期刊
PHYSICS IN MEDICINE AND BIOLOGY
卷 66, 期 3, 页码 -出版社
IOP Publishing Ltd
DOI: 10.1088/1361-6560/ab98d0
关键词
architectural distortion; digital breast tomosynthesis; computer aided detection; mammary gland distribution; deep learning
资金
- National Key R&D Program of China [2018YFC1704206, 2016YFB0200602, 2019YFC0121903, 2019YFC0117301]
- Fundamental Research Funds for the Central Universities [19LGYJS63]
- NSFC [81971691, 81801809, 81830052, 11401601]
- Natural Science Foundation of Guangdong Province, China [2019A1515011168, 2018A0303130215]
- Science and Technology Innovative Project of Guangdong Province [2016B030307003, 2015B010110003, 2015B020233008]
- Science and Technology Planning Project of Guangdong Province [2015B020233008, 2015B020233002, 2017B020210001]
- China Medical Research Fund of GuangDong Province [A2017496]
- Guangzhou Science and Technology Creative Project [201604020003]
- Guangdong Province Key Laboratory of Computational Science Open Grant [2018009]
- Construction Project of Shanghai Key Laboratory of Molecular Imaging [18DZ2260400]
- Clinical Research Startup Program of Southern Medical University by High-level University Construction Funding of Guangdong Provincial Department of Education [LC2016ZD018]
- Nanfang Hospital, Southern Medical University [2019C017]
- Clinical Research Program of Nanfang Hospital, Southern Medical University [2018CR040]
A deep-learning-based model using mammary gland distribution as prior information was proposed to improve the performance of Computer Aided Detection (CADe) for breast lesions. The model achieved significantly better results compared to existing methods, demonstrating the effectiveness of incorporating gland distribution and deep learning techniques in CADe for Architectural Distortion (AD) detection.
Computer aided detection (CADe) for breast lesions can provide an important reference for radiologists in breast cancer screening. Architectural distortion (AD) is a type of breast lesion that is difficult to detect. A majority of CADe methods focus on detecting the radial pattern, which is a main characteristic of typical ADs. However, a few atypical ADs do not exhibit such a pattern. To improve the performance of CADe for typical and atypical ADs, we propose a deep-learning-based model that used mammary gland distribution as prior information to detect ADs in digital breast tomosynthesis (DBT). First, information about gland distribution, including the Gabor magnitude, the Gabor orientation field, and a convergence map, were produced using a bank of Gabor filters and convergence measures. Then, this prior information and an original slice were input into a Faster R-CNN detection network to obtain the 2-D candidates for each slice. Finally, a 3-D aggregation scheme was employed to fuse these 2-D candidates as 3-D candidates for each DBT volume. Retrospectively, 64 typical AD volumes, 74 atypical AD volumes, and 127 normal volumes were collected. Six-fold cross-validation and mean true positive fraction (MTPF) were used to evaluate the model. Compared to an existing convergence-based model, our proposed model achieved an MTPF of 0.53 +/- 0.04, 0.61 +/- 0.05, and 0.45 +/- 0.04 for all DBT volumes, typical + normal volumes, and atypical + normal volumes, respectively. These results were significantly better than those of 0.36 +/- 0.03, 0.46 +/- 0.04, and 0.28 +/- 0.04 for a convergence-based model (p 0.01). These results indicate that employing the prior information of gland distribution and a deep learning method can improve the performance of CADe for AD.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据