Abstract
Background: This study aims to build a focal adhesion-related genes-based prognostic signature (FAS) to accurately predict gastric cancer (GC) prognosis and identify key prognostic genes related to gastric cancer. Results: Gene expression and clinical data of gastric cancer patients were sourced from Gene Expression Omnibus and The Cancer Genome Atlas. Subsequently, the GEO dataset was randomly distributed into training and test cohorts. The TCGA dataset was used to validate the external cohort. Lasso Cox regression was used to detect OS-related genes in the GEO cohort. A risk score model was established according to the screened genes. A nomogram, based on the clinical characteristics and risk score, was generated to predict the prognosis of gastric cancer patients. Using time-dependent receiver operating characteristic (ROC) and calibration performances, we evaluated the models' validity. The patients were grouped into a high- or low-risk group depending on the risk score. Low-risk patients exhibited higher OS than high-risk patients (entire cohort: p < 0.001; training cohort: p < 0.001, test cohort: p < 0.001). Furthermore, we found a correlation between high-risk gastric cancer and extracellular matrix (ECM) receptor interaction, high infiltration of macrophages, CD44, and HLA-DOA. Conclusion: The generated model based on the genetic characteristics of the focal adhesion prognostic gene can aid in the prognosis of gastric cancer patients in the future.