Tu1659 Forward Viewing EUS Versus Standard Oblique Viewing EUS in the Performance of EUS-Guided Fine Needle Aspiration of Solid Lesions: a Prospective Randomized Controlled Study

Abstract

Tu1658 Does EUS Add to the Value of High-Resolution CT Staging in Pancreatic Carcinoma? Jayaprakash Sreenarasimhaiah*, Deepak Agrawal Medicine/Gastroenterology, University of Texas Southwestern, Dallas, TX Background: Pancreatic carcinoma often presents at an advanced stage at diagnosis. Distant metastatic disease is easily detected by computerized tomography (CT). Additionally, the presence of vascular invasion or celiac node (CN) involvement may deem a tumor unresectable. Computerized tomography (CT) and endoscopic ultrasound (EUS) are frequently used to determine resectability. Aims: To determine if EUS provides additional detail to CT staging of pancreatic carcinoma in regard to vascular invasion and CN involvement. Methods: The endoscopic database was examined retrospectively for a 5 year-period between July 2007 and July 2012. Consecutive patients with pancreatic carcinoma proven by EUS-FNA (fine needle aspiration) were included. Patients with CN visualized by EUS were examined and compared to pancreas protocol CT scan performed with 2 and 5mm image slices. The detection of invasion into great vessels adjacent to the tumor was also examined for each modality. Results: 117 patients with EUS-biopsy-proven carcinoma were identified. Mean age was 60 years with 59% males. Mean size of largest tumor dimension was 3.9cm. Tumors were located in the pancreas head or uncinate process in 88 (75%), tail in 10 (9%), and in the body or neck in 19 (16%). 30/117 (26%) had visible CN on EUS while 10/117 (9%) had visible CN on CT scan. In 21/30 nodes seen on EUS, size was O1cm and had morphology suspicious for malignancy. EUSFNA demonstrated 11/21 CN with malignant, 4/11 atypical or non-diagnostic, and 6/ 11 benign cytology. 4/11 (41%) patients with malignant CN did not have visible nodes on CT. In 21 patients with abnormal CN, EUS identified 13 with vascular invasion, 4 with portal vein abutment, and 4 without invasion. CT imaging was consistent with EUS in identifying vascular invasion in all 21 patients. Conclusions: Celiac lymphadenopathy was detected three-fold more commonly by EUS than CT scan. Malignant CN was missed in approximately 40% of cases that EUS identified. While vascular invasion was identified similarly by both modalities, its presence did not predict CN. Thus, complete staging of pancreatic carcinoma should include EUS despite findings by high-resolution CT imaging. AB548 GASTROINTESTINAL ENDOSCOPY Volume 81, No. 5S : 2015 HIGH-RESOLUTION