Extracellular vesicles (EVs)-derived miRNAs are valuable non-invasive biomarkers for early cancer diagnosis. However, most current methods detect either total EVs-miRNAs or those from a single subpopulation, overlooking subpopulation heterogeneity. Moreover, EVs isolation and miRNA analysis are often conducted in separate workflows, increasing sample handling and compromising reproducibility. Herein, we presented an integrated, single-tube platform for subpopulation-specific miRNA profiling, in which multivalent aptamer-based EVs capture and CRISPR/Cas12a-mediated signal amplification were co-localized on a rationally designed DNA scaffold. The multivalent aptamer system achieved a capture efficiency of 84.3%, significantly outperforming monovalent aptamers (56.4%). Upon in situ lysis of captured EVs subpopulation, the released miRNAs preferentially encountered adjacent recognition probes due to the increased local concentration, thereby efficiently initiating catalytic hairpin assembly (CHA). The resulting abundant duplex subsequently activated Cas12a, achieving a limit of detection (LOD) as low as 1.42 × 10particles/μL. Critically, the modularity of the platform allowed for straightforward reconfiguration to target distinct EVs subpopulations by exchanging the capture aptamer. We demonstrated this flexibility by profiling miR-21 and miR-155 across two specific subpopulations (CD63and MUC1). The results revealed both cell-line-specific expression patterns and marked heterogeneity across CD63/MUC1-captured EVs fractions. Meanwhile, MUC1subpopulation outperformed CD63in clinical diagnosis, with higher AUC values for both miR-21 (0.85 vs 0.75) and miR-155 (0.83 vs 0.73). Overall, our study highlighted the critical importance of EVs subpopulation heterogeneity in cancer diagnostics and provided more informative biomarker readouts. 4 + + + +
