Updated FTP location. Clinical monitoring and prediction on drug ADME behaviors is of great essence due to the variability from individual patients. To date, pharmacogenomic approaches have allowed genotypic detection on a few metabolizing enzymes, mainly limited to cytochrome P450 polymorphisms. However, it is inadequate to hinge on genotype to sketch the whole real-time phenotype. Fortunately, the booming field of biofluid extracellular vesicles (EVs) offers an attractive option in clinical liquid biopsy. And the proteins inside EVs are able to reveal the phenotypes of parental cells from lesions or tissues, for example, EV metabolizing enzymes to reflect metabolism in liver. With liver tissue EV proteome as foundation, herein, we utilized a tandem mass tag-carrier strategy to characterize ADME proteins from plasma EVs. Mined from 44 identifiable and prevalent ADME proteins, 34 stringently quantified proteins emerged. As a result, previously-annotated mouse cytochrome P450 3A11 (Cyp3a11, homolog to human CYP3A4) and UDP glucuronosyltransferase 2A3 (Ugt2a3) were kinetically induced upon daily rifampicin dosage. Besides, dasatinib (Sprycel), a tyrosine kinase inhibitor to treat several subtypes of leukemia, was proved to timely elevate Cyp3a11 abundance in mouse plasma EVs, but less strong than rifampicin. Altogether, EV-based liquid biopsy is a promising direction to describe ADME activities in an interindividual manner.