moderately and poorly differentiated in 2/22 (9%), 6/22 (27%) and 13/22 (59%) respectively.In 17 patients GC samples were reevaluated for histological characteristics, 14 were MSIHand 3 failed in MSI testing, however IHC stainings for the affected MMR protein werenegative in all these patients. Six (35%), 4 (23%) and 7 (41%) GC samples showed intestinaltype, diffuse type and mixed type characteristics respectively. In 13 GC samples, the tissuesurrounding GC could be evaluated. Intestinal metaplasia was diagnosed in 5 (38%) samplesand atrophic gastritis in 2 (15%) samples. Pre-malignant gastric lesions were evenly distrib-uted throughout intestinal, diffuse and mixed type gastric cancer samples (p=0.40). Conclu-sions: This study demonstrates that GC develops along similar histopathologic pathways assporadic gastric cancer in LS. Pre-malignant lesions in tissue surrounding GC were evidentin >50% of samples. These data support upper GI surveillance endoscopy in LS subjects.

Sa1228

Family Communication in Lynch Syndrome Families: Experiences With theFamily-Linked ApproachCeline H. Leenen, Mariska den Heijer, Conny van der Meer, Ernst J. Kuipers, Moniquevan Leerdam, Anja Wagner

Background: Lynch syndrome (LS) is a genetic predisposition for colorectal carcinoma,endometrial cancer and other malignancies and is caused by inherited germline mutationsin mismatch repair (MMR) genes. Current genetic counselling practice relies on diagnosedfamily members to inform their biological family about LS, referred to as the family-linkedapproach. From previous research is known that the uptake of genetic testing by familymembers of LS carriers is not optimal. However, little is known about communication withinfamilies regarding the diagnosis of LS in their family. The objective of this study thereforewas to investigate family members' experiences with the family-linked procedure. Methods:85 LS carriers and non-carriers from LS families diagnosed in 2003-2008 at the ErasmusMC, were invited by phone and mail to participate. Furthermore, they were asked to approachother family members, especially family members who had not opted for genetic testing, toparticipate in this study. A questionnaire was sent to all family members who consented toparticipate, addressing their experiences with the family-linked procedure. Tested and non-tested family members were compared using the Fisher's exact test. Results: Of 85 eligiblesubjects from 22 LS families, 62 subjects (23 LS carriers, 32 non-carriers, and 7 non-testedfamily members) completed the questionnaire (73% response rate). The median age ofrespondents was 63 years (range 23-86 years, 28% males). Of these, 46/62 (74%) had beeninformed by a family member, 15/62 (24%) were index patients, and informed by thegeneticist, 1 respondent was informed by the general practioner. Of respondents informedby their relatives, 13/46 (28%) were informed within one week after diagnosis in the family,in most cases by the index patient. 27/46 (69%) family members did not report to experienceburden when informed by a relative, 2/46 reported severe burden. However, 38/46 (83%)reported that relations with their relative(s) had not changed. Of all respondents, 41/62 (66%)agreed with the current approach of being informed by family members. Of respondents whodid not agree, 13/17 (76%) preferred the medical specialist at the hospital to inform familymembers. Finally, 43/62 (69%) respondents agreed with the statement that physicians shouldalso inform patients about LS in their family. No differences regarding family relations andopinion towards informing family members were found between tested and non-testedfamily members. Conclusion: These results indicate that the majority of members of LSfamilies do not experience burden when informed by their relatives about LS. The family-linked approach does not affect family relations. However, the majority of LS family membersprefer an active approach from physicians in the LS information process in their families.

