Densely Deep Supervised Networks with Threshold Loss for Cancer Detection in Automated Breast Ultrasound

Na Wang; Cheng Bian; Yi Wang; Min Xu; Chenchen Qin; Xin Yang; Tianfu Wang; Anhua Li; Dinggang Shen; Dong Ni

doi:10.1007/978-3-030-00937-3_73

Densely Deep Supervised Networks with Threshold Loss for Cancer Detection in Automated Breast Ultrasound

Na Wang, Cheng Bian, Yi Wang, Min Xu, Chenchen Qin, Xin Yang, Tianfu Wang, Anhua Li, Dinggang Shen, Dong Ni

Department of Brain and Cognitive Engineering

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

26 Citations (Scopus)

Abstract

Automated breast ultrasound (ABUS) is a new and promising tool for diagnosing breast cancer. However, reviewing ABUS images is extremely time-consuming and oversight errors could happen. We propose a novel 3D convolutional network for automatic cancer detection in ABUS. Our contribution is twofold. First, we propose a threshold loss function to provide voxel-level adaptive threshold for discriminating cancer and non-cancer, thus achieving high sensitivity with low FPs. Second, we propose a densely deep supervision (DDS) mechanism to improve the sensitivity significantly by utilizing multi-scale discriminative features of all layers. Both class-balanced cross entropy loss and overlap loss are employed to enhance DDS performance. The efficacy of the proposed network is validated on a dataset of 196 patients with 661 cancer regions. The 4-fold cross-validation experiments show our network obtains a sensitivity of 93% with 2.2 FPs per ABUS volume. Our proposed novel network can provide an accurate and automatic cancer detection tool for breast cancer screening by maintaining high sensitivity with low FPs.

Original language	English
Title of host publication	Medical Image Computing and Computer Assisted Intervention – MICCAI 2018 - 21st International Conference, 2018, Proceedings
Editors	Alejandro F. Frangi, Gabor Fichtinger, Julia A. Schnabel, Carlos Alberola-López, Christos Davatzikos
Publisher	Springer Verlag
Pages	641-648
Number of pages	8
ISBN (Print)	9783030009366
DOIs	https://doi.org/10.1007/978-3-030-00937-3_73
Publication status	Published - 2018
Event	21st International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2018 - Granada, Spain Duration: 2018 Sept 16 → 2018 Sept 20

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	11073 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Other

Other	21st International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2018
Country/Territory	Spain
City	Granada
Period	18/9/16 → 18/9/20

Bibliographical note

Funding Information:
Acknowledgments. This work was supported in part by the National Natural Science Foundation of China under Grant 61571304, Grant 61701312, and in part by the Shenzhen Peacock Plan under Grant KQTD2016053112051497.

Publisher Copyright:
© 2018, Springer Nature Switzerland AG.

ASJC Scopus subject areas

Theoretical Computer Science
General Computer Science

UN SDGs

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

Access to Document

10.1007/978-3-030-00937-3_73

Cite this

Wang, N., Bian, C., Wang, Y., Xu, M., Qin, C., Yang, X., Wang, T., Li, A., Shen, D., & Ni, D. (2018). Densely Deep Supervised Networks with Threshold Loss for Cancer Detection in Automated Breast Ultrasound. In A. F. Frangi, G. Fichtinger, J. A. Schnabel, C. Alberola-López, & C. Davatzikos (Eds.), Medical Image Computing and Computer Assisted Intervention – MICCAI 2018 - 21st International Conference, 2018, Proceedings (pp. 641-648). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11073 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-030-00937-3_73

Densely Deep Supervised Networks with Threshold Loss for Cancer Detection in Automated Breast Ultrasound. / Wang, Na; Bian, Cheng; Wang, Yi et al.
Medical Image Computing and Computer Assisted Intervention – MICCAI 2018 - 21st International Conference, 2018, Proceedings. ed. / Alejandro F. Frangi; Gabor Fichtinger; Julia A. Schnabel; Carlos Alberola-López; Christos Davatzikos. Springer Verlag, 2018. p. 641-648 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11073 LNCS).

