Background/Aims: To investigate the accuracy of multidetector computed tomography (MDCT) in preoperatively determining the surgical resectability of pancreatic adenocarcinomas. Materials and Methods: Multidetector computed tomography, surgery, and pathological results of 274 patients with pancreatic adenocarcinoma were evaluated retrospectively. MDCT findings were compared with surgical and pathological findings to determine the sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy of MDCT in determining surgical resectability. Results: A total of 124 of 274 (56%) patients (83 males, mean age: 60 years) underwent laparoscopy and/or laparotomy. The sensitivity, specificity, PPV, NPV, and accuracy of MDCT in determining the surgical resectability of pancreatic adenocarcinomas were 100%, 72%, 78%, 100%, and 86%, respectively. Liver metastases in 9 cases, peritoneal metastases in 3 cases, and vascular invasion in 5 cases, which were determined during surgery, were not reported by MDCT. On re-review of the MDCT images of these 17 patients, no metastatic lesions could be seen in 9 patients with liver metastases and in 2 of 3 patients with peritoneal metastases. In 1 patient, a peritoneal implant of a diameter of 8 mm was missed on MDCT. There was no vascular invasion according to Lu criteria on the MDCT images in the 5 cases that had vascular invasion in the surgical exploration. Conclusion: The accuracy of MDCT is high in the preoperative determination of surgical resectability of pancreatic adenocarcinomas, but the detection of small liver and peritoneal metastases and accurate determination of vascular invasion are still major problems. Surgeons should be aware of the limitations of preoperative MDCT.