In adenocarcinoma of the esophagus and esophagogastric junction for prognostication and treatment allocation, one prerequisite is accurate pretreatment staging. This staging, we hypothesized, would be improved by the use of positron emission tomography (PET). After 55 patients suitable for radical esophageal resection were staged with PET, spiral computed tomography (CT), and endoscopic ultrasonography (EUS), results were compared with histopathology and with survival. Accuracy in detecting locoregional lymph node metastasis did not differ significantly between EUS (72%), PET (60%), and CT (58%). Adding PET to standard staging failed to improve the accuracy of N staging (P=0.250). In M staging, accuracy between CT (75%) and PET (76%) did not differ. The accuracy of combined studies of CT and PET and of EUS, CT, and PET were 87% (P=0.016 versus CT) and 91% (P=0.031 versus EUS and CT), respectively. Of the 55 patients, 19 (35%) had metastatic lesions. By combined use of CT and EUS and by combined use of CT, EUS, and PET, 8 and 14 (P=0.031), respectively, could be detected. In nodal disease without distant metastases, PET did not improve the prediction of survival. However, positive PET for distant metastasis by either positive EUS or CT predicts well the poor survival of these patients. The staging value of PET by itself in adenocarcinoma of the esophagus is limited because of low accuracy for nodal and the lack of specificity for distant disease prognosis. Adding PET to standard staging does, however, improve detection of stage IV disease and its associated poor survival.