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

Wang, N, Bian, C, Wang, Y, Xu, M, Qin, C, Yang, X, Wang, T, Li, A, Shen, D & Ni, D 2018, Densely Deep Supervised Networks with Threshold Loss for Cancer Detection in Automated Breast Ultrasound. in AF Frangi, G Fichtinger, JA Schnabel, C Alberola-López & C Davatzikos (eds), Medical Image Computing and Computer Assisted Intervention – MICCAI 2018 - 21st International Conference, 2018, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 11073 LNCS, Springer Verlag, pp. 641-648, 21st International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2018, Granada, Spain, 18/9/16. https://doi.org/10.1007/978-3-030-00937-3_73

Wang N, Bian C, Wang Y, Xu M, Qin C, Yang X et al. Densely Deep Supervised Networks with Threshold Loss for Cancer Detection in Automated Breast Ultrasound. In Frangi AF, Fichtinger G, Schnabel JA, Alberola-López C, Davatzikos C, editors, Medical Image Computing and Computer Assisted Intervention – MICCAI 2018 - 21st International Conference, 2018, Proceedings. Springer Verlag. 2018. p. 641-648. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-030-00937-3_73

Wang, Na ; Bian, Cheng ; Wang, Yi et al. / Densely Deep Supervised Networks with Threshold Loss for Cancer Detection in Automated Breast Ultrasound. Medical Image Computing and Computer Assisted Intervention – MICCAI 2018 - 21st International Conference, 2018, Proceedings. editor / Alejandro F. Frangi ; Gabor Fichtinger ; Julia A. Schnabel ; Carlos Alberola-López ; Christos Davatzikos. Springer Verlag, 2018. pp. 641-648 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{9014beeb8ab44e58a584b8251db002fb,

title = "Densely Deep Supervised Networks with Threshold Loss for Cancer Detection in Automated Breast Ultrasound",

abstract = "Automated breast ultrasound (ABUS) is a new and promising tool for diagnosing breast cancer. However, reviewing ABUS images is extremely time-consuming and oversight errors could happen. We propose a novel 3D convolutional network for automatic cancer detection in ABUS. Our contribution is twofold. First, we propose a threshold loss function to provide voxel-level adaptive threshold for discriminating cancer and non-cancer, thus achieving high sensitivity with low FPs. Second, we propose a densely deep supervision (DDS) mechanism to improve the sensitivity significantly by utilizing multi-scale discriminative features of all layers. Both class-balanced cross entropy loss and overlap loss are employed to enhance DDS performance. The efficacy of the proposed network is validated on a dataset of 196 patients with 661 cancer regions. The 4-fold cross-validation experiments show our network obtains a sensitivity of 93% with 2.2 FPs per ABUS volume. Our proposed novel network can provide an accurate and automatic cancer detection tool for breast cancer screening by maintaining high sensitivity with low FPs.",

author = "Na Wang and Cheng Bian and Yi Wang and Min Xu and Chenchen Qin and Xin Yang and Tianfu Wang and Anhua Li and Dinggang Shen and Dong Ni",

note = "Funding Information: Acknowledgments. This work was supported in part by the National Natural Science Foundation of China under Grant 61571304, Grant 61701312, and in part by the Shenzhen Peacock Plan under Grant KQTD2016053112051497. Publisher Copyright: {\textcopyright} 2018, Springer Nature Switzerland AG.; 21st International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2018 ; Conference date: 16-09-2018 Through 20-09-2018",

year = "2018",

doi = "10.1007/978-3-030-00937-3_73",

language = "English",

isbn = "9783030009366",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Verlag",

pages = "641--648",

editor = "Frangi, {Alejandro F.} and Gabor Fichtinger and Schnabel, {Julia A.} and Carlos Alberola-L{\'o}pez and Christos Davatzikos",

booktitle = "Medical Image Computing and Computer Assisted Intervention – MICCAI 2018 - 21st International Conference, 2018, Proceedings",

