also had lower degree of activation of left ventrolateral pre-frontal cortex (Mean change, R Vs NR: 12.1±14.1% Vs 19.8±25.5%, p=0.017). Conclusion: Treatment response in IBS patients is associated with reduced activation at left amydala, ACC and pre-frontal cortex. Reduced activation in these regions with no difference in perception thresholds to visceral stimuli suggests that treatment response is mediated by modulation in central autonomic and affective pathways rather than decreased visceral hypersensitivity.
T1865
PURPOSE: Improved spatial resolution of novel MR scanners allows HR data acquisition. In this presentation we created v-MRCPs of clinical cases and evaluate its efficacy. METHODS and MATERIALS: One hundred patients (22-81 year) scheduled for MRCP were allocated to this study. The source data (0.8mm, 96 slices) were acquired in coronal planes with a navigator gated respiratory-triggering using a 1.5T scanner. Using commercially available 3D software created movies of the sequential endoscopic images of the pancreaticobiliary(PB) system. Three orthogonal radiological planes of the source MRCP were integrated into the movie, which are showing the conic extent of the endoscopic viewing field. Qualitative assessment was performed with a five-point scoring by board-certified endoscopists. RESULTS: The thin-slice data acquisition allows non-cleaved rendering of v-MRCPs. Image quality was assessed as excellent in 96 cases. The divergence of the accessory pancreatic duct(aPD) from the main pancreatic duct(MPD) could be identified in 82 cases[Figure]. Sequential fly-through images could be created even in the distal segment of the all stenotic lesions (34 cases). No complicated segmentation was necessary because of high selectivity for the PB system of the hydrography data. The radiological planes are useful as external images like X-ray fluoroscopy. CONCLUSION: The practical creation of virtual cholangiopancreatoscopy could be performed by a combination of HR-MRCP data and interactive visualization. An accurate grasp of the inner surface of the PB system could be facilitated by vMRCPs without any invasive procedures.
T1863 Magnetic Resonance Enteroclysis in the Diagnosis of Complicated Celiac Disease Stijn J. Van Weyenberg, Abdulbaqi Al-Toma, Maarten A. Jacobs, Cornelis Van Kuijk, Chris J. Mulder, Jan Hein Van Waesberghe Introduction The majority of patients with celiac disease (CD) show clinical and histological improvement after introduction of a gluten free diet. However, some patients do not and suffer from so called refractory celiac disease (RCD). Subdivision of RCD can be made according to the amount of aberrant T-cells. Five-year survival of patients with RCD I is comparable to the general population, but is only 60% for patients with RCD II. This difference is mainly caused by the occurance of enteropathy associated T-cell lymphoma (EATL). T-cell flowcytometry is not widely available. Double balloon endoscopy to visualize RCD II associated mucosal lesions or EATL is also not widely available and is an invasive procedure. Therefore a minimal invasive method is needed to identify CD patients with a high risk of having or developing malignant complications. In other small bowel condidtions, conventional small bowel radiology is dissapointing and magnetic resonance imaging seems superior to computed tomography. We therefore investigated the use of magnetic resonance enteroclysis (MRE) in patients with suspected RCD. The purpose of this study was to evaluate CD-related abnormalities on MRE and to identify individual or grouped findings that differentiate between patients regarding (pre-) malignant complications Methods Retrospective evaluation of 29 consecutive MRE studies in patients with CD. All studies were perfomed after jejunal delivery of contrasting agent. Studies were evaluted by two experienced radiologist blinded to results of previous imaging and final diagnosis. Studies were evaluated for multiple known CD-related findings. Small bowel histology and T-cell flow cytometry were considered the standard of reference. Mean follow-up was 24.3 months Results Final diagnoses were as follows: uncomplicated CD (n=12), RCD I (n=5), RCD II (n=8), EATL (n=4). No single MRE-finding discriminated between the subtypes of CD or EATL. From the MRE-data a four-point MRE-score was derived, consisting of 1) the presence of ≤ 9 jejunal folds per 5 cm, 2) bowel wall thickening, 3) mesenterial fat infiltration, and 4) splenic volume <120 cm3. If the MRE score was ≥ 3 the results of MRE in discriminating high risk patients (RCD II or EATL) from low risk patient (uncomplicated CD or RCD I) were as follows: sensitivity 83.3%, specificity 94.1%, positive predictive value 90.9%, negative predictive value 88.9%, and overall diagnostic accuracy 89.7%. Conclusion RCD and EATL are rare conditions. Despite the limited sample size, our study suggest that MRE is a promising tool in identifying CD patients with a high risk of developing lymphoma. T1864 MRI Findings of the Pelvic Floor Correlates with Function Kumaran Thiruppathy, Stuart A. Taylor, Dave R. Chatoor, Richard Cohen, Anton V. Emmanuel
v-MRCPs image and the conic extent of the endoscopic viewing field(above)
