RESUMO
Four-hundred eighty-five patients underwent US examination; 183 of them had gastric cancer, 239 colorectal cancer, 38 pancreatic cancer, 11 esophageal cancer, and 14 had gastric lymphoma. All patients underwent surgery. In 95 cases fine-needle biopsy under US guidance was performed. Lymphadenopathies were classified by the criteria proposed by Yoshinaka et al., type I: poorly-defined borders, diffuse internal echoes; type II: well-defined borders, diffuse internal echoes; type III: well-defined borders, notchings, strong internal echoes. Twenty/twenty-nine type I, 66/98 type II, and 39/43 type III adenopathies were found to be neoplasm-positive. Of 73 patients with adenopathy from gastric cancer, 9 were type I, 42 were type II, and 22 were type III (183 patients examined); of 9 patients with adenopathy from esophageal cancer, 7 were type II and 2 were type III (11 patients examined); of 48 patients with adenopathy from colorectal cancer, 5 were type I, 28 were type II, and 15 were type III (239 patients examined); of 29 patients with adenopathy from pancreatic cancer, 7 were type I, 18 were type II, and 4 were type III (38 patients examined); finally, of 11 patients with adenopathy from gastric lymphoma, 8 were type I, and 3 were type II (14 patients examined). The relationship between US and pathology was possible from a statistical point of view only. Type I lymphadenopathies seem to suggest lymphomatous involvement, whereas type III ones suggest metastatic involvement. US is a valid approach method, which must be supported by other investigation techniques--e.g., CT and lymphography--in order to avoid high false-negative percentages.