Introduction-The accuracy of endoscopic ultrasound (EUS)for the diagnosis and staging of pancreatic ductal adenocarcinoma (PDA) has been confirmed. Chemoradiotherpay (CRT) induces tumor changes which can limit the accuracy of EUS. The aim of our study was to analyze the efficacy of EUS following neoadjuvant CRT comparing findings with the pathology results. Patients and methods-From November 1996 to October 2003, 45 patients with histologically proven and EUS-staged PDA were treated with neoadjuvant CRT and radical surgery. All were restaged before surgery using both EUS and computed tomography. Fifteen patients were found to hove developed distant metastases. Thirty patients finally underwent pancreaticoduodenectomy (N = 24) or distal pancreatectomy (N = 6). Results - Following CRT, tumor stage was correctly assessed in 12 patients (40% ). The most frequent misinterpretation was overestimation of tumor size (N = 13, 43.3% ). Locoregional vascular invasion of veins was suspected by EUS in 13 patients (43.3% ) but surgical findings and the histological examination were both negative. Node status was correctly assessed in 27 patients (90% ) but nodal involvement was found on the histological specimen in only 3 patients. Conclusion - Preoperative EUS after neoadjuvant CRT for PDA does not enable reliable definitive selection of patients for surgery, probably due to radiation- induced pancreatic changes. Introduction-The accuracy of endoscopic ultrasound (EUS) for the diagnosis and staging of pancreatic ductal adenocarcinoma (PDA) has been confirmed. Chemotherapy of tumor changes which can limit the accuracy of EUS. The aim of our study was to analyze the efficacy of EUS following neoadjuvant CRT comparing findings with the pathology results. From November 1996 to October 2003, 45 patients with histologically proven and EUS-staged PDA were treated with neoadjuvant CRT and radical surgery. All were restaged before surgery using both EUS and computed tomography. Fifteen patients were found to hove developed distant metastases. Thirty patients finally underwent pancreaticoduodenectomy (N = 24) or distal pancreatectomy (N = 6). Results - Following CRT, tumor stage was correctly assessed in 12 patients (40% ). The most frequent misinterpretation was overestimation of tumor size (N = 13, 43.3%). Locoregional vascular invasion of veins was suspected by EUS in 13 patients (43.3%) but surgical findings and the histological examination were both negative. Node status was correctly assessed in 27 patients (90%) but nodal involvement was found on the histological specimens in only 3 patients. Conclusion - Preoperative EUS after neoadjuvant CRT for PDA does not enable reliable definitive selection of patients for surgery, probably due to radiation-induced pancreatic changes.