Deep learning algorithms for automated detection of Crohn's disease ulcers by video capsule endoscopy.

BACKGROUND AND AIMS: The aim of our study was to develop and evaluate a deep learning algorithm for the automated detection of small-bowel ulcers in Crohn's disease (CD) on capsule endoscopy (CE) images of individual patients. METHODS: We retrospectively collected CE images of known CD patients and control subjects. Each image was labeled by an expert gastroenterologist as either normal mucosa or containing mucosal ulcers. A convolutional neural network was trained to classify images into either normal mucosa or mucosal ulcers. First, we trained the network on 5-fold randomly split images (each fold with 80% training images and 20% images testing). We then conducted 10 experiments in which images from n - 1 patients were used to train a network and images from a different individual patient were used to test the network. Results of the networks were compared for randomly split images and for individual patients. Area under the curves (AUCs) and accuracies were computed for each individual network. RESULTS: Overall, our dataset included 17,640 CE images from 49 patients: 7391 images with mucosal ulcers and 10,249 images of normal mucosa. For randomly split images results were excellent, with AUCs of .99 and accuracies ranging from 95.4% to 96.7%. For individual patient-level experiments, the AUCs were also excellent (.94-.99). CONCLUSIONS: Deep learning technology provides accurate and fast automated detection of mucosal ulcers on CE images. Individual patient-level analysis provided high and consistent diagnostic accuracy with shortened reading time; in the future, deep learning algorithms may augment and facilitate CE reading.

The Role of Small-Bowel Capsule Endoscopy in the Diagnostic Algorithm of Complicated Perianal Disease.

INTRODUCTION: Complicated perianal disease (cPD) may be the sole presentation of Crohn's disease (CD). The role of small-bowel capsule endoscopy (SBCE) in the diagnostic algorithm of cPD is unclear. We aimed to evaluate the role of SBCE as a diagnostic tool, in patients with cPD, after a negative standard workup for CD. METHODS: A multicenter, retrospective, cross-sectional study, in patients with cPD, and negative standard workup for CD (ileocolonoscopy and cross-sectional imaging), who underwent SBCE for suspected CD. Demographics, biomarkers, and the Lewis Score (LS) were recorded and analyzed. An LS ≥ 135 was considered a positive SBCE for diagnosing CD. RESULTS: Ninety-one patients were included: 65 (71.4%) males; median age: 37 (29-51) years; cPD duration: 25.1 (12.5-66.1) months. Positive SBCE: 24/91 (26.4%) patients. Fecal calprotectin (FC) positively correlated with LS (r = 0.81; p < 0.001). FC levels of 100 µg/g and 50 µg/g had a sensitivity of only 40% and 55% to rule out small-bowel CD, with a negative predictive value (NPV) of only 76% and 80%, respectively. CONCLUSIONS: SBCE contributed to CD diagnosis in a quarter of patients with cPD after a negative standard workup. FC levels correlated with the degree of inflammation defined by the LS. However, the NPV of FC was low, suggesting that SBCE should be considered for patients with cPD even after a negative standard workup.