T1205 How Relevant is Acoustic Radiation Force Impulse Elastography (ARFI) for the Evaluation of Liver Fibrosis?

variance (CV) of the G1 peak. A second case was called tetraploid by IC, aided by visualisation of the 4N nuclei. This case was labeled as diploid by FC with a peak at 4N gated out as debris. All controls were diploid at both centres. Of the matched cases 67% were aneuploid, 9% tetraploid, and 24% diploid. CV of the G1 peak was lower by FC than ICDA (p = 0.04). There was no significant difference in the DNA index of the aneuploid peak between the two methods. Conclusion: These data demonstrate ICDA is a reliable technique for the detection of DNA ploidy abnormalities compared to FC when nuclei are extracted from thick FFPE sections. Results would suggest that due to a lower CV FC is better at measuring near diploid aneuploid populations, but may miss tetraploid populations which are readily seen on ICDA. This work is supported by the CRUK Experimental Cancer Medicine Centre at UCL and the NIHR Comprehensive Biomedical Research Centre at UCLH.

T1204 Which is the Best Noninvasive Ultrasound Method for the Evaluation of Liver Fibrosis? Ioan Sporea, Roxana Sirli, Alina Popescu, Mirela Danila, Diana Nicolita, Alexandra E. Deleanu Background: Currently, validated methods for the non-invasive evaluation of fibrosis in chronic hepatitis are transient elastography (TE) and serological tests (FibroTest-ActiTest). A new ultrasound based noninvasive method became available in the last years: Acoustic Radiation Force Impulse (ARFI). The aim of this study is to try to find which method that uses ultrasound waves is the best for the evaluation of liver stiffness (LS). Patients and methods: Our study included 71 subjects (54 with HCV and 17 with HBV chronic hepatitis) in which we compared TE and ARFI to the liver biopsy (LB) (evaluated according to the Metavir scoring system, considered to be the “gold standard”). In each patient we performed LS measurement by means of TE (FibroScan®, EchoSens) and RT-E (by using a Siemens Acuson S2000TM ultrasound system). 10 LS measurements were perfomed both by means of TE and ARFI; median values were calculated expressed in kiloPascals and meters/second, respectively. TE, ARFI and LB were performed in the same session. Results: On LB, from the 71 patients, 6 (8.4%) subjects had F1, 25 (35.2%) subjects had F2, 24 (33.8%) patients had F3 and 16 (22.5%) patients had F4. A direct, strong, linear correlation (Spearman r= 0.707) was found between TE measurements and fibrosis (p<0.0001) and a weaker one between ARFI and fibrosis (rho=0.469; p<0.0001). TE measurements were also correlated with ARFI measurements r=0.532, p<0.0001. By comparing the AUROC curves, TE and ARFI had similar predictive values for the presence of significant fibrosis (F≥2 Metavir): AUROC ARFI=0.649, AUROC TE=0.731 (p=0.476); and cirrhosis (F=4 Metavir): AUROC ARFI=0.868, AUROC TE= 0.936 (p=0.294) Conclusion: LS measurements assessed by means of TE correlate better than those assessed by means of ARFI to the histological fibrosis in patients with HBV and HCV chronic hepatitis. Both methods have excellent predictive values for the presence of cirrhosis.

