MET Expression Level in Lung Adenocarcinoma Loosely Correlates with MET Copy Number Gain/Amplification and Is a Poor Predictor of Patient Outcome.

MET amplification has been associated with shorter survival in cancer patients, however, the potential correlation of MET overexpression with either MET amplification or patient outcome is controversial. The aim of this study was to address these questions by correlating MET expression level with MET copy number and patient outcome in a cohort of 446 patients who had a lung adenocarcinoma: 88 with MET amplification, 118 with polysomy 7, and 240 with negative results by fluorescence in situ hybridization. MET expression assessed by immunohistochemistry was semi-quantified by expression level: absent (0+), weak (1+), moderate (2+) and strong (3+); or by H-score: 0-99, 100-199, and ≥200. MET expression level or H-score was positively but weakly correlated with MET copy number or MET/CEP7 ratio. Strong expression of MET (3+ or H-score ≥ 200) was associated with a shorter overall survival, but it was not an independent hazard for survival by multivariant analysis. We conclude that MET expression is loosely correlated with MET copy number gain/amplification. Strong expression of MET does not independently predict patient outcome.

1. Introduction

MET, located on chromosome 7q31, is an oncogenic receptor tyrosine kinase for hepatocyte growth factor. Activation of MET pathway, either by MET amplification or a splice site alteration in exon 14, plays important roles on tumor cell survival, proliferation, metastasis, and drug resistance [1,2,3]. Evidence from preclinical and clinical trials suggests MET activation serves as a primary oncogenic driver in a subset of patients with non-small cell lung cancer (NSCLC) and as a secondary driver of acquired resistance to targeted therapy in EGFR-mutant [3] or ALK-positive patients [4]. Recent investigations have shown that a subset of patients harboring MET exon 14 mutation or MET amplification can benefit from MET inhibitors [5,6,7].

Fluorescence in situ hybridization (FISH) has been commonly used to detect MET copy number change or MET amplification. MET amplification often exhibits a high copy number of MET (≥15 copies or clusters of MET signals) and/or a MET/CEP7 (centromere of chromosome 7) ratio ≥ 1.8 [8,9] or ≥2.0 [10,11] by FISH. MET amplification has been detected in 1% to 5.8% of treatment-naïve NSCLC patients [9,10,11,12,13,14,15,16,17] and in 5% to 26% of patients with EGFR-mutant NSCLC who develop resistant to EGFR inhibitors [13,14,15,16,18,19,20]. MET amplification has been associated with an inferior outcome [9,11,12,21,22,23,24]. Immunohistochemistry (IHC) is a widely used assay to assess gene expression level, it is more readily available than FISH in most pathology laboratories and is more cost-effective than FISH. The frequency of MET overexpression in unselected NSCLCs ranges from 15% to 70%, mainly depending on the criteria used to “define” overexpression [17,24,25,26,27,28,29,30,31]. Some studies have suggested that MET express level can serve as a potential predictive marker for overall survival (OS) and/or progression-free survival, and the likelihood that patient responds to MET kinase inhibitor [24,30,31,32,33]. On the contrary, other studies have not found an association between MET expression level and patient survival [25,27,28,29]. Others have suggested that MET IHC is not a good screening test for MET amplification or MET exon 14 mutation in lung cancer [17,26]. Previous clinical trials that focused on MET pathway–directed targeted therapy in unselected or MET-overexpressing NSCLC patients have yielded largely negative results [3].

In this retrospective study, we included 446 patients with lung adenocarcinoma, we evaluated the correlations between MET copy number gain /amplification (by FISH) and MET expression level (by IHC). We also evaluated whether overexpression of MET can predict patient outcome.

2. Materials and Methods

2.1. Study Cohort

The study group included patients who were diagnosed with lung adenocarcinoma at our institution during a six-year period (January 2014 to December 2019), and who had been evaluated for MET copy number/amplification by FISH and MET expression by IHC. Patient follow-up and outcomes were obtained by electronic chart review. This study was approved by the Institutional Review Board at The University of Texas MD Anderson Cancer Center and was conducted in accord with the Declaration of Helsinki.

