Cancer-derived small extracellular vesicles (csEVs) play critical roles in the genesis and development of various cancers. However, accurate detection of low-abundance csEVs remains particularly challenging due to the complex clinical sample composition. In the present study, we constructed a Programmable Isothermal Cascade Keen Enzyme-free Reporter (PICKER) for the reliable detection and acquisition of the relative abundance of csEVs in total sEVs (tsEVs) by integrating dual-aptamer recognition (cancer-specific protein EpCAM and tetraspanin protein CD63) with a catalytic hairpin assembly (CHA) amplification. By employing this strategy, we were able to achieve a detection limit of 420 particles/μL csEVs. Particularly, we proposed a novel particle ratio index of csEV against tsEV (PRcsEV/tsEV) to greatly eliminate errors from inconsistent centrifugation, which was calculated from the fluorescence ratio produced by csEVs and tsEVs. The PICKER showed a 1/10,000 discrimination capability by successfully picking out 1.0 × 103 csEV from 1.0 × 107 tsEV per microliter. We also found that the PRcsEV/tsEV value increased proportional to the stages of breast cancer by analyzing EVs from clinical patients' plasma. Taken together, we established a PICKER strategy capable of accurately discriminating csEVs, and the proposed PRcsEV/tsEV had been proven a potential indicator of breast cancer staging, paving the way toward facilitating cancer diagnosis and precision therapeutics.