S1114

Association of Large Hyperplastic Polyps with Synchronous AdvancedColorectal NeoplasiaDan Li, Chengshi Jin, Charles McCulloch, Barry M. Berger, Thomas F. Imperiale,Jonathan P. Terdiman

Background & Aims: Hyperplastic polyps may increase the risk for colorectal cancer. Thepresent study sought to determine the association between the presence of large hyperplasticpolyps in the colorectum and synchronous advanced colorectal neoplasia. Methods: Among4,714 asymptomatic persons who underwent screening colonoscopy, cases of advancedcolorectal neoplasia (tubular adenoma >/=1cm, adenoma with any villous histology, adenomawith carcinoma in situ or high grade dysplasia, or invasive adenocarcinoma) were comparedwith controls without advanced neoplasia for the presence of large (>/= 1 cm) hyperplasticpolyps. Candidate predictors of advanced neoplasia were age, sex, family history of colorectalcancer, body mass index, the presence and number of small tubular adenomas (<1cm) andsmall hyperplastic polyps (<1cm), and the presence of large hyperplastic polyps. Independentpredictors of advanced neoplasia were determined by multivariate logistic regression analysis.Results: We identified 467 cases and 4, 247 controls. In total, 109 subjects (2.3% of thestudy population) had large hyperplastic polyps. Independent predictors of advanced colonicneoplasia were increasing age (OR=4.51, 95% CI 1.43-14.3, P=0.01 for subjects >/=80 yearsvs. 50 to 54 years), 3 or more non-advanced tubular adenomas (OR=2.33, 95% CI 1.37-3.96, P=0.0017), and large hyperplastic polyps (OR = 3.24, 95% CI 2.05-5.13, P<0.0001).Right-sided (OR=3.38, 95% CI 1.81-6.32) and left-sided (OR=2.66, 95% CI 1.37-5.19)large hyperplastic polyps had a similar association with advanced colorectal neoplasia (rightvs. left P=0.622). BMI, male sex, colorectal cancer in a first degree-relative, and the presenceof multiple small hyperplastic polyps did not associate independently with advanced colonicneoplasia. Conclusions: Although found in a small percentage of average risk subjectsundergoing screening colonoscopy, large hyperplastic polyps are strongly and independentlyassociated with synchronous advanced colorectal neoplasia. Our results suggest that largehyperplastic polyps are important lesions that may merit complete removal and subsequentcolonoscopic surveillance.

S1115

A Prognostic/Diagnostic DNA Fluorescent in Situ Hybridization Probe Set forIdentifying Patients At Risk and (Early) Neoplasia in Barrett's EsophagusRygiel M. Agnieszka, Francesca Milano, Wytske Westra, Brenda Elzer, Annet Schaap,Jacques J. Bergman, Kausilia K. Krishnadath

BACKGROUND and AIM: The progression of Barrett's esophagus (BE) to esophageal aden-ocarcinoma (EAC) is often characterized by the accumulation of genetic abnormalities. Theseabnormalities may serve as useful prognostic or diagnostic clinical markers to improvesurveillance of BE patients. The goal was to comprehensively evaluate a panel of geneticmarkers with potential prognostic and/or diagnostic value for dysplasia and EAC in BE.MATERIAL and METHODS: Fluorescent in situ hybridization with DNA probes for thelocus specific regions (LSI) of 9p21 (p16), 17p13.1 (p53), 17q11.2 (Her-2/neu), 8q24 (c-myc), 20q13 and for centromeric regions (CEP) of chromosome 7, 17 was applied on brushcytology specimens of a total of 183 BE patients with different stages of dysplasia or EACas defined by histology. RESULTS: We found that gains of chromosome 7 /17 and loss of17p13.1 (p53) and 9p21 (p16) were present in low frequencies of 5-30% in non-dysplasia(ND) which further significantly increased to 9-36% in low grade (LGD), 46-88% in highgrade dysplasia (HGD) and 60-86% in EAC cases. Amplification of 17q11.2 (Her-2/neu),8q24 (c-myc), 20q13 loci was not observed in ND and LGD but was present in 5-45% ofHGD and 25-33% of EAC cases. The frequency of 17q11.2 (Her-2/neu) and 8q24 (c-myc)amplifications combined with that of gains of chromosome 7 and 17 resulted in the bestsensitivity of 89% and specificity of 88% to detect HGD or EAC. CONCLUSIONS: Gainsof the chromosome 7, 17 and loss of 9p21 (p16) and 17p13.1 (p53) are potentially prognosticmarkers since they are already present in ND cases and correlate with increasing severityof dysplasia, while gain of chromosome 7 and 17 and amplification of the 17q11.2 (Her-2/neu) and 8q24 (c-myc) are useful as diagnostic markers for identifying HGD and/or EAC.The combined DNA FISH probe set including CEP 7, 17, and LSI probes for 9p21(p16),17p13.1 (p53), 17q11.2 (Her2/neu) and 8q24 (c-myc) seems to be an suitable surveillancetool for early identification of BE patients at risk, and for diagnosing (early) neoplasia inBE. In the future, adding DNA FISH markers to current surveillance programs may highlyimprove its efficiency.

