학술논문
Current evidence and future directions on improving the endoscopic recognition of early colorectal carcinoma using artificial intelligence – a scoping review.
Document Type
Article
Author
Source
Subject
*ARTIFICIAL intelligence
*COLORECTAL cancer
*CONVOLUTIONAL neural networks
*RECOGNITION (Psychology)
*
*
*
Language
ISSN
0013-726X
Abstract
This scoping review examines the current evidence and future directions for using artificial intelligence (AI) to improve the endoscopic recognition of early colorectal carcinoma (CRC). The review includes 24 articles published in the past 5 years that used AI to distinguish CRC from colorectal polyps on endoscopy or endocytoscopy imaging. The studies used various AI algorithms, with convolutional neural network architectures being the most common. The diagnostic performances of the AI systems varied, but overall showed promise. However, there is a need for further research, particularly in real-time performance and multicenter external validation. Future research should focus on developing AI systems that can differentiate CRC from premalignant lesions and provide information on submucosal invasion depth. [Extracted from the article]