Background
Breast cancer is the most common cancer among women worldwide [
1]. The clinical application of targeted therapies, such as tamoxifen and trastuzumab, has decreased the mortality of breast cancer in recent years. However, epidemiological studies show that more than 400,000 patients worldwide die from breast cancer each year [
2]. Breast cancer is a heterogeneous disease that has been classified into five molecular subtypes: luminal A, luminal B, human epidermal growth factor receptor-2 (HER-2) overexpressing, basal-like, and normal-like [
3]. Current therapeutic regimens for breast cancer are designed according to clinical pathological factors and molecular typing. However, patients with the same clinical stage and molecular type often display markedly different treatment responses and overall outcomes, which lead to treatment failure [
4‐
7]. Therefore, the identification of new prognostic factors and potential therapeutic targets is necessary to improve individual treatment strategies.
The tyrosine kinase c-Met, a key regulator of invasive growth, is overexpressed in certain aggressive cancer cells [
8]. c-Met, also called MET and hepatocyte growth factor receptor (HGFR), is a plasma membrane protein that transduces signals from the extracellular matrix to the cytoplasm and is activated by binding to HGF [
9]. c-Met is involved in uncontrolled survival, growth, angiogenesis and metastasis of cancer cells [
10]. Crizotinib, a dual tyrosine kinase inhibitor of ALK and c-Met kinases, has shown promising results in the treatment of lung adenocarcinoma [
11]. Tivantinib, a c-Met inhibitor is being tested in patients with MET-high hepatocellular carcinoma in an ongoing Phase III clinical trial [
12]. c-Met was shown to be involved in the development of herceptin and endocrine therapy resistance in breast cancer [
13,
14]. However, no evidence-based clinical data are available for c-Met inhibitors in breast cancer treatment. Despite the fact that the prognostic role of c-Met in breast cancer has been discussed since the 1990s [
15,
16], there is no consensus on its impact. Some studies suggest that c-Met is a stronger prognostic indicator of poor prognosis than traditional markers such as Her2/neu and epidermal growth factor receptor (EGFR) [
17‐
19], whereas others show no statistically significant relation between c-Met and prognosis in breast cancer [
20,
21]. In recent years, c-Met was reported to be associated with favorable prognosis in breast cancer patients [
22,
23]. Therefore, systematic studies are necessary to obtain high level evidence-based results of the prognostic value of c-Met for the identification of patients who would benefit from c-Met targeted therapy and to guide future clinical trials.
In the present study, we enrolled and combined all eligible published studies analyzing the relationship between c-Met expression and relapse free survival (RFS) or overall survival (OS) in breast cancer to clarify the relationship between c-Met expression and prognosis in breast cancer. c-Met plays a critical role in early-stage invasion of cancer cells [
24], and crosstalk of c-Met signaling pathways with estrogen receptor (ER) and HER-2 signaling pathways has been reported [
13,
25]. To validate the prognostic role of c-Met in different subtypes breast cancer, we performed a subgroup analysis in lymph node negative and different molecular subtypes of breast cancer.
Discussion
In recent years, the development of target-based therapies has improved the prognosis of cancer patients. However, only a subset of patients benefits from the use of specific drugs, and the development of resistance often results in clinical treatment failure. The identification of novel targets is a challenging task for the medical oncologist, and valuable prognostic markers might become potential therapeutic targets in the future. The trasmembrane tyrosine kinase receptor c-Met plays a vital role in cell survival, growth and metastasis [
8]. c-Met is overexpressed in a variety of carcinomas and is associated with resistance to herceptin and gefitinib, and it represents an attractive target for antitumor treatment [
13,
43]. c-Met overexpression has been reported in 14–53.6 % of patients with breast cancer [
20,
39,
40]. Evidence of the influence of c-Met expression on survival outcomes in breast cancer is inconclusive. In the present study, we analyzed 21 studies published between 1998 and 2014 and comprising a total of 6010 cases. The results of the meta-analysis showed that c-Met overexpression is a statistically significant adverse predictor of RFS and OS in unselected breast cancer. These results provide evidence supporting future trials evaluating the effect of c-Met inhibitors in breast cancer.
Originally, Iressa, a selective EGFR inhibitor, showed promising results among Asian patients, but not in Western populations, suggesting a possible role of ethnic differences between Asian and Western lung cancer patients [
44]. The differences in the characteristics of breast cancer between Asian and Western countries have also been discussed for several years [
45]. In the present study, we performed a subgroup analysis according to ethnicity. In the Western patient group, there were 8 studies analyzing RFS and 13 studies analyzing OS according to c-Met expression. Our results showed that c-Met is a predictor of poor prognosis (both RFS and OS) in Western patients. In the Asian patient group, four studies analyzing c-Met expression according to OS/RFS were identified. The results showed that there was a trend toward increased recurrence and mortality in c-Met overexpressing patients, although the difference did not reach statistical significance. Further analysis including a larger number of patients and studies is necessary to evaluate the prognostic role of c-Met in Asian breast cancer patients, and to determine whether c-Met status has a different influence on the prognosis of Asian and Western breast cancer patients.
