ABSTRACT
Gastric cancer stands as the fourth leading cause of cancer-related deaths globally, primarily comprising adenocarcinomas, categorized by anatomic location and histologic type. Often diagnosed at advanced stages, gastric cancer prognosis remains poor. To address the critical need for accurate tumoral markers for gastric cancer diagnosis, we conducted a study to assess classical markers like CEA and CA-19-9 alongside the novel marker miR-106. Our investigation revealed distinct dynamics of these markers compared to non-cancerous groups, although no disparities were observed across different disease stages. Univariable and multivariable logistic regression analyses demonstrated that elevated levels of miR-106, CEA and CA 19-9 were predictive of a positive histopathological exam, with the respective odds ratios of 12.032 (95% CI: 1.948-74.305), 30 (95% CI: 3.141-286.576), and 55.866 (95% CI: 4.512-691.687). Subsequently, we utilized predicted probabilities from regression models to construct receiver operating characteristic (ROC) curves, identifying CA 19-9 as the optimal predictor for gastric adenocarcinoma diagnosis when considering age and gender, with an area under the curve (AUC) of 0.936 (p < 0.001). Hence, classical markers exhibit superior performance compared to the novel marker miR-106 in predicting gastric adenocarcinoma.
