2D Convolutional Neural Networks for 3D Digital Breast Tomosynthesis Classification

Yu Zhang; Xiaoqin Wang; Hunter Blanton; Gongbo Liang; Xin Xing; Nathan Jacobs

doi:10.1109/BIBM47256.2019.8983097

2D Convolutional Neural Networks for 3D Digital Breast Tomosynthesis Classification

Yu Zhang, Xiaoqin Wang, Hunter Blanton, Gongbo Liang, Xin Xing, Nathan Jacobs

Computer Science

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

24 Scopus citations

Abstract

Automated methods for breast cancer detection have focused on 2D mammography and have largely ignored 3D digital breast tomosynthesis (DBT), which is frequently used in clinical practice. The two key challenges in developing automated methods for DBT classification are handling the variable number of slices and retaining slice-to-slice changes. We propose a novel deep 2D convolutional neural network (CNN) architecture for DBT classification that simultaneously overcomes both challenges. Our approach operates on the full volume, regardless of the number of slices, and allows the use of pre-trained 2D CNNs for feature extraction, which is important given the limited amount of annotated training data. In an extensive evaluation on a real-world clinical dataset, our approach achieves 0.854 auROC, which is 28.80% higher than approaches based on 3D CNNs. We also find that these improvements are stable across a range of model configurations.

Original language	English
Title of host publication	Proceedings - 2019 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2019
Editors	Illhoi Yoo, Jinbo Bi, Xiaohua Tony Hu
Pages	1013-1017
Number of pages	5
ISBN (Electronic)	9781728118673
DOIs	https://doi.org/10.1109/BIBM47256.2019.8983097
State	Published - Nov 2019
Event	2019 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2019 - San Diego, United States Duration: Nov 18 2019 → Nov 21 2019

Publication series

Name	Proceedings - 2019 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2019

Conference

Conference	2019 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2019
Country/Territory	United States
City	San Diego
Period	11/18/19 → 11/21/19

Bibliographical note

Publisher Copyright:
© 2019 IEEE.

Keywords

Breast Cancer
Convolutional Neural Networks
Full-Field Digital Mammography
Full-Volume Digital Breast Tomosynthesis

ASJC Scopus subject areas

Biochemistry
Biotechnology
Molecular Medicine
Modeling and Simulation
Health Informatics
Pharmacology (medical)
Public Health, Environmental and Occupational Health

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

Access to Document

10.1109/BIBM47256.2019.8983097

Cite this

Zhang, Y., Wang, X., Blanton, H., Liang, G., Xing, X., & Jacobs, N. (2019). 2D Convolutional Neural Networks for 3D Digital Breast Tomosynthesis Classification. In I. Yoo, J. Bi, & X. T. Hu (Eds.), Proceedings - 2019 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2019 (pp. 1013-1017). Article 8983097 (Proceedings - 2019 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2019). https://doi.org/10.1109/BIBM47256.2019.8983097

Zhang, Yu ; Wang, Xiaoqin ; Blanton, Hunter et al. / 2D Convolutional Neural Networks for 3D Digital Breast Tomosynthesis Classification. Proceedings - 2019 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2019. editor / Illhoi Yoo ; Jinbo Bi ; Xiaohua Tony Hu. 2019. pp. 1013-1017 (Proceedings - 2019 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2019).

@inproceedings{51fbe334b0894667a98614749a0e642d,

title = "2D Convolutional Neural Networks for 3D Digital Breast Tomosynthesis Classification",

abstract = "Automated methods for breast cancer detection have focused on 2D mammography and have largely ignored 3D digital breast tomosynthesis (DBT), which is frequently used in clinical practice. The two key challenges in developing automated methods for DBT classification are handling the variable number of slices and retaining slice-to-slice changes. We propose a novel deep 2D convolutional neural network (CNN) architecture for DBT classification that simultaneously overcomes both challenges. Our approach operates on the full volume, regardless of the number of slices, and allows the use of pre-trained 2D CNNs for feature extraction, which is important given the limited amount of annotated training data. In an extensive evaluation on a real-world clinical dataset, our approach achieves 0.854 auROC, which is 28.80% higher than approaches based on 3D CNNs. We also find that these improvements are stable across a range of model configurations.",

keywords = "Breast Cancer, Convolutional Neural Networks, Full-Field Digital Mammography, Full-Volume Digital Breast Tomosynthesis",

author = "Yu Zhang and Xiaoqin Wang and Hunter Blanton and Gongbo Liang and Xin Xing and Nathan Jacobs",

note = "Publisher Copyright: {\textcopyright} 2019 IEEE.; 2019 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2019 ; Conference date: 18-11-2019 Through 21-11-2019",

year = "2019",

month = nov,

doi = "10.1109/BIBM47256.2019.8983097",

language = "English",

series = "Proceedings - 2019 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2019",

pages = "1013--1017",

editor = "Illhoi Yoo and Jinbo Bi and Hu, {Xiaohua Tony}",

booktitle = "Proceedings - 2019 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2019",