Sa1229

Hereditary Colorectal Cancer Type X- Should Extracolonic Surveillance BeRecommended?Revital Kariv, Hana Strul, Guy Rozner, Rakefet Shtoyerman, Yael Etzion-Daniel, ZamirHalpern, Irit Solar

Background: Hereditary colorectal cancer (CRC) type X defines families who fulfill theAmsterdam criteria but with tumors that are microsatellite stable and/or do not demonstratemismatch repair (MMR) deficiency. Few studies have suggested that these families do notsignificantly demonstrate extracolonic tumors and that surveillance colonoscopies shouldbe performed less often than in Lynch syndrome, every 3-5 years with no need for extracolonicsurveillance. Methods: Familial CRC registry in a large referral center in Israel was searchedfor families who fulfill Amsterdam and Bethesda criteria and that in which Lynch syndromewas ruled out by Tumor testing for microsatellite instability (MSI) and MMR immunohisto-chemistry. Pedigree was studied for detailed cancer history. Results: 308 probands fulfilledclinical criteria for hereditary CRC, tumors from 106 families were tested for microsatelliteinstability (MSI) and IHC for MMR proteins. 40 families were identified as suspected forLynch syndrome by MSI-H tumors and/or tumor MMR deficiency by IHC. 10 familiesfulfilled Amsterdam criteria but had MSS tumors, thereby defined as hereditary CRC typeX. Extracolonic tumors appeared in 9/10(82%) families with wide spectrum of tumors thatmostly overlapped Lynch syndrome tumor spectrum: stomach, lymphoma, endometrium,kidney, melanoma, pancreas, leukemia, brain, testes, breast, prostate. The average numberof family members with cancer in type X CRC was 5 (3-10) with an average of 3 familymembers with CRC and 2 (1-5) with extracolonic cancers. Age range for the earliest cancerdiagnosis in family was 16-50 (average 39) with an average of 2 family members with anycancer before 50y. 4/10 families were of Ashkenazi origin, 3/10 were from North Africa, 2/10 non Ashkenazi Russian and 1/10 from Iraq. Among 33/56 (59%) families with MSStumors and positive Bethesda criteria, family members demonstrated extracolonic tumors.Conclusions: profile of families with hereditary CRC type X and of those with MSS tumorsand Bethesda criteria in Israel favors extracolonic surveillance due to multiple extracolonictumors in family members. Further research is needed in regard to colonic and extracolonicage of surveillance onset and intervals.

S-259 AGA Abstracts

Sa1230

Pedigrees Obtained by a Genetic Counselor Compared With Self-ReportedHealth History Questionnaire Demonstrate Strong Agreement in First-DegreeRelatives but Moderate to Poor Agreement in Second- and Third-DegreeRelativesAimee L. Lucas, Jeanine Genkinger, Sheila Kumar, Lauren G. Khanna, Joanna Martinez-Gomez, Harold Frucht

BACKGROUND: Approximately 15% of pancreatic cancers (PCs) have a hereditary compon-ent. A careful family history (FH) including third-degree relatives is integral to assessing apatient's risk for syndromic malignancy. FH of malignancy can be assessed either throughself-completed questionnaires or by use of a trained genetic counselor (GC.) We hypothesizethat GC and health history questionnaire (HHQ) pedigrees would demonstrate agreementin close family members, but poor agreement in more distant relatives. METHODS: This isa retrospective study of patients at high-risk of PC evaluated at a PC prevention programbetween 2007 and 2010. Patients met with a GC for full pedigree analysis, as well ascompleted a HHQ pedigree. GC pedigrees and HHQs for each subject were examined fordemographic information, total number of first-, second- and third-degree family membersreported, with PC or with any cancer. Statistical analysis was performed using a paired t-test and overall agreement estimated by calculation of kappa value κ. RESULTS: CompleteGC pedigrees and HHQs were available for 50 patients. The mean age for the GC groupwas 54.98 years and for the HHQ group was 54.45 years (p-value=.45). Frequencies ofdifferences in variables revealed GC documented more cancers than HHQ. Near perfectagreement was demonstrated between the reporting of the total number of first-degreerelatives (κ=.95; 95% CI; .89-1.00) and first degree relatives with PC (κ= .93; 95% CI .83-1.00), and moderate agreement between total number of cancers in first-degree relatives(κ= .52; 95% CI, .34-.70). Substantial agreement was seen between GC and HHQ pedigreesfor second-degree relatives with PC (κ= .74; 95% CI, .57-.91) but poor agreement betweenthe total number of second-degree relatives reported (κ= .07; 95% CI, .03-.16) and totalnumber of cancers in second-degree relatives (κ= .40; 95%CI, .23-.56). Poor or no agreementwas found between GC and HHQ pedigrees for total number of third-degree relatives (κ=-.01; 95% CI, -.04-.02), total number of cancers in third-degree relative (κ= .27; 95% CI,.09-.46), and number of PC in third-degree relatives (κ= -.03; 95%CI, -.07-.02). CONCLU-SIONS: GC and HHQ pedigrees demonstrate good agreement through first degree relatives,although agreement decreases in the second- and third-generation pedigree. The similarityin reported FH of PC in second-degree relatives may be a reflection of our patient populationand referral bias. More cancers are documented with GC than HHQ. Improved FH of cancermay stratify patients into high-risk groups, and suggest a more appropriate means of cancerscreening and early detection. While the ability to review a HHQ with family over timeseems appealing, our data suggest that it may be more prudent to consider the use of a GCto obtain more complete FH information in a high-risk patient population.

