Mota Joana, João Almeida Maria, Mendes Francisco, Martins Miguel, Ribeiro Tiago, Afonso João, Cardoso Pedro, Cardoso Helder, Andrade Patricia, Ferreira João, Macedo Guilherme, Mascarenhas Miguel
Precision Medicine Unit, Department of Gastroenterology, São João University Hospital, 4200-427 Porto, Portugal.
WGO Gastroenterology and Hepatology Training Center, 4200-427 Porto, Portugal.
Diagnostics (Basel). 2024 Sep 19;14(18):2072. doi: 10.3390/diagnostics14182072.
Colon capsule endoscopy (CCE) enables a comprehensive, non-invasive, and painless evaluation of the colon, although it still has limited indications. The lengthy reading times hinder its wider implementation, a drawback that could potentially be overcome through the integration of artificial intelligence (AI) models. Studies employing AI, particularly convolutional neural networks (CNNs), demonstrate great promise in using CCE as a viable option for detecting certain diseases and alterations in the colon, compared to other methods like colonoscopy. Additionally, employing AI models in CCE could pave the way for a minimally invasive panenteric or even panendoscopic solution. This review aims to provide a comprehensive summary of the current state-of-the-art of AI in CCE while also addressing the challenges, both technical and ethical, associated with broadening indications for AI-powered CCE. Additionally, it also gives a brief reflection of the potential environmental advantages of using this method compared to alternative ones.
结肠胶囊内镜检查(CCE)能够对结肠进行全面、无创且无痛的评估,尽管其适应证仍然有限。冗长的阅片时间阻碍了其更广泛的应用,这一缺点有可能通过整合人工智能(AI)模型来克服。采用AI的研究,特别是卷积神经网络(CNN),与结肠镜检查等其他方法相比,在将CCE用作检测结肠某些疾病和病变的可行选择方面显示出巨大的前景。此外,在CCE中应用AI模型可能为微创全肠道甚至全内镜解决方案铺平道路。本综述旨在全面总结CCE中AI的当前技术水平,同时解决与扩大AI驱动的CCE适应证相关的技术和伦理挑战。此外,它还简要思考了与替代方法相比使用该方法潜在的环境优势。