Su1272 Percutaneous SPYGLASS™ Cholangioscopy for Assessment of Biliary Strictures

Su1271

Background: ERCP is challenging in the Roux-en-Y gastric bypass (RYGB) patient. ERCP can be performed using device-assisted enteroscopy (DAE), although the major papilla is frequently inaccessible, and ERCP with a forward-viewing enteroscope can be daunting. Laparoscopic-assisted ERCP is challenging and carries morbidity associated with surgery. Single-session transgastric ERCP using DAE to access the remnant has been reported, although this approach is time consuming and success relies on the ability of DAE to access the remnant. Aim: To assess the technical feasibility of EUS-guided percutaneous gastric access for single-session ERCP. Methods: This was a prospective case series. Patients with history of RYGB and failed DAE for ERCP were included. Demographics, procedure details, and outcomes were recorded. A standardized management strategy was employed. A linear echoendoscope was advanced to the gastric pouch and a 22-gauge needle was used to access the gastric remnant. Contrast was injected with fluoroscopic confirmation of positioning. Carbon dioxide was insufflated via the EUS needle and gastric remnant distension was visualized fluoroscopically. External markers were placed to indicate the location of the remnant. The abdominal wall was prepped for gastrostomy. A T-tag needle with contrastfilled syringe was introduced into the gastric remnant using a bubble test and contrast injection to confirm position. Four T-tags were placed to affix the remnant to the abdominal wall. An introducer needle was inserted through the center of the affixed area. A 0.035" guidewire was advanced into the remnant through the needle, which was then withdrawn. A scalpel was used to make a 1 cm incision along the wire. A fully covered esophageal stent was advanced over the wire, angled toward the pylorus, and deployed under fluoroscopic visualization. A dilation balloon was inserted into the stent and inflated to 18 mm. A standard duodenoscope was inserted through the stent and advanced to the major papilla to carry out ERCP. After ERCP, a G-tube was placed into the gastrostomy, and the stent removed. Results: Five patients underwent EUS-guided transgastric access for ERCP. In all cases, access and single-session ERCP were successful. Mean time for remnant access was 24.8 ±3.9 minutes. Mean time for closure after ERCP was 13.2 ±1.7 minutes. Two patients required suture placement at the G-tube site. One patient required repeat ERCP, which was carried out successfully via the G-tube tract. There were no major access, ERCP, or G-tube related complications. G-tubes were removed 4 to 7 weeks after placement, with removal of residual T-tags as needed. Conclusion: EUS-guided percutaneous gastric access with single-session ERCP via a transgastric esophageal stent appears safe, effective, and timeefficient. Additionally, it allows access for early repeat ERCP as needed.

Su1270 School-Based Study of Recurrent Abdominal Pain (RAP) in Siberian Adolescents: Prevalence, Structure According Questionnaire on Pediatric Gastrointestinal Symptoms ROME III Version (QPGS-RIII) and Association With Psychologic Problems According Strengths and Difficulties Questionnaire (SDQ) Sergey Tereshchenko, Arina Vityutneva, Nina Gorbacheva Several studies have used the QPGS-RIII criteria to estimate the rates of various functional gastrointestinal disorders (FGIDs) among children and adolescents with primary symptoms of RAP. In a population based study, children with RAP had significantly more psychologic problems than children without RAP. The majority of those studies have been performed in regions with low prevalence rates of childhood Helicobacter pylori (Hp) infection, however, such data are limited in Russia, where prevalence of Hp infection in adolescents is high (25-67 %). METHODS: 459 urban (Krasnoyarsk) adolescents aged 12-18 were screened for RAP (criteria were as follow: (1) "More than two episodes of RAP per month in the last two months" OR (2) "Abdominal pain intensity according 7-items Likert-type pain scale ≥ 4"). 75 screen positive adolescents were tested with QPGS-RIII and self report version of SDQ questionnaires [1, 2]. Two-tailed exact Fisher and Kruskal-Wallis tests were used. RESULTS: The prevalence of RAP according assigned screening criteria was 16.3 % (75/459) with higher level in girls, than in boys (21.9 % vs 11.0 %, p=0.002) and in 12-14 age group, than in 15-18 age group (27.4 % vs 11.7 %, p<0.001). The prevalence of RAP according pediatric Rome III criteria for non-cyclic pain-related FGIDs (RAP frequency ≥ 1 a week with symptoms duration ≥ 2 months) was 5.4 % (25/459) with the same gender-age tendencies but without statistical significance (p>0.1). The rates of FGIDs in accordance to QPGS-RIII scoring instructions were as follows: functional dyspepsia - 0.4 % (2/459), irritable bowel syndrome - 2.6 % (12/459), abdominal migraine (AM) - 2.0 % (9/459), functional abdominal pain - 1.3 % (6/459), functional abdominal pain syndrome - 0.65 % (3/459). The total FGIDs prevalence (any of them) was 6.5 % (30/459) and non-cyclic pain-related FGIDs rate (with exception of AM) was 4.6 % (21/459). Functional constipation was registered in 2.8 % (13/459) screen positive adolescents. We have found a strong positive association between RAP frequency and SDQ total difficulties score (p<0.001), hyperactivity score (p= 0.065) and, especially, emotional symptoms score (p<0.001, fig. 1). No differences have been found in conduct problems, peer problem, and prosocial behaviour scores. CONCLUSION: Thus, prevalence of RAP and FD in Siberian adolescents is not higher than ones previously reported in regions with low prevalence rates of childhood Hp infection. We suppose that Hp do not play causative role in adolescent's RAP in majority of cases and other factors can be more important, especially emotional problems. REFERENCES: 1. Goodman R. // Journal of the American Academy of Child and Adolescent Psychiatry. 2001. - Vol. 40, N 11. - P. 1337-1345. 2. Rasquin A., Di Lorenzo C., Forbes D. et al. // Gastroenterology. - 2006. - Vol. 130, N 5. - P. 1527-1537.