Introduction: The role of the levator ani, in particular puborectalis, in the aetiology of faecal incontinence (FI) remains uncertain. Levator ani not only forms the pelvic floor, but the puborectalis component creates the anorectal angle and merges with the upper fibres of the external anal sphincters. Rapidly advancing imaging techniques have allowed us to identify abnormalities of the pelvic floor to which anorectal symptoms have been tentatively attributed. Pelvic dynamometry allows us to assess physical strength of the pelvic floor. The aim of the study was to establish the relationship between symptoms, dynamometry and MRI findings. Methods: 17 female patients (mean age 43, range 27-56) with urge faecal incontinence were studied. All completed a Wexner incontinence score, and underwent pelvic floor dynamometery and static and dynamic MRI of the pelvic floor. From the static MRI, the thinnest and thickest part of the (i) superficial, (ii) bulk of the external sphincter and (iii) puborectalis were identified; hereafter, for brevity, the thinnest measurements of each of these muscles are used for analysis. The dynamic scan allowed anorectal angle to be measured on valsalva. Results: Patients had a mean Wexner score of 12 (+/-6.5). Mean thickness of the superficial part of external sphincter (0.23cm +/-0.12) did not correlate with symptoms (p>0.74) but did correlate with pelvic floor strength on dynamometry (r=0.58, p<0.02). External sphincter bulk thickness (mean 0.33 cm+/-0.1) had a negative correlation with FI scores (r=-0.53, p<0.03) and a positive correlation with pelvic floor strength (r=0.59, p<0.02). Puborectalis thickness (mean 0.47cm +/-0.1) had no relationship with FI scores (r=-0.28, p=0.27), but did correlate with pelvic floor strength (r=0.81, p<0.0003). Dynamic MRI revealed that anorectal angle correlated with FI scores (r =0.55 p<0.02) and was negatively correlated with pelvic floor strength (r=-0.62 p<0.01). Conclusion: Anatomical variations in thickness of the superficial and bulk of external sphincter and puborectalis correlated with pelvic floor strength. External sphincter bulk and anorectal angle did show correlations with symptoms. In summary, we have objectively quantified the relationship between the structure and strength of puborectalis in faecally incontinent patients. Secondly, we have shown that the thickness of the bulk of the external sphincter contributes to resisting urge incontinence, whilst the role of puborectalis may be in maintain the anorectal angle.
T1866 Factors Predictive of Early Neoplasia in Barrett's Lesions Identified with Autofluorescence Imaging: A Stepwise Multi-Centre Structured Assessment Wouter L. Curvers, Rajvinder Singh, Michael B. Wallace, Louis-Michel Wong Kee Song, Krish Ragunath, Herbert C. Wolfsen, Fiebo J. ten Kate, Paul Fockens, Jacques J. Bergman Introduction Autofluorescence imaging (AFI) is a novel imaging technique that may improve the detection of early neoplasia (EN) in Barrett's esophagus (BE). However, AFI is associated with a high false positive (FP)-rate ranging from 40-81%. The aim of this study was to perform a stepwise structured assessment by endoscopists from 4 international centres with experience in AFI to evaluate which features may predict the presence of EN in AFI positive areas. Methods Orientation phase: AFI and WLE images of 10 areas with corresponding histology were evaluated by two AFI-experts in an unblinded manner for features that might predict the presence of EN. A scoring list with potentially relevant factors was then developed. Learning phase: 82 areas of BE were selected from our prospective imaging database based on image quality and correlation of the AFI and WLE images. These areas were subsequently evaluated by the same two experts, now blinded for histology. In case of disagreement a decision was reached by consensus. Based on the results, features that were significantly associated with the presence of EN were identified and subsequently incorporated in a validation scoring list. Validation phase: The same 82 areas were evaluated using the validation scoring list by 5 endoscopists with experience in AFI, again blinded for the histology results. Results Learning phase: 63 areas were scored as AFI-positive and 26 of these contained non- dysplastic BE (FP-rate 41%). AFI-positive areas had a significantly higher rate of violet/ brown color versus green color for AFI-negative areas. In multivariate analysis we found that the best model for predicting EN in AFI-positive areas included the following factors: opaque AFI-intensity (p<0.01), proximity of the gastric folds > 1cm (p=0.03) and subtle abnormalities of the mucosa on WLE (p<0.01). Validation phase: 308 of 410 evaluations (5x82 areas) were scored as AFI-positive of which 130 had non-dysplastic BE (FP-rate 42%). Kappa values for inter-observer agreement were 0.48 for AFI-intensity, 0.51 for proximity of gastric folds and 0.56 for appearance of the mucosa on WLE. In multivariate analysis we confirmed that the factors of the model were independent predictive factors of EN containing AFI-positive areas: Opaque AFI intensity (p=0.01), proximity of the gastric folds > 1cm (p<0.01) and subtle abnormalities of the mucosa on WLE (p<0.01). Conclusion Violet/ brown areas identified during AFI as abnormal are more likely to contain EN if they have an opaque AFI intensity, are not located close to the gastric folds and show subtle abnormalities on subsequent inspection with white light endoscopy.
A-579
AGA Abstracts
AGA Abstracts
Virtual MR-Cholangiopancreatoscopy(V-Mrcps) By a Combination of HighResolution(Hr) Data Acquisition and Interactive Visualization Tetsuya Yamagishi, Takashi Kawai