Objective: Increased serum human epidermal growth factor receptor 2 levels have been found in metastatic breast cancer patients and are correlated with human epidermal growth factor receptor 2 overexpression in tumor cells. However, the prevalence of serum human epidermal growth factor receptor 2 in gastric cancer patients has not been elucidated. Methods: We retrospectively analyzed formalin-fixed paraffin-embedded tumor tissues and serum samples from 96 advanced gastric cancer patients. Human epidermal growth factor receptor 2 expression and gene amplification in tumor cells were determined by immunohistochemistry and fluorescence in situ hybridization. Serum human epidermal growth factor receptor 2 levels were measured using a chemiluminescent immunoassay. Human epidermal growth factor receptor 2 positivity in tumor cells was defined as immunohistochemistry 2+ with fluorescence in situ hybridization positive or immunohistochemistry 3+ with any fluorescence in situ hybridization results. Results: All tissue samples and serum samples were successfully measured. Nineteen patients (20%) were human epidermal growth factor receptor 2-positive in tumor cells. The median serum human epidermal growth factor receptor 2 level was 9.3 ng/ml (range, 5.0-332.4 ng/ml), and serum human epidermal growth factor receptor 2 levels were significantly separated according to human epidermal growth factor receptor 2 status in tumor cells (P < 0.0001, Wilcoxon's rank sum test); median serum human epidermal growth factor receptor 2 levels in human epidermal growth factor receptor 2-negative patients and -positive patients were 8.9 (range, 5.0-20.5) and 24.0 (range, 9.7-332.4), respectively. There were 15 serum human epidermal growth factor receptor 2-positive patients (16%) using a cutoff value of 15 ng/ml. The sensitivity and the specificity of serum human epidermal growth factor receptor 2 with respect to human epidermal growth factor receptor 2 positivity in tumor cells were 53 and 94%, respectively. Conclusions: Serum human epidermal growth factor receptor 2 measurements cannot be substituted for tissue human epidermal growth factor receptor 2 diagnosis in advanced gastric cancer patients. However, serum human epidermal growth factor receptor 2 levels are associated with human epidermal growth factor receptor 2 overexpression in tumor cells. Further investigations of clinical significance of serum human epidermal growth factor receptor 2 as a predictive marker and a therapy-monitoring marker are warranted.