Lymph node status is the best indicator of prognosis in breast cancer. Additional makers are necessary to predict prognosis in patients with lymph node negative breast cancer. C-Met expression is higher and more frequently positive in metastatic lymph nodes than in the primary tumor [
19]. In the present analysis, three studies provided data on OS in lymph node negative patients. The meta-analysis results showed that c-Met overexpression was associated with a 2.04-fold increased risk of mortality (combined HR 2.04, 95 % CI 1.48–2.80;
P < 0.0001) in lymph node negative breast cancer. These results demonstrate that c-Met might act at the early stages of breast cancer, and its expression should be detected on postoperative pathology to predict prognosis and guide the postoperative treatment.
Breast cancer is divided into five molecular subtypes based on the status of ER, PR, HER-2 and Ki67 [
3]. In the present study, we performed subgroup analysis according to molecular subtypes. Four studies provided data on RFS in the hormone-receptor positive subgroup. The meta-analysis results showed that c-Met overexpression was associated with a 1.41-fold increased risk of recurrence (combined HR 1.41, 95 % CI 1.11–1.79;
P = 0.005) in the hormone-receptor positive group. Endocrine therapy is the most important systemic treatment for hormone-receptor positive breast cancer at all stages [
46]. C-Met and the Ron receptor tyrosine kinase, a member of the c-Met family of receptors, are associated with resistance to breast cancer endocrine therapy in vitro [
14,
47]. Overexpression of HER-2 is associated with resistance to endocrine therapy in breast cancer [
48]. Zagouri et al. showed that c-Met was not a prognostic factor in ER- and HER2-positive breast carcinomas [
20]. In addition, the prognostic value of c-Met was shown to be independent of HER2/neu [
19]. Consequently, c-Met might influence the prognosis of hormone-receptor positive patients by mediating resistance to endocrine therapy, especially in the hormone-receptor positive/HER-2 negative subgroup in a Her-2 independent manner. This subgroup is likely to benefit from combined treatment with c-Met inhibitors and estrogen inhibition therapy in the future. However, additional studies are needed to confirm these results.
Functional crosstalk of c-Met with HER-2 has been reported to enhance cell invasion in Madin-Darby canine kidney (MDCK) epithelial cells in vitro [
49]. In breast cancer cells, this crosstalk is involved in the development of Herceptin resistance in vitro [
13]. In the present analysis, two studies provided RFS data in HER-2 positive patients. The meta-analysis results showed that c-Met overexpression was associated with poor prognosis, but the findings did not reach statistical significance (combined HR 1.20, 95 % CI 0.91–1.59;
P = 0.20). Additionally, Minuti et al. found that c-Met is associated with shorter time to progression (TTP) in HER2-positive metastatic breast cancer [
36]. Thus, additional studies are necessary to explore the clinical interaction of c-Met and Her-2.
According to currently available data, TNBC is the most aggressive subtype of breast cancer, and no targeted therapy is currently available [
39]. TNBC could be further subclassified into basal-like breast cancer (BLBC) and quintuple-negative breast cancer (QNBC), and c-Met is involved in the development of BLBC [
33]. Five studies provided RFS data in the TNBC subgroup. The meta-analysis results showed that c-Met overexpression increased recurrence risk by 2.31-fold in TNBC (combined HR 2.31, 95 % CI: 1.53–3.48,
P < 0.0001), which was the highest risk in this study. The results indicate that c-Met could be a therapeutic target, thereby providing new treatment options for TNBC.
Quality assessment according to REMARK guidelines was performed for all 21 included studies. The studies fulfilled, on average, 14 items (range, 10–18 items) of the guidelines. Sensitivity and sub-group analyses were performed to ensure that the results were reliable and valid. However, our meta-analysis had several limitations. First, the results of sub-analysis were less powerful because the combined HR of some subgroups was calculated on the basis of 2–5 studies with a relative small patient sample size. Second, c-Met was detected by five different methods, although most studies detected c-Met by IHC (excluding the molecular subtype groups). In addition, there were differences in the criteria for c-Met positivity in IHC detection. Third, the funnel plot analysis showed some asymmetry, suggesting the possibility of publication bias in unselected patients and Western patients. The trim and fill sensitivity analysis did not change the general results, suggesting that the results were not influenced by the unpublished negative studies or the small sample size. Additional high-quality data are necessary to draw more reliable conclusions.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
All authors participated in the study design. SY performed the literature search, data extraction, statistical analysis and drafted the manuscript. XJ, HZ and KL participated in data extraction. All authors read and approved the final manuscript.