271 Background: Circulating rare cells (CRCs) such as circulating tumor cells (CTCs), tumor-associated macrophages (TAMs) and cancer-associated fibroblasts (CAFs) play major roles in tumor progression. However, there are limited reports characterizing single CRCs comprehensively because of technical limitations including isolation and analysis of rare cells. We aimed to test the feasibility of the automated picking and mRNA profiling platforms of CRCs for evaluating the phenotypic heterogeneity of CRCs and the relationship between clinical outcomes and CRCs in patients with pancreatic cancer. Methods: CRCs were collected from 3ml of whole blood in 4 patients with pancreatic cancer using filter-based lab-on-a-disc platform. The filter membrane was transferred to ALS CellCelector (ALS, Germany) which is picking single nucleated CD45 negative cells to 96-well plate. The single cells were analyzed with the BioMark HD (Fluidigm, USA) which is digital PCR-based mRNA profiling for 48 markers such as epithelial, mesenchymal, stem-like, macrophage, fibroblast, oncogenic and immunogenic markers. Results: CRCs were detected in 3 of 4 patients (75%). Sixty-six-year-old female who had large pancreatic cancer with liver and lung metastasis presented 13 single rare cells including 7 epithelial CTCs with hematopoietic features (CD45 and CD14 positive), 3 mesenchymal CTCs, 2 stem-like CTCs, and 1 CAF. One epithelial CTC with hematopoietic features and one stem-like CTC were detected in 51-year-old male and 61-year-old female with metastatic pancreatic cancer, respectively. There was no CTC in 80-year-old male with localized pancreatic cancer. Conclusions: The automated picking and mRNA profiling platforms showed the feasibility of showing tumor heterogeneity and predicting clinical outcomes in patients with pancreatic cancer.