TY - JOUR
T1 - ABUSDet
T2 - A Novel 2.5D deep learning model for automated breast ultrasound tumor detection
AU - Song, Xudong
AU - Lu, Xiaoyang
AU - Fang, Gengfa
AU - He, Xiangjian
AU - Fan, Xiaochen
AU - Cai, Le
AU - Jia, Wenjing
AU - Wang, Zumin
N1 - Publisher Copyright:
© 2023, The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature.
PY - 2023/11
Y1 - 2023/11
N2 - Automated Breast Ultrasound is a highly advanced breast tumor detection modality that produces hundreds of 2D slices in each scan. However, this large number of slices poses a significant burden for physicians to review. This paper proposes a novel 2.5D tumor detection model, named “ABUSDet,” to assist physicians in automatically reviewing ABUS images and predicting the locations of breast tumors in images. At the core of this approach, a sequence of data blocks partitioned from a pre-processed 3D volume are fed to a 2.5D tumor detection model, which outputs a sequence of 2D tumor candidates. An aggregation module then clusters the 2D tumor candidates to produce the ultimate 3D coordinates of the tumors. To further improve the accuracy of the model, a novel mechanism for training deep learning models, called “Deliberate Training,” is proposed. The proposed model is trained and tested on a dataset of 87 patients with 235 ABUS volumes. It achieves sensitivities of 77.94%, 75.49%, and 65.19% at FPs/volume of 3, 2, and 1, respectively. Compared with the 2D and 3D object detection models, the proposed ABUSDet model achieves the highest sensitivity with relatively low false-positive rates. Graphical abstract: [Figure not available: see fulltext.]
AB - Automated Breast Ultrasound is a highly advanced breast tumor detection modality that produces hundreds of 2D slices in each scan. However, this large number of slices poses a significant burden for physicians to review. This paper proposes a novel 2.5D tumor detection model, named “ABUSDet,” to assist physicians in automatically reviewing ABUS images and predicting the locations of breast tumors in images. At the core of this approach, a sequence of data blocks partitioned from a pre-processed 3D volume are fed to a 2.5D tumor detection model, which outputs a sequence of 2D tumor candidates. An aggregation module then clusters the 2D tumor candidates to produce the ultimate 3D coordinates of the tumors. To further improve the accuracy of the model, a novel mechanism for training deep learning models, called “Deliberate Training,” is proposed. The proposed model is trained and tested on a dataset of 87 patients with 235 ABUS volumes. It achieves sensitivities of 77.94%, 75.49%, and 65.19% at FPs/volume of 3, 2, and 1, respectively. Compared with the 2D and 3D object detection models, the proposed ABUSDet model achieves the highest sensitivity with relatively low false-positive rates. Graphical abstract: [Figure not available: see fulltext.]
KW - 2.5D tumor detection
KW - Automated breast ultrasound (ABUS)
KW - Breast cancer
KW - Deliberate training
UR - http://www.scopus.com/inward/record.url?scp=85168324685&partnerID=8YFLogxK
U2 - 10.1007/s10489-023-04785-0
DO - 10.1007/s10489-023-04785-0
M3 - Article
AN - SCOPUS:85168324685
SN - 0924-669X
VL - 53
SP - 26255
EP - 26269
JO - Applied Intelligence
JF - Applied Intelligence
IS - 21
ER -