Search

Background and study aims  Small bowel ulcerations are efficiently detected with deep learning techniques, whereas the ability to diagnose Crohn's disease (CD) in the colon with it is unknown. This study examined the ability of a deep learning framework to detect CD lesions with pan-enteric capsule endoscopy (CE) and classify lesions of different severity. Patients and methods  CEs from patients with suspected or known CD were included in the analysis. Two experienced gastroenterologists classified anonymized images into normal mucosa, non-ulcerated inflammation, aphthous ulceration, ulcer, or fissure/extensive ulceration. An automated framework incorporating multiple ResNet-50 architectures was trained. To improve its robustness and ability to characterize lesions, image processing methods focused on texture enhancement were employed. Results  A total of 7744 images from 38 patients with CD were collected (small bowel 4972, colon 2772) of which 2748 contained at least one ulceration (small bowel 1857, colon 891). With a patient-dependent split of images for training, validation, and testing, ulcerations were diagnosed with a sensitivity, specificity, and diagnostic accuracy of 95.7 % (CI 93.4-97.4), 99.8 % (CI 99.2-100), and 98.4 % (CI 97.6-99.0), respectively. The diagnostic accuracy was 98.5 % (CI 97.5-99.2) for the small bowel and 98.1 % (CI 96.3-99.2) for the colon. Ulcerations of different severities were classified with substantial agreement (κ = 0.72). Conclusions  Our proposed framework is in excellent agreement with the clinical standard, and diagnostic accuracies are equally high for the small bowel and colon. Deep learning approaches have a great potential to help clinicians detect, localize, and determine the severity of CD with pan-enteric CE., Competing Interests: Competing interests The authors declare that they have no conflict of interest., (The Author(s). This is an open access article published by Thieme under the terms of the Creative Commons Attribution-NonDerivative-NonCommercial License, permitting copying and reproduction so long as the original work is given appropriate credit. Contents may not be used for commercial purposes, or adapted, remixed, transformed or built upon. (https://creativecommons.org/licenses/by-nc-nd/4.0/).)

Aim: To investigate whether immune mediated diseases (IMD) are more frequent in patients with inflammatory bowel disease (IBD)., Methods: In this population based registry study, a total of 47325 patients with IBD were alive and registered in the Danish National Patient Registry on December 16, 2013. Controls were randomly selected from the Danish Civil Registration System (CRS) and matched for sex, age, and municipality. We used ICD 10 codes to identify the diagnoses of the included patients. The IBD population was divided into three subgroups: Ulcerative colitis (UC), Crohn's disease (CD) and Both the latter referring to those registered with both diagnoses. Subsequently, odds-ratios (OR) and 95%CI were obtained separately for each group and their respective controls. The use of Bonferoni post-test correction adjusted the significance level to P < 0.00125. P -values were estimated using Fisher's exact test., Results: There were significantly more women than men in the registry, and a greater percentage of comorbidity in the IBD groups ( P < 0.05). Twenty different IMDs were all significantly more frequent in the IBD group. Sixteen were associated with UC versus twelve with CD. In both UC and CD ORs were significantly increased ( P < 0.00125) for primary sclerosing cholangitis (PSC), celiac disease, type 1 diabetes (T1D), sarcoidosis, asthma, iridocyclitis, psoriasis, pyoderma gangrenosum, rheumatoid arthritis, and ankylosing spondylitis. Restricted to UC ( P < 0.00125) were autoimmune hepatitis, primary biliary cholangitis, Grave's disease, polymyalgia rheumatica, temporal arteritis , and atrophic gastritis. Restricted to CD ( P < 0.00125) were psoriatic arthritis and episcleritis. Restricted to women with UC ( P < 0.00125) were atrophic gastritis, rheumatoid arthritis, temporal arteritis, and polymyalgia rheumatica. Restricted to women with CD were episcleritis, rheumatoid arthritis, and psoriatic arthritis. The only disease restricted to men ( P < 0.00125) was sarcoidosis., Conclusion: Immune mediated diseases were significantly more frequent in patients with IBD. Our results strengthen the hypothesis that some IMDs and IBD may have overlapping pathogenic pathways., Competing Interests: Conflict-of-interest statement: The authors report no conflict of interest.

