Circulating pancreatic ductal adenocarcinoma (PDAC) derived small extracellular vesicles (sEVs) are nano-sized membranous vesicles secreted from PDAC cells and released into surrounding body fluids, such as blood. The use of plasma-derived sEVs for cancer diagnosis is particularly appealing in biomedical research because the sEVs reflect some key features (e.g. genetic and phenotypic status) related to the organs from which they originate. For example, the surface membrane proteins and their expression level on sEVs were reported to be related to the presence and progression of PDAC. However, difficulty in sEVs isolation and lack of ultrasensitive assays for simultaneous analysis of multiple protein biomarkers on patient plasma-derived sEVs hinder their application in the clinic. In our previous study, we have demonstrated the application of magnetic beads (MBs) and surface-enhanced Raman scattering (SERS) assay for phenotypic analysis of cancer cells-derived sEVs using different cell lines. To further demonstrate the clinical application of the proposed assay, we have profiled the sEVs' phenotypes (relative expression of biomarker Glypican 1, EpCAM and CD44V6) of healthy donors and PDAC patients to enable simultaneous detection of multiple surface membrane proteins on plasma-derived sEVs. We discovered that the PDAC sEVs' phenotype signatures had high accuracy for PDAC diagnosis (100%) and showed strong correlation with cancer stages, which were further validated by the imaging techniques (e.g. computerized tomography and magnetic resonance imaging) and also the correlation of cancer stages with CA19-9 (gold standard biomarker) and the sEVs' phenotype signatures. The present proof-of-concept study thus provides an initial investigation of using the proposed SERS assay for PDAC diagnosis and early cancer stage prediction in the clinic.