State Key Laboratory of Oncology in South China, Guangzhou, China
Department of Medical Oncology, Sun Yat-sen University Cancer Center, Guangzhou, China
Corresponding author: Yu-Hong Li, MD, PhD, Department of Medical Oncology, Sun Yat-sen University Cancer Center, 651 Dongfeng East Rd, Guangzhou, China 510060; Fax: (011) 86-20-87343535; email@example.com
Several large studies have reported an extremely low incidence of MET gene amplification (GA) in patients with radically resected gastric cancer. This study was conducted to evaluate the prevalence and prognostic role of MET in patients with recurrent/metastatic gastric cancer who received chemotherapy.
MET GA and protein expression of recurrent/metastatic gastric cancer samples were evaluated by fluorescence in situ hybridization and immunohistochemistry (IHC), respectively.
This retrospective study included 232 patients with recurrent/metastatic gastric cancer. MET GA and strong protein expression (IHC3+) were observed in 8.3% (19 of 230 samples) and 9.6% (22 of 229 samples) of samples, respectively. A significant correlation was observed between MET GA and protein expression (r = 0.378; P < .001). MET GA was correlated with poor performance status (P < .001) and poorly differentiated tumors (P = .0015). Both MET GA and IHC 3+ expression were associated with a substantially shorter median overall survival (OS) and progression-free survival (PFS). The median OS and PFS for patients with MET GA versus those without MET GA were 5.7 months versus 15.5 months (P < .001) and 3.6 months versus 6.9 months (P < .001), respectively. The median OS and PFS for patients with MET IHC 3+ expression versus IHC 0 to 2+ expression were 6.3 months versus 15.1 months (P < .001) and 3.6 months versus 7.0 months (P < .001), respectively.
Gastric cancer is one of the most common causes of cancer-related deaths worldwide. In China, gastric cancer is the third most commonly diagnosed malignancy. The majority of patients present with inoperable locally advanced or metastatic disease. Although systemic chemotherapy is the primary treatment option for these patients, the prognosis remains dismal with a median survival of < 1 year.[3-6] The ToGA (Trastuzumab for Gastric Cancer) trial demonstrated that the combination of the human epidermal growth factor receptor 2 (HER2)-targeting monoclonal antibody trastuzumab and cytotoxic drugs could improve the median overall survival (OS) in patients with HER2-positive advanced gastric cancer. However, < 20% of patients with advanced gastric cancer are HER2 positive. Therefore, the development of new molecularly targeted agents is urgently needed.
MET is a proto-oncogene located on chromosome 7q21-31 that codes for the hepatocyte growth factor (HGF) receptor. Unregulated activation of the HGF/MET pathway has been observed in many human cancers, including gastric cancer. Previous studies have shown that activation of the HGF/MET pathway can reduce apoptosis and promote cancer cell proliferation, invasion, angiogenesis, and metastasis.[9, 10] In patients with gastric cancer, the most common cause of HGF/MET activation results from MET gene amplification (GA). Earlier reports have described MET GA and protein overexpression in approximately 20%[11-13] and 50%,[14, 15] respectively, of gastric cancer samples. However, these results were variable because of the small sample sizes, different methodologies used, and different populations tested. Recently, several studies performed in large consecutively enrolled cohorts have demonstrated that an increase in MET gene copy number was associated with an advanced stage of disease and worse clinical outcome.[16-19] Nevertheless, these studies mainly included patients with gastric cancer who had undergone radical resection.
The HGF/MET pathway is an attractive target for cancer treatment. Drugs that target the HGF/MET pathway, including small molecular tyrosine kinase inhibitors (TKIs) and monoclonal antibodies, have been developed. Promising results from ongoing clinical trials for several types of cancers have been reported.[20-22] Preclinical studies have found that gastric cancer cells with a high level of MET GA were susceptible to the MET-specific TKI PHA-665752. There is also a report in which among 4 patients with MET GA who were treated with the MET-specific TKI crizotinib, 2 experienced tumor shrinkage. In addition, the results of a phase 2 study demonstrated that the addition of rilotumumab (AMG 102), a human HGF antibody, to the regimen of epirubicin, cisplatin, and capecitabine improved progression-free survival (PFS), especially in patients with high MET expression. Another humanized MET antibody, onartuzumab (MetMAb; Genentech, San Francisco, Calif), is currently being tested in combination with 5-fluorouracil, folinic acid, and oxaliplatin in patients with metastatic HER2-negative gastric cancer in a phase 3 study (ClinicalTrials.gov identifier: NCT01662869).
