A Multi-Center Cohort-Based circRNA Diagnostic Model for Detection of Gastric Cancer

BACKGROUND: Gastric cancer (GC) remains one of the most detrimental diseases to human health. Owing to the subtle nature of early symptoms and the absence of robust and effective screening biomarkers, most patients are diagnosed at an advanced stage. Herein, our objective is to establish a non-invasive diagnostic strategy based on circular RNAs (circRNAs) to facilitate the detection of GC. METHODS: We conducted a comprehensive genome-wide screening to identify key circRNAs, which were subsequently validated via RT-qPCR and translated into a plasma-based liquid biopsy analysis. The Chi-square test was applied to evaluate the relationship between circRNA expression and clinicopathological parameters. Receiver Operating Characteristic (ROC) curves were employed to evaluate the diagnostic efficacy of circRNAs in GC. A logistic regression model was established for the prediction of GC and validated in independent clinical cohorts. RESULTS: In the discovery phase, we identified 2 circRNA candidates, hsa_circ_0001185 and hsa_circ_0005265, which were subsequently found to be significantly upregulated in the serum of GC patients through liquid biopsy analysis. The Chi-square test revealed that elevated expression of these circRNAs was significantly correlated with differentiation grade, lymph node metastasis, and TNM stage. ROC analysis demonstrated that hsa_circ_0001185 and hsa_circ_0005265 effectively discriminated GC patients from non-diseased controls, with AUC values of 0.909 and 0.853, respectively. The Diagnostic Model for GC (GC-DM) we developed exhibited an AUC of 0.915, which was subsequently validated in two independent cohorts. CONCLUSION: We developed a GC diagnostic model based on hsa_circ_0001185 and hsa_circ_0005265, demonstrating robust non-invasive diagnostic potential for the detection of GC patients.