Background and Aim: Capsule endoscopy (CE) has a high sensitivity for diagnosing small bowel Crohn's disease, but video analysis is time-consuming. The quick view (qv) function is an effective tool to reduce time consumption. The aim of this study was to determine the rate of missed small bowel ulcerations with qv-CE compared with standard view and the diagnostic accuracy of qv-CE in suspected Crohn's disease., Methods: This study consisted of two parts: (i) 12 small bowel segments with Crohn's disease of varying severity were selected for a detailed analysis of the number and type of lesions visualized with CE and qv-CE, and (ii) a blinded study of the diagnostic accuracy of qv-CE including 40 patients with suspected Crohn's disease. Ileocolonoscopy and CE served as gold standard., Results: Part 1: CE visualized 171 ulcerations compared with 102 detected with qv-CE (miss rate 40%, P = 0.02). Part 2: qv-CE identified 15 of 16 patients with small bowel Crohn's disease corresponding to a 94% sensitivity, and overall, 39 out of 40 patients were classified correct (diagnostic accuracy 98%). Qv-CE was false negative in one patient because of a leap of 3 min and 20 s in the terminal ileum. Reading times varied from 5 to 18 min (median 10)., Conclusion: Despite a significant number of missed lesions, qv-CE is a safe and time-reducing method for diagnosing small bowel Crohn's disease. To avoid false negative cases, we recommend viewing the terminal ileum in standard view., (© 2013 Journal of Gastroenterology and Hepatology Foundation and Wiley Publishing Asia Pty Ltd.)

It has recently become clear that the transcriptional output of the human genome is far more abundant than previously anticipated, with the vast majority of transcripts not coding for protein. Utilizing whole-genome tiling arrays, we analyzed the transcription across the entire genome in both normal human bronchial epithelial cells (NHBE) and NHBE cells exposed to the tobacco carcinogen NNK. Our efforts focused on the characterization of non-coding transcripts that were greater than 300 nucleotides in length and whose expression was increased in response to NNK. We identified 12 Long Stress-Induced Non-coding Transcripts that we term LSINCTs. Northern blot analysis revealed that these transcripts were larger than predicted from the tiling array data. Quantitative real-time RT-PCR performed across a panel of normal cell lines indicates that these transcripts are more abundantly expressed in rapidly growing tissues or in tissues that are more prone to cellular stress. These transcripts that have increased expression after exposure to NNK also had increased expression in a number of lung cancer cell lines and also in many breast cancer cell lines. Collectively, our results identified a new class of long stress responsive non-coding transcripts, LSINCTs, which have increased expression in response to DNA damage induced by NNK. LSINCTs interestingly also have increased expression in a number of cancer-derived cell lines, indicating that the expression is increased in both, correlating cellular stress and cancer., (Copyright 2010 Elsevier Inc. All rights reserved.)

Recent studies with tiling arrays have revealed more genomic transcription than previously anticipated. Whole new groups of non-coding transcripts (NCTs) have been detected. Some of these NCTs, including miRNAs, can regulate gene expression. To date, most known NCTs studied have been relatively short, but several important regulatory NCTs, including XIST, MALAT-1, BC1 and BC200, are considerably larger in length and represent a novel class of long, non-coding RNA species. Whole-genome tiling arrays were utilized to identify novel long NCTs across the entire human genome. Our results have identified a new group of long (>400 nt), abundantly expressed NCTs and have found that a subset of these are also highly evolutionarily conserved. In this report, we have begun to characterize 15 long, conserved NCTs. Quantitative real-time RT-PCR was used to analyze their expression in different normal human tissue and also in breast and ovarian cancers. We found altered expression of many of these NCTs in both cancer types. In addition, several of these NCTs have consistent mutations when sequences of normal samples were compared with a panel of cancer-derived cell lines. One NCT was found to be consistently mutated in a panel of endometrial cancers compared with matched normal blood. These NCTs were among the most abundantly expressed transcripts detected. There are probably many long, conserved NCTs, albeit with lower levels of expression. Although the function of these NCTs is currently unknown, our study indicates that they may play an important function in both normal cells and in cancer development.