From Tumor Pathways to Blood Signature: A Machine Learning-Validated 5-miRNA Signature for the Early Detection of Gastric Cancer

Background: Gastric cancer (GC) remains a leading cause of global cancer mortality, with late-stage diagnosis contributing significantly to poor patient outcomes. Circulating microRNAs (miRNAs) offer promise due to their stability in biofluids and established roles in carcinogenesis. However, existing miRNA biomarker candidates for GC suffer from inconsistent validation across studies, limited specificity, and insufficient mechanistic links to gastric tumor biology. We addressed this by integrating tissue and blood transcriptomics to identify GC-specific miRNAs, which were then validated using machine learning. Methods: Dysregulated genes (DEGs) and miRNAs (DEMs) were identified from tissue mRNA (GSE54129, GSE113255) and blood miRNA/mRNA datasets (GSE106817, GSE174302). Pathway enrichment (Reactome) revealed GC-specific pathways shared between tissue DEGs and blood DEM targets. Targets of 59 DEMs were enriched in these pathways in the blood miRNA dataset. From these, a 5-miRNA panel was selected using 10 machine learning feature selection methods (e.g., Gini Index, Information Gain) and validated using Random Forest and Naïve Bayes classifiers on discovery (GSE106817) and external (GSE164174) datasets. Results: Integration identified 59 GC-specific extracellular miRNAs linked to 39 enriched pathways (e.g., signaling, metabolism). The 5-miRNA panel (hsa-miR-124-3p, hsa-miR-23a-3p, hsa-miR-22-3p, hsa-miR-29b-3p, hsa-miR-92a-3p) achieved near-perfect discovery performance (RF: AUC=98.50%, ACC=98.36%) and high external validation (AUC=95.30%, ACC=89.24%). Conclusion: Our pipeline bridges tissue pathology and circulating miRNA profiles, yielding a highly specific 5-miRNA Blood Signature with clinical diagnostic potential for GC.