TY - JOUR
T1 - A global benchmark of algorithms for segmenting the left atrium from late gadolinium-enhanced cardiac magnetic resonance imaging
AU - Xiong, Zhaohan
AU - Xia, Qing
AU - Hu, Zhiqiang
AU - Huang, Ning
AU - Bian, Cheng
AU - Zheng, Yefeng
AU - Vesal, Sulaiman
AU - Ravikumar, Nishant
AU - Maier, Andreas
AU - Yang, Xin
AU - Heng, Pheng Ann
AU - Ni, Dong
AU - Li, Caizi
AU - Tong, Qianqian
AU - Si, Weixin
AU - Puybareau, Elodie
AU - Khoudli, Younes
AU - Géraud, Thierry
AU - Chen, Chen
AU - Bai, Wenjia
AU - Rueckert, Daniel
AU - Xu, Lingchao
AU - Zhuang, Xiahai
AU - Luo, Xinzhe
AU - Jia, Shuman
AU - Sermesant, Maxime
AU - Liu, Yashu
AU - Wang, Kuanquan
AU - Borra, Davide
AU - Masci, Alessandro
AU - Corsi, Cristiana
AU - de Vente, Coen
AU - Veta, Mitko
AU - Karim, Rashed
AU - Preetha, Chandrakanth Jayachandran
AU - Engelhardt, Sandy
AU - Qiao, Menyun
AU - Wang, Yuanyuan
AU - Tao, Qian
AU - Nuñez-Garcia, Marta
AU - Camara, Oscar
AU - Savioli, Nicolo
AU - Lamata, Pablo
AU - Zhao, Jichao
N1 - Funding Information:
The authors would like to thank Nvidia, MedTech CoRE New Zealand, and Arterys for providing prizes for the winners of the 2018 LA Segmentation Challenge. Z.X. and J.Z. are grateful for Nvidia for donating Titan-X Pascal GPU for algorithm development and testing, and the NIH/NIGMS Center for Integrative Biomedical Computing (CIBC) at the University of Utah for providing the LGE-MRI dataset. This work was funded by the Health Research Council of New Zealand [ #16/385 ].
Funding Information:
The authors would like to thank Nvidia, MedTech CoRE New Zealand, and Arterys for providing prizes for the winners of the 2018 LA Segmentation Challenge. Z.X. and J.Z. are grateful for Nvidia for donating Titan-X Pascal GPU for algorithm development and testing, and the NIH/NIGMS Center for Integrative Biomedical Computing (CIBC) at the University of Utah for providing the LGE-MRI dataset. This work was funded by the Health Research Council of New Zealand [#16/385].
Publisher Copyright:
© 2020 Elsevier B.V.
Copyright:
Copyright 2020 Elsevier B.V., All rights reserved.
PY - 2021/1
Y1 - 2021/1
N2 - Segmentation of medical images, particularly late gadolinium-enhanced magnetic resonance imaging (LGE-MRI) used for visualizing diseased atrial structures, is a crucial first step for ablation treatment of atrial fibrillation. However, direct segmentation of LGE-MRIs is challenging due to the varying intensities caused by contrast agents. Since most clinical studies have relied on manual, labor-intensive approaches, automatic methods are of high interest, particularly optimized machine learning approaches. To address this, we organized the 2018 Left Atrium Segmentation Challenge using 154 3D LGE-MRIs, currently the world's largest atrial LGE-MRI dataset, and associated labels of the left atrium segmented by three medical experts, ultimately attracting the participation of 27 international teams. In this paper, extensive analysis of the submitted algorithms using technical and biological metrics was performed by undergoing subgroup analysis and conducting hyper-parameter analysis, offering an overall picture of the major design choices of convolutional neural networks (CNNs) and practical considerations for achieving state-of-the-art left atrium segmentation. Results show that the top method achieved a Dice score of 93.2% and a mean surface to surface distance of 0.7 mm, significantly outperforming prior state-of-the-art. Particularly, our analysis demonstrated that double sequentially used CNNs, in which a first CNN is used for automatic region-of-interest localization and a subsequent CNN is used for refined regional segmentation, achieved superior results than traditional methods and machine learning approaches containing single CNNs. This large-scale benchmarking study makes a significant step towards much-improved segmentation methods for atrial LGE-MRIs, and will serve as an important benchmark for evaluating and comparing the future works in the field. Furthermore, the findings from this study can potentially be extended to other imaging datasets and modalities, having an impact on the wider medical imaging community.
AB - Segmentation of medical images, particularly late gadolinium-enhanced magnetic resonance imaging (LGE-MRI) used for visualizing diseased atrial structures, is a crucial first step for ablation treatment of atrial fibrillation. However, direct segmentation of LGE-MRIs is challenging due to the varying intensities caused by contrast agents. Since most clinical studies have relied on manual, labor-intensive approaches, automatic methods are of high interest, particularly optimized machine learning approaches. To address this, we organized the 2018 Left Atrium Segmentation Challenge using 154 3D LGE-MRIs, currently the world's largest atrial LGE-MRI dataset, and associated labels of the left atrium segmented by three medical experts, ultimately attracting the participation of 27 international teams. In this paper, extensive analysis of the submitted algorithms using technical and biological metrics was performed by undergoing subgroup analysis and conducting hyper-parameter analysis, offering an overall picture of the major design choices of convolutional neural networks (CNNs) and practical considerations for achieving state-of-the-art left atrium segmentation. Results show that the top method achieved a Dice score of 93.2% and a mean surface to surface distance of 0.7 mm, significantly outperforming prior state-of-the-art. Particularly, our analysis demonstrated that double sequentially used CNNs, in which a first CNN is used for automatic region-of-interest localization and a subsequent CNN is used for refined regional segmentation, achieved superior results than traditional methods and machine learning approaches containing single CNNs. This large-scale benchmarking study makes a significant step towards much-improved segmentation methods for atrial LGE-MRIs, and will serve as an important benchmark for evaluating and comparing the future works in the field. Furthermore, the findings from this study can potentially be extended to other imaging datasets and modalities, having an impact on the wider medical imaging community.
KW - Convolutional neural networks
KW - Image segmentation
KW - Late gadolinium-enhanced magnetic resonance imaging
KW - Left atrium
UR - http://www.scopus.com/inward/record.url?scp=85095460279&partnerID=8YFLogxK
U2 - 10.1016/j.media.2020.101832
DO - 10.1016/j.media.2020.101832
M3 - Article
AN - SCOPUS:85095460279
SN - 1361-8415
VL - 67
JO - Medical Image Analysis
JF - Medical Image Analysis
M1 - 101832
ER -