Document resource
Background Gastric cancer (GC) is one of the most common and lethal malignancies worldwide. Early detection significantly improves prognosis, yet gastroscopy is invasive and impractical for population-wide screening. Non-invasive biomarkers are therefore needed to identify high-risk individuals and guide targeted endoscopic referral. Circulating miRNAs are stable in serum and represent promising liquid biopsy candidates. However, aging systematically reshapes miRNA profiles, and conventional biomarker studies rarely account for this, limiting their specificity. GC patients exhibit miRNA aging patterns distinct from those of healthy individuals. We therefore developed an miRNA-based aging clock to capture this dysregulation and evaluate its utility for non-invasive GC detection.Methods We analyzed public serum miRNA datasets (HC, n=912; GC, n=1,471) and an independent Wenzhou cohort (HC, n=54; GC, n=58). Age-associated miRNAs were identified in public HC samples using Pearson and Spearman correlation (adjusted P<0.001). These miRNAs were used to train a 40-miRNA aging clock on 70% of the public dataset (n=644) and validated in the remaining 30% (n=268). We then applied this clock to all samples and defined the miRNA Aging Dysregulation Index (miRADI) as the ratio of predicted to chronological age. A 6-miRNA GC classifier was further developed by applying Lasso-based stability selection to the clock-derived miRNAs. Diagnostic performance was assessed by AUC in both cohorts, with CA19-9 comparison in the Wenzhou cohort.Results The aging clock accurately predicted chronological age in healthy validation samples (Pearson R=0.74; MAE=4.86 years). When applied to GC patients, the clock revealed significant molecular age deceleration compared to HC across all cohorts (P<0.0001). We quantified this deviation as miRADI, which alone discriminated GC from HC with AUC >0.80. The 6-miRNA classifier achieved an internal AUC of 0.97 and an external AUC of 0.90. Notably, it identified stage I GC in the Wenzhou cohort with AUC=0.93, significantly outperforming CA19-9 (AUC=0.81). Target enrichment analysis confirmed links to aging-related and oncogenic pathways.Conclusions We developed a serum miRNA aging clock that captures shared biology between aging and GC. The miRADI index provides a non-invasive risk metric, and the 6-miRNA classifier reliably detected early-stage GC, outperforming CA19-9. This framework may support risk stratification and guide targeted endoscopic referral.