Su1272 Percutaneous SPYGLASS™ Cholangioscopy for Assessment of Biliary Strictures Pernilla M. D'Souza, Aducio Thiesen, Richard J. Owen, Philippe Sarlieve, Dermot McNally, Gurpal Sandha Background: Assessment of biliary strictures with per oral cholangioscopy is limited in patients with roux-en-Y hepaticojejunostomy and in those where ERCP cannulation fails. Percutaneous trans-hepatic cholangioscopy (PTC) is beneficial in this clinical scenario. Aim: Our aim is to assess the utility of percutaneous trans-hepatic SpyGlass™ cholangioscopy (PTSC) for biliary strictures in patients where conventional per oral cholangioscopy is not feasible. Methods:All patients that underwent PTSC are reviewed with respect to operating characteristics and procedural outcomes. At the time of referral, all patients had an 8.5 French (Fr) PTC drain in situ. Percutaneous brushings from the area of interest were either unsatisfactory or non-diagnostic. In preparation for PTSC, the percutaneous tract was serially dilated and a 12 Fr sheath to accommodate the SpyGlass™ system (10 Fr) was placed immediately prior to the procedure. Results: Three patients underwent PTSC (Table 1). Two operators were required to operate the SpyGlass™ cholangioscope, one to advance the cholangioscope while anchoring the 12 Fr sheath and the other to operate the controls and obtain biopsies. Maneuverability was slightly more challenging than with per oral cholangioscopy. SpyBite® biopsies were obtained from area of interest and diagnosis achieved in all 3 cases (Table 1). There were no complications. Conclusion:PTSC for the assessment of biliary strictures is feasible and safe resulting in high diagnostic accuracy. However, it is a two-operator procedure and is somewhat more challenging than conventional per oral cholangioscopy. Table 1: Summary of patients undergoing PTSC

Figure 1. Emotional symptoms scores according Strengths and Difficulties Questionnaire in adolescents with different RAP frequency.

S-421

AGA Abstracts

AGA Abstracts

EUS-Guided Percutaneous Gastric Remnant Access for Single-Session ERCP in the Gastric Bypass Patient Nitin Kumar, Marvin Ryou, Christopher C. Thompson

