GIE Senior Associate Editor David L. Diehl, MD, highlights this article from the February issue: “Identification of patients with malignant biliary strictures using a cholangioscopy-based deep learning artificial intelligence (with video)” by Neil B. Marya, MD, et al.

The assessment of indefinite biliary strictures is difficult because results of intraductal biopsy and cytologic brushing can be inconclusive. It was hoped that evaluation of specific cholangioscopic features could predict malignancy and improve diagnostic accuracy, but interobserver variability is a difficult problem to overcome.

With ongoing developments in artificial intelligence (AI)—assisted analysis of endoscopic imaging, it would be interesting to see whether “computer vision” would be more useful than human vision in analyzing malignant features of cholangioscopic imaging of biliary strictures.

The research group of Marya et al used convolutional neural networks (CNNs) to analyze endoscopic features of biliary strictures that correlate with malignancy. The remarkable and potentially practice-changing results were that the CNN was significantly more accurate for the diagnosis of malignancy in biliary strictures than brush cytology or forceps biopsy. (See the graphical abstract below.)

It would be quite useful if this AI technology could be used “on the fly” during a cholangioscopy procedure. It is feasible to see this technology go from “nice to have” to “must have” in the clinical evaluation of malignant biliary strictures. It might even become an early breakout application of AI in endoscopic practice.

Graphical Abstract

Read the full article online.

The information presented in Endoscopedia reflects the opinions of the authors and does not represent the position of the American Society for Gastrointestinal Endoscopy (ASGE). ASGE expressly disclaims any warranties or guarantees, expressed or implied, and is not liable for damages of any kind in connection with the material, information, or procedures set forth.