It has been reported that HER2 expression is different in gastric and breast cancers, and a gastric cancer scoring system (GCSS) has recently been suggested. We investigated HER2 protein expression using GCSS and a breast cancer scoring system (BCSS) and correlated it with HER2 gene amplification. HER2 status was evaluated in 1091 cases by analyzing tissue microarrays constructed using 2 different cores from each case. Polyclonal (HercepTest) and monoclonal (Pathway) antibodies were used for immunohistochemistry (IHC), and results were scored by BCSS and GCSS. Gene amplification was evaluated by automated dual-color silver-enhanced in situ hybridization (SISH) in all cases and correlated with the results from fluorescence in situ hybridization (FISH) in 590 cases. The concordance between the IHC results using polyclonal and monoclonal antibodies was high (κ = 0.785). The results of dual-color SISH and FISH showed very high concordance as well (κ = 0.918). GCSS was significantly more sensitive for detecting SISH positivity than was BCSS in both antibodies (polyclonal, P = .003; monoclonal, P < .001), but specificity was higher in BCSS than GCSS (polyclonal, P = .004; monoclonal, P < .001). It has been recently shown that HER2-overexpressing patients with unresectable gastric cancer benefited from trastuzumab therapy. Because IHC is recommended before gene amplification studies in HER2 testing, GCSS should be used for evaluating HER2 expression in gastric cancers. Copyright © 2012 Elsevier Inc. All rights reserved.