臨床研究等提出・公開システム

Study of AI-based COmputer-aided diagnosis in colonoscopy for Polyp charactErization: a multi-center randomized controlled trial (AI-SCOPE Trial)

RCT for CADx in colonoscopy

Misawa Masashi

Showa University Northern Yokohama Hospital

Chigasaki-chuo 35-1, Tsuduki-ku, Yokohama-city

mmisawa@med.showa-u.ac.jp

Misawa Masashi

Showa University Northern Yokohama Hospital

Chigasaki-chuo 35-1, Tsuduki-ku, Yokohama-city

+81-45-949-7000

mmisawa@med.showa-u.ac.jp

Recruiting

Jan. 14, 2025

Interventional

randomized controlled trial

open(masking not used)

no treatment control/standard of care control

parallel assignment

diagnostic purpose

Adult patients 18 years of age or older scheduled for colonoscopy and consenting to the study

1. patients with a history of colorectal resection (excluding appendectomy)
2. Patients who are known to have a micro-polyp that is expected to be involved in the primary endpoint before the examination, and who have specific characteristic information (location, size, morphology, pathological diagnosis, etc.) that could affect the endoscopic diagnosis of the lesion.
3. Patients with known intra-abdominal mass or intestinal mass
Patients with inflammatory bowel disease, polyposis, or hereditary colorectal cancer
5. patients who cannot be treated with anticoagulants or antiplatelet agents in accordance with the Guidelines for Gastrointestinal Endoscopy for Patients on Antithrombotic Agents (2012) and Supplement (2017), and whose lesions cannot be resected
6. patients who have an emergency endoscopy (gastrointestinal bleeding, sigmoid colon torsion repair)
7. pregnant women and lactating mothers
8. Other patients who are judged by the physician in charge or the examining physician to be unsuitable for the study, including those with underlying medical conditions.

Both

Colorectal polyp

The intervention group (EndoBRAIN: EB group) uses EndoBRAIN-X, which is approved by the Pharmaceutical Affairs Law, and the results are referred to the physician for endoscopic diagnosis. The control group (standard colonoscopy: SC) will undergo colonoscopy without EndoBRAIN-X and the physician will perform the endoscopic diagnosis as usual.

Percentage of lesions for which pathology can be omitted

Sensitivity, specificity, positive predictive value, negative predictive value, and positive predictive value for neoplastic lesions (including SSL) of 5 mm or less
(Also examine by confidence level)
Sensitivity, specificity, positive predictive value, negative predictive value, and positive predictive value for neoplastic lesions less than 5 mm (not including SSL)
(Also examine by confidence level)
Percentage of tumor-free resections: Percentage of patients without adenoma or SSL in the resected lesion (Repici A et al.
(Repici A et al. Gut 2022;71(4):757-765)
Sensitivity, specificity, positive predictive value, negative predictive value, and positive predictive value for SSL in serrated lesions (also by confidence level)
Agreement rates of surveillance intervals based on JGES, ASGE, and ESGE guidelines.
Percentage of cases with shortened or prolonged surveillance intervals due to misdiagnosis
Percentage of lesions with discrepancies between physician and AI diagnosis
Endoscope removal time (time elapsed from cecum to anus including observation and treatment time)
Medical costs associated with the examinations (comparison assuming that pathological diagnosis and endoscopic treatment were omitted).
Learning curve of examiner's diagnostic ability
Evaluation by proficiency level
Sensitivity, specificity, positive diagnostic rate, negative predictive value, and positive predictive value of EB-X's diagnostic support results

+81-3-3784-8129

No

None