AG

AA

bst

ract

scases not requiring amplification, and ,10days in 97.5% requiring F/DDISH. Conclusion:Overall HER-positivity rate was 25.8%, identical to that reported in the benchmark ToGAtrial. Our case series was approximately twice the size of the UK ToGA cohort, encompassingreferrals from across the UK. However, when compared with ToGA, subgroup analysisshowed comparable but higher rates in D cancers (10% vs 6.1%), lower in both INT/mixedcancers (30.4/13.5% vs 32.2/20.4%), showed opposite proportions of GOJ/stomach cancers(28.6/31.3% vs 33.1/20.9%), and grouping proximal (OE+GOJ) cancers together, a higherratio compared to stomach cancers was seen (0.51 vs 0.33). Positivity in Sq cancer is ofuncertain significance and warrants further research into new potential GI applications forHER2 targeted therapies.

Membranous HER2 staining pattern in gastroesophageal junction adenocarcinoma.

Dual-silver in-situ hybridisation (DDISH) schematic showing (A) comparative HER2 andchromosome 17 (C17) gene copy numbers in normal tissue and (B) abnormal ratio of HER2/C17 in a HER2 positive gastric adenocarcinoma.

Mo1135

The Genetic Polymorphism of Pre-MicroRNA 938, Targeting IL-17A 3'-UTR, IsAssociated With the Risk for DAP-kinase Hypermethylation in Helicobacterpylori Infected Japanese SubjectsTomiyasu Arisawa, Hideto Yamada, Tomomitsu Tahara, Masakatsu Nakamura, TomoeNomura, Ranji Hayashi, Kazuhiro Matsunaga, Toshimi Otsuka, Nobuyuki Toshikuni,Hisakazu Shiroeda, Mikihiro Tsutsumi, Tomoyuki Shibata

[Background] CpG island aberrant methylation is shown to be an important mechanism ingene silencing. Death-associated protein kinase (DAP-kinase) participates in various apoptosissystem. Methylation of DAP-kinase has been reported in many cancers including gastriccancer. On the other hand, interleukin-17A (IL-17A) plays an important role in tissueinflammation by inducing release of proinflammatory and neutrophil-mobilizing cytokines.

S-588AGA Abstracts

We have already reported the close association of IL-17A gene polymorphism with thegastric carcinogenesis (Hum Immunol, 2008). In this study, we investigated the effect ofpolymorphisms in IL-17A gene 3'-UTR (*1249C.T, rs3748067) and pri-microRNA938(rs2505901T.C), targeting IL-17A gene 3'-UTR, on the methylation of DAP-kinase. [Materi-als and Methods] Non cancerous gastric mucosa samples were obtained from 295, including189 Helicobacter pylori (HP) infected subjects. Methylation status of DAP-kinase gene wasdetermined by methylation-specific polymerase chain reaction. Genotyping was performedby PCR-SSCP methods. [Results] Methylation of DAP-kinase was seen in 174 subjects. Themean age and HP positive ratio were significantly higher in methylated group (M group)than non methylated group (non-M group) (p=0.0036 and 0.0030, respectively), whereasthe male/female ratio was not different between two groups. The distribution of IL17A*1249C.T genotype in M group was 158CC, 10CT and 6TT, whereas the distribution innon-M group was 101CC, 14CT and 6TT. The distribution of rs2505901T .C genotype inM group was 106TT, 61TC and 7CC, whereas the distribution in non-M group was 66TT,51TC and 4CC. Overall, there was no significant association of IL17A *1249C .T andrs2505901T.C with DAP-kinase methylation by logistic regression analysis after adjustmentfor age, gender and HP infection status. In HP infected subjects, however, rs2505901 CT+CCgenotype had a significant reduced risk for DAP-kinase methylation (OR, 0.532; 95%CI,0.287-0.986; p=0.045), although there was no significant relationship between IL17A*1249C.T polymorphism and DAP-kinase methylation. In addition, in the HP infectedsubjects younger than 60 years old, rs2505901 C carrier had a more reduced risk for DAP-kinase methylation (OR, 0.414; 95%CI, 0.178-0.966; p=0.041). [Conclusions] Our resultssuggest that rs2505901, located in pri-miR938 gene, C allele carrier has a significant reducedrisk for DAP-kinase methylation under the influence of HP infection, especially in youngersubjects, whereas IL-17A *1249C.T polymorphism does not affect DAP-kinase methylation.

