Capsule Endoscopy for Crohn's Disease: Current Status of Diagnosis and Management

Influence of training and expertise on deep neural network attention and human attention during a medical image classification task

In many different domains, experts can make complex decisions after glancing very briefly at an image. However, the perceptual mechanisms underlying expert performance are still largely unknown. Recently, several machine learning algorithms have been shown to outperform human experts in specific tasks. But these algorithms often behave as black boxes and their information processing pipeline remains unknown. This lack of transparency and interpretability is highly problematic in applications involving human lives, such as health care. One way to "open the black box" is to compute an artificial attention map from the model, which highlights the pixels of the input image that contributed the most to the model decision. In this work, we directly compare human visual attention to machine visual attention when performing the same visual task. We have designed a medical diagnosis task involving the detection of lesions in small bowel endoscopic images. We collected eye movements from novices and gastroenterologist experts while they classified medical images according to their relevance for Crohn's disease diagnosis. We trained three state-of-the-art deep learning models on our carefully labeled dataset. Both humans and machine performed the same task. We extracted artificial attention with six different post hoc methods. We show that the model attention maps are significantly closer to human expert attention maps than to novices', especially for pathological images. As the model gets trained and its performance gets closer to the human experts, the similarity between model and human attention increases. Through the understanding of the similarities between the visual decision-making process of human experts and deep neural networks, we hope to inform both the training of new doctors and the architecture of new algorithms.

Current and Emerging Approaches to the Diagnosis and Treatment of Crohn's Disease Strictures

The management and understanding of Crohn's disease (CD) continues to evolve quickly. Intestinal strictures were previously thought to be an inevitable result of irreversible fibrosis caused by chronic inflammation. However, increased understanding of the dynamic nature of strictures and of the pathophysiology of this condition has highlighted emerging targets for potential treatment. In the diagnosis of strictures, a distinction must be made between inflammatory and fibrotic types, as the former may respond to medical therapy. Emerging technologies, such as dual-energy computed tomography enterography and iodine density, have allowed more accurate characterization of strictures. Surgical and endoscopic treatment remains the mainstay for fibrotic strictures, but developments in systemic and intralesional biologic therapy have shown efficacy. This article reviews the pathophysiology of this debilitating complication of CD as well as current and emerging diagnostics and treatments.

Artificial Intelligence Research and Development for Application in Video Capsule Endoscopy

Artificial intelligence (AI) research for medical applications has expanded quickly. Advancements in computer processing now allow for the development of complex neural network architectures (eg, convolutional neural networks) that are capable of extracting and learning complex features from massive data sets, including large image databases. Gastroenterology and endoscopy are well suited for AI research. Video capsule endoscopy is an ideal platform for AI model research given the large amount of data produced by each capsule examination and the annotated databases that are already available. Studies have demonstrated high performance for applications of capsule-based AI models developed for various pathologic conditions.

Indian guidelines on imaging of the small intestine in Crohn's disease: A joint Indian Society of Gastroenterology and Indian Radiology and Imaging Association consensus statement

The Indian Society of Gastroenterology (ISG) Task Force on Inflammatory Bowel Disease and the Indian Radiological and Imaging Association (IRIA) developed combined ISG-IRIA evidence-based best-practice guidelines for imaging of the small intestine in patients suspected to have or having Crohn's disease. The 29 consensus statements, developed through a modified Delphi process, are intended to serve as reference for teaching, clinical practice, and research.

A Small Bowel Ulcer due to Clopidogrel with Cytomegalovirus Enteritis Diagnosed by Capsule and Double-Balloon Endoscopy

We report the first case of small bowel ulcers due to clopidogrel in a 74-year-old man. He presented with diarrhea and melena after having been taking low-dose aspirin (LDA) and clopidogrel. There was no evidence of bleeding in the stomach, duodenum, or colon. Capsule endoscopy showed multiple ulcers and erosions in the small intestine. Double-balloon endoscopy revealed multiple ulcers throughout the ileum. Examination of the biopsy specimen showed cytomegalovirus infection. His LDA was discontinued and he was prescribed ganciclovir. However, the small bowel ulcers were aggravated. Therefore, clopidogrel was discontinued. The small bowel ulcers subsequently healed completely, forming scars.

Imaging of the small intestine in Crohn's disease: Joint position statement of the Indian Society of Gastroenterology and Indian Radiological and Imaging Association

The Indian Society of Gastroenterology (ISG) Task Force on Inflammatory Bowel Disease and the Indian Radiological and Imaging Association (IRIA) developed combined ISG-IRIA evidence-based best-practice guidelines for imaging of the small intestine in patients with suspected or known Crohn's disease. These 29 position statements, developed through a modified Delphi process, are intended to serve as reference for teaching, clinical practice, and research.