AGA Abstracts

colitic mice and the DSS-induced colitis model, suggesting that miRNA signatures could be a powerful tool for discriminating among various types of intestinal inflammation (e.g., chronic vs. acute and CD vs. UC). These murine data may help us establish miRNA patterns that can be used to assess the diagnosis, prognosis, and therapeutic responses of IBD patients. Conclusion: We used a mouse model to establish a putative miRNA signature of colitis, and established that it could be a useful tool for following the evolution of intestinal inflammation in mice. Future work is warranted to adapt this concept to human serum samples and identify IBD specific miRNA patterns in human patients. Su1275 Circulating Protein Signatures As Biomarkers for Inflammatory Bowel Disease Management Emilie Viennois, Mark Baker, Bo Xiao, Saravanan Ayyadurai, Hamed Laroui, Didier Merlin Background and Aims: Inflammatory bowel diseases (IBDs), principally ulcerative colitis (UC) and Crohn's disease (CD), are chronic and progressive inflammatory disorders of the gastrointestinal tract. In IBD, protein serological biomarkers could be powerful tools for assessing disease activity, performing early-stage diagnosis and managing treatment. The interleukin-10 knock-out (IL10-/-) mouse is a powerful model that develops a time-dependent IBD that predominates in the colon and shares histopathological features with human UC. Here, we used IL10-/- mice to investigate the protein expression profiles in circulating blood at different time points and correlate them with development of the disease. Methods: Serum samples were collected from IL10-/- mice once a week for 15 weeks after weaning. Circulating protein profiles were investigated using 2-dimensional differential gel electrophoresis (2D-DiGE), which allows the identification, quantification and statistical analysis of differentially accumulated proteins. The resulting putative biomarkers were validated in serum samples using ELISA. Results: The 2D-DiGE analysis allowed us to identify proteins that underwent differential accumulation during the development of intestinal inflammation, such as hemopexin, transferrin, haptoglobin and peroxiredoxin 2, which accumulated as inflammation progressed, and alpha-1beta-glycoprotein, which decreased during the establishment of colitis. Conclusions and perspectives: Based on the murine data, we can establish a protein signature that suggests colonic inflammation. A large collection of human serum samples containing UC, CD and control patients is currently being used to validate this combination of proteins as a biomarker signature for intestinal inflammation, and to assess its potential use as a management tool for IBD patients.

Su1273 Spatial and Temporal Stability of Paneth Cell Phenotypes in Crohn's Disease: Implications for Prognostic Cellular Biomarker Development Ta-Chiang Liu, Feng Gao, Dermot P. McGovern, Thaddeus S. Stappenbeck Background: Paneth cells are an intestinal epithelial cell lineage that modulate the microbiome through the production and secretion of anti-microbial peptides. We recently demonstrated that morphologic defects of Paneth cell secretory granules (herein termed "Paneth cell phenotype") in the terminal ileum correlated with poorer clinical outcome in Crohn's Disease (CD) patients. We used uninvolved tissues from surgical resections for these studies. To determine if Paneth cell phenotypes can be utilized as a prognostic biomarker for CD we need to validate its use in a wider clinical setting. Methods: Paneth cell phenotypes were determined using lysozyme immunofluorescence stained sections. Each Paneth cell was classified as normal or abnormal (disordered, diminished, diffuse, or excluded granule). We first analyzed paired involved and uninvolved resection specimens (n=46 cases) to test the relationship with disease state. We next analyzed cases with multiple ileal resections over time (n=36) to test if phenotypes were stable over time. To determine the minimum number of crypts needed for accurate phenotyping, we first generated 'virtual biopsies' by randomly selecting a specific number of crypts (range: 10 to 60 crypts per case) from surgical resection specimens that each had a minimum of 250 crypts analyzed (n=85 cases) and then used mathematical modeling to determine the minimum number of crypts needed for accurate Paneth cell phenotyping. These results were supported by analyzing cases that had resection and biopsy specimens collected within 1 year (n=26 cases). Correlation analysis was performed using Pearson correlation (Graphpad Prizm software). P < 0.05 was considered to be significant. Results: We found that the Paneth cell phenotypes in areas uninvolved by active/chronic inflammation strongly correlated with that in involved areas in the same patient (P < 0.0001), suggesting that disease activity state does not affect the phenotype. Importantly, in the same cohort of patients, the Paneth cell phenotype remained stable over time (P < 0.0001). Using mathematical modeling, we showed that a minimum of 40 crypts were required to generate accurate Paneth cell phenotyping results, demonstrating that a typical biopsy sample could be used for this assay. As further validation of the ability to use biopsies to accurately analyze Paneth cell phenotypes, we found that Paneth cell phenotypes were significantly correlated in biopsy specimens and surgical resection specimens collected from the same patient within a 1-year period (P = 0.0004). Conclusions: Accurate Paneth cell phenotypes can be assessed using biopsy materials, thus allowing further clinical applications of this novel stratification platform.