Sa1231

Palb2 Seems Not to Be Involved in Pancreatic Cancer and/or Breast CancerDevelopment in a Dutch Cohort of Familial Pancreatic Cancer-Families andFamilies With Clustering of Both Pancreatic Cancer and Breast CancerFemme Harinck, Irma Kluijt, Saskia van Mil, Quinten Waisfisz, Theo A. van Os, Cora M.Aalfs, Anja Wagner, Maran Olderode, Rolf H. Sijmons, Jan-Werner Poley, Ernst J. Kuipers,Paul Fockens, Marco J. Bruno

Introduction It has recently been discovered that carriers of a PALB2-mutation not onlyhave an increased risk of breast cancer (BC), but also of pancreatic cancer (PC). Thus farthese mutations have mainly been found in PC-patients from families affected by both PCand BC. Since the prevalence of gene mutations varies between different populations, westudied the prevalence of PALB2-mutations in a Dutch cohort of familial-PC (FPC) patientsand in families affected by both PC and BC. Methods The presence of germline mutationsin PALB2 was evaluated by direct sequencing of the entire coding region and intron-exonboundaries on genomic DNA isolated from whole blood. DNA originated from (1) FPC-patients: i.e. PC-patients with ≥1 first degree relative (FDR) with PC, (2) BC-patients fromFPC-families: i.e. families with PC in either ≥2 FDRs, ≥3 relatives of any degree, or 2relatives with one <50 years at diagnosis, (3) BC- patients from families that did not fulfillFPC criteria but with ≥1 FDR or second degree relative with PC, and (4) PC-patients witha personal history of BC. None of the patients carried a BRCA1 or BRCA2 mutation. ResultsPALB2 mutation analysis was performed in a total of 66 patients from 58 families including24 FPC-patients, 7 BC-patients from FPC-families, 32 BC-patients with a family history ofPC, and 3 PC-patients with a personal history of BC. Of the 24 FPC-patients, eight (33.3%)had at least two affected FDR with PC and 15 (65.2%) had at least three affected relativeswith PC. Six BC-patients from FPC-families (85.7%) had at least one FDR affected by PC.Among the 32 BC-patients from non-FPC families, 24 (75.0%) had a FDR affected by PC.In none of these 66 patients a PALB2 mutation was found. Conclusions Based on thesedirect sequencing data, PALB2 appears not to be involved in the development of FPC andfamilial clustering of PC and BC in BRCA1/2-negative families in this Dutch cohort ofpatients. Consequently, the search for novel susceptibility genes must continue to unravelthe genetics in PC.

Sa1232

Quality of Life After Surgery for Colon Cancer in Patients With LynchSyndrome; Partial Versus (Sub)Total ColectomyJasmijn F Haanstra, Wouter H. de Vos tot Nederveen Cappel, Steven Vanhoutvin,Annemieke Cats, Juda Vecht, Jan H. Kleibeuker, Alexandra Langers, Paul C. van deMeeberg, Evelien Dekker, Jerry H. Bergmann, Hans F. Vasen, Fokko Nagengast, Peter vanDuijvendijk

