Entropy-Driven Nucleic Acid Amplifier Based on Spatial Confinement as a “Booster” for Detection of Extracellular Vesicle MicroRNAs to Diagnose Gastric Cancer and Monitor Therapeutic Response

Gastric cancer (GC) continues to pose a significant global health burden with persistent diagnostic challenges, especially in the detection of early-stage GC. Herein, a strand displacement reaction-mediated nucleic acid amplifier based on the spatial confinement (SC-SDR) effect as a “booster” is constructed to detect extracellular vesicle-derived microRNAs (EVs-miRNAs). Constraining the reactant and fuel strands in a limited space using a T-shaped DNA structure results in a significant improvement in the reaction kinetics and sensitivity because of the high local strand concentrations, ultimately enabling the detection of EVs-miRNAs at the femtomolar level. SC-SDR is conjugated onto a hydrophobic tether to aid delivery into EVs, allowing for the in situ detection of EVs-miRNAs. Four EVs-miRNAs act as biomarkers in combination with a random forest (RF) algorithm for use in GC diagnostics, prognostics, and early warning. In a cohort of 58 patients with GC, this diagnostic model effectively identifies 51 of the 58 cases, showing a satisfactory accuracy of 87.93%. This diagnostic efficiency outperforms that of conventional biomarkers (CEA and CA19-9), which exhibit accuracies of only 25.86% (15/58) and 17.24% (10/58), respectively. The longitudinal analysis of EVs-miRNA expression in GC patients before and after surgery and in patients with gastric intraepithelial neoplasia (GIN) reveals the dual utility of this approach as both a robust prognostic biomarker for GC progression and a promising predictive marker for GIN development. Overall, this study highlights the combined power of SC-SDR and machine learning for the analysis of EVs-miRNAs, paving the way for the clinical diagnosis and prognostication of GC.