Relationships of capsule endoscopy Lewis score with clinical disease activity indices, C-reactive protein, and small bowel transit time in pediatric and adult patients with small bowel Crohn's disease

Relationships between the capsule endoscopy Lewis score (LS) and clinical disease activity indices and C-reactive protein (CRP) are controversial in adult patients with Crohn's disease (CD). Also, data on pediatric patients are relatively less. However, correlation between LS and small bowel transit time (SBTT) remains investigational. The aim of the present study was to explore the correlations between LS and clinical disease activity indices, CRP, SBTT in pediatric, and adult patients with small bowel CD.Retrospective, single-center study on consecutive inpatients with established small bowel CD was conducted. The clinical disease activity index was determined using the abbreviated Pediatric Crohn's Disease Activity Index (aPCDAI) in patients aged <18 years and the Harvey-Bradshaw Simple Index (HBI) in adults. Spearman's rank correlation coefficient was used to assess the correlations of LS with aPCDAI, HBI, CRP, and SBTT, respectively.150 patients were enrolled (30 children and adolescents). In pediatric patients, correlations between LS and aPCDAI, CRP were moderate (r1 = 0.413; r2 = 0.379; P1 = .023; P2 = .044). There was no correlation between LS and SBTT (r = -0.029; P = .88). In adults, weak correlations were found between LS and HBI, SBTT (r1 = 0.213; r2 = 0.237; P1 = .019; P2 = .009). Correlation between LS and CRP was moderate (r = 0.326; P < .001). Strong correlations were found between CRP and HBI, aPCDAI (r1 = 0.522; r2 = 0.650; P < .001). The follow-up patients were all in clinical remission after treatment within 4 months, whereas only a minority reached mucosal healing. HBI, aPCDAI, CRP, and LS in all patients were reduced after treatment, whereas difference in CRP in pediatric patients and difference in LS in adults between baseline and follow-up were not found to be statistically significant. Also, the average SBTT at baseline was not found to be different from that at follow-up in all patients.The role of capsule endoscopy should be emphasized both in pediatric and adult patients with small bowel CD. Furthermore, the small bowel transit time may not be affected by the grade of small intestinal inflammation.

Accuracy of computed tomographic features in differentiating intestinal tuberculosis from Crohn's disease: a systematic review with meta-analysis

Abdominal computed tomography (CT) can noninvasively image the entire gastrointestinal tract and assess extraintestinal features that are important in differentiating Crohn's disease (CD) and intestinal tuberculosis (ITB). The present meta-analysis pooled the results of all studies on the role of CT abdomen in differentiating between CD and ITB. We searched PubMed and Embase for all publications in English that analyzed the features differentiating between CD and ITB on abdominal CT. The features included comb sign, necrotic lymph nodes, asymmetric bowel wall thickening, skip lesions, fibrofatty proliferation, mural stratification, ileocaecal area, long segment, and left colonic involvements. Sensitivity, specificity, positive and negative likelihood ratios, and diagnostic odds ratio (DOR) were calculated for all the features. Symmetric receiver operating characteristic curve was plotted for features present in >3 studies. Heterogeneity and publication bias was assessed and sensitivity analysis was performed by excluding studies that compared features on conventional abdominal CT instead of CT enterography (CTE). We included 6 studies (4 CTE, 1 conventional abdominal CT, and 1 CTE+conventional abdominal CT) involving 417 and 195 patients with CD and ITB, respectively. Necrotic lymph nodes had the highest diagnostic accuracy (sensitivity, 23%; specificity, 100%; DOR, 30.2) for ITB diagnosis, and comb sign (sensitivity, 82%; specificity, 81%; DOR, 21.5) followed by skip lesions (sensitivity, 86%; specificity, 74%; DOR, 16.5) had the highest diagnostic accuracy for CD diagnosis. On sensitivity analysis, the diagnostic accuracy of other features excluding asymmetric bowel wall thickening remained similar. Necrotic lymph nodes and comb sign on abdominal CT had the best diagnostic accuracy in differentiating CD and ITB.

Current role of capsule endoscopy in Crohn’s disease

Capsule endoscopy (CE) currently plays an important role in Crohn’s disease (CD). It is a noninvasive technique that has led to a breakthrough in the endoscopic diagnosis of diseases of the small intestine. Its superior diagnostic performance and excellent safety profile lead to its considerable acceptance on the part of the patient. This paper reviews current indications of CE in three stages of clinical practice: Suspected CD, unclassified colitis and its extensive role in diagnosed CD. The diagnostic and therapeutic impact of the results of CE on the monitoring of this disease is also reviewed. Knowledge of its applications, the interpretation of its results in an appropriate context and the existence of a validated endoscopic activity index could change the way in which these patients are managed. The definition of mucosal healing and postoperative recurrence by means of endoscopic scoring systems will endow CE with new applications in the management of CD in the near future.