Lynch syndrome (LS) is an autosomal dominant disorder caused by mutations in MMR-genes. Mutation carriers have a 60 to 85% lifetime risk of developing colorectal cancer(CRC). Most patients participate in intensive colonoscopic surveillance programs. If CRC is

AG

AA

bst

ract

s

AG

AA

bst

ract

sdetected while under surveillance, generally, subtotal colectomy is recommended to reducethe risk of metachronous CRC as compared to partial colectomy. As the risk of CRC underintensive surveillance diminishes, recommendations for surgery should be reconsidered ifa difference in quality of life (QoL) after both types of procedures exists. Aim of this studywas to investigate the QoL and functional outcome in LS patients after partial (PC) andsubtotal colectomy (STC). Methods A nationwide cross-sectional study in the Netherlands wasperformed with two QoL questionnaires (SF-36 and EORTC QLQ CR-38) and a questionnaireabout functional outcome (COREFO). Results Questionnaires were sent to 193 LS patientswho underwent surgery for CRC. The response rate was 71%. 137 patients were included( 64 patients in PC group and 73 patients in the STC group). None of the scales of the SF-36 showed a significant difference between both groups. The EORTC QLQ CR-38 presentedmore problems with defecation after STC (p= 0.01). Analysis of the COREFO revealed thatafter STC patients have a significant higher stool frequency compared to PC patients (p=<0.01). Conclusion Although stool frequency is higher after subtotal colectomy comparedto partial colectomy, general QoL does not differ between both types of surgery in patientswith LS. Therefore, subtotal colectomy is still the treatment of choice in young patients (e.g.< 60 years) with CRC, detected under surveillance. Patients should be informed aboutpossible differences in functional outcome.

Sa1233

Long-Term Risk of Neoplastic Change in Ileal Pouches Created for FamilialAdenomatous PolyposisRajesh Pendlimari, Eric J. Dozois, Jennifer Y. Wang, Daniel Leonard, Maureen C.Mckenna, Robert R. Cima, Heidi Chua, David W. Larson

Purpose: Neoplastic change in ileal reservoirs has been reported in patients with familialadenomatous polyposis (FAP), but little is known regarding the development of neoplasticchange and the natural history of these changes in a large group of patients with long-termfollow-up. We aim to determine the incidence and progression of neoplastic change in theileal pouch of FAP patients at our institution who had long-term follow-up of their ilealreservoir. Methods: All patients who underwent surgery for FAP with creation of ilealreservoir between 1972 and 2007 were reviewed using a prospectively maintained ilealpouch database. Demographic, surgical, endoscopic and pathologic data was collected. Onlypatients with at least 2 years of follow-up and had undergone at least one postoperativepouchoscopy were included. Results: 117 patients were identified with a median age of 26years (range, 4 - 60 years), 52 (44%) were male. Type of ileal reservoirs included; J-pouch(n=104, 88.9%), Kock pouch (n=9, 7.7%), S-pouch (n=3, 2.6%), W-pouch (n= 1, 0.8%).Median follow-up was 125 mos (range, 25 - 423 mos). Polyps concerning for neoplasticchange were biopsied in 33 patients and revealed; low grade dysplasia (n=30), non-dysplasticpolyps (n=2)and adenocarcinoma (n=1), No patients had high-grade dysplasia. Polyp mor-phology in patients with low grade dysplasia was tubular adenoma (n=22) and tubulovillous(n=8). Median time to development of dysplasia was 149 mos (range, 15 - 405 mos).Adenocarcinoma developed in one patient after 284 mos. The risk of dysplasia at 10, 20and 25 years was 17%, 45% and 69%, respectively (Figure 1). Conclusion: In FAP patientswith ileal reservoirs the risk of low-grade dysplasia increases significantly with time. Progres-sion to high-grade dysplasia and cancer is rare, even with long-term follow-up. FAP patientswith ileal reservoirs necessitate regular follow-up due to high-risk of neoplastic change..

