Skip to navigation Skip to content

Introduction

  • Home
  • Introduction
  • News & Announcements

News & Announcements

Title Research Team of Gastroenterology and Hepatology, Develops AI-based Next-Generation Endoscopic System

Hospital ANAM

Date 2021-07-07

Read 846

Research Team Led by Hyuk Soon Choi and Jae Min Lee, 

Professors of Gastroenterology and Hepatology, 

Develops AI-based Next-Generation Endoscopic System 



 

A joint study of a team led by Hyuk Soon Choi and Jae Min Lee, professors of gastroenterology and hepatology, together with KAIST professor Jaegul Choo's team, developed an AI-based next-generation endoscopic system. 
  
 Gastrointestinal endoscopy plays an important role in the diagnosis and treatment of gastrointestinal diseases. Patients tend to flock to hospitals with rich experiences and advanced technologies in search for receiving accurate and safe procedures. Quality control becomes much more important when it comes to different endoscopy procedures done for screening. Among others, endoscopic retrograde cholangiopancreatography, or ERCP, is a very complicated and difficult procedure so it is mainly performed by skilled doctors in university hospitals.  

 
 Leading institutions at home and abroad are engaged in the development of technologies that can upgrade the endoscopic procedures and help them offer the best possible treatments to patients. A joint study of a team led by Hyuk Soon Choi and Jae Min Lee, professors of gastroenterology and hepatology, together with KAIST professor Jaegul Choo's team, developed an AI and deep learning-based novel endoscopic image assessment system presenting a new approach to endoscopic examination and treatment.

 

 AI-driven system for endoscopic image assessment is expected to help enhance the public health as it can improve the accuracy of analyzing and classifying images. It may also be used as quality control measures for cancer screening or endoscopic lab certification. This study done by the Anma Hospital team was meaningful in that it has shown promising results in leading the development of technologies in medicine.

 

 The study team led by professor Choi used AI to identify the location of stomach by endoscopic image. The AI-driven quality control system was developed through convolutional neural networks (CNNs) using deep learning. The CNN model successfully classified the upper gastrointestinal (UGI) tract images with 97.58% accuracy, 97.42% sensitivity, 99.66% specificity, 97.50% positive predictive value (PPV), and 99.66% negative predictive value (NPV). This AI-based assessment technology allows thorough and accurate examination of the upper gastrointestinal tract which consists of the esophagus, stomach, and duodenum.

 

 Meanwhile, professor Lee's team conducted artificial intelligence-assisted analysis of endoscopic retrograde cholangiopancreatography image for identifying ampulla and cannulation difficulty. Highly skilled doctors trained the AI to identify ampulla and to predict the level of difficulty for selective cannulation. The model that they finally developed detected the ampulla of Vater with 76.2% precision, 78.4% recall, and identified more than 70% easy-to-cannulate cases. It also demonstrated its potential to be used in specialized endoscopic procedures.

 

 Professor Choi said, “I am very happy as our competency as a research team has been recognized. I have a plan to conduct researches on various areas such as minimally invasive endoscopic instruments as well as AI. Professor Lee said, “I hope AI research will facilitate further development of specialized endoscopy, a procedure to diagnose and treat problems in the gallbladder, bile ducts, and pancreas. I will do my best so that Anam Hospital will be at the forefront of developing related technologies.” Dr. Yoon Tae Jeen, head of the department of gastroenterology said, “Collaboration among those experts in medicine, computer science and other related industries will facilitate further development of AI programs, which in turn will boost emergence of new technologies or advancements of existing ones applied in the medical field.”

 

 The study of professor Choi and his team, "Development of artificial intelligence system for quality control of photo documentation in esophagogastroduodenoscopy," was published in Surgical Endoscopy while the study of professor Lee and his team titled, "Artificial intelligence-assisted analysis of endoscopic retrograde cholangiopancreatography image for identifying ampulla and difficulty of selective cannulation" was introduced in Scientific Reports.


List