}

TY - GEN

T1 - Densely Deep Supervised Networks with Threshold Loss for Cancer Detection in Automated Breast Ultrasound

AU - Wang, Na

AU - Bian, Cheng

AU - Wang, Yi

AU - Xu, Min

AU - Qin, Chenchen

AU - Yang, Xin

AU - Wang, Tianfu

AU - Li, Anhua

AU - Shen, Dinggang

AU - Ni, Dong

N1 - Funding Information: Acknowledgments. This work was supported in part by the National Natural Science Foundation of China under Grant 61571304, Grant 61701312, and in part by the Shenzhen Peacock Plan under Grant KQTD2016053112051497. Publisher Copyright: © 2018, Springer Nature Switzerland AG.

PY - 2018

Y1 - 2018

N2 - Automated breast ultrasound (ABUS) is a new and promising tool for diagnosing breast cancer. However, reviewing ABUS images is extremely time-consuming and oversight errors could happen. We propose a novel 3D convolutional network for automatic cancer detection in ABUS. Our contribution is twofold. First, we propose a threshold loss function to provide voxel-level adaptive threshold for discriminating cancer and non-cancer, thus achieving high sensitivity with low FPs. Second, we propose a densely deep supervision (DDS) mechanism to improve the sensitivity significantly by utilizing multi-scale discriminative features of all layers. Both class-balanced cross entropy loss and overlap loss are employed to enhance DDS performance. The efficacy of the proposed network is validated on a dataset of 196 patients with 661 cancer regions. The 4-fold cross-validation experiments show our network obtains a sensitivity of 93% with 2.2 FPs per ABUS volume. Our proposed novel network can provide an accurate and automatic cancer detection tool for breast cancer screening by maintaining high sensitivity with low FPs.

AB - Automated breast ultrasound (ABUS) is a new and promising tool for diagnosing breast cancer. However, reviewing ABUS images is extremely time-consuming and oversight errors could happen. We propose a novel 3D convolutional network for automatic cancer detection in ABUS. Our contribution is twofold. First, we propose a threshold loss function to provide voxel-level adaptive threshold for discriminating cancer and non-cancer, thus achieving high sensitivity with low FPs. Second, we propose a densely deep supervision (DDS) mechanism to improve the sensitivity significantly by utilizing multi-scale discriminative features of all layers. Both class-balanced cross entropy loss and overlap loss are employed to enhance DDS performance. The efficacy of the proposed network is validated on a dataset of 196 patients with 661 cancer regions. The 4-fold cross-validation experiments show our network obtains a sensitivity of 93% with 2.2 FPs per ABUS volume. Our proposed novel network can provide an accurate and automatic cancer detection tool for breast cancer screening by maintaining high sensitivity with low FPs.

UR - http://www.scopus.com/inward/record.url?scp=85053855553&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-00937-3_73

DO - 10.1007/978-3-030-00937-3_73

M3 - Conference contribution

AN - SCOPUS:85053855553

SN - 9783030009366

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 641

EP - 648

BT - Medical Image Computing and Computer Assisted Intervention – MICCAI 2018 - 21st International Conference, 2018, Proceedings

A2 - Frangi, Alejandro F.

A2 - Fichtinger, Gabor

A2 - Schnabel, Julia A.

A2 - Alberola-López, Carlos

A2 - Davatzikos, Christos

PB - Springer Verlag

T2 - 21st International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2018

Y2 - 16 September 2018 through 20 September 2018

ER -

Densely Deep Supervised Networks with Threshold Loss for Cancer Detection in Automated Breast Ultrasound

Abstract

Publication series

Other

Bibliographical note

ASJC Scopus subject areas

UN SDGs

Access to Document

Other files and links

Fingerprint

Cite this