Sa1234

Clinical and Molecular Features of the Hyperplastic Polyposis SyndromeCristina Sanchez-Fortun, Carla Guarinos, Lucía Pérez-Carbonell, Cristina Alenda, ArtemioPayá, Estefanía Rojas, Nuria Acame, Xavier Bessa, José-Luis Soto, Francisco Rodriguez-Moranta, Joaquin Cubiella, David Nicolás-Pérez, Luis Bujanda, Alberto Herreros deTejada, Maite Herraiz, Josep-Maria Reñé, Luisa De-Castro, Anna Serradesanferm,Fernando Fernández-Bañares, Fernando Martínez, Ana Guerra, Virginia Piñol, JudithBalmaña, Marcos Bustamante, María-Luisa Rincón, Angel Ferrandez, Cristina Rodríguez-Alonso, Angeles Pizarro, Rodrigo Jover

OBJECTIVE: Hyperplastic polyposis syndrome (HPS) is a rare condition that has been relatedwith an increased colorectal cancer (CRC) risk. The aim of this study is to describe theclinical and molecular features and the risk of cancer in patients with HPS and their relatives.METHODS: EPIPOLIP study is a multicenter nationwide project aimed to investigate causesof multiple colonic polyps, that includes patients from 21 spanish hospitals presentingbetween 10 and 100 colonic polyps. In this particular study we have selected those patientssatisfying the WHO diagnostic criteria of HPS. Demographic, clinical, familial and patholo-gical data were collected. Pathological specimens of 178 polyps from 10 patients wereanalyzed for BRAF V600E somatic mutation or KRAS somatic mutation. RESULTS: A totalof 32 patients were identified (15 male, 17 female) with a mean age at diagnosis of 53 ±5.7 years (Range 44-66). Twelve patients (37,5%) showed overweight (BMI > 25) and 25

S-260AGA Abstracts

patients (65,6%) were smokers. CRC was detected at diagnosis or during the surveillancein 6 patients (18,8%). Three of these patients had TNM stage I CRC, 1 patient stage II, and2 patients stage III CRC. CRC was located in the rectosigma in 5 out of 6 cases. Familyhistory of CRC was reported in 15 patients (46,9%). A median of 25,5 hyperplastic polyps(range 7-128) per person were reported. Twenty-two patients (68,8%) developed hyperplasticpolyps > 10 mm. Nine patients (29%) had serrated adenomas. Among these patients, themedian number of serrated adenomas was 8 (range 1-41) per person. Twenty-two patients(68,8%) present also adenomatous polyps, with a median of 2 adenomas (range 0-24) perperson. Molecular analysis revealed the presence of BRAF or KRAS mutation in 123 out of178 of the analyzed polyps (69.1%). All the 10 studied patients with HPS showed BRAF(9 patients) or KRAS (1 patient) mutated polyps. The mean percentage of BRAF or KRASmutated polyps was 74.3% (range 47-100%). CONCLUSIONS: Patients with hyperplasticpolyposis syndrome have frequently familial or personal history of CRC. Serrated andconventional adenomas are frequently combined with hyperplastic polyps. Most of the polypsin HPS show somatic BRAF and KRAS mutation that could be used as a biomarker forthis disease.

Sa1235

The Efficacy of Endoscopic Ampullectomy in Familial Adenomatous PolyposisSyndromeEunjeong Jang, Lawrence R. Zukerberg, Robert D. Odze, Manish Gala, Peter B. Kelsey,David G. Forcione, William R. Brugge, Brenna C. Bounds, Sapna Syngal, Daniel C.Chung

