Abstract
OBJECTIVE: This study aims to investigate the correlation between the tumor-stroma ratio (TSR) and peritoneal metastasis (PM) in gastric cancer (GC) and constructs a diagnostic model based on preoperative examination data. METHODS: To determine the feasibility of obtaining TSR in GC patients through preoperative examinations, the consistency of TSR between endoscopic biopsy tissues and postoperative histopathological tissues was evaluated. Additionally, the correlation between TSR and PM in GC was analyzed using Gene Expression Omnibus (GEO) datasets. To validate TSR's clinical potential in diagnosing PM, 640 GC patients from two medical centers were enrolled. A training cohort of 330 patients evaluated TSR and synchronous PM correlation, and a validation cohort of 310 patients was used. An additional cohort of 510 patients was established to investigate TSR and metachronous PM. A diagnostic model based on preoperative data was developed and a nomogram constructed. RESULTS: The TSR shows good consistency between endoscopic biopsy tissues and postoperative histopathological tissues. A significant correlation between TSR and PM was observed. The TSR-based model, combined with CA125, CA724 and Borrmann type, exhibited strong diagnostic effectiveness and considerable predictive efficacy, with an Area Under the Curve (AUC) of 0.85 in the training cohort, 0.73 in the external validation cohort, and 0.72 in the metachronous PM cohort. CONCLUSION: The TSR emerges as a crucial marker for PM in GC, with the developed model, based on TSR and preoperative examination data, demonstrating substantial diagnostic and predictive capabilities.