2.2. FISH Analysis

FISH analysis for MET/CEP7 (MET labeled as red and CEP7 labelled as green, Biocare Medical, LLC., Concord, CA) was performed on formalin-fixed paraffin-embedded (FFPE) tissue sections as a part of clinical workup for all newly diagnosed NSCLC at our institution as described previously [9]. After hybridization and post-wash, DAPI was applied to each slide. Fifteen fields or whole tissue section were scanned by Metafer automated imaging system (MetaSystems). Technologists analyzed sixty tumor cells (with relatively large nuclei) by counting signals of both MET and CEP7 and calculated the MET/CEP7 ratio for each specimen. For cells with clusters of MET signals, the copy number of MET was recorded as 20 for calculation purpose. MET amplification (MET-amp) was considered when one or more of these criteria were met: the MET/CEP7 ratio was ≥1.8; >10% of cells showed clusters of MET signals; or MET copy ≥ 15. Polysomy 7 was considered if MET/CEP7 ratio < 1.8 and MET copy ≥ 5 but <15. The remaining cases were designated as MET negative (MET-neg) [9] (Figure 1).

Figure 1

Fluorescence in situ hybridization (FISH) analysis using MET/CEP7 probes (×60). (A): MET negative; (B): Polysomy 7 (copy number gain detected in both MET and CEP7); (C): MET amplification (copy number gain in MET, not in CEP7, MET/CEP7 ratio > 1.8); (D): MET amplification (clusters of MET signals). FISH probe signals: MET in red and centromere 7 (CEP7) in green.

2.3. Immunohistochemistry Stain

MET IHC was performed on FFPE tissue sections using anti-total MET rabbit monoclonal antibody (clone SP44, Roche Diagnostics, dilution 1:1). MET expression intensity (level) and percentage of cells were evaluated and recorded by two pathologists independently. In tumors that exhibited intra-tumor heterogeneous staining intensity, the predominant (≥50% of tumor cells) staining pattern was recorded. The intensity was classified as absent (IHC0+), weak (IHC1+), moderate (IHC2+) and strong (IHC3+) (Figure 2, ×20). In addition to the staining intensity, H-score was also assessed. The H-score was calculated by a formula including both intensity and percentage of cells: (3× percentage of cells with IHC3+) + (2× percentage of cells with IHC2+) + (1× percentage of cells with IHC1+). H-score ranged from 0 to 300.

Figure 2

MET expression level by immunohistochemistry stain (×20). (A): Absent (0+); (B): Weak (1+); (C): Moderate (2+); (D): Strong (3+).

2.4. Statistical Analysis

Spearman correlation analysis was used to analyze the correlation coefficient (r) between MET expression level (IHC or H-score) and MET copy number (or MET/CEP7 ratio). Chi-square was used to compare categorical variables. Overall survival (OS) was estimated by Kaplan-Meier method from the date of diagnosis of lung adenocarcinoma to the date of death from any cause or censored at time of last follow-up for living patients. Multivariate cox proportional hazard regression analysis was used to assess the relationship between survival and age, tumor stage, MET FISH groups and MET overexpression (IHC3+ or H-score ≥ 200). The difference was considered as significant when p < 0.05.

3. Results

3.1. Patients

During this study period, there were 1987 patients who had FISH for MET/CEP7 tested on the first diagnostic specimen of lung adenocarcinoma in our institution, including 112 (5.6%) patients with MET/CEP ratio ≥ 1.8 (MET-amp), 146 (7.3%) with polysomy 7, and 1729 (87%) with a MET-neg FISH results.9 Among the 258 patients with MET-amp or polysomy 7, 206 (88 with MET-amp and 118 with polysomy 7) had materials available for IHC. In addition, we included a comparable number (n = 240) of patients from MET-neg group who had IHC available and who showed similar distribution of age, gender, tumor stage, and diagnostic year compared to the groups of MET-amp and polysomy 7.

