Post written by Vinay Jahagirdar, MD, from the Department of Internal Medicine, University of Missouri Kansas City School of Medicine, Kansas City, Missouri, USA.

The focus of our study was to assess the pooled diagnostic accuracy parameters of deep machine learning by means of convolutional neural network (CNN) algorithms in predicting the severity of ulcerative colitis (UC) using endoscopic imaging.

Interobserver variations exist when assigning scores based on endoscopic appearance. Scores depend on the interpretation of imaging by the observer.

Increasingly, machine learning has been used in the field of medical image recognition and detection. Individual studies have used CNN-based deep learning models to classify the severity of UC based on endoscopic images. We wanted to consolidate existing literature and present pooled estimates for diagnostic accuracy parameters.

Based on our analysis, artificial intelligence (AI) demonstrated a pooled accuracy of 91.2%, sensitivity of 83.9%, specificity of 92.3%, positive predictive value of 86.5%, and negative predictive value of 89.4% for predicting the severity of UC, albeit with expected heterogeneity.

Further real-life clinical studies are needed to establish the role of AI in accurately staging UC because this carries implications for disease prognostication and therapeutic planning.

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.