T1207 Transoral Sutured Revision of Dilated Gastrojejunostomy for Treatment of Weight Regain in Roux-en-Y Gastric Bypass Patients: A Retrospective Review of 243 Procedures in 186 Consecutive Patients With 2 Year Kaplan-Meier Estimates Marvin K. Ryou, Song Yu, Ian T. Greenwalt, Bechien U. Wu, David B. Lautz, Christopher C. Thompson Background Dilation of the gastrojejunostomy (DGJ) is a potential mechanism for weight regain after Roux-en-Y gastric bypass (RYGB). Endsocopic dilated gastrojejunostomy reduction (DGJR) via transoral suturing is a less invasive and potentially effective alternative to surgical revision. Aim To assess the safety and efficacy of DGJR using time-to-event analysis for weight regain. Methods 243 consecutive DGJR procedures performed for weight regain in 186 patients from 2005-07 were retrospectively evaluated. DGJR included mucosal ablation and placement of interrupted stitches at the anastomosis. Inclusion criteria were ≥6 months post-RYGB; DGJ ≥2cm. Exclusion criteria were fistulae and marginal ulcers. Patient characteristics, serial weights, stoma sizes, and complications were obtained from medical records. Given variable loss to follow-up, a time-to-weight regain analysis was performed using the Kaplan-Meier method. Patients were censored when they regained their baseline weight or if they had a second procedure. Univariate and multivariate Cox proportional hazards regression analyses were performed to identify success predictors. A subgroup analysis of 45 patients were matched to 45 controls not having undergone DGJR based on age, sex, diabetes, time since RYGB, and % excess weight loss after RYGB. Results The mean age was 46 years, 91% female, mean initial BMI was 52.3, mean weight regain from nadir weight was 53 lbs. DGJR technical success was achieved in 98.7%. Average stoma reduction was 2.25 to 0.73 cm (67.5%) using an average of 3.7 stitches. There were 6 overnight admissions (2.5%) for abdominal pain (3) and GI bleeding (3). Kaplan-Meier estimates showed 75% of patients were above their baseline weight or avoided a 2nd procedure at 6 months; 66% at 12 months; 50% at 16 months. Cox proportional hazards regression analysis showed that % EWL post RYGB predicted weight loss/stabilization at 12 months. In a subgroup matched cohort analysis with controls, 35% of sutured patients were below their baseline weight at 2 years compared to 10% of controls (P < 0.001). Conclusion DGJR appears safe and Kaplan-Meier analysis shows that DJGR significantly alters weight trends compared to matched controls up to 2 years post-procedure.

T1205 How Relevant is Acoustic Radiation Force Impulse Elastography (ARFI) for the Evaluation of Liver Fibrosis? Ioan Sporea, Roxana Sirli, Alina Popescu, Mirela Danila, Simona I. Bota Background: Currently, validated methods for the non-invasive evaluation of fibrosis in chronic hepatitis are transient elastography (TE) and serological tests (FibroTest-ActiTest). A new ultrasound based noninvasive method became available in the last years: Acoustic Radiation Force Impulse (ARFI). The aim of this study is to find the relevance of Acoustic Radiation Force Impulse Elastography (ARFI) for the evaluation of liver fibrosis. Patients and methods: Our study included 183 subjects (94 females and 89 males, mean age 46.9+/15.9): 70 subjects without fibrosis (healthy volunteers - considered F0 Metavir), 8 subjects with F1, 26 subjects with F2, 25 patients with F3 (all patients underwent liver biopsy) and 54 patients with liver cirrhosis (16 with liver biopsy and 38 with clinically, ultrasonographic and/or endoscopic signs of cirrhosis). In each patient we performed ARFI (by using a Siemens Acuson S2000TM ultrasound system). Ten ARFI measurements were made in each patient and a median value was calculated, measured in meters/second. Results: From the 183 subjects, ARFI valid measurements were obtained in 179 subjects (invalid determinations in 2.2%). A direct, strong, linear correlation (Spearman rho=0.663) was found between ARFI measurements and fibrosis (p<0.0001). The mean values of ARFI according to fibrosis groups were: F0 1.31+/-0.48m/s; F1 1.52+/-1.02m/s; F2 1.61+/-0.68m/s; F3 1.76+/-0.76m/ s; F4 2.81+/-0.71m/s) For predicting significant fibrosis (F≥2 Metavir) the AUROC for ARFI was 0.819, for a cut-off value of 1.4 m/s, the Se was 75.7%, Sp was 75.3%%, with 80.4% PPV and 69.9% NPV. For predicting severe fibrosis (F≥3 Metavir) the AUROC for ARFI was 0.839, for a cut-off value of 1.7 m/s, the Se was 76.9%, Sp was 91.2%, with 87% PPV and 83.8% NPV. For predicting cirrhosis (F=4 Metavir), the optimized cut-off was 1.8 m/ s (AUROC=0.907, with 94.4% Se, 86.5 Sp, 75% PPV and 97.3% NPV). Conclusions: Liver stiffness evaluation by means of ARFI can be performed in the vast majority of patients (97.8%). ARFI is an accurate method for predicting cirrhosis (AUROC 0.907).

