Deep Learning and Device-Assisted Enteroscopy: Automatic Detection of Gastrointestinal Angioectasia

Device-assisted enteroscopy is an important complement of capsule endoscopy for the assessment and treatment of small bowel diseases. Additionally, device-assisted enteroscopy, combining both antegrade and retrograde approaches, enable a pan-endoscopic and interventive evaluation of the gastrointestinal tract. This exam is mainly performed in the investigation of obscure gastrointestinal bleeding after positive findings in capsule endoscopy.

We have developed a pioneer deep learning algorithm for real-time application to device-assisted enteroscopy, including push enteroscopy, single-balloon enteroscopy, and double-balloon enteroscopy, for the automatic detection of angioectasia. This is the first ever described AI algorithm for application to device-assisted enteroscopy. The development of this system complements our pool of solutions for pan-endoscopic assessment of the gastrointestinal tract, including a solution for interventive enteroscopy.

The article can be found here.

In Press // Medicina 2021, 57(12), 1378 // December 2021