TY - GEN
T1 - Improving brain tumor segmentation in multi-sequence MR images using cross-sequence MR image generation
AU - Zhao, Guojing
AU - Zhang, Jianpeng
AU - Xia, Yong
N1 - Publisher Copyright:
© Springer Nature Switzerland AG 2020.
PY - 2020
Y1 - 2020
N2 - Accurate brain tumor segmentation using multi-sequence magnetic resonance (MR) imaging plays a pivotal role in clinical practice and research settings. Despite their prevalence, deep learning-based segmentation methods, which usually use multiple MR sequences as input, still have limited performance, partly due to their insufficient ability to image representation. In this paper, we propose a brain tumor segmentation (BraTSeg) model, which uses cross-sequence MR image generation as a self-supervision tool to improve the segmentation accuracy. This model is an ensemble of three image segmentation and generation (ImgSG) models, which are designed for simultaneous segmentation of brain tumors and generation of T1, T2, and Flair sequences, respectively. We evaluated the proposed BraTSeg model on the BraTS 2019 dataset and achieved an average Dice similarity coefficient (DSC) of 81.93%, 87.80%, and 83.44% in the segmentation of enhancing tumor, whole tumor, and tumor score on the testing set, respectively. Our results suggest that using cross-sequence MR image generation is an effective self-supervision method that can improve the accuracy of brain tumor segmentation and the proposed BraTSeg model can produce satisfactory segmentation of brain tumors and intra-tumor structures.
AB - Accurate brain tumor segmentation using multi-sequence magnetic resonance (MR) imaging plays a pivotal role in clinical practice and research settings. Despite their prevalence, deep learning-based segmentation methods, which usually use multiple MR sequences as input, still have limited performance, partly due to their insufficient ability to image representation. In this paper, we propose a brain tumor segmentation (BraTSeg) model, which uses cross-sequence MR image generation as a self-supervision tool to improve the segmentation accuracy. This model is an ensemble of three image segmentation and generation (ImgSG) models, which are designed for simultaneous segmentation of brain tumors and generation of T1, T2, and Flair sequences, respectively. We evaluated the proposed BraTSeg model on the BraTS 2019 dataset and achieved an average Dice similarity coefficient (DSC) of 81.93%, 87.80%, and 83.44% in the segmentation of enhancing tumor, whole tumor, and tumor score on the testing set, respectively. Our results suggest that using cross-sequence MR image generation is an effective self-supervision method that can improve the accuracy of brain tumor segmentation and the proposed BraTSeg model can produce satisfactory segmentation of brain tumors and intra-tumor structures.
KW - Brain tumor segmentation
KW - Deep learning
KW - MR image generation
KW - Multi-sequence MR imaging
UR - http://www.scopus.com/inward/record.url?scp=85085509599&partnerID=8YFLogxK
U2 - 10.1007/978-3-030-46643-5_3
DO - 10.1007/978-3-030-46643-5_3
M3 - 会议稿件
AN - SCOPUS:85085509599
SN - 9783030466428
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 27
EP - 36
BT - Brainlesion
A2 - Crimi, Alessandro
A2 - Bakas, Spyridon
PB - Springer
T2 - 5th International MICCAI Brainlesion Workshop, BrainLes 2019, held in conjunction with the Medical Image Computing for Computer Assisted Intervention, MICCAI 2019
Y2 - 17 October 2019 through 17 October 2019
ER -