人工智能技术在宫颈细胞筛查中的应用进展和挑战

子宫颈癌是女性最常见的恶性肿瘤之一，通过人类乳头瘤病毒(human papillomavirus，HPV)检测和宫颈细胞学筛查，进行早期诊断和早期治疗能够控制宫颈癌的发病和死亡。由于缺乏宫颈细胞筛查人员，使其收效甚微。人工智能(artificial intelligency，AI)技术应用宫颈癌筛查，可望提供最佳的解决方案。通过文献复习和归纳，本文阐述了AI辅助宫颈癌筛查的进展，包括不同的AI算法模型的利弊，人工筛查与AI辅助筛查之间不同的人机交互筛查工作模式和应用于场景；分析了目前AI辅助宫颈癌筛查的结果和应用优势；例举了开发AI辅助宫颈癌筛查中遇到的问题和挑战。旨在为开发和利用AI辅助宫颈细胞筛查提供借鉴和思考，促进AI辅助宫颈癌筛查产品早日落地和应用，减少我国宫颈癌的发病率和死亡率。

Applicational progress and challenges of the artificial intelligence-aided cervical cancer cytological screening

Cervical cancer is one of the most common malignant tumors in women. Early detection and treatment are critical to reduce its mobility and motality. Cytological screening combined with HPV test is the best way for its early detection. However, the early diagnosis is impeded due to severely lack of cytopathologists. The application of artificial intelligency (AI) technology in cervical cancer screening will provide the best solution to enhance the screening efficiency and quality. We reviewed literatures of the AI-aided cervical cancer screening, described its progress of AI algorithm models, human screening and AI-aided screening interactive models in the cervical cytology; described the Prons and Cons of different machine and deep learning algorithms based on the bright and dark rules; analyzed available results of the AI-aided cervical cancer screening, and diacussed problems and challenges in exploring and applying of the AI-aided cervical cancer screening products. The purpose of this review is to provide insights for the research and development of the AI-aided cervical cancer screening to promote its application and implementation, which will contribute to reduce the mobility and motality of cervical cancer.