Gastroesophageal cancer (GEC) is an aggressive disease characterized by a high frequency of metastasis and poor overall survival rates. GEC presents HER2 overexpression in 5 to 25% of tumors eligible for HER2-targeted therapy. HER2 evaluation requires protein levels and copy number alteration analyses by immunohistochemistry (IHC) and in situ hybridization (FISH or SISH), respectively. These are semiquantitative methodologies that need an expert and well-trained pathologist. Therefore, the use of new surrogate methods for HER2 evaluation in cancer, such as gene expression analysis, might improve GEC HER2 classification. We evaluated HER2 positivity in GEC through conventional IHC and SISH analyses and investigated the potential application of HER2 mRNA expression by quantitative PCR to categorize GEC samples as HER2-positive or HER2-negative. Among 270 GEC samples, 10.9% were HER2-positive by IHC and SISH analyses. HER2 mRNA was overexpressed in HER2-positive GEC samples and presented high accuracy in distinguishing those tumors from HER2-negative GEC. Nevertheless, HER2 mRNA analysis was not capable of classifying HER2-equivocal GEC samples into HER2-positive or -negative according to SISH data. Quantitative PCR analysis showed HER2 overexpression in HER2-positive GEC samples. Nevertheless, HER2 mRNA analysis failed to classify HER2-equivocal GEC according to SISH data.