Prognostic value of c-MET in head and neck cancer: A systematic review and meta-analysis of aggregate data.

Objectives: The hepatocyte growth factor (HGF)/mesenchymal-epithelial transition factor (c-MET) ligand/receptor axis has been implicated in pathogenesis of malignant diseases including squamous cell carcinoma of the head and neck (SCCHN). Overexpression of c-MET has been reported as a common molecular abnormality in SCCHN, although its prognostic and predictive value remains to be validated.
Methods: We systematically searched literature for studies evaluating c-MET expression on immunohistochemistry in newly diagnosed, non-metastatic SCCHN. The c-MET expressing cases were classified into three categories according to predefined cut-off values for positivity. Our aim was to assess the prevalence of c-MET expression and its relationship with selected clinicopathological variables.
Results: Twenty-eight studies with 2019 cases were included. Relative frequencies of c-MET expression above cut-off levels I, II, and III were 81.8%, 63.8%, and 46.2%, respectively. Differences between these three values were statistically significant (p<1.0×10 ^-6 ). Above cut-off level II, c-MET positivity was associated with worse overall survival (p=4.0×10 ^-6 ), positive nodal status (p=1.0×10 ^-4 ), higher disease stage (p=7.0×10 ^-4 ), older age (p=2.1×10 ^-3 ), disease recurrence (p=2.0×10 ^-2 ), and primary tumour localization in the oral cavity (p=2.3×10 ^-2 ). Above cut-off level III, c-MET positivity was associated with worse disease-free or progression-free survival (p=9.0×10 ^-6 ), p16 negativity (p=2.4×10 ^-4 ), worse overall survival (p=4.0×10 ^-4 ), positive epidermal growth factor receptor (EGFR) status (p=7.2×10 ^-4 ), and larger primary tumours (p=4.6×10 ^-3 ).
Conclusion: In SCCHN, immunohistochemical overexpression of c-MET above cut-off levels III and particularly II was associated with inferior survival outcomes and advanced disease. Moreover, it represents a promising predictive biomarker for c-MET targeting, yet the optimal scoring method remains to be defined.
(Copyright © 2017 Elsevier Ltd. All rights reserved.)