Background: Occult dissemination of tumor cellular and nuclear material into the bloodstream is well-recognized and has been reported for a variety of tumor types, including pancreatic ductal adenocarcinoma (PDAC). These cells and nuclear material can be detected in the blood as circulating tumor cells (CTCs) or cell-free DNA (cfDNA) as a means of "liquid biopsy" and their presence is referred to as "tumoremia." Limited data suggest that various biopsy procedures and surgery can also precipitate tumoremia. There are no published data indicating the risk of hematogenous tumor seeding following EUS FNA and no data clarifying whether any factors may be associated with its occurrence. In PDAC patients, we aimed to prospectively evaluate peripheral blood to determine the risk of EUS FNA-induced tumoremia and to identify the associated clinical, laboratory, and tumor demographic features. Methods: Peripheral blood from 35 patients with PDAC was collected before and following primary tumor EUS FNA. Plasma cfDNA concentrations and KRAS 12/13 mutation status were evaluated. Tumor seeding was defined by a !2fold increase in cfDNA and/or conversion from circulating tumor deoxyribonucleic acid (ctDNA) KRAS mutant non-detectable to detectable. Factors that were evaluated to determine their correlation with tumoremia included age, gender, BMI, ECOG status, serum glucose level, CA 19-9 concentration, tumor location, tumor diameter, tumor stage, resectability status, number of FNA passes, infiltration of major arteries and veins, presence of IPMN, mortality, survival duration, pre-FNA cfDNA level, and the tumor allelic ratio. Results: Based on our pre-defined metrics, either a !2-fold increase in the total cfDNA concentration (nZ7; 20.6%), and/or conversion in ctDNA mutant detection status (nZ3; 8.8%) primary tumor EUS FNA-induced hematogenous tumor seeding occurred in 10/35 (28.6%) patients. A broad spectrum of clinical, laboratory, and tumor demographic features were evaluated, none of which statistically correlated with a !2-fold increase in cfDNA, other than the number of major arteries and/or veins infiltrated by the primary tumor. (Table ) The only feature that correlated with the conversion in KRAS mutation status was the CA 19-9 level. For the combined group (nZ10), who met either metric for FNA-induced tumoremia, there were no characteristics that correlated with these endpoints. Conclusions: Our data indicate that the occurrence of hematogenous tumor seeding associated with EUS FNA at the present time, cannot be predicted based upon an evaluation of an extensive array of clinical, laboratory, and tumor related characteristics. Additional studies are needed to determine the clinical impact of EUS FNAinduced tumor seeding as well as to verify whether any predictive features can be identified to help predict its occurrence.