BACKGROUND & AIMS: Endoscopic ampullectomy is increasingly performed as primarytherapy for ampullary adenomas (AA) in familial adenomatous polyposis (FAP) patients.The aim of this study was to define procedure-associated morbidities and long-term outcomeof FAP patients with ampullary adenomas. METHODS: We performed a retrospective chartreview of 26 FAP patients (9 male, 17 female) who underwent endoscopic ampullectomyfor AA at two tertiary institutions between 2000 and 2010. The clinical records, genetic testresults, endoscopic examinations, and histopathology were reviewed. The severity of duo-denal polyposis was graded according to Spigelman's classification. After ampullectomy,patients were followed with endoscopic exams according to the recommendations of indi-vidual endoscopists. The severity of acute pancreatitis was assessed by Ranson's criteria.RESULTS: The mean age at the time of endoscopic ampullectomy was 38.9 ± 13.7 (range:20-76) years. According to Spigelman's Classification, 20/26 (77%) patients had stage IIduodenal polyposis. There was no procedure-related mortality. Complications of endoscopicampullectomy were acute pancreatitis (19%), abdominal pain (4%), and bleeding (4%).None of the patients with acute pancreatitis had more than two Ranson's severity criteria.All were successfully managed with conservative measures. Pancreatic duct stents were notplaced in 5 cases, and the rate of pancreatitis was higher in cases without a stent (33% vs.15%, p = NS). The mean adenoma size was 1.03 ± 0.6 cm. Three adenomas (12%) containedhigh-grade dysplasia. The mean follow-up was 30.5 months (range 1-117 months). Long-term follow-up was available for 24 patients, and recurrences were observed in 7 (29%) ata mean of 22.1 (range 11-57) months after ampullectomy. Among these seven, one withhigh-grade dysplasia and Spigelman's stage III polyposis underwent a Whipple's operationfor a recurrence. One patient had a 2nd ampullectomy 14 months after the first. Two patientswere treated with APC fulguration at 19 and 57 months. The remaining 3 recurrences weremanaged by forceps excision, and follow-up exams demonstrated no residual adenoma.Histologic margins were positive for adenomatous tissue in 64% of initial ampullectomies,but there was no correlation between a positive margin and adenoma recurrence. Importantly,no ampullary cancers were observed in the follow-up period. CONCLUSIONS: Endoscopicampullectomy is a safe and effective procedure for AA in FAP patients. None of the complica-tions in our experience was severe. The adenoma recurrence rate of 29% emphasizes theneed for continued surveillance.

Sa1236

High BRCA 1/2 Mutation Rate With Less Significant Breast Cancer AmongFamilies With Pancreatic Cancer Predisposition: A New Indication for BRCA1/2 Testing?Sheila Kumar, Sunil Amin, Kristin Engelstad, Aimee L. Lucas, Lauren G. Khanna,Caroline Hwang, Wendy Chung, Harold Frucht

BACKGROUND: BRCA 1/2 mutations predispose to breast, ovarian, and prostate cancerand are often tested in patients with a familial clustering of such cancers. Recently, BRCA1/2 mutations have been associated with increased risk of pancreatic cancer (PC). It is likelythat the phenotypes of BRCA1/2 mutations vary. We hypothesize that individuals fromfamilies with PC predisposition may have less familial breast cancer when compared toindividuals ascertained through a breast cancer prevention program. METHODS: This is aretrospective case control study comparing pedigrees of patients seen at a PC preventionprogram with those at a breast cancer prevention program at the same institution. 60 patientsin the PC program were tested for BRCA1/2 mutations from 2006-2010. 9 were excludedfrom analysis because of known family history of a mutation and 2 because of lack of familyhistory. 49 patients were then matched for age, presence or absence of cancer at time ofpresentation and Ashkenazi Jewish status, with 49 patients tested for BRCA1/2 in the breastcancer prevention program. Pedigrees for each subject and matched control were examinedfor number of first-degree (FDR) and second-degree relatives (SDR) with history of pancreas,breast, ovarian, or prostate cancer. Statistical comparisons were performed using the chi-square test and Fischer's exact test. RESULTS: Mean age of the PC prevention group was61.2 years, 51% were male, 59.2% had PC, 87.8% were Ashkenazi Jewish, and 16.3% werepositive for BRCA1/2 mutations. PC prevention patients had a significantly higher percentageof PC in FDR (9.4% vs 2.8%, p<0.001), while patients seen in the breast program had asignificantly higher percentage of breast cancer in FDR and SDR (12.8% vs 5.7%, p=0.004;5.7% vs 1.8%, p<0.001). There was no difference between ovarian and prostate cancer inFDR within the 2 populations (0.8% vs 0.4%, p=0.61; 2.3% vs 1.7%, p=0.66). There wasno difference in FDR with breast cancer among the PC prevention patients who were positivefor BRCA1/2 and those who were negative (6.4% vs 5.5%; p=0.733). CONCLUSION: