Abstract
INTRODUCTION: The peritoneum is the most common metastatic site of gastric cancer and is associated with a dismal prognosis. However, there is no reliable biomarker for predicting peritoneal metastasis (PM). MATERIALS AND METHODS: Whole-exome sequencing (WES) was performed on formalin-fixed, paraffin-embedded (FFPE) samples from 63 patients with stage I-III gastric cancer and circulating tumor DNA (ctDNA) samples from 10 patients with stage IV gastric cancer. Differentially expressed genes (DEGs) were identified between the PM and non-PM groups and analyzed by multiple bioinformatics analyses. Univariate and multivariate Cox regression analyses were used to identify the risk factors for PM and a risk score model was developed. RESULTS: The number of mutant genes and the tumor mutation burden (TMB) in the PM group were higher than those in the non-PM group (p < 0.05). There was a significant positive correlation between the number of mutant genes and the TMB (R(2) = 0.9997). The risk of PM was significantly higher in the high TMB group than in the low TMB group (p = 0.045). Forty-nine DEGs were identified as associated with PM in gastric cancer. CDC27 mutations were associated with a higher risk for PM and poor survival. The CDC27 mutations were located in the Apc3 region, the TPR region, and the phosphorylation region, and new mutation sites were not included in the TCGA database. Multivariable Cox regression analysis demonstrated that pathological T stage, poor tumor differentiation, Borrmann type, and CDC27 mutations were independent predictive factors of PM. A risk score model was constructed that demonstrated good performance. CONCLUSION: Through WES, we identified 49 DEGs relevant to PM in gastric cancer. CDC27 mutations were independently associated with PM by statistical and bioinformatics analyses. A risk score model was built and was demonstrated to effectively discriminate gastric cancer patients with and without PM.