BACKGROUND AND AIMSCurrently, there are no biomarkers with high accuracy in the detection of pancreatic cancer. This article aims to evaluate the performance of a novel diagnostic model based on a combination of biomarkers in the washing fluid obtained by EUS-FNA with imaging examination in pancreatic cancer.METHODSThis study included 59 patients with pancreatic lesions who underwent EUS-FNA and were categorized into malignant and benign groups on the basis of pathology diagnosis. The levels of CEA, CA19-9, CA125, CA724, CYFRA 21-1, IMP3, SMAD4, and S100P in EUS-FNA washing fluid were detected by ELISA. We attempted to construct a new diagnostic method by combining the above biomarkers with EUS, CT, MRI, and PET-CT.RESULTSCEA, CA19-9, CA724, CA125, and CYFRA 21-1 showed statistical significance in the diagnosis of pancreatic cancer (AUC > 0.7, p < 0.05). CA724 had a specificity of up to 100% in the group with positive EUS diagnosis. If at least two positive imaging results (EUS/CT/MRI/PET-CT) combined with at least one tumor marker (CEA/CA199/CA724) in series, the sensitivity was 88.57% and the specificity was 91.67%.CONCLUSIONSCombining the tumor markers CEA, CA19-9, CA724, CA125, and CYFRA 21-1 in the washing fluid of EUS-FNA with commonly used imaging methods can help distinguish benign and malignant pancreatic lesions.