Abstract Pancreatic cancer is an almost uniformly lethal malignancy, in large part because is it rarely detected in its early, curable stages. We recently conducted a study to evaluate the use of proteins associated with extracellular vesicles (EVs) isolated using an alternating current electrokinetic (ACE) system for the early detection of pancreatic ductal adenocarcinoma. Because earlier detection of pancreatic cancer greatly improves outcomes, we sought to devise a blood-based test with high specificity for detecting early-stage PDAC based on identification of EV-bound biomarkers. EVs were purified from control or stage I-II pancreatic ductal adenocarcinoma (PDAC) patient plasma using an ACE microelectrode array. Following collection of EVs onto the electrodes, contaminating plasma elements were washed away, and the electric field was turned off to release EVs into buffer. The isolated EV/exosome samples were then tested using bead-based multiplex immunoassay kits to determine relative concentrations of biomarkers of interest. Sample cohorts were divided into a training set (controls, N=146; PDAC cases, N=33) to develop a logistic regression-based classification model, and a validation set (controls, N=139; PDAC cases, N=35) to generate a probability-based score for predicting the presence of PDAC. For the training set, the algorithm identified six protein biomarkers associated with the presence of stages I-II cancer in PDAC patient plasma, with a receiver-operating characteristic (AUC) of 0.997 (95% CI: 0.978-1.000) compared to control plasma samples. For the validation set, the algorithm had a sensitivity of 82.9% (95% CI: 67.3% - 91.9%) at a specificity of 99.3% (95% CI: 96.0% - 99.9%), generating an AUC = 0.978 (95% CI=0.954-0.995). The exosomal PDAC assay shows feasibility for developing minimally invasive liquid biopsies that allow low-risk, routine screening of high-risk patients. Those patients identified by the test would be able to undergo additional testing to determine if there is PDAC present. The additional screening and testing of these patients may increase the likelihood of curative treatment. Further work in large datasets is warranted. Citation Format: Juan P. Hinestrosa, Jean M. Lewis, Heath I. Balcer, Razelle Kurzrock, Scott Lippman, Rajaram Krishnan. Blood-based extracellular vesicle biomarker test for detection of early-stage pancreatic cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 5960.