M1605 The Incidence and Classification of Esophagogastric Junction Adenocarcinoma in Korea Jun-Won Chung, Hwoon-Yong Jung, Kee Wook Jung, Jin-Ho Kim, Kee Don Choi, Kwi- Sook Choi, Ho June Song BACKGROUND: The incidence of adenocarcinoma of the esophagogastric junction (AEG) is increasing in Western countries. It is not well known whether similar changes are occurring in Asia. The present study examined the incidence of AEG in Korea, and characterized the tumors according to Siewert's classification. METHODS: The study retrospectively examined records of 16811 patients diagnosed with esophageal cancer and gastric adenocarcinoma between 1992 and 2006. Patients were divided into three 5-year cohorts (cohort 1 [1992- 1996]: n=2734, cohort 2 [1997-2001]: n=5727, cohort 3 [2002-2006]: n=8340). Cohorts were compared in terms of the ratio of AEG versus ESC plus GNCA. Siewert's classification was used to categorize tumors into one of three types. Demographic features, 5-year survival rate and stage were compared between the three tumor types. RESULTS: The numbers of patients diagnosed with AEG, esophageal squamous cell carcinoma (ESC) and gastric non- cardiac adenocarcinoma (GNCA) were 610, 1450 and 14751, respectively. The ratio of AEG versus ESC plus GNCA did not change over time (0.037, 0.034 and 0.038 for cohorts 1, 2 and 3, respectively; p=0.040). The incidence of stage I, II cancer increased over time, as did the 5-year survival rate (p<0.01). Of the 610 AEG patients, 23 (3.7%) had type I, 47 (7.7%) had type II and 540 (88.5%) had type III tumors. The three types differed in terms of mean patient age, prevalence of undifferentiated tumor, ratios of early stage cancer and 5-year survival rate. CONCLUSION: Unlike Western countries, the incidence of AEG did not increase over the 15 years to 2006 in Korea. Three AEG types were identified according to Siewert's classification, and the distribution of AEG types and the survival rates associated with each type differed from that reported in Western studies. M1606 Type I (Barrett's) Adenocarcinoma As a Distinct Paradigm in Junctional Tumours: Rationale for Simplifying the AEG Classification System Narayanasamy Ravi, Bussa Gopinath, Patrick J. Byrne, John V. Reynolds Objective: To apply biological criteria to the Siewert topographic classification of adenocarcin- oma of the esophago-gastric junction (AEG), and to suggest a simpler classification that is consistent with pathogenesis and outcome. Background Data: The AEG classification is based on topography, and AEG I and AEG II have been combined in current European trials. A consensus exists that all AEG I tumours originate in Barrett's epithelium, this is rare in AEG II or AEG III, and the topographical classification may require refinement based on biological, metabolic and molecular characterization. Methods: A prospective cohort (n=407) of AEG tumours [130 type I, 145 type I, 132 type III] was evaluated. Clinical and pathological variables associated with prognosis were compared across AEG groups [pT, pN, stage, differentitation, R status,venous and perineural involvement, number of positive nodes, percent of positive nodes, and tumour length] . Results: Compared with type I tumours, type II tumours had significantly (p<0.05) more advanced pT and pN, number and percent node positivity, poor differentiation, radial margin involvement, and tumour length (7/10). Comparing II with III, tumour length, pT and number of positive nodes were significantly (p<0.05) less in II (3/10). 88% of AEG I had associated Barrett's compared with 18% and 2% for AEG II and III respectively. The median survival of 26 months for AEG I compared with 18 and 16 months respectively in AEG II and AEG III (p=0.037). Conclusions: AEG II tumours have more in common with AEG III than AEG I. These data support the view that AEG I, predominantly Barrett's adenocarcinoma, are a unique group, and it begs the question whether II and III together rather than I and II is the most appropriate pairing for clinical trials, and whether AEG I and AEG II/III may represent an improved modification of the current topographical classification. M1607 GERD, Barrett's Esophagus (BE), and Esophageal Adenocarcinoma (EA) in South Carolina in 2006: Geographic Analysis of Prevalence and Effect of Availability of Endoscopy Centers Lubin F. Arevalo, Marcelo F. Vela BACKGROUND: A GERD-BE-EA sequence has been described. Early BE recognition can enable surveillance and prompt detection of EA. Analyses of GERD, BE, and EA prevalence in relation to endoscopist availability are lacking and can be performed by software that evaluates geographically referenced data. AIMS: Asses geographic variation of GERD, BE, and EA prevalence in South Carolina (SC) counties in 2006, examine geographic relationships among disease states and evaluate effect of availability of endoscopy centers. METHODS: 2006 no. of GERD, BE, and EA cases (ICD-9 search), no. of Endoscopy Centers (ECs), and population demographics for each county obtained from the SC Office of Statistics. Geo- graphic distribution and relationships among GERD/BE/EA prevalence and ECs studied by Geographic Information System for Epidemiological Studies (ArcView 3 software). RESULTS: Prevalence: South Carolina 2006 Prevalence was 2590/100.000 for GERD; 792/100.000 for BE (M:F ratio 1.8:1, W:NW ratio 4.3:1); 366/100.000 for EA (M:F ratio 4.4:1, W:NW ratio 4.6:1). GIS analysis GERD/BE/EA (maps not shown): (1) BE prevalence higher in counties with high GERD prevalence, (2) EA prevalence higher in counties with high BE prevalence, (3) EA prevalence higher in counties with low GERD prevalence. GIS analysis BE/EA and ECs: (1) EA prevalence higher in counties with low number ECs (see map), (2) BE prevalence