Mo1136

An Association Between Genetic Polymorphisms of Ripk2 (Ile259thr and*351a>G) and the Susceptibility to Gastric Cancer in JapanTomiyasu Arisawa, Tomomitsu Tahara, Masakatsu Nakamura, Hideto Yamada, TomoeNomura, Ranji Hayashi, Kazuhiro Matsunaga, Toshimi Otsuka, Nobuyuki Toshikuni,Hisakazu Shiroeda, Tomoyuki Shibata

[Background] RICK protein, a serine-threonine kinase, functions downstream of the patternrecognition receptors NOD1 and NOD2 to mediate NF-kB and MAPK activation. Recently,several reports have been shown a significant association between cancer risk and geneticpolymorphisms of both NOD2 and RIPK2 encoding RICK. There has been no investigationfor an association of RIPK2 polymorphisms with gastric cancer risk, although NOD2 polymor-phisms are reported to be associated with gastric cancer risk. The rs16900627 (*351A .G),a constituent polymorphism in one of the large LD block with rs7844627 etc. in RIPK2, isidentified. We investigated an association between gastric cancer risk and RIPK2 polymor-phism rs16900627. In addition, an association of minor polymorphism rs2230801(Ile259Thr) was also investigated. [Material and Methods] The study was performed in 524patients with gastric cancer (GC cases) and 697 subjects with no evidence of gastric malignan-cies (controls) on upper gastro-duodenal endoscopy. We employed the PCR-SSCP methodto detect the gene polymorphism. [Results] The mean age, male/female ratio and Helicobacterpylori (HP) positive ratio in GC cases were significantly higher than those in controls. Thedistributions of rs16900627A.G and rs2230801T.C in controls were 554AA, 130AG and13GG (HWE p=0.12), and 681TT, 16TC and 0CC (HWE p=1.0) respectively, whereas thedistributions in GC cases was 395AA, 115AG and 14GG, and 517TT, 7TC and 0CC,respectively. The genotype frequency of rs16900627 GG homozygous was significantlyhigher in controls than in GC cases (p=0.029). Overall, rs16900627 AG+GG genotype hada significantly increased risk for the development of gastric cancer by logistic regressionanalysis after adjustment for gender, age and HP infection status (dominant genetic model:OR, 1.38; 95%CI, 1.03-1.84; p=0.032). Using number of rs16900627G allele as a co-variate,this allele was also significantly correlated to the increased risk for gastric carcinogenesis(OR, 1.37; 95%CI, 1.06-1.77; p=0.016). No significant association was found betweengastric cancer risk and rs2230801. In addition, rs16900627 minor allele was significantlyassociated with intestinal type of gastric cancer (dominant genetic model: OR, 1.56; 95%CI,1.11-2.20; p=0.011 and number of G allele: OR, 1.53; 95%CI, 1.13-2.07; p=0.0062), whereasno significant association with diffuse type of gastric cancer was found. [Conclusion] Ourresults provided the first evidence that RIPK2 gene polymorphism is significantly associatedwith the susceptibility to gastric cancer development. The rs16900627A .G minor allele isclosely associated with an increased risk for the development of gastric cancer, especiallyintestinal type of cancer.

Mo1137

Racial Differences in Young Patients With Colorectal Cancer: A Population-Based Study in the United StatesJyothsna Talluri, Siddesh V. Besur, Siva K. Talluri

Introduction: Colorectal cancer is one of the leading causes of cancer deaths in the UnitedStates.Studies report an increase in the incidence of colorectal cancer in patients less than50 years. A higher proportion of the patients in this young age group are African-American.Racial differences in tumor characteristics and survival of young African-Americans withCaucasians in United States was not previously studied. The objective of this study is toevaluate the differences in characteristics of African-Americans and Caucasians in youngpatients (less than 50 years) with colorectal cancer. The secondary objective is to comparedifferences in colorectal cancer-specific survival between young Caucasians and African-Americans. Methods: We used National Cancer Institute's Surveillance, Epidemiology, andEnd Results (SEER) database for this study. It contains information about patient and tumorcharacteristics at diagnosis, survival. Seventeen registries are enrolled in the SEER databaseand covers approximately 28% of the US population. We analyzed a retrospective cohortof colorectal cancer patients included in the SEER database. We included all patients lessthan 50 years with primary colorectal cancer in SEER database from 1973-2009. Differencesin demographic and tumor characteristics (tumor size, grade, lymph node status and stage)at diagnosis were compared using the chi-square tests. We compared the survival time

