A retrospective study of computer-aided detection system for detection improvement of adenomas
10.3760/cma.j.cn321463-20240311-00047
- VernacularTitle:计算机辅助检测系统提高腺瘤检出率的回顾性研究
- Author:
Pengju WANG
1
;
Longsong LI
;
Bo ZHANG
;
Yaxuan CHENG
;
Fanqi MENG
;
Wen XIAO
;
Ningli CHAI
Author Information
1. 解放军总医院第一医学中心消化内科医学部，北京 100853
- Keywords:
Artificial intelligence;
Colonic polyps;
Colonoscopy;
Computer-aided detection
- From:
Chinese Journal of Digestive Endoscopy
2024;41(6):443-448
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To assess the efficacy of a novel spatial-temporal polyp detection system in colonoscopy.Methods:This research was a retrospective comparative study. Eight hundred and thirty-three participants who underwent computer-aided detection (CADe) colonoscopy at the First Medical Center of Chinese PLA General Hospital between March and June 2023 were enrolled to the experimental group, while 770 individuals who received conventional colonoscopy from March to June 2022, in the identical operation room were to the control group. The primary outcome was the adenoma detection rate (ADR), and the secondary outcomes were the polyp detection rate (PDR), adenomas per colonoscopy (APC), and polyps per colonoscopy (PPC).Results:The ADR [29.3% (244/833) VS 21.7% (167/770), χ2=12.133, P<0.001] and PDR [47.9% (399/833) VS 37.9% (292/770), χ2=16.241, P<0.001] were significantly higher in the experimental group than those in the control group. Adenomas ≤5 mm [23.5% (196/833) VS 16.1% (124/770), χ2=13.808, P<0.001] and flat-type adenomas [15.1% (126/833) VS 7.3% (56/770), χ2=24.519, P<0.001] were detected in a significantly higher proportion of subjects in the experimental group than those in the control group. There were significant difference in APC [0 (0,1) VS 0 (0,1), Z=-3.698, P<0.001] and PPC [0 (0,1) VS 0 (0,1), Z=-4.424, P<0.001] between the experimental and control groups. The use of CADe system significantly increased both ADR [29.5% (167/566) VS 18.9% (89/472), χ2=15.709, P<0.001] and PDR [47.3% (268/566) VS 33.3% (157/472), χ2=21.123, P<0.001] in junior endoscopists. However, in senior endoscopists, there was no statistical significant difference in ADR [28.8% (77/267) VS 26.2% (78/298), χ2=0.502, P=0.479] or PDR [49.1% (131/267) VS 45.3% (135/298), χ2=0.800, P=0.371] with or without CADe system. Conclusion:The use of CADe system significantly increases overall polyp and adenoma detection in clinical practice, especially in the detection of diminutive and flat-type lesions. Junior endoscopists gain greater advantages from the use of CADe system than their senior peers.
