Hybrid AI-assistive diagnostic model permits rapid TBS classification of cervical liquid-based thin-layer cell smears

• Published in: Nature communications Vol. 12; no. 1; p. 3541
• Main Authors: Zhu, Xiaohui, Li, Xiaoming, Ong, Kokhaur, Zhang, Wenli, Li, Wencai, Li, Longjie, Young, David, Su, Yongjian, Shang, Bin, Peng, Linggan, Xiong, Wei, Liu, Yunke, Liao, Wenting, Xu, Jingjing, Wang, Feifei, Liao, Qing, Li, Shengnan, Liao, Minmin, Li, Yu, Rao, Linshang, Lin, Jinquan, Shi, Jianyuan, You, Zejun, Zhong, Wenlong, Liang, Xinrong, Han, Hao, Zhang, Yan, Tang, Na, Hu, Aixia, Gao, Hongyi, Cheng, Zhiqiang, Liang, Li, Yu, Weimiao, Ding, Yanqing
• Format: Journal Article
• Language: English
• Published: London Nature Publishing Group UK 10.06.2021; Nature Publishing Group; Nature Portfolio
• Subjects: 631/114/2397; 631/1647/245/2226; 631/67/2322; 692/699/67/2322; Adaptive sampling; Adaptive systems; Artificial Intelligence; Cancer; Cervical cancer; Cervix; Classification; Computer Simulation; Criteria; Cytology; Decision trees; Deep Learning; Diagnosis; Diagnostic systems; Early Detection of Cancer; Female; Humanities and Social Sciences; Humans; Image Processing, Computer-Assisted; multidisciplinary; Prospective Studies; Quality control; Retrospective Studies; Sample preparation; Science; Science (multidisciplinary); Segmentation; Specimen Handling - methods; Target detection; Uterine Cervical Neoplasms - diagnosis; Uterine Cervical Neoplasms - diagnostic imaging; Uterine Cervical Neoplasms - pathology; Uterine Cervical Neoplasms - physiopathology; Vaginal Smears - methods
• ISSN: 2041-1723; 2041-1723
• DOI: 10.1038/s41467-021-23913-3

Technical advancements significantly improve earlier diagnosis of cervical cancer, but accurate diagnosis is still difficult due to various factors. We develop an artificial intelligence assistive diagnostic solution, AIATBS, to improve cervical liquid-based thin-layer cell smear diagnosis according to clinical TBS criteria. We train AIATBS with >81,000 retrospective samples. It integrates YOLOv3 for target detection, Xception and Patch-based models to boost target classification, and U-net for nucleus segmentation. We integrate XGBoost and a logical decision tree with these models to optimize the parameters given by the learning process, and we develop a complete cervical liquid-based cytology smear TBS diagnostic system which also includes a quality control solution. We validate the optimized system with >34,000 multicenter prospective samples and achieve better sensitivity compared to senior cytologists, yet retain high specificity while achieving a speed of <180s/slide. Our system is adaptive to sample preparation using different standards, staining protocols and scanners. Technical advancements have significantly improved early diagnosis of cervical cancer, but accurate diagnosis is still difficult due to various practical factors. Here, the authors develop an artificial intelligence assistive diagnostic solution to improve cervical liquid-based thin-layer cell smear diagnosis according to clinical TBS criteria in a large multicenter study.