Currently, clinical trials for the development of drugs targeting the HGF/MET pathway primarily focus on patients with recurrent/metastatic disease. An accurate understanding of the role and definition of MET status with respect to prognosis in patients with recurrent/metastatic gastric cancer is important because this could possibly provide information for clinical trials of anti-MET compounds. The increasing evidence indicates that MET GA is more frequent among patients with advanced gastric cancer and plays a critical role in cancer transformation and progression.[14, 17, 18, 24, 26-28] However, in patients with recurrent/metastatic disease, the prevalence of MET amplification and protein expression and their correlation with treatment outcome remain elusive. Therefore, we sought to investigate the frequencies and relationship of MET GA as evaluated by fluorescence in situ hybridization (FISH) and protein expression as evaluated by immunohistochemistry (IHC) to determine whether MET status is associated with clinical outcome in a large cohort of patients with recurrent/metastatic gastric cancer.
MATERIALS AND METHODS
This retrospective study included 232 patients with histologically confirmed, inoperable, locally advanced, recurrent, or metastatic adenocarcinoma of the stomach or gastroesophageal junction (GEJ) who were referred to Sun Yat-sen University Cancer Center between April 2005 and June 2012. Written informed consent according to the institutional guidelines was provided by all patients, and formalin-fixed paraffin-embedded primary tumor tissues were used for genetic testing. Clinical and pathological data including age, performance status, primary tumor site, tumor differentiation, and metastatic sites were collected as summarized in Table 1.
Table 1. Percentage of Tumors That Tested Positive for MET by FISH or Immunohistochemistry According to Patient Demographics and Tumor Characteristics
All data were presented as number (percentage) except age.
Bold type indicates statistical significance.
Laurean classification could be assessed in 229 patients, 227 patients, and 229 patients of the total number of patients, number of MET FISH patients, and number of MET IHC patients, respectively.
Tumor differentiation could be assessed in 209 patients, 207 patients, and 208 patients of the total number of patients, number of MET FISH patients, and number of MET IHC patients, respectively.
Response rate could be assessed in 182 patients, 180 patients, and 180 patients of the total number of patients, number of MET FISH patients, and number of MET IHC patients, respectively, according to Response Evaluation Criteria in Solid Tumors (RECIST) criteria.
All patients received first-line fluoropyrimidine-based systemic chemotherapy for metastatic disease. A total of 17 patient (7.3%) received monotherapy and 215 patients (92.7%) received combination therapy, of whom 88 (40.9%) received a fluoropyrimidine/platinum, 79 (36.7%) received a fluoropyrimidine/taxane, 20 (9.3%) received a fluoropyrimidine/irinotecan, and 28 (13.1%) received a fluoropyrimidine/taxane/platinum triple-drug regimen. A total of 109 patients (47.0%) received second-line (or more) chemotherapy. Seven patients with HER2-positive disease received anti-HER2 therapy (as first-line therapy in 5 patients and as second-line therapy in 2 patients). All patients were followed until October 2012. The dates of disease progression and death were recorded.
MET FISH and IHC
MET FISH and IHC analysis were performed and assessed independently. Pathologists were blinded to the clinical and molecular characteristics of the patients. FISH assay was performed as described previously. Briefly, 4-μm thick serial sections from each tissue block were submitted to dual-color FISH assay using a MET/chromosome 7 centromere (CEP7) probe cocktail (Kreatech Diagnostics, Amsterdam, the Netherlands). FISH signals for each locus-specific probe were assessed under an Olympus BX51 TRF microscope (Olympus, Tokyo, Japan) equipped with a triple-pass filter (4′,6-diamidino-2-phenylindole [DAPI]/green/orange; Vysis, Abbott Molecular, Abbott Park, Ill). Two independent observers scored at least 100 nonoverlapping interphase nuclei to determine the number of MET gene-specific (red) and CEP7-specific (green) signals. Using the University of Colorado Cancer Center criteria for epidermal growth factor receptor gene, the MET gene status was classified into 6 groups as follows: 1) disomy (≤ 2 copies in < 90% of cells); 2) low trisomy (≤ 2 copies in ≥ 40% of cells, 3 copies in 10%-40% of cells, and ≥ 4 copies in < 10% of cells); 3) high trisomy (≤ 2 copies in ≥ 40% of cells, 3 copies in ≥ 40% of cells, and ≥ 4 copies in < 10% of cells); 4) low polysomy (≥ 4 copies in 10%-40% of cells); 5) high polysomy (≥ 4 copies in ≥ 40% of cells); and 6) GA (defined by the presence of tight MET gene clusters and a MET gene-to-chromosome ratio of ≥ 2 or ≥ 15 copies of MET per cell in ≥ 10% of analyzed cells). The second through fifth criteria were defined as polysomy. GA was regarded as a positive FISH result and others as a negative result.
