Comprehensive analysis of risk factors associated with submucosal invasion in patients with early-stage gastric cancer.

Background: Submucosal invasion in early-stage gastric cancer (GC) is a critical determinant of prognosis and treatment strategy, significantly influencing the risk of lymph node metastasis and recurrence. Identifying risk factors associated with submucosal invasion is essential for optimizing patient management and improving outcomes.

Aim: To comprehensively analyze clinical, imaging, and endoscopic characteristics to identify predictors of submucosal invasion in patients with early-stage differentiated GC.

Methods: A retrospective study was conducted at our institution from January 2019 to January 2023, including 268 patients diagnosed with early-stage differentiated GC who underwent surgical resection or endoscopic submucosal dissection. Data were collected on demographic, clinical, imaging, and endoscopic characteristics, with endoscopic images reviewed independently by two gastroenterologists. Statistical analysis included univariate and multivariate logistic regression to identify significant predictors of submucosal invasion, and receiver operating characteristic (ROC) curve analysis to evaluate the predictive value of continuous variables.

Results: A total of 268 patients were included, with 178 males and 90 females, and a mean age of 61.5 ± 9.8 years. Univariate analysis showed that male gender, history of alcohol consumption, smoking, and computed tomography-detected gastric wall thickening were more prevalent in patients with submucosal invasion. Significant endoscopic predictors included tumor location in the upper two-thirds of the stomach, depressed morphology, marginal elevation, and high color differences on white-light endoscopy (WLE) and linked color imaging (LCI). Multivariate analysis identified upper stomach location [odds ratio (OR): 5.268], depressed type (OR: 5.841), marginal elevation (OR: 4.132), and LCI color difference ≥ 18.1 (OR: 4.479) as significant predictors. ROC analysis showed moderate predictive value for lesion diameter, WLE, and LCI color differences (area under the curve: 0.630, 0.799, and 0.760, respectively).

Conclusion: Depressed-type lesions, marginal elevation, location in the upper two-thirds of the stomach, and significant color differences on LCI are high-risk indicators for submucosal invasion. These findings suggest that such lesions warrant more aggressive intervention to prevent disease progression and improve patient outcomes.