Endosonography for malignant lymphadenopathy: Caveats and pitfalls

Endoscopic utrasonography (EUS) is a sensitive tool for detecting lymphadenopathy, although not equivalent to histology. We report our prospective experience evaluating for malignant adenopathy in patients with GI malignancies. 128 patients with confirmed malignancies and who had not received prior chemo- or radiotherapy, underwent EUS (Olympus GF-UM20) followed by complete surgical staging. Lymph nodes were considered to be malignant by EUS criteria if they exhibited the following features: round to oval, hypoechoic, homogeneous, well-demarcated, and proximity to a primary tumor. Nodal staging by EUS had the following performance characteristics: Primary: n Sensitivity Specificity Accuracy Esophagus 32 86% 73% 81% Pancreas 39 47%*90% 69% Rectal 24 67% 78% 75% Other (a) 33 65% 93% 79% Total 128 67% 85% 76% (a) Stomach, Ampulla, Bile Duct, Gallbladder, Duodenum*Fisher p<0.01 vs Esophagus Ten patients had false-positive nodes (Esophagus-3, Rectum-4, Pancreas-2, Stomach-1). Lymph node pathology on these cases was reviewed: 4-lipid granuloma (Pancreas-2, Esophagus-1, Stomach-1), 1 large caseating granuloma (Esophagus), 8 reactive (Esophagus-3, Pancreas-2, Rectum-3), 2 with anthracotic pigment (Esophagus), 1 no specific pathologic change (Rectum). Conclusions: 1. Using these criteria, EUS is approximately 75% accurate for malignant lymphadenopathy. 2. EUS is more Sensitive for peri-esophageal than peripancreatic adenopathy. 3. Granulomas (lipid and other types), in addition to reactive and anthracotic changes, may be important causes of false positive EUS N-staging.