OBJECTIVES: The main objective is to compare the accuracy of EUS and CEH EUS for the diagnosis of pancreatic cancer (PC). The secondary objective is to evaluate the accuracy of EUS FNA and to determine to what extent EUS and CEH EUS findings are affected by endosonographer subjectivity. METHODS: A prospective single-centre study was conducted in patients with pancreatic lesions detected on CT. The patients were examined by EUS, CEH EUS and EUS FNA. The obtained results were compared with the final diagnosis that was based on cytology and further clinical findings and on histopathological findings from subjects who underwent surgery. A second reading of the EUS and CEH EUS images was performed by the endosonographer, who was blinded to clinical data of patients. RESULTS: We examined 116 patients, 73 had a final diagnosis of PC, 14 had NETs and 20 had other tumours. The sensitivity, specificity, NPV, PPV, and accuracy of EUS for diagnosis of PC were 83.1, 62.5, 83.1, 70.7 and 78.6%, for CEH EUS 94.5, 61.7, 84.1, 84 and 84.1% and for EUS FNA 87.6, 91.2, 95.5, 77.5 and 88.8, respectively. The inter-observer agreement for EUS marker of PC was good (κ = 0.75), and that for CEH EUS was average (κ = 0.59 for arterial phase and κ = 0.68 for washout in venous phase). CONCLUSION: CEH EUS is a non-invasive method that allows more accurate identification of PC than EUS. The subjectivity of CEH EUS evaluation is worse than that of EUS but acceptable.

b Department of Neurosurgery and Neurooncology 1st Faculty of Medicine Charles University Military University Hospital Prague Czech Republic

c Department of Pathology Military University Hospital Prague Czech Republic

Department of Internal Medicine 1st Faculty of Medicine Charles University Military University Hospital Prague Czech Republic

References provided by Crossref.org