Su1276 αE Integrin Expression As a Predictive Biomarker for Induction of Clinical Remission by Etrolizumab: Analysis of a Phase II Trial in Moderate-to-Severely Active Ulcerative Colitis Mary E. Keir, Gaik Wei Tew, Diana Luca, Jeffrey Eastham-Anderson, Lauri Diehl, Jackson G. Egen, Severine Vermeire, John C. Mansfield, Christopher A. Lamb, Brian G. Feagan, Julian Panes, Daniel C. Baumgart, Stefan Schreiber, Iris Dotan, William Sandborn, Gert De Hertogh, John A. Kirby, Gert A. Van Assche, Paul J. Rutgeerts, Sharon O'Byrne Background: Etrolizumab is a humanized monoclonal antibody to the β7 integrin subunit that blocks α4β7:MAdCAM-1 and αEβ7:E-cadherin interactions and is effective at inducing remission in patients with moderate to severe ulcerative colitis (UC). Methods: Patients enrolled in the Phase 2 study of etrolizumab in UC were assessed for baseline biopsy gene expression of MAdCAM-1, αE and α4 integrin by qPCR (n=106) as well as baseline αE+ cells by immunohistochemistry (n=76). Baseline measures were evaluated for enrichment using remission at week 10, defined as a total Mayo clinic score (MCS) of ≤ 2 with no individual subscore > 1. Remission at week 14-16 in patients enrolled in an open-label extension of the study was evaluated using partial MCS of ≤ 2 with no individual subscore > 1. Results: αE gene expression and αE+ cells were log-normally distributed in baseline biopsies. No difference in baseline clinical characteristics was observed between αE high and αE low groups. Enrichment in response to treatment with 100 mg/dose etrolizumab as measured by remission was observed in patients with higher than median levels of αE gene expression at screening (38% αE high vs. 13% αE low). No enrichment was observed in α4 high patients (27% α4 high vs. 24% α4 low) or MAdCAM-1 high patients (27% MAdCAM-1 high vs. 24% MAdCAM-1 low). Enrichment in response to treatment as measured by remission was observed in TNF antagonist Naive patients with high levels of αE gene expression at screening (67% αE high vs. 17% αE low). Similarly, enrichment of remission in response to treatment was observed in patients with higher than median levels of αE+ cells per total cells (all comers, 50% αE high vs. 7% αE low; TNF antagonist Naive patients, 67% αE+ high vs. 25% αE+ low). In data collected as part of an open-label extension of the Phase 2 study, patients with an inadequate response to TNF antagonists (TNF-IR) with high levels of αE gene expression in baseline biopsies that did not respond at the primary endpoint showed higher remission (18% αE high vs. 0% αE low) and response (53% αE high vs. 8% αE low) to continued treatment at a week 14-16 time point. Enrichment of remission (22% αE+ high vs. 0% αE+ low) and response (56% αE+ high vs. 17% αE+ low) was also observed at week 14-16 in TNF-IR patients with baseline high levels of αE+ cells per total cells. Conclusion: These results suggest that higher than median levels of αE gene expression and/or αE+ cells per total cells in the inflamed colon are associated with increased clinical benefit of etrolizumab treatment in patients. Thus αE gene expression or αE+ cells in colonic biopsies may be predictive biomarkers to identify UC patients most likely to benefit from etrolizumab treatment.

Su1274 Longitudinal Study of Circulating miRNA Biomarkers in Inflammatory Bowel Disease Emilie Viennois, Mark Baker, Bo Xiao, Saravanan Ayyadurai, Hamed Laroui, Didier Merlin Background: Inflammatory bowel diseases (IBDs), principally ulcerative colitis (UC) and Crohn's disease (CD), are chronic and progressive inflammatory disorders of the gastrointestinal tract. The current serum biomarkers for IBD, which include lipocalin2 and calprotectin, are limited by low specificity and poor predictive power. Serological assessments of IBD activity and early-stage diagnosis could be markedly improved through the use of a noninvasive serological biomarker signature, such as miRNA. MiRNAs represent a class of naturally occurring small non-coding RNAs whose plasma concentrations can emulate tissueand disease-specific expression profiles. The interleukin-10 knockout (IL10-/-) mouse is a well-described model that reproducibly develops a time-dependent IBD that predominates in the colon and shares histopathological features with human UC. This model can be used to investigate miRNA expression profiles in circulating blood. Here, we hypothesized that the time-course of the disease and its lesions can be correlated with specific miRNA profiles. Methods: Serum samples were collected from wild-type (WT) and IL10-/- mice once a week for 15 weeks, and miRNA profiles were analyzed using high-throughput technology, allowing us to study small changes during the progression of intestinal inflammation. Results: Among the 104 miRNAs tested, mmu-miR-29b, -122, -150, -192, -194, -146a and -375 were found to increase with disease progression, while mmu-miR-335-5p, -148a, -152, -140, -331-3p, -195, -140* and -199b were observed to decrease. Statistical analyses revealed that among the miRNAs that were differentially expressed in colitic versus non-colitic mice, six could be used as a profiling set that could very efficiently determine the inflammation status of the intestinal mucosa. Moreover, the miRNA profiles were not identical between IL10-/-

AGA Abstracts

S-422