Background: c-Met gene amplification has been identified as one of the acquired resistance to epidermal growth factor receptor (EGFR) tyrosine kinase inhibitors (TKIs) and de novo activation in advanced non-small-cell lung cancer (NSCLC). However, it is not clear whether c-Met overexpression could be as the biomarker for de novo.Methods: Advanced NSCLC patients with de novo c-Met expression were detected by immunohistochemistry (IHC) .≥50% tumor cells with moderate to high intensity staining were defined as c-Met positive. Gene copy numbers have been detected by FISH ( By Cappuzzo scoring system ≥5 copies were positive or MET/CEP7 ratio ≥1.8 was defined as c-MET amplification ).The statuses of EGFR, ALK, KRAS and ROS1were also tested. Results: From January 2013 to December 2014, 24 eligible patients with c-Met IHC overexpression received crizotinib treatment (3 female, median age 59 years), with 19 evaluable for response.Eleven of them achieved partial response (PR), 3 were stable disease (SD) and 5 were progressive disease (PD).All responders had high c-Met IHC status, and 8 with FISH positive. (Table 1).Adverse events of grade 3 QT prolongation have been found in 1 patient. For one death with interstitial lung disease, causality to crizotinib was not ruled out. The other most frequent drug-related AEs were grade1-2, including nausea (14/19), anorexia (14/19), vomiting(10/19),visual impairment (6/19). EGFR, ALK, KRAS and ROS1 were all negative . Accrual of advanced patients is ongoing. Conclusions: c-Met overexpression could be as a biomarker for de novo c-Met amplified NSCLC. c-Met inhibitor against de novoc-Met overexpressed NSCLC is a good strategy. IHC seems not worse than FISH in predicting efficacy for c-Met inhibitor.