臨床研究等提出・公開システム

The validation study of detectability and diagnostic accuracy of AI-aided endoscopic diagnosis system for colonoscopy; multicentric study (CAD-COLON study)

The validation study of detectability and diagnostic accuracy of AI-aided endoscopic diagnosis system for colonoscopy (CAD-COLON study)

358

Mean age AI-first 61.63 +- 9.89 / Standard CS-first 61.44 +- 10.01 Sex ratio AI-first Male 76.4%,Female 23.6% / Standard CS-first Male 76.8%,Female 23.2%

Allocation AI-first179 / Standard CS-first179 Per protocol set2 AI-first172 / Standard CS-first174 Per protocol set1 AI-first171 / Standard CS-first173

None

<Primary endpoint> 1.Adenoma miss rate(AMR):AI-first13.8% / Standard CS-first36.7%(P<0.0001) <Secondary endpoint> 1.Polyp miss rate(PMR):AI-first14.2% / Standard CS-first40.6%(P<0.0001) 2.Adenoma detection rate(ADR):AI-first64.5% / Standard CS-first53.6%(P=0.036) 3.Polyp detection rate(PDR):AI-first69.8% / Standard CS-first60.9%(P=0.084) 4.Mean number of adenomas per procedure(MAP):AI-first1.42+-2.01 / Standard CS-first1.25+-1.80%(P=0.412)

A total of 358 patients were enrolled and 179 patients were assigned to the SC-first group or CADe-first group. The AMR of the CADe-first group was significantly lower than that of the SC-first group (13.8%vs.36.7%,P<0.0001). Similar results were observed for the polyp miss rate (14.2%vs.40.6%,P<0.0001). The adenoma detection rate of CADe-assisted colonoscopy was 64.5%, which was significantly higher than that of standard colonoscopy (53.6%;P=0.036).

July. 20, 2021

July. 03, 2021

https://doi.org/10.1007/s00535-021-01808-w

No

None

https://jrct.niph.go.jp/latest-detail/jRCTs032190061

Tamai Naoto

The Jikei University School of Medicine

3-25-8, Nishi-Shimbashi, Minato-ku, Tokyo

tamai-naoto@jikei.ac.jp

Kamba Shunsuke

The Jikei University School of Medicine

3-25-8, Nishi-Shimbashi, Minato-ku, Tokyo

+81-3-3433-1111

kanba@jikei.ac.jp

Complete

July. 24, 2019

Interventional

randomized controlled trial

single blind

no treatment control/standard of care control

parallel assignment

diagnostic purpose

In patients scheduled for colonoscopy / treatment, patients who meet the following criteria.
1. 40-80 years old
2. Patients who gave written informed consent based on their own intention after receiving sufficient explanation about this study
3. Informed consensus has been obtained that endoscopic resection should be performed if a lesion is found
4. The examination purpose is for screening of colon cancer or surveillance after endoscopic treatment of colorectal lesions

1. Patients with known inflammatory bowel disease and colonic stenosis
2. Patients with known familial adenomatous polyposis syndrome
3. Patients with known colorectal cancer/polyp
4. Patients who have had colon surgery in the past (except for appendectomy and surgery on the anus)
5. Patients with high risk of bleeding after endoscopic treatment, and difficult management of anticoagulation or antiplatelet agent.
6. Patients with severe organ failure (cirrhosis of Child C, heart failure of ACC / AHA stage D)
7. Pregnant patient
8. Patients who are judged as ineligible for registration by attending doctors

Both

Colorectal polyps

Detection and diagnosis support for colorectal lesions to doctors by AAE system
1. Standard colonoscopy-first-group: Standard colonoscopy is performed first, followed by colonoscopy under the support of the AAE system
2. AAE-first-group: Colonoscopy under the support of the AAE system is performed first, followed by normal colonoscopy
All colorectal polyps detected in any group are removed

Colorectal adenoma/cancer/polyp

Artificial Intelligence/Detection and diagnosis support

D015179

D003113

Adenoma miss rate

Adenoma miss rate per patient, Adenoma detection rate, Polyp miss rate, Polyp miss rate per patient, Polyp detection rate, Mean number of adenomas per procedure

+81-3-3433-1111

June. 03, 2019

none