OBJECTIVE To evaluate the use of dual-phase helical computed tomography (CT) (with or without CT angiography) to assess resectability in patients suspected to have pancreatic cancer. METHODS Tumor resectability was prospectively evaluated in 89 patients who later underwent surgery for suspected pancreatic cancer. Helical CT scans were obtained in the vascular phase and a phase of maximal hepatic enhancement. CT angiograms were produced with multiprojection volume reconstruction and maximum-intensity projection. CT results were correlated with surgical and histopathologic results. RESULTS Helical CT allowed detection of pancreatic cancer in 74 of 76 cases (97%). There were six false-positive results (positive predictive value, 92%). For prediction of irresectability, helical CT had an accuracy of 91%, negative predictive value of 79%, and sensitivity of 91%. Helical CT allowed detection of liver metastases in 21 of 28 cases (75%), nodal involvement in 13 of 24 cases (54%), and vascular invasion in 35 of 40 cases (88%). CT angiography demonstrated 30 of the 35 cases of vascular invasion detected with helical CT (86%). CONCLUSIONS Use of dual-phase helical CT improves prediction of resectability in patients with pancreatic cancer. CT angiography cannot show all of the findings seen on helical scans.