prospectively; images of white light imaging (WLI) and NBI were evaluated and compared. After examination, forceps biopsy was tried for all lesions. Definition of successful direct POC with NBI was advancing the slim endoscope into the strictured segment and complete inspection under WLI and NBI. Reference standards were histopathologic proof of malignancy, or the clinical course during follow-up in cases without malignancy. RESULTS; The overall success rate of NBI by direct POC was 92.9% (26/28). The sensitivity of NBI for the diagnosis of malignant biliary stricture was 93.3% (14/15). Identification of surface vessels and distal margins of the lesions by NBI observation was significantly higher than with conventional WLI (P0.05). Forceps biopsy during direct POC was successfully performed in 82.1% (23/28). Procedure related complications were not observed in any patients. CONCLUSIONS; NBI by direct POC using an ultra- slim upper endoscope seems to be a feasible and useful endoscopic procedure that can provide more information to endoscopic diagnosis of biliary strictures. 415e Photodynamic Therapy in Bile Duct Cancer - Can the Preliminary Positive Results Be Confirmed? Katja Wirths, Philip A. Hilgard, Alexander Dechene, Juergen F. Riemann, Guido Gerken, Thomas Zoepf Introduction: In limited patient series photodynamic therapy (PDT) could demonstrate significant prolongation of survival. The aim of this study was to evaluate, if these results can be confirmed in larger unselected patient series under normal clinical conditions.Methods: 53 consecutive patients (36m; 17f) with bile duct cancer (BDC), who were not eligible for surgery received PDT. 48 hours prior to laser irradiation patients 2mg/kg b.w. Photofrin II was applied. Light application was performed using a 635nm diode laser system with a cylindrical diffuser tip of 4cm length. Patients were treated periinterventional with antibiotics for 14 days and were provided with at least two plastic bile duct endoprostheses.Results: Median age was 70 years (range 43-81), 36 male and 17 female patients. 81% showed Bismuth type IV tumor, 11% Bismuth type I and 4% Bismuth type II and III respectively. The median interval between diagnosis and first PDT treatment was 63 days. Median follow up was 360 d [range 39- 1380]. 32/53 patients died during follow up. Median survival of the patients was 630 d [95% confidence interval 490-716d]. 24% of patients developed cholangitis during the follow up, which could be treated by stent exchange and antibiotics.Conclusion: The impressive impact of PDT on the overall survival in patients with advanced bile duct cancer were by this study confirmed in a rather unselected patient cohort under normal clinical conditions. Further randomized studies should address the value of a combination of chemotherapy or radiotherapy with photodynamic therapy. 415f FOXP3 Upregulation in Biopsy Specimens of the Duodenal Papilla Has Prognostic Significance in Cases With Ampullary Tumors Kensuke Kubota, Shingo Kato, Seitaro Watanabe, Takeshi Shimamura, Noritoshi Kobayashi, Atsushi Nakajima Background: Pre-endoscpic diagnosis in patients with ampullary tumors is sometimes difficult. Endoscopic biopsy combined with immunohistochemistry may improve the accuracy. Evaluation of the expression status of Forkhead/ winged helix family of transcription factor P3 regulatory T cells (FOXP3) has been reported as being useful for the selection of therapy in patients with pancreaticobiliary cancers. Aim: To investigate the usefulness of pre-endoscopic diagnosis and the prognostic value of the expression status of FOXP3 and ki-67 in endoscopic biopsy specimens of the duodenal papilla in patients with ampullary tumors. Methods: The data of 54 patients were studied. The median tumor size was 2 cm (range 5-60mm). We classified the 54 patients into two groups: Group A (n=28), consisting of patients who underwent endoscopic resection (ESP) (n=17) or TMN Stage I patients (n=11) who were potential candidates for ESP, and Group B (n=26), consisting of patients who were contraindicated for ESP because of invasion of the ampullary cancers beyond the muscularis propria of the duodenum and/or the pancreas. FOXP3 and ki-67- positive cells were counted in high-power-fields (HPFs) in the biopsy specimens. The two groups were statistically compared in terms of the biopsy discrepancies before therapy and the FOXP3 and ki-67 expression status, and the prognostic value of the latter was also examined. We assessed a cutoff value of FOXP3 expression based on the results of the study. Results: 1) The overall pre- endoscopic diagnosis rate was 79.9% (41/54). Pre-endoscopic discrepancies were significantly more frequent in Group A (32.1%, 10/28) than in group B (13.4%, 4/26) (p0.05). 