S1116

Incidence of Right-Sided Colorectal Cancers After Breast and OtherGynecological Cancers: A Population-Based StudyLinda Y. Tang, Zoann Nugent, Alain Demers, Syed A. Rasul, Harminder Singh

Estrogen levels, which are involved in the development of breast and other gynecologicalcancers, may also be responsible for higher incidence of right-sided colorectal neoplasia inwomen. No previous study has evaluated the risk of right-sided colorectal cancer (CRC)after breast and other gynecological cancers. Increased incidence of right-sided CRC mayjustify a preference for CRC screening by colonoscopy. Our aim was to determine theincidence of right-sided CRC after the diagnosis of gynecological cancers. We were alsointerested in determining the current endoscopic practice after diagnosis of gynecologicalcancers. Methods: All cases of breast, endometrial, and ovarian cancers diagnosed between1956 and 2005 were identified from the Manitoba Cancer Registry and followed-up untildiagnosis of CRC, death, migration out of the province, or December 31, 2005. Age-standardized incidence ratios (SIR) for all CRC and right-sided CRC (cecum, ascendingcolon and hepatic flexure) were calculated to compare the observed CRC incidence to thatexpected in the general population. Stratified analysis was performed to determine the riskafter different durations of follow-up (<1 year, 1-5 years, 5-10 years, >10 years) after theprimary cancer diagnosis. Colonoscopy practice over the last twenty years was determinedby linkage to the provincial physicians' billing claims administrative database. Results: There

T : 11501$$CH204-02-08 16:47:06 Page 181Layout: 11501B : o

A-181 AGA Abstracts

were 32,568 cases of breast, endometrial or ovarian cancers diagnosed between 1956 and2005. The SIR for all CRC was 0.99 (95% CI 0.90-1.07) and for right-sided CRC was 1.04(95% CI 0.91-1.20). This risk did not change in the more recent years: after primarygynecological cancer diagnosis between 1985 and 2005, the SIR for all CRC was 0.98 (95%CI 0.86-1.12) and that for right-sided CRC was 1.08 (95% CI 0.87-1.34). The SIRs remainedequal to unity for different durations of follow-up for both all CRCs and right-sided CRCs(SIRs for right-sided CRC: <1 year 1.09 (95%CI 0.59-1.82), 1-5 years 1.29 (95%CI 0.99-1.68), 5-10 years 1.16 (95%CI 0.87-1.54), and >10 years 0.88 (95%CI 0.71-1.09). Similarrates of colonoscopy were performed one year prior to the breast cancer diagnosis and forfollow-up of 10 years. There was an increase in the rate of colonoscopies performed withinone month of the diagnosis of uterine and ovarian cancers, which returned to the baselinethereafter. Conclusions: There is no increase in the overall risk for CRC or for right-sidedCRC after the diagnosis of breast, endometrial or ovarian cancer. CRC screening strategy, afterthe diagnosis of a gynecological cancer, should be similar to that for the general population.

