| Unique ID issued by UMIN | UMIN000060514 |
|---|---|
| Receipt number | R000069227 |
| Scientific Title | A Multicenter Retrospective Observational Study on Changes in Detection Accuracy According to Version of Upper Endoscopy Image Diagnosis Support Program for Detection |
| Date of disclosure of the study information | 2026/01/29 |
| Last modified on | 2026/01/29 20:54:39 |
A Multicenter Retrospective Observational Study on Changes in Detection Accuracy According to Version of Upper Endoscopy Image Diagnosis Support Program for Detection
Changes in detection accuracy depending on the version of upper endoscopy AI
A Multicenter Retrospective Observational Study on Changes in Detection Accuracy According to Version of Upper Endoscopy Image Diagnosis Support Program for Detection
A Multicenter Retrospective Observational Study on Changes in Detection Accuracy According to Version of Upper Endoscopy Image Diagnosis Support Program for Detection
| Japan |
Esophageal disease
| Gastroenterology |
Malignancy
NO
The purpose of this study is to verify the performance of the new version 2.0 of the EW10-EG01 upper endoscopy image diagnosis support program for detection, developed by Fujifilm, from a clinical perspective. Specifically, the EW10-EG01 ver. 1.1 and ver. 2.0 programs will be applied offline and compared using case videos collected in a previous study (UMIN000049887). As a result, we aim to evaluate the effectiveness of the device in improving detection accuracy.
Efficacy
Confirmatory
By observation mode (WLI/BLI/LCI), frequency of detection frames (average number per case and average number per unit examination time). However, detection frames generated for target lesions are excluded.
Sensitivity for target lesions
Number of detection frames generated for each type of detected object (residue, bubbles, mucosal folds, benign polyps, inflammation, erosions, ulcer scars, others)
Observational
| 20 | years-old | <= |
| Not applicable |
Male and Female
Case videos were recorded for 181 cases (215 lesions) registered in the multi-institutional prospective observational study "Studying the usefulness of an endoscopic image diagnosis support system using artificial intelligence for esophageal cancer" (primary study UMIN000049887).
Cases of patients who expressed their intention to be excluded from the study after opting out
Cases whose case videos are unsuitable for analysis
181
| 1st name | TAKAO |
| Middle name | |
| Last name | MAEKITA |
Wakayama Medical University Hospital
Central endoscope section
6410012
811-1 Kimiidera, Wakayama City, Wakayama Prefecture, Japan
0734410627
maekita@wakayama-med.ac.jp
| 1st name | TAKAO |
| Middle name | |
| Last name | MAEKITA |
Wakayama Medical University Hospital
Central endoscope section
6410012
811-1 Kimiidera, Wakayama City, Wakayama Prefecture, Japan
0734410627
maekita@wakayama-med.ac.jp
Wakayama Medical University
Wakayama Medical University
Other
Wakayama Medical University
811-1 Kimiidera, Wakayama City, Wakayama Prefecture, Japan
0734410627
maekita@wakayama-med.ac.jp
NO
| 2026 | Year | 01 | Month | 29 | Day |
Unpublished
Preinitiation
| 2025 | Year | 12 | Month | 23 | Day |
| 2025 | Year | 12 | Month | 23 | Day |
| 2028 | Year | 03 | Month | 31 | Day |
The necessary case data (including case videos) have already been collected in the primary study(UMIN000049887). Patient information will be removed from the case data and collected from each participating institution at Wakayama Medical University.
Fujifilm Corporation will perform offline image analysis on the collected case videos using both the EW10-EG01 (ver. 1.1) and EW10-EG01 (ver. 2.0) software, extracting images of areas where the upper endoscopy imaging support program for detection showed a positive result. Patient information will not be sent to Fujifilm Corporation.
Images of areas where the upper endoscopy imaging support program for detection showed a positive result will be sent via Wakayama Medical University to each participating institution, where the detected area will be identified as cancer or a false positive.
The identification results from each participating institution will be compiled at Wakayama Medical University, and the evaluation items described below (such as the frequency of detection frames excluding target lesions and sensitivity to target lesions) will be evaluated.
Contributing Researchers
Mikitaka Iguchi, Wakayama Medical University
Shinya Taki, Wakayama Medical University
Fumiaki Kuwashima, Wakayama Medical University
Masateru Takao, Wakayama Medical University
Yosuke Mukai, Wakayama Medical University
Masayuki Kitano, Wakayama Medical University
Collaborating Research Institutions
Yoshiro Kawahara, Okayama University Hospital
Motohiko Kato, Keio University
Akihito Nagahara, Juntendo University Hospital, Juntendo University School of Medicine
Kazuhiro Furukawa, Nagoya University Hospital
Shiro Oka, Hiroshima University
Yuichi Teramura, Fujifilm Corporation
| 2026 | Year | 01 | Month | 29 | Day |
| 2026 | Year | 01 | Month | 29 | Day |
Value
https://center6.umin.ac.jp/cgi-open-bin/icdr_e/ctr_view.cgi?recptno=R000069227