UEG's educational platform for gastroenterology education bringing recordings, courses, guidelines, and more together in one place.
Most popular content
Rate for better recommendations!
Your feedback helps us show you more of what you like and less of what you dislike!
Log in to access this content.
Free for all myUEG account holders. Your access level is set automatically based on your occupation. Medical professionals get full access to all content. If you are a non-medical user, you can only access UEG Week content from congresses you attended.
Not sure what you can access? Learn more about account types.
Organoid technology in precision medicine in inflammation
1 University Medical Center Schleswig-holstein, Campus Kiel, Kiel, Germany
Event
Session
Published
Log in to access this content.
Free for all myUEG account holders. Your access level is set automatically based on your occupation. Medical professionals get full access to all content. If you are a non-medical user, you can only access UEG Week content from congresses you attended.
Not sure what you can access? Learn more about account types.
Introduction
We have reported ahead of the world that AI could detect esophageal squamous cell carcinoma(SCC) in endoscopic still images and movies with high sensitivity and in amazing speed. However, that report was investigated retrospectively in non-clinical setting. AI system would be more helpful for non-expert endoscopists who can miss ESCCs in white light imaging and NBI endoscopy without iodine staining.There have been no reports of AI support for the detection of ESCC in clinical practice.
Aims & Methods
We conducted the RCT to reveal the effect of AI system on the improvement of ESCC detection in clinical trial. This study was prospective, single-center, exploratory RCT. We randomly assigned patients to two groups, AI support group and endoscopist alone group. The primary outcome was detection rates of ESCC in non-experts, and the secondary outcomes were detection rates of ESCC in experts and in all endoscopists. Experts were defined as endoscopists trained more than 1 year at cancer-specialized hospital, and the rests were defined as non-experts. We observed the esophagus with white light imaging, following with NBI. After observation of upper GI tract, iodine was stained to whole esophagus. High risk patients of ESCC or head and neck SCC(HNSCC) were enrolled.
Results
Total of 325 patients were enrolled in this study, and 5 patients were excluded due to stenosis. 320 patients were analyzed in each group (152 patients in AI support group and 168 patients in endoscopist alone group). 90% of patients were for surveillance after treatment of ESCC or HNSCC. About 60% were performed by non-experts in each group. About details of detected ESCCs, median size was around 10mm, almost all lesions showed 0-IIb or 0-IIc. In non-experts, detection rate of ESCC was 47% in AI support group, and 45% in endoscopist alone group. Similarly, in experts, detection rate of ESCC was 87% in AI support group, and 57% in endoscopist alone group. There was no significant difference between both groups in experts and non-experts. In all endoscopists, detection rate was 57% in AI support group, and 50% in endoscopist alone group. In addition, we performed additional AI analysis of still images in newly discovered ESCCs. The sensitivity was 65%(25/40 lesions) and the positive predictive value was 40%. In endoscopist alone group, the detection rate of ESCC would be expanded to be 1.5 times higher when AI support with still images added.
| Details of detected ESCCs | |||
| Patients | Endoscopist + AI | Endoscopist alone | P value |
| Male/Female | 13 / 2 | 14 / 2 | |
| Age, median (range), year | 74 (56-84) | 73 (63-85) | |
| Endoscopist, n (%) Expert / Non-expert | 4 (27) / 11 (73) | 7 (44) / 9 (56) | |
| Size, median (range), mm | 10 (3-25) | 8.5 (2-20) | |
| Morphology, n (%) 0-I / 0-IIa / 0-IIb / 0-IIc | 1 (5) / 0 (0) / 13 (65) / 6 (30) | 0 (0) / 1 (5) / 10 (50) / 9 (45) | |
| Detection rate of ESCC among non-experts, % | 47 | 45 | 0.93 |
| Detection rate of ESCC among experts, % | 87 | 57 | 0.20 |
| Detection rate of ESCC among all endoscopists, % | 57 | 50 | 0.70 |
Conclusion
AI system could not improve the detection rate of ESCC in this clinical trial. Further research would be needed to improve clinical effectiveness.
References
Horie Y, et al. Gastrointest. Endosc. 2019; 89:25-32
Shiroma S, et al. Sci Rep. 2021; 11(1): 7759
RANDOMIZED CONTROLLED TRIAL OF ARTIFICIAL INTELLIGENCE DIAGNOSTIC SYSTEM IN CLINICAL PRACTICE TO DETECT ESOPHAGEAL SQUAMOUS CELL CARCINOMA
MASATO WATANABE 1, Toshiyuki Yoshio 1, Eisuke Nakao 2, Yohei Ikenoyama 1, Ken Namikawa 1, Yoshitaka Tokai 1, Shoichi Yoshimizu 1, Yusuke Horiuchi 1, Akiyoshi Ishiyama 1, Toshiaki Hirasawa 1, Junko Fujisaki 1, TOMOHIRO TADA 3
1 Cancer Institute Hospital, Japanese Foundation for Cancer Reseach, Tokyo, Japan
2 Tane General Hospital, Osaka, Japan
3 Tada Tomohiro Institute of Gastroproctcology, Saitama, Japan
Event
Submission format
Session
Citation
Published
Log in to access this content.
Free for all myUEG account holders. Your access level is set automatically based on your occupation. Medical professionals get full access to all content. If you are a non-medical user, you can only access UEG Week content from congresses you attended.
Not sure what you can access? Learn more about account types.
When should we think of and how should we diagnose pancreatic exocrine insufficiency?
1 Israelitic Krankenhaus, Hamburg, Germany
Event
Published
Log in to access this content.
Free for all myUEG account holders. Your access level is set automatically based on your occupation. Medical professionals get full access to all content. If you are a non-medical user, you can only access UEG Week content from congresses you attended.
Not sure what you can access? Learn more about account types.
Molecular classification of GI cancers
1 Lithuanian University of Health Sciences, Kaunas, Lithuania
Event
Published
Log in to access this content.
Free for all myUEG account holders. Your access level is set automatically based on your occupation. Medical professionals get full access to all content. If you are a non-medical user, you can only access UEG Week content from congresses you attended.
Not sure what you can access? Learn more about account types.
Simple solution to vexing problems
1 Mayo Clinic, Rochester, United States
Event
Published
Log in to access this content.
Free for all myUEG account holders. Your access level is set automatically based on your occupation. Medical professionals get full access to all content. If you are a non-medical user, you can only access UEG Week content from congresses you attended.
Not sure what you can access? Learn more about account types.
The future of microbiome therapeutics
1 Rambam Health Care Campus, Haifa, Israel
Event
Published
Log in to access this content.
Free for all myUEG account holders. Your access level is set automatically based on your occupation. Medical professionals get full access to all content. If you are a non-medical user, you can only access UEG Week content from congresses you attended.
Not sure what you can access? Learn more about account types.
Discussion
1 Erasmus University MC Dept. of Gastronterology, Rotterdam, Netherlands
2 Israelitic Krankenhaus, Hamburg, Germany
3 Umeå University Hospital, Umeå, Sweden, Umeå, Sweden