significantly change management in at least 2 cases: 1) possible hyperechoic liver metastasis (pancreatic cancer case) not seen on prior imaging demonstrating typical centripetal filling and peripheral nodular pattern of a hemangioma, avoiding FNA; 2) EUS-indeterminate (cyst vs. mass), but CEH-EUS-neoplastic, mediastinal lesion surgically confirmed to be neoplastic. No device-/agent-related adverse events occurred. Conclusions: CEH-EUS achieves excellent imaging with minimal added procedure time. It is accurate, distinguishes mucoid cysts and vascular lesions from solid lesions and changes management significantly in selected cases. W1260 Advanced Translumenal Access Device for Pseudocyst Drainage William R. Brugge, Kenneth F. Binmoeller, Janak N. Shah, John Lunsford, Hoang G. Phan, Fiona Sander Introduction: EUS-guided pseudocyst drainage provides internal drainage without the need for surgical intervention. Current methods require multi-step access through the GI lumen wall into the pseudocyst, creation of a cystgastrotomy tract, and placement of multiple 7-10F plastic stents. Access and dilation steps are often tedious and difficult to perform due to the lack of a suitable stent delivery devices. Herein we describe a fully integrated translumenal access device that creates and dilates an access tract, then facilitates placement of two guidewires. Aim: The aim of this study is to evaluate the deployment of a new Translumenal Access Device (TAD) in a benchtop model. Description: The TAD is a 140 cm long, 2.5 mm diameter device that coaxially incorporates a 19g trocar, a 3cm long balloon and two separate 0.035 in guidewire lumens. The TAD handle provides precise and independent control of both the balloon and trocar. The 19g trocar tip incorporates an ‘‘extending blade’’ that creates an incision significantly larger than its constrained profile, thereby allowing easy passage of the balloon catheter directly over it. The 3 cm balloon is activated in two stages: 4-6 atm of pressure inflates a 20mm diameter distal anchoring balloon flange to maintain pseudocyst access, while 10-12 atm of pressure inflates a 10 mm diameter dilation balloon section. When the balloon catheter is retracted proximally, the distal anchoring balloon flange ensures apposition of the cyst and GI walls while placing the dilation portion of the balloon in an ideal tract dilation position. Methods: The following bench-top force measurements were made using a test fixture housing ex-vivo porcine stomach tissue: (a) the force required for the TAD trocar to puncture tissue, (b) the force required for the TAD balloon catheter to advance through punctured tissue, and (c) the force required to pull the distal anchoring balloon flange through the punctured tissue. Results: The force required for the TAD trocar to puncture tissue averaged 0.64 psi, similar to the 19g control needle. The force required to advance the TAD balloon catheter through the tissue punctured by the TAD trocar ranged from 5.7-10.9 times less than the force required to advance the same catheter through tissue punctured by the 19g control needle. The distal anchoring balloon required more than 2 lbs of force to dislodge the catheter, indicating it’s effectiveness in securing a stable operating platform. Conclusion: This advanced translumenal access device provides significant conceptual improvements over current access and dilation methods. W1261 Accuracy of EUS, EBUS, and Combined EUS/EBUS for Non-Small Cell Lung Cancer Evaluation in Patients with a Negative CT and PET of the Mediastinum Laith H. Jamil, Noelia Cubero De Frutos, Kanwar R. Gill, Seth A. Gross, Jorge M. Pascual, Massimo Raimondo, Timothy A. Woodward, Julia Crook, John Odell, Michael B. Wallace Purpose: The presence of mediastinal lymph nodes (MLNs) metastasis in patients with suspected Non-Small Cell Lung Cancer (NSCLC) is a critical determinant of therapy and prognosis. Combined Endoscopic Ultrasound (EUS) and Endobronchial Ultrasound (EBUS) with fine needle aspiration (FNA) has recently been shown to be highly accurate in evaluating MLNs in patients suspected of having NSCLC [Wallace et al. JAMA, 2008]. EUS detects approximately 61% of malignant metastases in CT negative patients [Wallace et al. Ann Thor Surg 2004] but has low negative predictive value (NPV), thus requiring further mediastinoscopy if EUS is negative. Current guidelines state that in patients suspected of NSCLC, who have a negative CT and PET of the mediastinum, do not need invasive staging unless there is a central tumor or N1 LN enlargement (Silvestri et al Chest. 2007, Detterbeck et al Chest. 2007). The diagnostic value of combined EUS/EBUS with FNA has not been well studied in patients who have no enlarged MLNs on CT scan of the chest and have a negative PET scan of the mediastinum. Methods: Prospective, double blind trial comparing EUS and EBUS FNA of patients with suspected NSCLC. EUS was performed by a gastroenterologist, and EBUS by a pulmonologist, in a back to back manner, each blinded to the other results. The subset of patients without enlarged (R1 cm) lymph nodes in the mediastinum on CT, who also had a negative PET scan of the mediastinum, where included in this analysis. Accuracy for each procedure and the combination was compared to the reference standard which included any pathology proven (by FNA or surgery) malignant nodes, or negative nodes by surgery or at least 6 months clinical/CT follow up. Results: A total of 248 patients underwent EUS and EBUS evaluation for possible NSCLC and had a CT scan for review. 