Sensitive small extracellular vesicles associated circRNAs analysis combined with machine learning for precision identification of gastric cancer

Small extracellular vesicle associated circular RNAs (sEV-circRNAs) are emerging as promising biomarkers for gastric cancer diagnosis. Current research predominantly focuses on identifying these biomarkers through high-throughput sequencing. However, there has been insufficient exploration into the practical application of sEV-circRNAs for early gastric cancer diagnosis. In this study, we developed a sensitive electrochemical platform that leverages tetrahedron-Dox-AuNPs (TDA) tag and DNA tetrahedron-enhanced catalytic hairpin assembly (DT-CHA) to detect sEV-circRNAs. Based on the dual signal amplification of the TDA tag and DT-CHA, the platform can achieve low-concentration detection of the target, with a detection limit of 153.1 aM and a linear range from 1 fM to 1 nM. By profiling four sEV-circRNAs (circNRIP1, circRANGAP1, circCORO1C, and circSHKBP1) in a gastric cancer cohort and combining suitable ML diagnostic model, this platform performed well in distinguishing healthy donors from early GC patients. Thus, this confluence of a multi-biomarker approach with machine learning analysis, applied to plasma sEV-circRNAs, emerges as an important strategy for cancer screening.