We evaluated positon emission tomography (PET) with 18F-fluorodeoxyglucose (FDG) in 120 patients with intestinal malignancies, focusing on its diagnostic yield and influence on the surgical strategy. PET had a sensitivity of 67% and a specificity of 100% for metastases in 28 patients with cardio-esophageal carcinoma. PET detected 64% of 22 primary pancreatic carcinomas, and had a sensitivity of 70% and a specificity of 83% for metastases. In two cases, PET showed false-positive signs of peritoneal metastasis (not found at laparotomy). Among 70 patients with recurrent or metastatic colorectal carcinoma, eight had signs of local recurrence of rectal carcinoma treated by abdominoperineal resection; PET gave four true-positive, one false-negative, and three false-positive results. PET was better than computed tomography (CT) for the diagnosis of peritoneal metastasis, but its sensitivity was only 58%. The diagnostic value of PET for hepatic metastases (87%) was similar to that of CT (77%) and sonography (87%). The diagnostic sensitivity of PET for pulmonary metastases (82%) was similar to that of CT (84%). PET modified the surgical strategy in two (7%) of 28 patients with cardio-esophageal carcinoma, one (5%) of 22 patients with pancreatic carcinoma, and 22 (33%) of 70 patients with colorectal carcinoma (appropriately in 11 cases, inappropriately in 11 cases). These disappointing results suggest that PET must be thoroughly evaluated in this setting before being widely adopted.