Among the 446 patients included in this study cohort, 95 (21.3%) patients with stage I/II and 351 (78.7%) patients with stage III/IV disease. The sex and age distribution are shown in Table 1.

Table 1

Associations of MET expression level with MET copy number, cancer stage and outcome.

	Absent (0+)(n = 41)	Weak (1+)(n = 55)	Moderate (2+)(n = 231)	Strong (3+)(n = 119)	Total(n = 446)
Age (range)	68 (33–85)	68 (44–83)	67 (29–94)	65 (29–83)	66 (29–94)
Gender (M/F)	22/19	30/25	100/131	51/68	203/243
FISH analysis
MET-amp	3 (3.4%)	3 (3.4%)	31 (35.2%)	51 (58%)	88 (19.7%)
Polysomy 7	10 (8.5%)	8 (6.8%)	56 (47.5%)	44 (37.3%)	118 (26.5%)
MET-neg	28 (11.7%)	44 (18.3%)	144 (60%)	24 (10%)	240 (53.8%)
Cancer stage
I/II	15 (36.6%)	12 (21.8%)	59 (25.5%)	9 (7.6%)	95 (21.3)
III/IV	26 (66%)	43 (78.2%)	172 (74.5%)	110 (92.3%)	351 (78.7)
Outcome
Alive/Dead	23/18	30/25	139/92	63/56	255/191
Median OS (mon)	57.2	47.7	34.3	28.8

FISH: fluorescence in situ hybridization; M/F: male/female; mon: months; neg: negative; OS: overall survival.

3.2. Immunohistochemistry

MET IHC exhibits membranous and cytoplasmic reactivity, and the IHC pattern was relatively homogenous in most cases. The expression level was determined on the basis of the stain intensity and fraction of positive cells by two pathologists. In most cases (n = 392, 88%) observations by both pathologists were concordant. Discordance occurred in 54 patients: 18 with 1+ or 2+ and 36 with 2+ or 3+. The discordant cases were reviewed by the third pathologist. After consensus, 41 (9.2%) neoplasms were negative for MET (0+), 55 (12.3%) were weakly positive (1+), 231 (51.8%) were moderately positive (2+), and 119 (26.7%) were strongly positive (3+). H-score fell into 0–99 in 96 (21.5%) cases; 100–199 in 205 (46%) and ≥200 in 145 (32.5%) patients (Table 2).

Table 2

Correlation of MET FISH groups with MET expression (IHC level or H-score).

	MET-amp(n = 88)	Polysomy 7(n = 118)	MET-neg(n = 240)	Total(n = 446)	Correlation(r)	p
IHC0	3 (7%)	10 (24%)	28 (68%)	41	a 0.4431b 0.4401	<0.0001<0.0001
IHC1+	3 (5%)	8 (15%)	44 (80%)	55
IHC2+	31 (13%)	56 (24%)	144 (62%)	231
IHC3+	51 (43%)	44 (37%)	24 (20%)	119
H-score 0–99	7 (7%)	19 (20%)	70 (73%)	96	a 0.4381b 0.4298	<0.0001<0.0001
H-score 100–199	27 (13%)	41 (20%)	137 (67%)	205
H-score 200–00	54 (37%)	58 (40%)	33 (23%)	145

Correlation coefficient of MET expression with a MET copy number or b MET/CEP7 ratio.

3.3. Correlation of MET Expression with MET Copy Number or MET/CEP7 Ratio

MET expression level (IHC 0–3) and MET H-score were loosely and positively correlated with MET copy number (p < 0.0001) or MET/CEP7 ratio (p < 0.0001), with a correlation coefficient (r) ranged from 0.4298 to 0.4431 (Table 2). MET expression level (or H-score) was significantly different (p < 0.0001) among patients with MET-amp, polysomy 7 or MET-neg (Table 2, Figure 3). High proportion (58%) of cases in the MET-amp group showed strong (3+) MET expression versus 37% in the polysomy 7 group and 10% in the MET-neg group. Similarly, high proportion (61%) of cases in the MET-amp group showed H-score ≥ 200 comparing to 49% in polysomy 7 group and 14% in MET-neg group. Patients with MET-neg and polysomy 7 more commonly showed moderate (IHC2+) expression of MET, 60% and 47%, respectively.

