npj Digital Medicine (Feb 2023)
Towards precision medicine based on a continuous deep learning optimization and ensemble approach
- Jian Li,
- Linyuan Jin,
- Zhiyuan Wang,
- Qinghai Peng,
- Yueai Wang,
- Jia Luo,
- Jiawei Zhou,
- Yingying Cao,
- Yanfen Zhang,
- Min Zhang,
- Yuewen Qiu,
- Qiang Hu,
- Liyun Chen,
- Xiaoyu Yu,
- Xiaohui Zhou,
- Qiong Li,
- Shu Zhou,
- Si Huang,
- Dan Luo,
- Xingxing Mao,
- Yi Yu,
- Xiaomeng Yang,
- Chiling Pan,
- Hongxin Li,
- Jingchao Wang,
- Jieke Liao
Affiliations
- Jian Li
- Department of Ultrasound, The Affiliated Changsha Central Hospital, Hengyang Medical School, University of South China
- Linyuan Jin
- Department of Ultrasound, The Affiliated Changsha Central Hospital, Hengyang Medical School, University of South China
- Zhiyuan Wang
- Department of Ultrasound, The Affiliated Cancer Hospital of Xiangya School of Medicine, Central South University
- Qinghai Peng
- Department of Ultrasound, The Second Xiangya Hospital of Central South University
- Yueai Wang
- Department of Ultrasound, The First Affiliated Hospital of Hunan University of Traditional Chinese Medicine
- Jia Luo
- Department of Ultrasound, The First Affiliated Hospital of Hunan University of Traditional Chinese Medicine
- Jiawei Zhou
- Department of Ultrasound, The Second Xiangya Hospital of Central South University
- Yingying Cao
- Department of Ultrasound, The Affiliated Cancer Hospital of Xiangya School of Medicine, Central South University
- Yanfen Zhang
- Department of Ultrasound, The Affiliated Changsha Central Hospital, Hengyang Medical School, University of South China
- Min Zhang
- Department of Ultrasound, The Affiliated Changsha Central Hospital, Hengyang Medical School, University of South China
- Yuewen Qiu
- Department of Ultrasound, The Affiliated Changsha Central Hospital, Hengyang Medical School, University of South China
- Qiang Hu
- Department of Ultrasound, The Affiliated Changsha Central Hospital, Hengyang Medical School, University of South China
- Liyun Chen
- Department of Ultrasound, The Affiliated Changsha Central Hospital, Hengyang Medical School, University of South China
- Xiaoyu Yu
- Department of Ultrasound, The Affiliated Changsha Central Hospital, Hengyang Medical School, University of South China
- Xiaohui Zhou
- Department of Ultrasound, The Affiliated Changsha Central Hospital, Hengyang Medical School, University of South China
- Qiong Li
- Department of Ultrasound, The Affiliated Changsha Central Hospital, Hengyang Medical School, University of South China
- Shu Zhou
- Department of Ultrasound, The Affiliated Changsha Central Hospital, Hengyang Medical School, University of South China
- Si Huang
- Department of Ultrasound, The Affiliated Changsha Central Hospital, Hengyang Medical School, University of South China
- Dan Luo
- Department of Ultrasound, The Affiliated Changsha Central Hospital, Hengyang Medical School, University of South China
- Xingxing Mao
- Department of Ultrasound, The Affiliated Changsha Central Hospital, Hengyang Medical School, University of South China
- Yi Yu
- Department of Ultrasound, The People’s Hospital of Liuyang
- Xiaomeng Yang
- Department of Ultrasound, Huaihua First People’s Hospital
- Chiling Pan
- eBay Inc.
- Hongxin Li
- Guangzhou Yirui Zhiying Technology Co. Ltd.
- Jingchao Wang
- Guangzhou Yirui Zhiying Technology Co. Ltd.
- Jieke Liao
- Guangzhou Yirui Zhiying Technology Co. Ltd.
- DOI
- https://doi.org/10.1038/s41746-023-00759-1
- Journal volume & issue
-
Vol. 6,
no. 1
pp. 1 – 11
Abstract
Abstract We developed a continuous learning system (CLS) based on deep learning and optimization and ensemble approach, and conducted a retrospective data simulated prospective study using ultrasound images of breast masses for precise diagnoses. We extracted 629 breast masses and 2235 images from 561 cases in the institution to train the model in six stages to diagnose benign and malignant tumors, pathological types, and diseases. We randomly selected 180 out of 3098 cases from two external institutions. The CLS was tested with seven independent datasets and compared with 21 physicians, and the system’s diagnostic ability exceeded 20 physicians by training stage six. The optimal integrated method we developed is expected accurately diagnose breast masses. This method can also be extended to the intelligent diagnosis of masses in other organs. Overall, our findings have potential value in further promoting the application of AI diagnosis in precision medicine.