156 patients had their PET scan available for review. A total of 64 patients had a CT scan that was negative for pathological appearing MLNs (short axis !10 mm) and a negative PET of the mediastinum. Among the 64 CT/PET negative patients, 10 (15.6%) were found to have MLN tumor involvement by the reference standard (FNA or surgery). The estimated sensitivities of EUS-FNA, EBUS-FNA, and EUS plus EBUS were 60% (6/10), 70% (7/10), and 90% (9/10), respectively. The NPVs were 93% (54/58) for EUS-FNA, 95% (54/57) for EBUS-FNA, and 98% (54/55) for EUS-FNA plus EBUS-FNA. Conclusion: In patients with suspected NSCLC, who have no MLNs on CT scan of the chest, combined with a negative PET scan of the mediastinum, the combination of EUS-FNA and EBUS-FNA is moderately sensitive but provides a high NPV in this low prevalence population. W1262 Endoscopic Ultrasound Elastography (EUS-EG) for Gastrointestinal Stromal Tumor (GIST) Ryoji Miyahara, Yasumasa Niwa, Masaaki Kurahashi, Toshihiko Nagaya, Kakunori Banno, Hidezumi Tatematsu, Masanao Nakamura, Hiroki Kawashima, Akihiro Itoh, Naoki Ohmiya, Yoshiki Hirooka, Osamu Watanabe, Takafumi Ando, Hidemi Goto Objective: To diagnose a GIST qualitatively, it is difficult to apply EUS-aided imaging only. It has been determined, therefore, to adapt a surgical operation, depending upon the findings of diagnosing a tumor for size and tissue by EUS-FNA (Fine Needle Biopsy). We have reported the usefulness of a GIST diagnosis by EUS-FNA or fluorodeoxyglucose-positron emission tomography (FDG-PET). We will hereby report how useful the simple and non-invasive EUS-EG was in qualitatively diagnosing submucosal tumors of the gastrointestinal tract, considering that EUS- EG is applicable simultaneously with a conventional EUS inspection. METHOD: From July 2006 to Setpember 2008, we applied EUS-EG to a total of 25 cases; 19 gastrointestinal tract GISTs and 6 leiomyomas. In either case, we could diagnose the tissue by surgical operation or EUS-FNA. EUS-EG had the explored 25 cases display the relatively soft tissue in red thru green and the relatively hard one in blue. The EUS-EG image analysis software (made by HITACHI Mecido Corp.), moreover, was used to explore a strain standard deviation (the larger the value, the more various levels of hardness would intermingle). Results: 1) With the image analysis software, GIST and leiomyoma cases showed their respective strain root-mean-square deviations of 55.2 and 37.5. Thus, hardness inside the GIST was found to be significantly uneven (PZ0.002). 2) In a normal EUS image, both GIST and leiomyoma were displayed as a low echoic mass leading to the muscular tunics. With EUS-EG, 15 cases out of the 19 GISTs had a hard tissue colored in blue at the limbus while displaying the relatively soft tissue in green thru red at the center. This could be deemed as a distinctive image finding. Under a diagnosis on the assumption that an evenly hard tumor mass was taken for leiomyoma and an unevenly hard one for GIST, sensitivity was found to be 88.2% and specificity 42.8%. 3) GIST cases may be divided into two groups; one shows an MIB-1 index of 10% or more and the other less than 10%, with such index taken for an indicator of malignancy on the guidelines in Japan. Then, their respective strain standard deviations are found to be 60.6 and 52.4. Although there was no significant difference, those cases that had a higher level of malignancy showed a tendency to be more unevenly hard within the tumor. Conclusion: EUS-EG might well be considered capable of turning out as a simple and useful inspection technique in a qualitative GIST diagnosis. W1263 Impact of EUS-Guided Fiducial Markers in Management of Patients with Pancreatic Cancer (PANCAN) Undergoing Image Guided Radiation Therapy (IGRT) Shyam Varadarajulu, Jessica Trevino, Rojymon Jacob Background: During the course of radiotherapy (RT), both the patient’s position and anatomy can vary from the original treatment planning setup, resulting in insufficient dose to the targeted tumor volume and an overdose to surrounding normal tissues. As soft tissue cannot be visualized easily on the IGRT machine, RT is currently delivered based on skeletal anatomy landmarks. Although EUS-guided placement of fiducials in pancreatic masses is technically feasible, its direct impact on patient management is unclear. Aim: Evaluate the impact of EUS-guided fiducial markers in management of patients with PANCAN undergoing RT. Methods: In this prospective study, 10 (6male, mean age 62 yrs [range, 40-86]) consecutive patients with PANCAN underwent EUS-guided fiducial placements over 3 months. Following computerized tomography (CT) simulation, a RT treatment plan was prepared for all patients to be delivered over 25 days. A kilovoltage (kv) image was obtained before each treatment session after positioning the patient on the linear accelerator. This image revealed fiducials within the tumor which was then fused with the original image generated during initial treatment planning; RT was then delivered. Daily shifts (organ movement) were recorded by comparing the fused images revealing fiducials with images fused based on skeletal anatomy (current standard of care). All patients had CT evaluation halfway through RT and at 5 weeks following RT to check for fiducial migration. Results: EUS-guided fiducial placement Abstracts www.giejournal.org Volume 69, No. 5 : 2009 GASTROINTESTINAL ENDOSCOPY AB325