Figure 3

Correlation of FISH groups and immunohistochemistry results. (A): Proportion of patients with different level of MET expression (IHC0+~3+) in patients with MET-amp, Polysomy 7 and MET-neg. (B): Proportion of patients with MET H-score of 0–99, 100–199 and ≥200 in patients with MET amp, Polysomy 7 and MET neg. Presented as percentage of cases.

3.4. Correlation of MET Expression and Tumor Stage

MET expression level and tumor stage are summarized in Table 1; strong (IHC3+) expression was associated with advanced (III/IV) tumor stage (p < 0.0001). Similarly, H-score ≥ 200 was also associated with advanced tumor stage (p = 0.0021).

3.5. Correlation of MET Expression and Patient Outcome

As reported previously [9], MET-amp was significantly associated with inferior survival. The median overall survival of three FISH groups (MET-amp, polysomy 7 and MET-neg) was 21.4 months, 33 months and 49.8 months, respectively (p = 0.0005, Figure 4A). The clinical outcomes of patients in the MET IHC groups are summarized in Table 1. Median overall survival was 57.2 months for patients with tumors showing no MET expression, versus 47.7 months, 34.3 months, 28.8 months for patients with weak (1+), moderate (2+), and strong (3+) MET expression, respectively (p = 0.0966). Although these data show a trend that patients with a higher level of MET expression (1+~3+) had a shorter survival comparing to the patients without MET expression (0+), but the differences were not significant (p = 0.2694, 0.6625, and 0.1927 for IHC 1+, 2+, 3+ comparing to IHC 0+) (Figure 4B).

Figure 4

Overall survival (OS) by Kaplan-Meier analysis. (A): Comparison of OS among patients with MET-amp, polysomy 7 and MET-neg, patients in MET-amp group showed significantly inferior OS; (B): No significant difference of OS among patients of groups with of IHC-neg, IHC1+, IGHC2+ and IHC3+; (C): Patients with IHC3+ showed a significantly shorter OS comparing to patients with IHC0/1+/2+; (D): Patients with H- score ≥ 200 showed a significantly inferior OS comparing to patients with H score < 200.

We then analyzed whether a higher MET expression (IHC2+ or IHC3+) could predict patient’s survival. As shown in Table 3, patients with IHC3+ showed a significant shorter median OS comparing to the patients with IHC0/1+/2+ (28.8 vs. 36.3 months, p = 0.0463, Figure 4C); however, the significance was not seen among patients with IHC0/1+ and patients with IHC2+/3+ (p = 0.6630) (Table 3).

Table 3

Association of MET expression with survival by using IHC2+/IHC3+ or H-score ≥ 150 or ≥200 as cut-off (Kaplan-Meier method).

	No. of Patients	Median Survival	p
IHC0/1+	96	47.7	0.6630
IHC2+/3+	350	33
IHC0/1+/-2+	327	36.3	0.0463
IHC3+	119	28.8
H-score < 150	122	47.5	0.2950
H-score ≥ 150	324	32.8
H-score < 200	301	41.1	0.0110
H-score ≥ 200	145	27.8

We also analyzed whether a high H-score (≥150 or ≥200) could predict patient’s outcome. As shown in Table 3, patients with H-score ≥ 200 showed a significant shorted median OS (27.8 vs. 41.1 months, p = 0.0110, Figure 4D) comparing to patients with H-score < 200, but patients with H-score ≥ 150 and patients with H-score < 150 showed comparable OS (p = 0.2950) (Table 3).

Though MET-amp, IHC3+ and H-score ≥ 200 were significantly associated with inferior survival in the univariate analysis, only MET-amp remained (in addition to older age and advanced stage) to be prognostically significant by multivariate analysis when MET FISH, age (≥65 years), tumor stage (III/IV), and MET expression (either with a IHC3+ or with b H-score ≥ 200) were co-analyzed for OS (Table 4). When age, tumor stage, and MET expression were included (without MET FISH), only older age and advanced tumor stage were significant hazards to survival (p < 0.0001). High-level MET expression (either IHC3+ or H-score ≥ 200) was not a significant hazard to survival.

