A research team from the Faculty of Engineering and the Faculty of Medicine at Chulalongkorn University has launched "DeepGI," an innovative artificial intelligence (AI) technology designed for colonoscopy and gastrointestinal endoscopy to detect colorectal cancer with high accuracy. This development, a collaborative effort between the two faculties, has attracted significant international attention and recognition, and was presented at the ENDO 2024 conference in South Korea.
Professor Rungsun Rerknimitr, Ph.D. Acting Assistant to the President for Innovation at Chulalongkorn University, an Associate Professor in the Department of Internal Medicine, Faculty of Medicine, and a specialist in gastrointestinal endoscopy at the Center of Excellence in Gastrointestinal Endoscopy, King Chulalongkorn Memorial Hospital, revealed that at the ENDO 2024 conference in South Korea, organized by the World Endoscopy Organization (WEO) and the Korean Society of Gastrointestinal Endoscopy (KSGE), which had over 3,200 attendees, Chulalongkorn's Department of Gastroenterology was selected as one of three centers to perform live endoscopy. During this live transmission, four patient cases were presented, one of which utilized DeepGI technology to detect precancerous gastric mucosa.
DeepGI was developed by Associate Professor Peerapol Veerakul, Ph.D. and Assistant Professor Natawut Nupairoj, Ph.D. from the Department of Computer Engineering, Faculty of Engineering, in collaboration with Professor Rungsun Rerknimitr, Ph.D. from the Faculty of Medicine and the research team at the Center of Excellence in Gastrointestinal Endoscopy at King Chulalongkorn Memorial Hospital. The project was supported by ESM Solutions Co., Ltd., which also serves as the distributor of DeepGI, currently being used at King Chulalongkorn Memorial Hospital and tested in several other hospitals.
DeepGI boasts four significant advantages:
Deep Technology: Utilizes deep learning technology for real-time detection and classification of abnormal tissues with high accuracy.
Vendor Unlock: Compatible with all brands and models of gastrointestinal endoscopy equipment.
Extensible Future: Can be expanded to detect abnormalities in other parts of the body, such as the bile ducts and stomach.
Affordable: More cost-effective than other systems on the market, making it accessible for hospitals and government agencies, particularly beneficial for rural healthcare.
The demonstration of DeepGI at the ENDO 2024 conference was well-received, with attendees expressing interest in further research collaborations with Chulalongkorn University to expand the innovation internationally. Currently, India and Italy have shown considerable interest in this technology.
DeepGI has demonstrated an impressive 90% accuracy rate in detecting gastrointestinal abnormalities. Research indicates that DeepGI can enhance polyp detection efficiency by 16%, increase polyp detection rates by 1%, and reduce colorectal cancer risk by 3%. Additionally, DeepGI can detect early-stage gastric cancer (Gastrointestinal Metaplasia - GIM), which is increasingly prevalent due to unhealthy dietary habits.
Associate Professor Dr. Peerapol Veerakul stated that DeepGI will significantly improve medical team efficiency and reduce diagnostic errors. The system processes video images during endoscopy and analyzes abnormalities using real-time AI, providing immediate diagnostic results and enhancing physician confidence in patient treatment decisions.
Assistant Professor Natawut Nupairoj, Ph.D. mentioned that DeepGI has been continuously developed and is ready for practical use in various hospitals. Comparisons with other systems on the market have shown that DeepGI is effective in detecting abnormalities in both the colon and stomach, significantly preventing colorectal and gastric cancer.
For more information about DeepGI, please visit https://deepgi.co or access the service at the Center of Excellence in Gastrointestinal Endoscopy, King Chulalongkorn Memorial Hospital, Thai Red Cross Society, Bhumisirimangkalanusorn Building, 10th Floor, Zone A, or call 0-2256-4000 ext. 81001-2.