higher in counties with high number of ECs (map not shown). CONCLUSION: This geographic analysis generally supports a GERD-BE-EA sequence. Availability of endoscopy centers may impact BE diagnosis/prevalence in different SC geographic areas, with higher EA prevalence where endoscopists are not available. Geographic data may be useful to determine need of endoscopic centers in specific geographic areas. However, the relationships examined are complex and there are unmeasured factors. A-381 AGA Abstracts M1608 The Esophageal Adenocarcinoma and Barrett's Esophagus Registry Consortium: A Large Prospective Resource for Collaborative Multidisciplinary Studies Virender K. Sharma, Michael B. Wallace, Yvonne Romero Background: Studies to identify genetic mutations that play a role in progression of Barrett's esophagus (BE) to esophageal adenocarcinoma (ACA) have been challenging due to limita- tions in sample size and length of follow-up. Aim: To develop a registry that includes blood, fresh-frozen and formalin-fixed tissue, linked pathologic and clinical data, and serial validated symptom and quality of life questionnaires over time, to provide the necessary infrastructure to help clinicians and scientists address important questions with sufficient power to yield meaningful results. Methods: Our multidisciplinary collaborative group formed the Eso- phageal Adenocarcinoma and Barrett's Esophagus Registry Consortium starting 9/2001, which to date has expanded to all three Mayo Clinic campuses. Consecutive patients with either: 1) Long segment BE alone; 2) BE with ACA; 3) ACA alone; 4) both ACA and squamous cell carcinoma (SCCA) features; 5) ACA with SCCA and BE; 6) gastroesophageal junction ACA; and 7) SCCA alone, as controls, have been approached using an “ultra-rapid” identifica- tion recruitment method (within 7 days of diagnosis). Participants agree to provide baseline and annual questionnaires. Blood is collected at baseline, and tissue is collected at every clinically indicated endoscopic or surgical procedure. Results: As of 11/20/07, 3199 living patients met initial diagnostic entry criteria. 92 patients were excluded (e.g. age <18). Of the remaining 3107 patients, 1799 have been enrolled (58%), 675 (38%) declined participa- tion, and 563 are pending. The enrolled patients include: 953 long segment BE alone; 188 ACA alone; 310 BE with ACA; 3 ACA with SCCA; 3 ACA with SCCA and BE; 154 gastroesophageal junction ACA; 53 gastroesophageal junction with BE, and 122 SCCA alone. Blood and tissue have been collected at least once from 1512 and 816 unique participants, respectively. Conclusions: We have successfully created a mechanism for multi-center, rapid, systematic, and serial collection of fresh-frozen and formalin-fixed tissue, blood, and demo- graphic, symptom and risk factor data related to BE and ACA: the Esophageal Adenocarcinoma and Barrett's Esophagus Registry Consortium. This resource will allow us to apply a systems biology approach to adenocarcinoma of the esophagus, including identification and validation of genetic mutations important in the progression from Barrett's esophagus to adenocarcin- oma. It holds great promise to facilitate to bench to bedside discoveries. M1609 Decreased Expression of the RAS-GTPase Activating Protein RASAL1 Is Involved in Progression of Colorectal Neoplasms Miki Ohta, Motoko Seto, Dai Mohri, Yoshinari Asaoka, Motohisa Tada, Yasuo Tanaka, Hideaki Ijichi, Fumihiko Kanai, Takao Kawabe, Masao Omata [Background & Aims] Although the importance of RAS signaling in the progression of colorectal neoplasms is no longer disputable, it has also been accepted that many colorectal cancers (CRCs) still retain wild-type KRAS gene. The RAS activity is modulated with the RAS-GTPase activating proteins (RASGAPs); whether the aberrant dysfunction of RASGAP genes contributes to the progression of colorectal neoplasms also remains open to question. [Materials & Methods] The gene expressions of 12 RASGAPs in the colorectal cancer (CRC) cell lines (n=18) were analyzed using quantitative RT-PCR. The CRC cells with lacked RASAL1 expression were established using RNAi technique and used for assays. Immuno- histochemistry (IHC) of RASAL1 was preformed in the clinical colorectal neoplasms (aden- omas; n=88, adenocarcinomas; n=64) after approval from the medical ethics committee and acquisition of informed consent. [Results] It was found that only the expression levels of RASAL1 gene were significantly lower in CRCs with retained wild-type KRAS than those with active mutant KRAS among the 12 RASGAPs examined. The Knockdown of RASAL1 in the CRC with retained wild-type KRAS resulted in RAS activation and the acceleration of the tumor growth in xenograft model. IHC showed that decreased RASAL1 expression was found in almost half of the CRCs (46.9%) whereas less detected in the adenomas (9.1%). RASAL1 depression in CRC was significantly correlated to the distal-location (P =0.0066), wild-type KRAS (P =0.0010) and abnormal TP53 expression (P =0.0208) of the tumors. [Conclusion] RASAL1 depression is critically involved in the development of the colorectal neoplasms with retained wild-type KRAS. M1610 Aurora Kinase a Abrogates Drug-Induced Apoptosis Through Regulating p53 in AKT-Dependent Mechanism Altaf Dar, Abbes Belkhiri, M. Blanca Piazuelo, Pelayo Correa, Alexander I. Zaika, Antoni Castells, Wael M. El-Rifai Using comparative genomic hybridization, we (and others) have shown frequent high-level amplifications of chromosome arm 20q13 in many cancers including upper gastrointestinal adenocarcinomas (UGCs). This region harbors the gene, Aurora kinase A (AURKA). In this study, we investigated AURKA's molecular and biological bases in adenocarcinomas of the AGA Abstracts