between Caucasians and African-Americans using Kaplan-Meier method. We evaluated theeffect of race on survival time after adjusting for differences in stage, age and grade usingCox-proportional hazards regression. We used SPSS software (version 19.0) for the analysis.Results: We included 11,184 patients who were less than 50 years (9432 Caucasians and1752 African-Americans) in our study. The mean (SD) age in years at diagnosis was similarin Caucasians 42(6) and African-Americans 42(6). Sigmoid colon was the commonest siteaffected in both groups (Caucasians 24% and African-Americans 20%). Differences in tumorcharacteristics at diagnosis is shown in Table 1. Surgery was performed in most of thepatients in both groups (Caucasians 99% and African-Americans 100%). Mean survival waslonger in Caucasian patients as compared to African-Americans (15 years vs. 12 years) P, 0.001 Log Rank (Mantel-Cox). Young Caucasians had better 5-year colon cancer-specificsurvival compared to African-Americans (62% vs. 52% respectively). After adjusting fordifferences in stage of disease at presentation, age and grade, African-Americans surviveshorter time as compared to Caucasians (hazard ratio 95% CI 1.39 (1.28-1.51) (Figure 1).Conclusion: Young African-Americans had advanced stage of colorectal cancer at diagnosis.They had shorter survival time even after adjustment for differences of stage, grade and age.Table1: The characteristics of colorectal cancer in Caucasians and African-Americans

All differences statistically significant P ,0.05

Figure 1. Differences in Survival between Caucasians and African-Americans ( ,50 years)after adjusting for age, stage and grade.

Mo1138

Socioeconomic Status and Gender Are Associated With Inequalities inSurgical Treatment and Short-Term Outcomes in Patients With Non-MetastaticColorectal CancerVincent K. Dik, Mieke J. Aarts, Valery Lemmens, Peter D. Siersema

Background Low socioeconomic status (SES) and male gender are associated with a lowersurvival in colorectal cancer (CRC) patients. Factors responsible for these inequalities includelifestyle, access to and use of medical care, and presence of comorbidities. Furthermore,several studies have shown that low SES cancer patients undergo chemotherapy less fre-quently, but only a few studies have focused on the association between SES and surgicaltreatment. Aim To investigate whether gender and SES are associated with surgical treatmentand short-term outcomes after surgery in patients with stage I-III CRC. Methods All patientsdiagnosed with stage I-III CRC in the period 2003-2010 in the Eindhoven Cancer Registryarea in the Netherlands (~2.4 million inhabitants) were included. Demographic and clinicaldata were routinely collected from medical charts. Multivariable logistic analyses were usedto estimate adjusted odds ratios (OR) for the outcomes 1-year mortality, residual tumorafter surgery, ≥12 lymph nodes examined and anastomotic leakage/abscess formation, andfor undergoing surgical treatment of the primary tumor, emergency surgery, and minimalinvasive (laparoscopic/endoscopic/TEM) treatment. The multivariable analyses included age,gender, SES, comorbidities, T and N stage, tumor localization, histology, differentiationgrade, chemo- and radiotherapy, and for 1-year mortality, surgery of the primary tumor.Results A total of 9,598 stage I-III CRC patients were included (65% colon, 35% rectum).Female gender (OR 0.81, 95%-CI 0.70-0.93) and high SES (OR 0.72, 95%-CI 0.60-0.87)were independently associated with a decreased 1-year mortality. Gender and SES were notassociated with the likelihood to undergo surgical treatment. However, in patients who didreceive surgical treatment (97%), high SES was independently associated with an increasedlikelihood to undergo minimal invasive surgery compared to low SES (OR 1.37, 95%-CI1.11-1.68). Female CRC patients were less likely to develop anastomotic leakage/abscessformation (OR 0.77, 95%-CI 0.62-0.95) and a similar trend was observed for colon cancer

S-589 AGA Abstracts

patients with a high SES (OR 0.73, 95%-CI 0.51-1.04). Female rectal cancer patients lessfrequently had 12 or more lymph nodes examined compared to male patients (OR 0.80,95%-CI 0.67-0.97). There were no significant associations found with emergency surgicaltreatment and residual tumor after surgery. Conclusion Marked SES and gender specificdifferences exist with regard to surgical treatment and (short-term) outcomes in stage I-IIICRC. However, it remains to be elucidated whether these findings can be attributed to otherpatient- or physician-related factors. Meanwhile, low SES and male patients may benefitfrom specific attention of physicians with regard to the choice of surgical technique and themonitoring of post-operative complications.

