Search

Background and study aims The wide use of PET-CT for the staging of neoplasms has extended to enlarged lymph nodes of unknown origin. Nevertheless, upper abdominal and mediastinal nodes are accessible to endoscopic ultrasonography-fine needle aspiration (EUS-FNA), providing a cytological diagnosis, with a high diagnostic yield in previous reports. In this paper, we have compared the accuracy of both procedures in this particular setting. Patients and methods After the finding of a lymphadenopathy in a conventional CT, both PET-CT and EUS-FNA were performed. The endoscopist had no information about PET-CT results. FNA was performed after a systematic EUS exam, with a 25 G needle and no suction. We considered the pathologic results as the gold standard or, if not available, the patients' outcome as a surrogate marker. Results A total of 54 patients were included. Common locations of nodes included subcarinal space (33.3%), porta hepatis (31.5%), upper mediastinum (15%), peripancreatic (7.4%), and other locations (12.8%). Adequate specimens were obtained in 48/54 patients (89%). The most common diagnoses based on cytology were benign/reactive (42%), epidermoid carcinoma (8.4%), lymphoma (8.4%), and ductal adenocarcinoma of pancreatic origin (6.3%). PET was positive in 67% of patients. The sensitivity, specificity, positive predictive value, negative predictive value, and overall accuracy of EUS-FNA were 91.3, 100, 100, 92.5, and 95.8%, respectively. The same values for PET-CT were 75, 25, 50, 50, and 50%, respectively. Conclusion In our series, EUS-FNA has proven to be the best diagnostic procedure to accurately establish the etiology of isolated adenopathies, showing a much better diagnostic yield than PET-CT, the role of which should be re-evaluated in this setting.