S1117

Faecal Occult Blood Tests: Immunological Superior to Guaiac Based?Frank A. Oort, Jochim S. Terhaar sive Droste, Mike E. Craanen, René W. van der Hulst,Henk van Heukelem, Ruud J. Loffeld, Liesbeth Mutsaers, Roy van Wanrooy, Eric C.Wesdorp, Suzanne J. van de Reijt, Laura de Baaij, Gerrit A. Meijer, Chris J. Mulder

Introduction: Guaiac-based faecal occult blood tests (FOBT's) in a colorectal cancer screeningsetting are commonly hampered by a poor specificity and positive predictive value, resultingin many (futile) follow-up colonoscopies. Hence, immunological FOBT's with apparentlybetter clinical performance, absence of dietary restrictions and only one faecal samplerequired, have been proposed as a more efficient screening tool. Aim: To compare animmunology-based (OC sensor®, Eiken chemical Co, Japan) and a guaiac-based (hemoc-cult®, Beckman Coulter, Inc. USA) FOBT in consecutive patients undergoing colonoscopyin terms of clinical yield of colorectal cancer and advanced adenomas. Methods: All patientsaged >= 18 years and scheduled for a colonoscopy in participating hospitals (N=5) wereasked to perform both FOBT's in the week prior to colonoscopy. Exclusion criteria were: Ahistory of IBD, failure to complete both tests and an incomplete colonoscopy. A haemoglobinconcentration of >= 100ng/ml in the test sample was considered a positive result. McNemar'stest was used for the comparison of correlated proportions. P<0,05 was considered statisticallysignificant. Results: After appliance of exclusion criteria, 1331 eligible patients were included.Colorectal carcinoma and advanced adenomas (i.e. >= 1 cm in diameter and/or villousarchitecture and/or high-grade dysplasia) were found in 3,4% and 8,6% of the patients,respectively. Small adenomas, colitis and other lesions were identified in 55,2% of thepatients. No lesions were found in 32.8% of the patients. The hemoccult® test and OCsensor® test showed positivity rates of 5.9% and 11,9%, respectively. Test characteristicsfor both FOBT's are shown in Table 1. None of the observed differences between the testswere statistically significant. Conclusions: Although the sensitivity and specificity of the OC-sensor test in detecting colorectal cancer was high in this patient group, the sensitivity todetect high-risk, pre-cancerous lesions was disappointing. The sensitivity of the Hemocculttest is poor for both CRC and advanced adenomas. Moreover, the low positive predictivevalue of both tests in pre-cancerous patients might hamper the introduction of either oneof these tests in a screening setting. A larger cohort is currently being investigated.Test characteristics

*advanced neoplasia consists of advanced adenomas and cancers

S1118

Prospective Comparative Evaluation of An Office Based Rapid ImmunologicalTest with a Guaiac Based Fecal Occult Blood Test for Colorectal CancerScreening in General Population with Average-RiskYogesh Shastri, Stefan Loitsch, Roland Nowak, Nada Povse, Jürgen Stein

BACKGROUND: Immunological fecal occult blood test(IFOBT) has established itself as anew tool for colorectal cancer (CRC) screening. Also the simpler, cheaper and more conveni-ent newer office based immunochemical strip test have been validated for CRC screeningin population with above average risk for CR neoplasia (Hoepffner N, Shastri YM- AlimentPharmacol Ther. 2006;23(1):145-54; Shastri YM- Gastroenterology 2007 Apr; 132 (4):(Suppl 1) A-317). This new office based IFOBT has not yet been evaluated as a CRC screeningmarker in average risk general population. AIM: To compare the performance characteristicsof an office based IFOBT with that of Guaiac based FOBT as a screening biomarker forcolorectal neoplasia in general population. METHODS: All employees of German Lufthansaairline older than 45 years were invited to participate in CRC screening test with Guaiacbased FOBT Haemoccult (Beckman Coulter Inc., USA) and the office based rapid IFOBTstrip test (Prevent ID® CC, Preventis GmbH, Germany). Out of 3223 subjects invited toparticipate in the study, 2206 (68.4%) submitted their stool samples. There were 1175(53.4%) females and 1012 males (45.9%) with average age of 51 yrs. The positivity ratefor the hemoccult test was 6.5% while that for IFOBT it was 2.7%. Out of these positivepatients, 121 (59.3%) underwent colonoscopy. Clinically significant lesions were definedas adenomas with high grade dysplasia, polyps larger than 10 mm, villous adenomas or 3or more adenomas. Each Guaiac based FOBT costs about 1 €, IFOBT strip test 3-5 €

while an ELISA based IFOBT costs around 15-20 €. RESULTS: Patient and performance

AG

AA

bst

ract

s