In Iran, gastric cancer is the second most common cancer and the leading cause of cancer death and is an important health problem because of high risk areas in north and northwestern regions . However about 80% of the patients were presented in advanced stages and did not gain any survival benefit from conventional therapy [3–5]. Therefore new strategy based on the screening of high risk population and achieving new prognostic factors to predict the behavior of gastric tumor as well as personalized target therapy are required.
The era of target therapy has started in 1990 by Tamoxifen and until now more than 38 agents have been introduced for various cancers, among them anti-HER2/neu and anti c-kit therapy are the known ones. While c-kit expression in gastric carcinoma is very limited, Her2-neu expressed in less than 20% of them [34, 35], and efforts for finding other molecular biomarker are currently in development.
c-MET is an oncogene that has gained attention as a prognostic marker, and an indicator of metastasis and poor prognosis as well as a good target for therapeutic inhibition . Until now many new anti-c-MET drugs have been invented and most are in preclinical and clinical testing [34, 36].
Previous studies have been shown c-MET overexpression in selected patients indicates that certain patients may be sensitive to targeted therapy . Moreover it has been confirmed that this protein overexpression is significantly associated with high level of c-MET mRNA and gene amplification [19, 28].
In the current study, c-MET overexpression was observed in 71% of gastric cancers and showed significant statistical relationships with tumor type and depth of tumor invasion as well as neural invasion and TNM staging. It also showed a relative relationship with vascular invasion and lymph node metastasis.
As far as we know this is the first time that c-MET oncogene is evaluated by Tissue microarray in Iran. Tissue microarray technology has many advantages including: facilitate the staining and the interpretation and reduces the intra – and inter observer variation of IHC interpretation and particularly saving the cost. We used a semi automatic instrument but a new simple method is also introduced by using mechanical pencil tips with much more low costs .
There were few studies using TMA to evaluate c-MET by IHC in gastric cancer [18, 21, 39] and all were performed in China.
Tang et al. studied c-MET expression in normal gastric mucosa, intestinal metaplasia and gastric cancer and found c-MET overexpression in 68.8% of 232 gastric cancers. Its expression was significantly higher than intestinal metaplasia. They did not find any relationship between c-MET expression in gastric cancer with tumor stage, grade of differentiation or tumor type .
In study by Zhao et al., c-MET expression was identified in cases of gastric carcinoma and matched normal gastric mucosa, as well as cases with chronic atrophic gastritis, intestinal metaplasia and dysplasia. c-MET overexpression was found in about 66% of gastric carcinoma which was significantly higher than normal mucosa, chronic gastritis, metaplasia and dysplasia. Its overexpression was associated with tumor type (more in intestinal than in diffuse type), grade of differentiation and lymph node metastasis .
Another study was also recently performed in China by Li et al., and they reported c-MET expression in 82.4% of 114 gastric cancer. They found significant association between c-MET expression and advanced clinical stage, lymph node metastasis and poor prognosis .
In contrast to Tang et al. , our findings are in agreement with Li et al.  and Zhao et al.  studies and in continuation of earlier studies that reported c-MET association with poor prognosis and local invasiveness.
In present study we did not assess survival rate because of the short term of study and limitation of patients' data, differences in patients' management and lack of regular patients' follow up.
Another difference between the current study and three above mentioned studies is in interpretation of c-MET reactivity. Zhao et al. and Li et al. used double score or product score (multiply intensity by area) with different criteria and cut offs, and Tang et al. used only the area of reactivity with the cut off of 10%; however we used only "intensity" of reaction.
Traditionally, some authors have used double score or H score (multiply intensity by area) in evaluation of IHC staining, but we more agree with Tolgay et al.  that due to the minute area (0.6 mm diameter) of the spots in TMA and homogeneity in their reactions, ignoring "area" is more appropriate in this setting.
In other studies using whole slide evaluation (without using TMA), a spectrum of criteria have been used to defining a positive reaction, but many used "area" with various cut offs of 5% to 30% [17, 19, 22, 28, 29], and few used a product score (intensity × area) .
It does not need to emphasis that with these various arbitrary criteria, either in TMA or in whole slide assessment, we will require a general consensus in regard to evaluation of c-MET expression in gastric cancer. It would not far that anti c-MET drugs become a part of cancer management and defining eligible patient would be our duty . In fact standardization of various types of antibodies used against c-MET would be another part of this work. Until now, antibodies commonly used are against extracellular alpha and beta subunits as well as against intra cytoplasmic c-terminal. It is not clarified that regardless of the name of producers, which type of antibody is more sensitive or specific.
Some studies reported more c-MET expression in diffuse tumors than in intestinal type [41–43]. In contrast to these results, Drebber et al. and Zhao et al.  reported more c-MET expression in intestinal tumors, and in agreement with them we also found a significant expression of c-MET in intestinal tumors than in diffuse.
TNM staging system according to The American Joint Committee on Cancer (AJCC)/International Union against Cancer (UICC) produced the most reliable system for predicting the survival of patients. Furthermore lymphatic and vascular invasion were considered as poor prognostic indicators [8, 31, 44, 45].
In the present study, c-MET expression was significantly related to depth of tumor invasion (pT), and TNM staging and at least implied as a marker of local invasiveness. As the depth of invasion and perforation of serosa were introduced as independent prognostic factors , thus c-MET expression could be regarded as a potential prognostic factor. This fact has been also affirmed by other studies [17, 19, 23, 28, 46].
c-MET expression in present study showed no relationship to tumor differentiation which was in agreement with Tang et al. results .
Several studies reported that c-MET expression had a remarkable relationship with lymph node metastasis [18, 19, 22, 26, 27, 39] while in our study there was a relative relationship (P = 0.065) between these two variables. One logical reason is that the average number of separated lymph nodes in our cases was 11, which was under the standard number of at least 15 per each gastrectomy sample , and another reason was related to type of surgery. The method of extended lymphadenectomy in gastrectomy which widely used in Japan and other countries did not advocate in Iran. For example in one study using extended lymphadenectomy an average of 39 lymph nodes was resected for each patient .
In the present study in agreement with previous studies, association of c-MET overexpression and advanced stage of gastric carcinoma was seen [12, 13, 17, 19, 28], while some authors reported discrepant results [20, 21, 29] .
Nakajima et al.  found c-MET as a negative independent prognostic factor and this finding was verified by others [26–28] .
In the present study we did not find any association between distant metastasis and c-MET expression, which was in agreement with Huang et al.  study and in contrast to Amemiya et al.  that showed all gastric cancers with distant metastasis overexpressed c-MET, but due to limited number of metastatic cases in our study, their findings cannot be ruled out.
In conclusion c-MET oncogene protein was frequently overexpressed in Iranian gastric carcinomas and it was related to clinicopathological characteristics such as tumor type, depth of invasion, neural invasion and TNM staging. It can also support the idea that c-MET is a potential marker for target therapy in Iranian gastric cancer.