T1208

T1206

High Throughput Quantification of Kinase Activities in Crohn's Disease Patients Demonstrates Active Suppression of P21rac Signalling in the NonInflamed Mucosa Kaushal Parikh, Gerard Dijkstra, Maikel P. Peppelenbosch

Comparison of Image Cytometry and Flow Cytometry for the Detection of DNA Ploidy Abnormalities in Barrett's Esophagus Jason M. Dunn, Peter S. Rabinovitch, Dahmane Oukrif, Stephen G. Bown, Marco Novelli, Laurence Lovat

Abstract Background & Aims: The most striking clinical difference between ulcerative colitis (UC) and Crohn's disease (CD) is that in CD there is discontinuity of inflammation, with ulcers being scattered throughout the gut, surrounded by apparently normal mucosa (so called skip lesions), whereas UC presents as a continuous colonic inflammation, ascending from the rectum. Although recent data from Genome Wide Association studies have increased our knowledge on the genetic basis of these diseases the signalling pathways especially with regard to these skip lesions remain poorly understood. In order to reveal the signal transduction pathways involved we aimed to generate comprehensive descriptions of mucosal kinase activity. Methods: Paired biopsies of inflamed and non-inflamed colonic mucosa were analyzed on novel peptide arrays exhibiting > 1000 pseudo substrates to characterize mucosal kinase activity in Toto. Results were compared to colonic biopsies obtained from non-IBD patients and validated using conventional biochemistry and immunohistochemistry. Results: Inflamed mucosa of Crohn's disease patients displays strong activation of the NF-kB pathway, STAT-related signalling, nitric oxide signalling and very strong activation of p21Rac-related signalling. In contrast, p21Rac signalling in the non-inflamed mucosa of CD patients is highly suppressed. Conclusions: Our results show that the non-inflamed mucosa in CD is highly different from the mucosa of non-IBD patients in its very active suppression of

Introduction: DNA ploidy abnormalities (aneuploidy/tetraploidy) are strong predictors of future cancer risk in Barrett's esophagus (BE) as measured by flow cytometry (FC). There is, however, variability in accuracy of diagnosis between laboratories, which may be explained by technical and sampling errors. Image cytometric DNA analysis (ICDA) is an optical technique allowing visualisation of abnormal nuclei and can be undertaken on archival tissue. Aim: To compare the accuracy of ICDA versus FC to detect DNA ploidy abnormalities. Methods: 48 formalin fixed paraffin embedded (FFPE) blocks from 35 patients were retrieved. 16 endoscopic mucosal resection specimens showed high grade dysplasia or intramucosal cancer. Of 32 esophagectomy specimens, 25 showed invasive adenocarcinoma and 7 blocks were cancer free margins used as controls. 40 μm sections were cut and deparaffinised with xylene, rehydrated through graduated ethanol, then denatured by protease XXIV. After filtration the nuclear suspension was separated for ICDA at UCL and FC at UW. ICDA was performed on Feulgen stained nuclear monolayers using a Fairfield Imaging system. Standardized methodology for FC was followed as previously described. Results: 44 samples were analyzed. 93% (41/44) were classified identically. 3 cases were disconcordant. One was near diploid aneuploid by FC but called diploid by IC due to a wider coefficient of

S-511

AGA Abstracts

AGA Abstracts

degree of steatosis. However, HRI is a better test to diagnose steatosis than ALT with better sensitivity and specificity.