}

Zhang, Y, Wang, X, Blanton, H, Liang, G, Xing, X & Jacobs, N 2019, 2D Convolutional Neural Networks for 3D Digital Breast Tomosynthesis Classification. in I Yoo, J Bi & XT Hu (eds), Proceedings - 2019 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2019., 8983097, Proceedings - 2019 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2019, pp. 1013-1017, 2019 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2019, San Diego, United States, 11/18/19. https://doi.org/10.1109/BIBM47256.2019.8983097

2D Convolutional Neural Networks for 3D Digital Breast Tomosynthesis Classification. / Zhang, Yu; Wang, Xiaoqin; Blanton, Hunter et al.
Proceedings - 2019 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2019. ed. / Illhoi Yoo; Jinbo Bi; Xiaohua Tony Hu. 2019. p. 1013-1017 8983097 (Proceedings - 2019 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2019).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - 2D Convolutional Neural Networks for 3D Digital Breast Tomosynthesis Classification

AU - Zhang, Yu

AU - Wang, Xiaoqin

AU - Blanton, Hunter

AU - Liang, Gongbo

AU - Xing, Xin

AU - Jacobs, Nathan

PY - 2019/11

Y1 - 2019/11

N2 - Automated methods for breast cancer detection have focused on 2D mammography and have largely ignored 3D digital breast tomosynthesis (DBT), which is frequently used in clinical practice. The two key challenges in developing automated methods for DBT classification are handling the variable number of slices and retaining slice-to-slice changes. We propose a novel deep 2D convolutional neural network (CNN) architecture for DBT classification that simultaneously overcomes both challenges. Our approach operates on the full volume, regardless of the number of slices, and allows the use of pre-trained 2D CNNs for feature extraction, which is important given the limited amount of annotated training data. In an extensive evaluation on a real-world clinical dataset, our approach achieves 0.854 auROC, which is 28.80% higher than approaches based on 3D CNNs. We also find that these improvements are stable across a range of model configurations.

AB - Automated methods for breast cancer detection have focused on 2D mammography and have largely ignored 3D digital breast tomosynthesis (DBT), which is frequently used in clinical practice. The two key challenges in developing automated methods for DBT classification are handling the variable number of slices and retaining slice-to-slice changes. We propose a novel deep 2D convolutional neural network (CNN) architecture for DBT classification that simultaneously overcomes both challenges. Our approach operates on the full volume, regardless of the number of slices, and allows the use of pre-trained 2D CNNs for feature extraction, which is important given the limited amount of annotated training data. In an extensive evaluation on a real-world clinical dataset, our approach achieves 0.854 auROC, which is 28.80% higher than approaches based on 3D CNNs. We also find that these improvements are stable across a range of model configurations.

KW - Breast Cancer

KW - Convolutional Neural Networks

KW - Full-Field Digital Mammography

KW - Full-Volume Digital Breast Tomosynthesis

UR - http://www.scopus.com/inward/record.url?scp=85084332273&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85084332273&partnerID=8YFLogxK

U2 - 10.1109/BIBM47256.2019.8983097

DO - 10.1109/BIBM47256.2019.8983097

M3 - Conference contribution

AN - SCOPUS:85084332273

T3 - Proceedings - 2019 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2019

SP - 1013

EP - 1017

BT - Proceedings - 2019 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2019

A2 - Yoo, Illhoi

A2 - Bi, Jinbo

A2 - Hu, Xiaohua Tony

T2 - 2019 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2019

Y2 - 18 November 2019 through 21 November 2019

ER -

Zhang Y, Wang X, Blanton H, Liang G, Xing X, Jacobs N. 2D Convolutional Neural Networks for 3D Digital Breast Tomosynthesis Classification. In Yoo I, Bi J, Hu XT, editors, Proceedings - 2019 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2019. 2019. p. 1013-1017. 8983097. (Proceedings - 2019 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2019). doi: 10.1109/BIBM47256.2019.8983097

2D Convolutional Neural Networks for 3D Digital Breast Tomosynthesis Classification

Abstract

Publication series

Conference

Bibliographical note

Keywords

ASJC Scopus subject areas

UN SDGs

Access to Document

Other files and links

Fingerprint

Cite this