Deep Learning for Fully Automatic Tumor Segmentation on Serially Acquired Dynamic Contrast-Enhanced MRI Images of Triple-Negative Breast Cancer

Zhan Xu; David E. Rauch; Rania M. Mohamed; Sanaz Pashapoor; Zijian Zhou; Bikash Panthi; Jong Bum Son; Ken Pin Hwang; Benjamin C. Musall; Beatriz E. Adrada; Rosalind P. Candelaria; Jessica W.T. Leung; Huong T.C. Le-Petross; Deanna L. Lane; Frances Perez; Jason White; Alyson Clayborn; Brandy Reed; Huiqin Chen; Jia Sun; Peng Wei; Alastair Thompson; Anil Korkut; Lei Huo; Kelly K. Hunt; Jennifer K. Litton; Vicente Valero; Debu Tripathy; Wei Yang; Clinton Yam; Jingfei Ma

doi:10.3390/cancers15194829

Deep Learning for Fully Automatic Tumor Segmentation on Serially Acquired Dynamic Contrast-Enhanced MRI Images of Triple-Negative Breast Cancer

Zhan Xu, David E. Rauch, Rania M. Mohamed, Sanaz Pashapoor, Zijian Zhou, Bikash Panthi, Jong Bum Son, Ken Pin Hwang, Benjamin C. Musall, Beatriz E. Adrada, Rosalind P. Candelaria, Jessica W.T. Leung, Huong T.C. Le-Petross, Deanna L. Lane, Frances Perez, Jason White, Alyson Clayborn, Brandy Reed, Huiqin Chen, Jia SunPeng Wei, Alastair Thompson, Anil Korkut, Lei Huo, Kelly K. Hunt, Jennifer K. Litton, Vicente Valero, Debu Tripathy, Wei Yang, Clinton Yam, Jingfei Ma

Research output: Contribution to journal › Article › peer-review

Abstract

Accurate tumor segmentation is required for quantitative image analyses, which are increasingly used for evaluation of tumors. We developed a fully automated and high-performance segmentation model of triple-negative breast cancer using a self-configurable deep learning framework and a large set of dynamic contrast-enhanced MRI images acquired serially over the patients’ treatment course. Among all models, the top-performing one that was trained with the images across different time points of a treatment course yielded a Dice similarity coefficient of 93% and a sensitivity of 96% on baseline images. The top-performing model also produced accurate tumor size measurements, which is valuable for practical clinical applications.

Original language	English (US)
Article number	4829
Journal	Cancers
Volume	15
Issue number	19
DOIs	https://doi.org/10.3390/cancers15194829
State	Published - Oct 2023

Keywords

deep learning
triple-negative breast cancer
tumor segmentation

ASJC Scopus subject areas

Oncology
Cancer Research

Access to Document

10.3390/cancers15194829

Cite this

Xu, Z., Rauch, D. E., Mohamed, R. M., Pashapoor, S., Zhou, Z., Panthi, B., Son, J. B., Hwang, K. P., Musall, B. C., Adrada, B. E., Candelaria, R. P., Leung, J. W. T., Le-Petross, H. T. C., Lane, D. L., Perez, F., White, J., Clayborn, A., Reed, B., Chen, H., ... Ma, J. (2023). Deep Learning for Fully Automatic Tumor Segmentation on Serially Acquired Dynamic Contrast-Enhanced MRI Images of Triple-Negative Breast Cancer. Cancers, 15(19), Article 4829. https://doi.org/10.3390/cancers15194829

Xu, Z, Rauch, DE, Mohamed, RM, Pashapoor, S, Zhou, Z, Panthi, B, Son, JB, Hwang, KP, Musall, BC, Adrada, BE , Candelaria, RP , Leung, JWT , Le-Petross, HTC , Lane, DL , Perez, F, White, J, Clayborn, A, Reed, B, Chen, H , Sun, J , Wei, P, Thompson, A, Korkut, A , Huo, L , Hunt, KK , Litton, JK , Valero, V , Tripathy, D, Yang, W, Yam, C & Ma, J 2023, 'Deep Learning for Fully Automatic Tumor Segmentation on Serially Acquired Dynamic Contrast-Enhanced MRI Images of Triple-Negative Breast Cancer', Cancers, vol. 15, no. 19, 4829. https://doi.org/10.3390/cancers15194829