Table 4

Multivariate cox proportional hazard regression analysis for overall survival.

	a Overall Survival		b Overall Survival
Variables	Hazard Ratio(95% CI)	p	Hazard Ratio(95% CI)	p
Age (≥65 vs. <65 years)	1.849 (1.372–2.491)	<0.0001	1.827 (1.355–2.462)	<0.0001
Stage (I/II vs. III/IV)	4.163 (2.560–6.770)	<0.0001	4.215 (2.585–6.873)	<0.0001
MET FISH
Polysomy 7 vs. MET-neg	1.046 (0.729–1.501)	0.588	1.100 (0.769–1.573)	0.602
MET-amp vs. MET-neg	1.407 (1.057–2.311)	0.045	1.547 (1.067–2.422)	0.027
Immunohistochemistry
H-score (≥200 vs. <200)	1.200 (0.865–1.664)	0.275
IHC (IHC0-2+ vs. IHC3+)			0.984 (0.694–1.396)	0.930

These were two analyses: H-score (left, a) or IHC (right, b) was included in separate analysis, while age, stage and MET FISH were included in both analyses.

4. Discussion

MET is an oncogenic receptor tyrosine kinase. In vitro and in mice, overexpression of MET alone is sufficient to induce oncogenic transformation [34,35]. Many studies have explored the potential of MET overexpression, as assessed by IHC, to serve as a prognostic marker and/or predictive marker for MET amplification, MET exon 14 mutation, or the response to MET kinase inhibitors. To date, the results of these studies have been contradictory (Table 5). Currently, no consensus on the definition of “positive” MET expression has been reached. Some studies used stain intensity of 2+ [27,28,29] or 3+ [36] (scale of 0+ to 3+); some studies used H-score (0–300), ≥20 [30], ≥150 [26], or ≥200 [17]; some studies multiplied the intensity score and the fraction score (range 0-12) and consider positive when score fell into 4–12 [24]. In large part related to this lack of consensus, the rate of MET expression has differed greatly in these studies, from 15% to 70% [17,24,25,26,27,28,29,30,31]. (Table 5). In the current study, we used IHC3+ or H-score ≥ 200 as cut-off, both predicted an inferior survival in the univariate analysis, but not in multivariate analysis.

Table 5

Current and Previous Studies on MET Overexpression and the Associations with MET Copy Number /Patient Survival.

Reference	Criteria for MET Overexpression	Positive Rate	Tissue	Antibody	Correlates with MET GCN?	Associates with Survival?
Bubendorf (2017) [28]	>50% exhibit 2+ staining	23.8%	TMA	SP44	Yes	Not correlated
Dziadziuszko (2012) [29]	METMab: ≥50% of cells with ≥2+	25% (44/174)	TMA	SP44	Yes	Not correlated
Guo (2019) [17]	H-score ≥ 200	39% (71/181)		SP44	Poor	NA
Mignard (2018) [26]	H-score ≥ 150 MetMab Score: 2+/3+	15/81(18.5%)14/81 (17%)		SP44	Poor	NA
	MetMab Score: 2+/3+
Park (2012) [24]	4–12 (of 0–12 scale)(intensity x fraction)	13.7 (52/380)		3D4 (Rabbit polyclonal	Yes	Significant shorter OS and DFS
Rivalland (2019) [27]	>50% exhibit 2+ staining	25% (193/763)	TMA	SP44	Yes	Not correlated
Tsuta (2012) [25]	Cytoplasmic/membrane, ≥10% cells	22.2%		SP44	NA	Not correlated
Weingertner (2015) [31]	METMab: ≥50% of cells ≥2+	44% (89/201)	TMA	SP44	High GCN often have overexpression; ~1/3 overexpression cases had high MET GCN	Only in non-smoke group (32/201)
	3+ in ≥10% of cells	28% (57/201)
	H-score: >140	42% (84/201)
Current study	IHC3+, ≥50% of cells	26.7%	Tissue section	SP44	Yes, but low correlation	Yes, but not an independent factor
	H-score ≥ 200	32.5%			Yes, but low correlation	Yes, but not an independent factor

GCN: gene copy number; OS: overall survival; DFS: disease-free survival; TMA: tissue microarray.

The design of this study is slightly different from the studies published previously. First, this study cohort included a large number of patients with MET-amp (n = 88) and polysomy 7 (n = 118), in contrast with most published studies in which a smaller number of patients with MET-amp and polysomy 7 were included (due to the low frequency of MET-amp or polysomy 7 in treatment-naïve NSCLC in general). This higher proportion of cases with MET-amp and polysomy 7 enabled us to do a better comparison of MET expression and MET copy number gain/amplification. Second, we analyzed the association of high-level expression of MET (IHC3+ or H-score ≥ 200) with survival by both univariate and multivariate analysis, though high-level of MET expression was associated with inferior survival by univariate analysis, it was not a significant hazard to survival by multivariate analysis when age and tumor stage were included in analysis. Third, we also analyzed the association of spectrum of stain intensities (0+~3+) with FISH results (MET copy number and MET/CEP7 ratio) and patient outcome.

We compared the association of MET copy number or MET/CEP7 ratio by FISH and MET expression by intensity or H-score in parallel, to explore whether IHC could possibly replace FISH. The results of this study showed a weakly positive correlation of MET expression with MET copy number or MET/CEP7 ratio (r ranged 0.4298 to 0.4431, p < 0.0001), which was in line with other studies [24,26,27,28,29]. In patients with MET-amp, ~58% of patients showed strong (3+) MET expression. However, IHC3+ was also seen in the polysomy 7 (~37%) and MET-neg (~10%) groups. Similar findings were also observed in H-score. The sensitivity of using IHC3+ was 43% and the specificity was 89% for predicting MET-amp in this cohort, which is similar to an earlier study by Mignard and colleagues [26]. Among the patients with polysomy 7 and MET-neg FISH results, IHC 2+ was the most common result, 47% in the polysomy 7 and 60% in MET-neg groups, respectively.

Multiple factors could contribute to the low concordance rate between MET expression and MET gene copy number (GCN). Biologically, MET expression is not only controlled by MET GCN, but also by transcriptional and/or translational regulation. Besides, intratumoral heterogeneity [15,37], sampling (small core biopsy or tissue microarray), tumor histology type and differentiation, and tumor load at a metastatic site also could contribute to a low concordance between MET expression and MET GCN. Among the cases with a strong MET expression (IHC3+ or H-score ≥ 200) but a negative FISH result, a common feature we have observed was low tumor infiltrations in the metastatic tissue biopsy specimens. We also noticed that tumors with MET-amp, but with 0+/1+ MET expression tended to be poorly differentiated.

MET amplification, IHC3+ or H-score ≥ 200 was significantly associated with inferior survival by univariate analysis, however, IHC3+ or H-score ≥ 200 was not an independent hazard to survival by multivariate analysis when age and stage were included (with or without MET FISH groups). This could be due to the fact that high expression of MET (IHC3+ or H-score ≥ 200) was highly associated with advanced tumor stage [33]. Thus, our findings support the interpretation that MET expression level is a poor predictor of patient survival [25,27,28,29].

In summary, the concordance between MET expression and MET GCN/amplification is low. Unlike MET amplification highly associates with a short survival, MET expression level is not significant hazard to survival. MET expression level by IHC can’t substitute FISH analysis for MET copy gain/amplification.

5. Conclusions

Apart from previous studies, this study cohort included much higher proportion of cases showing MET-amp (88, ~20%) and polysomy 7 (118, ~26%), in addition to 240 (54%) cases with MET-neg. We found that MET expression was weakly and positively associated with MET copy or MET/CEP7 ratio. High-level MET expression (IHC3+ or H-score ≥ 200), though seemed to be associated with an inferior survival, was not an independent hazard to survival.