Diagnosis of Prostate Cancer Metastasis via Extracellular Vesicles Isolated Using Two‐Phase Interface as Membrane‐Less Filter

Abstract Extracellular vesicles (EVs), nano‐sized particles secreted by cells, are increasingly recognized as promising biomarkers. However, the isolation and purification of EVs need improvement, impeding their practical application. Aqueous two‐phase systems (ATPS) offer a method to separate EVs with high purity and yield compared to other techniques, yet the unclear isolation mechanism limits efficiency. To elucidate the separation process and enhance ATPS‐based EV isolation, Kramers' theory and Fick's law are employed. The simulations and experiments reveal that the liquid–liquid interface in ATPS acts as a size cut‐off filter for EVs, functioning without a membrane. It is discovered that rapid transport of particles to the interface is crucial for fast isolation, but this transport in separated phases relies solely on diffusion, which slows the process. To address this, a vortex is introduced to enhance particle movement through convection, significantly improving efficiency. This method achieves over 80% recovery of EVs from blood plasma and removes more than 90% of low‐density lipoprotein, high‐density lipoprotein, and albumin within an hour. Applying this ATPS‐based membrane‐less filter to plasma from prostate cancer patients, concentrations of markers on EVs are quantified. Using machine learning, metastatic and non‐metastatic prostate cancer are distinguished with greater accuracy than the traditional PSA‐based method.