| Budget Amount *help |
¥4,030,000 (Direct Cost: ¥3,100,000、Indirect Cost: ¥930,000)
Fiscal Year 2018: ¥910,000 (Direct Cost: ¥700,000、Indirect Cost: ¥210,000)
Fiscal Year 2017: ¥3,120,000 (Direct Cost: ¥2,400,000、Indirect Cost: ¥720,000)
|
|---|
| Outline of Final Research Achievements |
In this study, we constructed and reported an endoscope automatic diagnosis system for evaluation of inflammatory activity of ulcerative colitis using artificial intelligence. As a pilot study, 22 835 ultra-magnified endoscopic images of 187 patients with ulcerative colitis taken from October 2016 to April 2018 were used. Of these, 12,900 in 87 cases were trained in artificial intelligence, and 9,953 to 525 sets of verification sample sets in 100 cases were created. The diagnostic ability of this system for histologic inflammation with a Geboes score of 3.1 or higher is 74% (95% CI: 65-81), 97% (95-99), 91% (88-93). Met. The diagnostic time per sample was about 0.4 seconds. The above results were reported in Gastrointestinal Endoscopy. (Maeda Y, et al. Gastrointest Endosc 2019.) Furthermore, by March 2019, about 45000 images were collected from 395 patients and the system was up-datad.
|