MET IHC staining was performed on formalin-fixed paraffin-embedded tissue samples using an automatic immunostainer (BenchMark XT; Ventana Medical Systems, Tucson, Ariz), according to the manufacturer's instructions. The primary antibody used was anti-total MET antibody (SP44; rabbit monoclonal, prediluted; Ventana Medical Systems). The staining intensity and percentage of positive cells were assessed. The MET IHC scoring algorithm for non-small cell lung cancer (NSCLC) was used to generate a MET IHC score from 0 to 3+ (Table 2).
Table 2. MET IHC Scoring Algorithm in Patients With Lung Cancer
Abbreviation: IHC, immunohistochemistry.
But <50% strong staining.
But <50% moderate or strong staining.
≥50% of tumor cells with strong membrane/cytoplasm staining
≥50% of tumor cells with moderate or strong membrane/cytoplasm staininga
≥50% of tumor cells with weak or moderate membrane/cytoplasm stainingb
No staining or ≤50% of tumor cells with membrane/cytoplasm staining of any intensity
The chi-square test or Fisher exact test (2-sided) was used to determine the correlations between MET GA/protein overexpression and clinicopathologic parameters. The association between MET FISH and IHC results was analyzed using the Pearson chi-square test. Survival curves were plotted using the Kaplan-Meier method, and compared using the log-rank test. All statistical tests were 2-sided, and statistical significance was defined as a P value < .05.
MET Amplification and Protein Expression
FISH analysis was successfully performed in 230 tissue samples. In all, 138 patients (60%), 73 patients (31.7%), and 19 patients (8.3%) demonstrated disomy, polysomy, and GA, respectively. Among 19 patients who exhibited MET GA, 4 demonstrated a cluster amplification, 5 had a MET/CEP7 ratio > 10, 4 had a MET/CEP7 ratio ranging from 5 to 10, and 6 had a MET/CEP7 ratio ranging from 2.5 to 5 (Figs. 1A-D).
MET protein expression was evaluated by IHC in 229 patients. Overall, 94 patients (41.0%) showed no IHC staining and 135 patients (59.0%) exhibited various staining intensity and percentage of MET protein (Table 3). According to the MET IHC scoring algorithm for NSCLC, 143 cases (62.4%), 22 cases (9.6%), 42 cases (18.3%), and 22 cases (9.6%) were scored as 0, 1+, 2+, and 3+, respectively (Figs. 1E-H).
Table 3. Profile of MET Protein Expression in Patients With Recurrent/Metastatic Gastric Cancera
MET immunohistochemistry was evaluated in 229 patients.
For those patients with heterogeneous intensity of MET, the predominant part was considered.
Correlation Between MET Amplification, Protein Expression, and Clinical Variables
Specimens from 227 patients were available for analysis of both MET GA by FISH and protein expression by IHC. Table 4 summarizes the association between MET GA and protein expression (using the NSCLC algorithm). Pearson correlation analysis revealed a significant correlation between MET GA and protein expression (r = 0.378; P < .001). Among the 19 cases with MET GA, 13 (68.4%) were scored as 3+, 5 (26.3%) were scored as 2+, and 1 (5.3%) was scored as 0.
Table 4. Correlation Between MET Gene Amplification Detected Using FISH and Protein Expression Detected Using IHCa, b
Abbreviations: FISH, fluorescence in situ hybridization; IHC, immunohistochemistry.
The scoring algorithm for non-small cell lung cancer was used for MET IHC.
A total of 227 specimens could be analyzed for both MET gene amplification by FISH and protein expression by IHC.