Mo1139

Coffee Consumption, Genetic Polymorphisms in CYP1A2 and NAT2, andColorectal Cancer RiskVincent K. Dik, Martijn G. van Oijen, Cuno Uiterwaal, Carla van Gils, Fränzel J. vanDuijnhoven, Stephane Cauchi, Philippe Froguel, Loic Yengo, Peter D. Siersema, HendrikB. Bueno-de-Mesquita

Background Coffee has been reported to reduce colorectal cancer (CRC) risk. However,findings of epidemiologic studies on this association are inconsistent, which may partly beexplained by differences in study population. Caffeine and other coffee components thatare possibly associated with CRC risk, are metabolized by the cytochrome P450 isoformCYP1A2 and N-acetyltransferase (NAT)2. Depending on genetic polymorphisms, large inter-personal variability can exist in CYP1A2 and NAT2 activity, resulting in an alteredmetaboliza-tion of coffee components that may influence CRC risk. Aim To investigate whether theassociation between coffee consumption and CRC risk is modified by genetic variation inCYP1A2 and NAT2. Methods A nested case-control study within the EPIC cohort study wasdone including 974 CRC cases and 1,591 controls, matched by sex, age at recruitment,study center and menopausal status. DNA was extracted from buffy coat samples andgenotyped using the Illumina Metabochip. Slow CYP1A2 activity was defined as the homozy-gous presence of the single nucleotide polymorphism (SNP) 164A .C. Slow NAT2 activitywas defined as the presence of two slow NAT2 alleles, which are characterized by thepresence of one or more of the SNPs 191G .A, 590G.A, 857G.A or 341T.C. For thisstudy 481C.T was used as a proxy for 341T.C (R2 0.89). Multivariable conditional logisticregression models were used to estimate odds ratios (OR) for the joint effects of differentlevels of coffee consumption with CYP1A2 and NAT2, adjusted for potential CRC riskfactors reported in literature. Results In comparison to non/low consumption, high coffeeconsumption was not associated with overall (OR 0.99, 95%-CI 0.79-1.24) or site-specificCRC. CYP1A2 activity was also not associated with CRC risk. However, a slow NAT2 statuswas associated with an increased CRC (OR 1.17, 95%-CI 1.00-1.37), and more specifically,rectal cancer risk (OR 1.34, 95%-CI 1.02-1.77). The joint effect model for coffee consumptionand CYP1A2 activity did not reveal significant associations with CRC risk (OR 1.04, 95%-CI 0.62-1.76 for high consumers with slow CYP1A2 activity vs. non/low consumers withrapid CYP1A2 activity). Furthermore, for coffee consumption and NAT2 activity combinedalso no association was found with CRC risk (OR 0.98, 95%-CI 0.71-1.36 for high consumingslow NAT2 acetylators vs. non/low consuming rapid NAT2 acetylators). Finally, separateanalyses for colon and rectal cancer did not reveal significant associations for high coffeeconsumption jointly with CYP1A2 or NAT2 activity. Conclusion In this nested case-controlstudy, coffee consumption was not associated with CRC risk and no effect modification byCYP1A2 and NAT2 activity was found. In contrast to previous studies, a slow NAT2acetylation status was associated with an increased rectal cancer risk.Table 1. Unadjusted odds ratio's for the main effects of coffee consumption, CYP1A2and NAT2.

All models were conditioned on the matching factors sex, age at recruitment, study centerand menopausal status.Table 2. Adjusted odds ratio's for the joint effects of CYP1A2 and NAT2 activity with coffeeconsumption for colorectal cancer risk within 974 CRC cases and 1,591 matched controls.

AG

AA

bst

ract

s