Ulcers and erosions are frequent findings in capsule endoscopy, and they’re often believed to be the source of bleeding in cases of obscure gastrointestinal bleeding. Our team developed a convolutional neural network for automatic detection and differentiation of ulcers and erosions and their classification according to a validated scale to predict their bleeding potential (Saurin’s scale). Our deep learning tool achieved an overall accuracy of 96%, with a sensitivity of 87% and a specificity of 96%. The authors are pleased to have a paper on this work accepted for publication in Techniques and Innovation in Gastrointestinal Endoscopy, journal with the American Gastroenterological Association’s seal of quality.
This the first artificial intelligence algorithm providing both detection of lesions and categorization according to a validated scale for estimating the bleeding potential of enteric lesions. This is important as gastrointestinal bleeding is the most common indication for capsule endoscopy.
The link for the article will posted here soon.
Accepted manuscript // July 2021