The distribution of MET amplification and protein expression according to various clinical and pathologic parameters was summarized in Table 1. Patients with MET GA had a higher frequency of poor performance status (P < .001) and poorly differentiated tumors (P = .015) compared with those without MET GA. Patients with MET GA also exhibited a higher frequency of metastatic disease at the time of initial diagnosis and diffuse or mixed-type (Lauren classification), and a greater number of metastases than those without MET GA, but the differences were not statistically significant. Similarly, patients with an IHC score of 3+ had a higher frequency of metastatic disease at the time of initial diagnosis and diffuse or mixed-type (Lauren classification), and a higher number of metastases, but none demonstrated statistical significance. No association was found between MET status and other clinical variables, including age, sex, primary tumor site, and metastatic sites.
The response rate to first-line chemotherapy was compared between patients with or without MET GA or strong protein expression. A total of 182 patients could be assessed by RECIST (Response Evaluation Criteria in Solid Tumors) criteria. Patients with MET GA had a significantly higher rate of disease progression compared with those without MET GA (P = .016). Patients with an IHC score of 3+ also demonstrated a higher rate of disease progression compared with those with an IHC score of 0 to 2+, although the difference did not reach the statistical significance (P = .079).
To evaluate the prognostic implication of MET amplification and protein overexpression, survival analyses were performed. The OS analysis was available for all patients, whereas PFS analysis was performed for 170 patients because of immeasurable disease or the patient being lost to follow-up. The results indicated that both MET amplification and strong protein expression were significantly associated with shorter OS and PFS. The median OS and PFS for patients with MET GA versus those without MET GA were 5.7 months versus 15.5 months (P < .001) and 3.6 months versus 6.9 months (P < .001), respectively. The median OS and PFS for patients with a MET IHC score of 3+ versus an IHC score of 0 to 2+ were 6.3 months versus 15.1 months (P < .001) and 3.6 months versus 7.0 months (P < .001), respectively (Fig. 2).
To the best of our knowledge, this is the first study of MET status assessment in a single institution cohort of 232 patients with recurrent/metastatic gastric cancer who had received systemic chemotherapy. The results indicated that the frequencies of MET GA and strong protein expression (IHC score 3+) were 8.3% and 9.6%, respectively, and that MET GA and an IHC score of 3+ were significantly associated with poor prognosis. These findings demonstrate that MET amplification and strong protein expression are not rare and likely play a pivotal role in recurrent/metastatic gastric cancer.
The prevalence of MET GA in patients with operable early-stage gastric cancer has been reported with varied results in several studies.[14, 16-19, 26-28] These studies used different methodologies, including Southern blot analysis, quantitative polymerase chain reaction (qPCR), and in situ hybridization (ISH) technologies (FISH or SISH). Earlier studies using Southern blot analysis or the qPCR assay reported relatively higher rates of MET GA.[14, 16, 17] Some of these studies reanalyzed the tumor samples with FISH assay and found that the true MET amplification incidence was quite low.[17, 19] However, a study from Japan recently reported that MET amplification by FISH could be confirmed in cases with a MET gene copy number of at least 4 by qPCR. Both Southern blot analysis and qPCR assays cannot discriminate polysomy of chromosome 7 from GA, and thus may overestimate the incidence of MET GA. ISH analysis has the advantage of direct visualization of the DNA copy number at the single-cell level, and thus provides more accurate information regarding the true copy number in tumor cells. Furthermore, ISH technology is technically more standardized and less affected by tissue variability, and therefore is widely used in clinical settings. Recently, low frequencies of MET GA in patients with operable gastric cancer have been reported by several studies using ISH assays. Janjigian et al reported that none of 38 Western patients with localized gastric cancer exhibited MET GA by FISH. A large cohort study of Korean patients who underwent surgical resection of primary gastric cancer found MET GA by SISH in 13 of 381 patients (3.4%). Lennerz et al screened a large series of 489 Western patients with stage 0 to IV gastroesophageal cancer and reported that only 10 patients (2%) exhibited MET GA by FISH. However, when focusing on patients with advanced gastric cancer, the incidence of MET GA appears to be higher. In the Korean study, MET GA was observed in 4 of 41 patients with stage IV disease (9.8%), whereas Lennerz et al found 6 of 109 patients with stage IV gastric and GEJ cancer (5.5%) demonstrated MET GA. In the current study, we enrolled patients with recurrent/metastatic gastric and GEJ cancer exclusively. Moreover, we used the standard FISH technology and similar MET GA definition. Our data demonstrate that 8.3% of patients with metastatic gastric cancer exhibited MET GA, which further supports the notion that MET GA is more frequent in patients with recurrent/metastatic gastric cancer, at least in Chinese population.
