Dataset Information

A deep learning framework for autonomous detection and classification of Crohn's disease lesions in the small bowel and colon with capsule endoscopy.

ABSTRACT: 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.

SUBMITTER: Majtner T

PROVIDER: S-EPMC8367448 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:BackgroundPatients with Crohn's disease (CD) undergo frequent endoscopic procedures, with visualization of the gastrointestinal mucosa central to treatment decision-making. Subsequently, a noninvasive alternative to optical colonoscopy (OC) would be welcomed. One such technology is capsule endoscopy, including the PillCam COLON 2 (PCC2), though research validating its use in ileocolonic CD is limited. This study aims to compare PCC2 with ileocolonoscopy (OC) in assessing mucosal CD through use of a standardized scoring system.MethodsAt an Australian tertiary hospital, same-day PCC2 and ileocolonoscopy results of 47 CD patients, with known nonstricturing disease, were prospectively collected and analyzed for correlation and agreement. Deidentified recordings were reported by a single expert gastroenterologist. Mucosal disease was quantified using the Simple Endoscopic Score for Crohn's Disease (SES-CD). The SES-CD results of paired endoscopic modalities were compared in total per bowel segment and per SES-CD variable.ResultsOf 47 PCC2 recordings, 68% were complete, fully assessing terminal ileum to rectum, and OC was complete in 89%. Correlation (r) between total SES-CD scores was strongest in the terminal ileum (r = 0.77, P < .001), with the SES-CD variable of "ulcer detection" showing the strongest agreement. The PCC2 (vs OC) identified additional ulcers in the terminal ileum; ascending, transverse, and descending colon; and rectum; scores were 5 (1), 5 (3), 1 (1), 2 (1), and 2 (2), respectively.ConclusionsThe PCC2 shows promise in assessing ileocolonic mucosa, especially in proximal bowel segments, with greater reach of visualization in the small bowel. Given the resource and safety considerations raised by the Coronavirus disease 2019 pandemic, capsule endoscopy has particular significance.This article aims to contribute to the limited body of research surrounding the validity of capsule endoscopy technology in assessing ileocolonic mucosa in Crohn's Disease patients. In doing so, an alternative option for patients enduring frequent endoscopies is given potential.

Project description:To assess the cleansing efficacy and safety of a new Colon capsule endoscopy (CCE) bowel preparation regimen.This was a multicenter, prospective, randomized, controlled study comparing two CCE regimens. Subjects were asymptomatic and average risk for colorectal cancer. The second generation CCE system (PillCam® COLON 2; Medtronic, Yoqneam, Israel) was utilized. Preparation regimens differed in the 1st and 2nd boosts with the Study regimen using oral sulfate solution (89 mL) with diatrizoate meglumine and diatrizoate sodium solution ("diatrizoate solution") (boost 1 = 60 mL, boost 2 = 30 mL) and the Control regimen oral sulfate solution (89 mL) alone. The primary outcome was overall and segmental colon cleansing. Secondary outcomes included safety, polyp detection, colonic transit, CCE completion and capsule excretion ? 12 h.Both regimens had similar cleansing efficacy for the whole colon (Adequate: Study = 75.9%, Control = 77.3%; P = 0.88) and individual segments. In the Study group, CCE completion was superior (Study = 90.9%, Control = 76.9%; P = 0.048) and colonic transit was more often < 40 min (Study = 21.8%, Control = 4%; P = 0.0073). More Study regimen subjects experienced adverse events (Study = 19.4%, Control = 3.4%; P = 0.0061), and this difference did not appear related to diatrizoate solution. Adverse events were primarily gastrointestinal in nature and no serious adverse events related either to the bowel preparation regimen or the capsule were observed. There was a trend toward higher polyp detection with the Study regimen, but this did not achieve statistical significance for any size category. Mean transit time through the entire gastrointestinal tract, from ingestion to excretion, was shorter with the Study regimen while mean colonic transit times were similar for both study groups.A CCE bowel preparation regimen using oral sulfate solution and diatrizoate solution as a boost agent is effective, safe, and achieved superior CCE completion.

Dataset Information

A deep learning framework for autonomous detection and classification of Crohn's disease lesions in the small bowel and colon with capsule endoscopy.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets