HER2 discordance between primary and metastatic gastric cancer: a systematic review and meta-analysis

Abstract Purpose Gastric cancer (GC) remains a major global health burden, with poor survival rates. HER2 is a key biomarker for targeted therapy, but discordance between primary tumors and metastases may impact treatment decisions. This meta-analysis evaluated the prevalence and clinical relevance of HER2 status differences between primary gastric cancers and metastatic lesions. Methods A systematic search of PubMed, Embase, and Cochrane was conducted for studies assessing HER2 status in matched primary and metastatic gastric cancer samples. Pooled proportions and 95% confidence intervals (CIs) were calculated using random-effects models. Subgroup analyses by metastatic site were performed. Statistical analyses were conducted using R software, version 4.2.3. Results This meta-analysis included 3,515 patients across 20 studies. The pooled proportion of HER2-positive expression in primary gastric tumors was 13% (95% CI: 12–15%; I 2 = 68.7%), while HER2-negative tumors accounted for 74% (95% CI: 72%-76%; I 2 = 94.9%). In metastatic sites, the overall pooled proportion of HER2-positive lesions was 18% (95% CI: 16%-20%; I 2 = 77.3%), with notable variation across anatomical locations: 38% in lung metastases, 31% in liver metastases, 19% in lymph nodes, and 7% in peritoneum. Conversely, HER2-negative metastases accounted for 82% overall (95% CI: 80%-84%; I 2 = 78.2%), with proportions of 93% in peritoneum, 81% lymph nodes, 69% in liver metastases, and 62% in lung. Conclusions Significant heterogeneity in HER2 expression between primary and metastatic gastric cancer underscores the need for reassessment of HER2 status in metastatic sites. Relying solely on primary tumor samples may lead to underestimation of HER2 positivity and suboptimal therapeutic decisions. Incorporating site-specific HER2 testing into clinical practice may enhance patient selection for HER2-targeted therapies and improve treatment outcomes. Graphic Abstract