MET protein expression is normally assessed by IHC. To provide more information for future clinical trials, we described the MET IHC expression profile, including the staining intensity and percentage of positive cells in patients with recurrent/metastatic gastric cancer. To the best of our knowledge, no standard antibody, staining procedure, or scoring system have been recommended to date for assessing MET expression in patients with gastric cancer. Lee et al defined positive MET IHC expression as ≥ 10% of tumor cells with moderate or strong staining intensity, whereas Janjigian et al defined positive IHC expression as ≥ 25% of tumor cells with moderate or strong staining intensity. A study of NSCLC defined positive IHC as ≥ 50% of cells having moderate or strong staining. Therefore, MET overexpression positivity varies widely in different studies. In the current study, we found that a MET IHC score of 3+ as defined by the lung cancer algorithm was significantly associated with FISH positivity and worse OS. Using this algorithm, we demonstrated that 9.6% of patients had strong MET protein expression (IHC score of 3+). However, whether this algorithm could be used to predict the efficacy of anti-MET therapy requires further investigation. Notably, we observed that approximately one-half of the patients with a MET IHC score of 3+ did not exhibit MET GA, indicating that MET overexpression might not correlate with MET GA. The study by Lee et al demonstrated that strong expression of MET protein (IHC score 3+) significantly correlates with MET GA by FISH. However, discordance between MET GA and protein expression has been observed by other studies.[19, 28] Another Korean study showed that the concordance rate between MET amplification by qPCR and protein expression was only 17.9% (32 of 179 cases). In addition, the study from the Memorial Sloan-Kettering Cancer Center demonstrated a high level of MET protein expression (63%) in 38 patients with locally advanced gastric adenocarcinoma, whereas none of these tumors was found to have MET GA by FISH. The mechanism of MET overexpression in the absence of gene amplification remains unclear. Some studies have suggested that deregulation of MET transcription or activation of other oncogene-related signaling pathways might play a role in MET overexpression.[33, 34] Another possible explanation for the discrepancy might be the different antibodies and scoring systems used in these various studies.
In the current study, we attempted to determine specific clinical and pathological characteristics that could identify patients with MET GA or protein overexpression. Our results indicate that patients with MET GA tend to have poor performance status and poorly differentiated tumors compared with those without MET GA. Both MET GA and strong protein expression demonstrated a tendency to be associated with metastatic disease at the time of initial diagnosis and a larger number of metastases, but the results were not statistically significant. We also observed that > 70% of patients with a MET IHC score of 3+ have diffuse or mixed-type tumors according to the Lauren classification, which is in keeping with a previous report of Korean patients.
We further explored the prognostic value of MET, which demonstrated that both MET GA and strong protein expression were significantly associated with worse survival in patients with recurrent/metastatic gastric cancer who received chemotherapy. This result is consistent with previous reports of patients with resected gastric cancer. A study by Graziano et al including 230 patients with stage II/III gastric cancer demonstrated that patients with MET GA by qPCR had significantly worse disease-free survival and OS. Lee et al also showed that MET GA by FISH and an IHC score of 3+ were significantly associated with poor survival. It is conceivable that oncogenes might play distinctive roles in different disease stages. The results of the current study suggest that MET GA is a prognostic factor for both early-stage and recurrent/metastatic disease. We also observed that patients with MET GA or an IHC score of 3+ had a much lower response rate and shorter PFS after systemic chemotherapy compared with patients without MET GA, suggesting that MET positivity (MET GA or an IHC score of 3+) might be an indicator of resistance to standard chemotherapy in patients with recurrent/metastatic gastric cancer. However, because of the lack of a control group, we could not determine whether MET is a predictive factor for chemotherapy.
In conclusion, the results of the current study demonstrated for the first time that MET amplification is not rare in patients with recurrent/metastatic gastric cancer. MET GA and strong protein expression appear to be significantly associated with unfavorable clinical outcomes. This information indicates that the current clinical development of anti-MET compounds for the treatment of recurrent/metastatic gastric cancer is a promising modality.
Shanghai Roche Pharmaceuticals Ltd provided a research grant and technical support.