Increased MET gene copy number negatively affects the survival of esophageal squamous cell carcinoma patients

This retrospective study was conducted in a cohort of 495 treatment-naive ESCC patients who underwent esophagectomy at Zhongshan Hospital between January 2007 and December 2010. Patients were included in the study if the following criterias were met: (1) underwent primary resection, (2) with no prior treatment, and (3) with available complete medical records. Patients were excluded from the study if they had disease progression within three months after surgery. Clinical and histopathological data, including sex, age, smoking status, tumor size, tumor location, differentiation, vessel or nerve invasion, pT stage, and pN stage, was obtained from the patients’ medical and pathological records. The pathologic tumor-node-metastasis (pTNM) stage was performed according to the 8th edition of the American Joint Committee on Cancer (AJCC) staging system. All patients were followed up every 3–6 months after tumor resection, and patients underwent follow-up examinations to identify possible tumor recurrence. Exam methods included endoscopy, computed tomography, magnetic resonance imaging, abdominal ultrasonography, and measurement of serum tumor marker levels.

Written informed consent was obtained from all patients, and the study was approved by the ethical committee of the Zhongshan Hospital, in accordance with the ethical standards of the World Medical Association Declaration of Helsinki.

TMA construction was performed as previously described [25]. Briefly, histological sections were examined by a pathologist, and representative tumor areas free from necrosis or hemorrhage were pre-marked in formalin-fixed paraffin-embedded (FFPE) donor blocks. Two or three core tissues (2 mm in width and 6 mm in length) from different representative areas per case were taken from the donor blocks and arranged in recipient blocks (tissue array blocks). Our TMAs contained the tumor samples, several normal esophagus and other control tissues.

MET gene status was evaluated using a commercially available FISH assay [26], with Vysis MET Spectrum Red FISH Probe (Abbott Molecular, Chicago, IL, USA) and control Vysis CEP7 Centromere Spectrum Green Probe (Abbott Molecular) on 4 μm-thick TMA sections. The signals of each sample were counted in at least 50 well-defined nuclei using a fluorescence microscope (BX43, Olympus, Tokyo, Japan) equipped with a Microscope Digital Camera (DP73, Olympus, Tokyo, Japan). An average MET gene copy number ≥ 5 and a MET/CEP7 ratio ≥ 2 (true MET amplification) were regarded as MET FISH positive [22].

The patients’ clinicopathological characteristics are summarized in Table 1. The patient group consisted of 408 men (82.4%) and 87 women (17.6%) with a median age of 61 years (range, 34–83 years). One hundred ninety-nine subjects (40.2%) were ever-smokers or smokers, whereas 296 (59.8%) were nonsmokers. The mean tumor size was 3.4 cm. By anatomic site, 47.9% of tumors were located in the lower esophagus, 47.0% in the middle esophagus, and 5.1% in the upper esophagus. The tumors were poorly differentiated in 40.2%, moderately differentiated in 56.0%, and well differentiated in 3.8%. Vessel and nerve invasion were identified in 110 (22.2%) and 178 (36.0%) tumors, respectively. There were 9.3% patients at pathologic stage T1, with 22.2, 68.3, and 0.2% at stages T2, T3, and T4, respectively. About pathologic N stages, there were 53.3, 25.9, 15.8, 5.1% patients at N0, N1, N2, and N3 stages respectively. According to the 8th edition of TNM staging, 38 patients (7.7%) were classified as having stage I disease, 234 patients (47.3%) as stage II, 193 patients (39.0%) as stage III, and 30 patients (6.1%) as stage IVa.

Among 495 patients, 28 (5.7%) cases were MET FISH positive (an average number of MET signals per nucleus ≥5.0), including 5 cases (1%) with true gene amplification (5 cases with MET: CEP7 ratio of ≥2.0) (Fig. 1c and d). Other specimens showed disomy or low polysomy (94.3%) (Fig. 1a and b).

The correlations between MET FISH-positivity and clinical pathologic characteristics are listed in Table 1. MET FISH-positivity was significantly associated with DFS (2.2% in patients without disease progression vs. 8.6% in patients with disease progression, P = 0.002) and OS (2.6% vs. 8.4%, P = 0.005). However, there were no statistically significant difference in sex (P = 0.638), age (P = 0.932), smoking (P = 0.919), tumor size (P = 0.434), tumor location (P = 0.941), differentiation (P = 0.957), vessel invasion (P = 0.194) and nerve invasion (P = 0.706), pT stage (P = 0.883), pN stage (P = 0.088), and clinical stage (P = 0.351).

The median follow-up time was 35.0 months (range 3–102 months). Two hundred sixty-nine patients (54.3%) had disease progression and two hundred sixty-one patients (52.7%) had died from esophageal cancer during the follow-up. The 5-year DFS and disease-specific OS rates for all patients were 44.1 and 44.4%, respectively.

Figure 2a and b reveals that a significantly poorer prognosis was observed in 28 patients with MET FISH-positivity, showing a median DFS or OS of 17.0 or 26.0 months, respectively, compared with 36.0 or 42.0 months in the group with MET FISH-negativity (P < 0.001 or P = 0.001). The 5-year DFS (17.9%) and OS (17.8%) rates for patients with MET FISH-positivity were significantly lower than the corresponding rates (45.7 and 46.0%) for patients with MET FISH-negativity. Univariate analysis indicated that MET FISH-positivity, differentiation, vessel invasion, nerve invasion and clinical stage had significant impacts on DFS, and MET FISH positive, vessel invasion, nerve invasion and clinical stage had significant impacts on OS (both P < 0.05). Multivariate analysis revealed MET FISH-positivity was an independent prognostic factor for DFS (HR, 1.953; 95% CI, 1.271–2.999; P = 0.002) and OS (HR, 1.926; 95% CI, 1.243–2.983; P = 0.003). Clinical stage was also found to be an independent prognostic factor for DFS and OS (Table 2).

In stage I-II patients, one hundred four patients (38.2%) had disease progression and one hundred one patients (37.1%) had died from esophageal cancer during the follow-up. In stage III-IVa patients, one hundred sixty-five patients (74.0%) had disease progression and one hundred sixty patients (71.7%) had died from esophageal cancer during the follow-up.

Figure 2c and d reveals that a significantly poorer prognosis was observed in 223 stage III-IVa patients, showing a median DFS of 20.0 months or OS of 25.0 months, respectively, compared with not-reached median survival in 272 stage I-II patients (P < 0.001). The 5-year DFS (23.7%) and OS (24.7%) rates for stage III-IVa patients, were significantly lower than the corresponding rates (60.8 and 60.4%) for stage I-II patients.

MET FISH-positivity was associated with DFS (P = 0.022) and OS (P = 0.046) in patients with stage I-II ESCC (Fig. 3a and b). In detail, a poorer prognosis was observed in 13 patients with MET FISH-positivity, with a median DFS or OS of 21.0 or 38.0 months, respectively, while those with MET FISH-negativity (n = 259) did not reach the median survival. MET FISH-positivity was also associated with DFS (P = 0.020) and OS (P = 0.024) in patients with stage III-IVa ESCC (n = 223) (Fig. 3a and b). In detail, a poorer prognosis was observed in 15 patients with MET FISH-positivity, with a median DFS or OS of 12.0 or 18.0 months, respectively, while those with MET FISH-negativity (n = 208), with a median DFS or OS of 20.0 or 25.0 months, respectively. What’s more, no statistical significance (DFS, P = 0.492 and OS, P = 0.344) was detected between stage I-II ESCC with MET FISH-positivity and stage III-IVa ESCC with FISH-negativity.