@article{f8c1b738910c4ba898f293189997e268,

title = "Deep Learning for Fully Automatic Tumor Segmentation on Serially Acquired Dynamic Contrast-Enhanced MRI Images of Triple-Negative Breast Cancer",

abstract = "Accurate tumor segmentation is required for quantitative image analyses, which are increasingly used for evaluation of tumors. We developed a fully automated and high-performance segmentation model of triple-negative breast cancer using a self-configurable deep learning framework and a large set of dynamic contrast-enhanced MRI images acquired serially over the patients{\textquoteright} treatment course. Among all models, the top-performing one that was trained with the images across different time points of a treatment course yielded a Dice similarity coefficient of 93% and a sensitivity of 96% on baseline images. The top-performing model also produced accurate tumor size measurements, which is valuable for practical clinical applications.",

keywords = "deep learning, triple-negative breast cancer, tumor segmentation",

author = "Zhan Xu and Rauch, {David E.} and Mohamed, {Rania M.} and Sanaz Pashapoor and Zijian Zhou and Bikash Panthi and Son, {Jong Bum} and Hwang, {Ken Pin} and Musall, {Benjamin C.} and Adrada, {Beatriz E.} and Candelaria, {Rosalind P.} and Leung, {Jessica W.T.} and Le-Petross, {Huong T.C.} and Lane, {Deanna L.} and Frances Perez and Jason White and Alyson Clayborn and Brandy Reed and Huiqin Chen and Jia Sun and Peng Wei and Alastair Thompson and Anil Korkut and Lei Huo and Hunt, {Kelly K.} and Litton, {Jennifer K.} and Vicente Valero and Debu Tripathy and Wei Yang and Clinton Yam and Jingfei Ma",

note = "Publisher Copyright: {\textcopyright} 2023 by the authors.",

year = "2023",

month = oct,

doi = "10.3390/cancers15194829",

language = "English (US)",

volume = "15",

journal = "Cancers",

issn = "2072-6694",

publisher = "Multidisciplinary Digital Publishing Institute (MDPI)",

number = "19",

}

TY - JOUR

T1 - Deep Learning for Fully Automatic Tumor Segmentation on Serially Acquired Dynamic Contrast-Enhanced MRI Images of Triple-Negative Breast Cancer

AU - Xu, Zhan

AU - Rauch, David E.

AU - Mohamed, Rania M.

AU - Pashapoor, Sanaz

AU - Zhou, Zijian

AU - Panthi, Bikash

AU - Son, Jong Bum

AU - Hwang, Ken Pin

AU - Musall, Benjamin C.

AU - Adrada, Beatriz E.

AU - Candelaria, Rosalind P.

AU - Leung, Jessica W.T.

AU - Le-Petross, Huong T.C.

AU - Lane, Deanna L.

AU - Perez, Frances

AU - White, Jason

AU - Clayborn, Alyson

AU - Reed, Brandy

AU - Chen, Huiqin

AU - Sun, Jia

AU - Wei, Peng

AU - Thompson, Alastair

AU - Korkut, Anil

AU - Huo, Lei

AU - Hunt, Kelly K.

AU - Litton, Jennifer K.

AU - Valero, Vicente

AU - Tripathy, Debu

AU - Yang, Wei

AU - Yam, Clinton

AU - Ma, Jingfei

PY - 2023/10

Y1 - 2023/10

N2 - Accurate tumor segmentation is required for quantitative image analyses, which are increasingly used for evaluation of tumors. We developed a fully automated and high-performance segmentation model of triple-negative breast cancer using a self-configurable deep learning framework and a large set of dynamic contrast-enhanced MRI images acquired serially over the patients’ treatment course. Among all models, the top-performing one that was trained with the images across different time points of a treatment course yielded a Dice similarity coefficient of 93% and a sensitivity of 96% on baseline images. The top-performing model also produced accurate tumor size measurements, which is valuable for practical clinical applications.

AB - Accurate tumor segmentation is required for quantitative image analyses, which are increasingly used for evaluation of tumors. We developed a fully automated and high-performance segmentation model of triple-negative breast cancer using a self-configurable deep learning framework and a large set of dynamic contrast-enhanced MRI images acquired serially over the patients’ treatment course. Among all models, the top-performing one that was trained with the images across different time points of a treatment course yielded a Dice similarity coefficient of 93% and a sensitivity of 96% on baseline images. The top-performing model also produced accurate tumor size measurements, which is valuable for practical clinical applications.

KW - deep learning

KW - triple-negative breast cancer

KW - tumor segmentation

UR - http://www.scopus.com/inward/record.url?scp=85173833518&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85173833518&partnerID=8YFLogxK

U2 - 10.3390/cancers15194829

DO - 10.3390/cancers15194829

M3 - Article

C2 - 37835523

AN - SCOPUS:85173833518

SN - 2072-6694

VL - 15

JO - Cancers

JF - Cancers

IS - 19

M1 - 4829

ER -

Deep Learning for Fully Automatic Tumor Segmentation on Serially Acquired Dynamic Contrast-Enhanced MRI Images of Triple-Negative Breast Cancer

Abstract

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this