Background: The procoagulant activity of tissue factor-bearing microvesicles (MV-TF) has been associated with the risk of developing venous thrombosis in cancer patients. However, MV-TF assays are limited either by i) a lack of specificity, ii) a low sensitivity, or iii) a lack of repeatability when high-speed centrifugation (HS-C) is used to isolate MV. Therefore, our objective was to develop a new hybrid "capture-bioassay" with improved reproducibility combining MV immunocapture from biofluids and measurement of their TF activity.Materials and methods: Factor Xa generation and flow cytometry assays were used to evaluate IMS beads performance, and to select the most effective capture antibodies. The analytical performance between IMS-based and HS-C-based assays was evaluated with various models of plasma samples (from LPS-activated blood, spiked with tumoral MV, or with saliva MV) and different biofluids (buffer, plasma, saliva, and pleural fluid).Results: Combining both CD29 and CD59 antibodies on IMS beads was as efficient as HS-C to isolate plasmatic PS + MV. The IMS-based strategy gave significantly higher levels of MV-TF activity than HS-C in tumor MV spiked buffer, and both pleural fluids and saliva samples. Surprisingly, lower TF values were measured in plasma due to TFPI (TF pathway inhibitor) non-specifically adsorbed onto beads. This was overcome by adding a TFPI-blocking antibody. After optimization, the new IMS-based assay significantly improved reproducibility of MV-TF bioassay versus the HS-C-based assay without losing specificity and sensitivity. In addition, this approach could identify the cellular origin of MV-TF in various biological fluids.Conclusion: Compared to HS-C, the IMS-based measurement of MV-TF activity in body fluids improves reproducibility and makes the assay compatible with clinical practice. It can facilitate future automation.