2) No significant difference in FOXP3 or ki-67expression was observed between group A and group B. 3) Poor prognosis, defined as local recurrence and/or metastasis within 3 years of therapy, was recognized in 7 out of the 26 (26.9%) cases in Group B. There was no difference in the Ki-67- expression level between the poor prognosis and good prognosis groups. In addition, FOXP3 upregulation was recognized significantly more frequently in the poor prognosis than in the good prognosis group (42.25.3/HPF vs. 27.4+8.9/HPF; p=0.0107). When the cutoff value for detecting poor prognosis was 39/HPF, the sensitivity and specificity were 71.4% and 100%, respectively. Conclusions: Discrepanccies in findings between pre- and post-endoscopic biopsy specimens were recognized in ampullary tumors, however, the presence of 39/HPF (cutoff value) of FOXP3-expressing cells in endoscopic biopsy specimens was reflective of a poor prognosis and predictive of recurrence after resection. 415g Screening for Familial Pancreatic Neoplasia:a Prospective, Multicenter Blinded Study of EUS, CT, and Secretin-MRCP (The NCI-Spore Lustgarten Foundation Cancer of the Pancreas CAPS 3 Study) Marcia I. Canto, Richard D. Schulick, Ihab R. Kamel, Elliot K. Fishman, Mark D. Topazian, Naoki Takahashi, Jeffrey H. Lee, Eric P. Tamm, Raghunandan Vikram, Sapna Syngal, John R. Saltzman, Koenraad J. Mortele, James J. Farrell, Daniel Margolis, Zhe Zhang, Gloria M. Petersen, Ralph H. Hruban, Michael G. Goggins High risk individuals (HRI) with a strong family history of pancreatic cancer (PC) and germline mutation carriers are at increased risk of PC. We determined the prevalence and characteristics pancreatic neoplasms at baseline screening of asymptomatic HRI by CT, secretin-magnetic resonance cholangiopancreatography (MRCP), and EUS. METHODS: 5 academic medical centers prospectively screened asymptomatic adults with EUS, CT, and MRCP performed within 2 days. Experts interpreted images in a standardized, blinded fashion. We offered surgery to HRI with suspected high grade cystic neoplasms or lesion progression. An expert pathologist reviewed the resected pancreata. RESULTS: We studied 216 HRI (195 FPC, 19 BRCA2 mutation carriers, 2 Peutz- Jeghers syndrome)(46% male, mean age 56 10 years) with EUS, CT, and MRCP. 36% (77/216) HRI had a normal pancreas by all 3 imaging tests, 93/216 (43%) HRI had at least one mass (85=cystic, 3=solid), 6 had an isolated dilated main pancreatic duct (MD), and the rest had nonspecific abnormalities. 56% of HRI with masses had multiple lesions, and 85% of these involved multiple locations. The majority of lesions (87.5%) were small 1 cm (mean 0.55 cm, range 0.2-2.3 cm). CT, MRI and EUS diagnosed 27%, 81% and 93% of HRI with 1 mass lesion. The concordance (percent agreement) for detection of any neoplastic-type lesion was higher between EUS and MRI (91%) than for EUS and CT (73%). There was a strong correlation between MRCP and EUS for the number (Spearman corr coefficient=0.82) and moderate agreement for location (kappa=0.43) of pancreatic masses. Among a total of 289 lesions detected, EUS and MRCP detected 229 and 218 lesions, respectively, compared to only 39 lesions detected by CT. The mean follow-up time was 9.6 months(range 0- 32).The final diagnoses were suspected/confirmed branch duct (BD) intraductal papillary mucinous neoplasm (IPMN)(n=85), combined IPMN (n=2), endocrine neoplasm(n=4),chronic pancreatitis( 5/9 EUS criteria, n=44), and indeterminate mass(n=2). Pancreatectomy was performed on 5 HRI (4 subtotal, 1 total): 4 had IPMN (2 mixed IPMN=1 HGD, 1 MGD; 2 BD-IPMN=1 HGD, 1 MGD) with associated multiple PanIN3, and 1 had BD-IPMN LGD and PanIN2. Only EUS visualized mural nodules(2) and MD wall thickening (2)in these patients. HRI with HGD had a mean maximum cyst size 1.7 and mean MD diameter=3.3 mm). CONCLUSION: Screening can detect prevalent lesions in 43% of asymptomatic HRI, including small noninvasive high grade IPMN and PanIN. EUS and MRI are superior to CT for detection of small (predominantly cystic) masses. EUS may also visualize features associated with high grade neoplasia. 415h Prospective Evaluation of Routine Cytology (RC), Digital Image Analysis (DIA) and Fluorescence in Situ Hybridization (FISH) At EUS FNA Michael J. Levy, Benjamin R. Kipp, Trynda N. Oberg, Michael B. Campion, Kevin C. Halling, Amy C. Clayton, Thomas J. Sebo, Lizhi Zhang, Michael Henry, Jonathan E. Clain, Ferga C. Gleeson, Elizabeth Rajan, Mark D. Topazian, Kenneth K. Wang, Gregory Gores Background: EUS FNA with RC analysis provides moderate diagnostic sensitivity. To optimize diagnosis, it is often necessary to obtain 5 or more needle passes which increases procedure time, medication use, personnel use, decreases room efficiency, and may heighten risks. New techniques are needed to increase diagnostic sensitivity with limited tissue sampling. Hypothesis: Molecular cytogenetic markers (DIA and FISH) enhance the diagnostic sensitivity of RC at EUS FNA.Aims: To prospectively determine the sensitivity, specificity, and accuracy of DIA, FISH, and composite test results with limited tissue sampling ( 2 passes) in patients undergoing EUS FNA.Methods: For patients undergoing EUS, 6 FNA samples were obtained when considered safe and technically feasible. Patients were blindly randomized to the first test performed (RC or DIA/FISH) and subsequent passes were alternated, thereby obtaining 3 Abstracts www.giejournal.org Volume 71, No. 5 : 2010 GASTROINTESTINAL ENDOSCOPY AB119