Mo1408 Diagnostic Accuracy & Strength of Agreement Between EUS and Histopathology in the Staging of Ampullary Tumor

Diagnostic Accuracy & Strength of Agreement Between EUS and Histopathology in the Staging of Ampullary Tumor Sandeep Lakhtakia, Eric Wee, Rajesh Gupta, Sekaran Anuradha, Mahesh G. Shetty, Rakesh Kalapala, Amitabh Monga, Arjunan Saravanan, Pradeep Rebala, Guduru V. Rao, D. Nageshwar Reddy Gastroenterology, Asian Institute of Gastroenterology, Hyderabad, India; Surgical Gastroenterology, Asian Institute of Gastroenterology, Hyderabad, India; Pathology, Asian Institute of Gastroenterology, Hyderabad, India INTRODUCTION: Accurate staging of an ampullary tumor is crucial to determine its resectability and the type of surgery. Small ampullary lesions which are not well visualized by CT scanning can be assessed in detail by endoscopic ultrasound (EUS). AIMS: To compare the EUS features of ampullary tumor with histopathological (HP) staging in loco-regional staging. METHODS: Data was collected retrospectively from a single center. We included patients with an ampullary tumor who underwent EUS evaluation (with either mechanical or electronic linear/radial EUS scopes) and surgical resection. Tumor (T) and nodal (N) TNM staging for EUS and HP were compared. Patients with prior gastrointestinal surgery, prior cancer treatment or metastatic disease were excluded. The strength of agreement and diagnostic accuracy between EUS and HP staging was calculated. Uni-variate logistic regression analysis of variance was performed to look for potential variables influencing staging. RESULTS: From 2009 to 2010, a total of 79 patients with ampullary tumors were identified. Of these, 28 did not undergo resection, 13 had palliative surgery only, and 12 had resection without EUS. The remaining 26 patients who underwent both EUS and Whipple’s surgery were included [(mechanical EUS scope: 17 (65.4%), electronic 9 (34.6%)]. The diagnosis was an adenocarcinoma in 25 (96.2%) and an adenoma in 1 (3.8%). The mean age was 52 years (73.1% males, 26.9% females). Nine (26.9%) patients had trans-papillary biliary stent placed prior to EUS and surgery. The tumor diameter assessment was not significantly different by EUS and HP (p 0.606). For T staging by HP staging: there were 2 (7.7%) T1, 11 (42.3%) T2, 12 (46.2%) T3 and 1 (3.8%) T4 tumors. EUS under-staged 5 (71.4%) and overstaged 2 (28.6%). The accuracy of EUS T staging was 73.1% with a kappa ( ) value of 0.564 (p 0.0001). The sensitivity, specificity, PPV, NPV of EUS respectively was: 50.0%, 91.7%, 33.3% and 95.7% for T1 tumors; 81.8%, 80.0%, 75.0% and 85.7% for T2; 75.0%, 92.9%, 90.0% and 81.3% for T3 tumors. For N staging by HP: 17 (65.4%) were N0 and 9 (34.6%) N1. EUS over-staged 3 (60.0%) lesions and under-staged 2(40.0%). The N staging diagnostic accuracy was 80.8% with a of 0.586 (p 0.003). The sensitivity, specificity, PPV, NPV for N0 disease was 82.4%, 77.8%, 87.5% and 70.0%, while for N1 was 77.8%, 82.4%, 70.0% and 87.5%. The type of echo-endoscope used (mechanical/electronic), trans-papillary biliary stent, T staging, size, grade of differentiation, presence of dilated common bile duct and dilated pancreatic duct did not influence the staging accuracy (p 0.05). CONCLUSION: EUS has a moderate strength of agreement with histopathology for both T and N staging, and a high diagnostic accuracy for nodal staging. Trans-papillary biliary stent and type of echo-endoscope did not influence the staging accuracy.