosine kinase inhibitors have been suggested or expected to have effects on MET amplification. For example, tepotinib and capmatinib, which are both approved by the FDA for NSCLC with MET exon14 skipping, have shown positive results for MET amplification. Improved anti activity of tepotinib plus gefitinib in patients with EGFR-mutant NSCLC and MET amplification was suggested [11], and limited efficacy of capmatinib in previously treated patients in advanced NSCLC with MET amplification was observed [12]. Savolitinib, a potent, selective MET TKI, plus osimertinib are undergoing trials to determine their effect on EGFR mutation-positive lung cancers with MET amplification [13]. Additionally, amivantamab is an EGFR–MET bispecific antibody with immune cell-directing activity that targets activating and resistant EGFR mutations and MET mutations and amplifications [14].

Recently, Lai and colleagues described MET amplification and polysomy (aneuploidy) using FISH and MET/CEP7 ratio in EGFR mutated NSCLC patients [15]. Among MET high group (defined by MET copy number equal to or more than 5.0 by FISH), MET amplification (defined by MET/CEP7 ratio equal to or more than 2.0) was observed only in 11.5%, and associated with suboptimal response to EGFR TKI, suggesting that MET amplification was another driver alteration, but it exists in limited population in MET high group. Roper and colleagues studied special and temporal heterogeneity of MET amplification in EGFR-positive NSCLC patients who received osimertinib [16]. In post-osimertinib biopsy samples, heterogeneity of MET amplification and polysomy were observed. Therefore, it is important to carefully interpret MET copy number gain result from the tissue-based assay. It may be ideal to use liquid biopsy to capture heterogeneity in the clinical setting, rather than multiple biopsied used in this study. Our study showed the feasibility of capturing MET focal amplification using liquid biopsy.

A major limitation of this study is the small sample size and the lack of complete clinical history, including treatment and response data. For example, in the MET alt dataset, we found NSCLC cases with EGFR driver and T790M with focal MET amp, suggesting resistance to EGFR TKI. It has been reported that NSCLC with secondary resistance to EGFR inhibitors tends to show MET amp [17,18]. However, we did not know the treatment history of the patients from whom the plasma was collected. Additionally, it is ideal for validating the proposed MET focal and non-focal amplification using FISH assays, the current golden standard to measure MET amplification. However, our de-identified data did not come with FISH results. Further study is warranted to validate the current proposed MET focal amplification definition by FISH assays in future studies.

5. Conclusions

This study describes an approach to distinguish focal and non-focal MET amplification using unfiltered data from comprehensive genomic profiling of cfDNA in advanced cancer patients. Focal MET amp accounted for ~30% of all MET amps, was found in 3.7% of patients with diverse cancers, and was associated with a higher plasma copy number. Clinical studies are warranted to assess the clinical utility of targeted therapies for tumors with focal MET amplification detected by NGS of cfDNA.