TY - JOUR
T1 - Multi-stage learning for segmentation of aortic dissections using a prior aortic anatomy simplification
AU - Chen, Duanduan
AU - Zhang, Xuyang
AU - Mei, Yuqian
AU - Liao, Fangzhou
AU - Xu, Huanming
AU - Li, Zhenfeng
AU - Xiao, Qianjiang
AU - Guo, Wei
AU - Zhang, Hongkun
AU - Yan, Tianyi
AU - Xiong, Jiang
AU - Ventikos, Yiannis
N1 - Funding Information:
This study was supported by the National Key R&D Program of China (2018AAA0102600 and 2018YFC0115400), the National Natural Science Foundation of China (81970404, 81770465 and 81911530224), the Beijing Municipal Science & Technology Commission (Z191100010618004), Beijing Natural Science Foundation (Z190014 and L192010), the Key R&D Program of Zhejiang Province (2019C03013) and the Beijing Nova Program (Z181100006218008). The authors would like to thank Shukun (Beijing) Network Technology Co. Ltd. for providing additional computing resources and technical support.
Funding Information:
This study was supported by the National Key R&D Program of China ( 2018AAA0102600 and 2018YFC0115400 ), the National Natural Science Foundation of China ( 81970404 , 81770465 and 81911530224 ), the Beijing Municipal Science & Technology Commission ( Z191100010618004 ), Beijing Natural Science Foundation ( Z190014 and L192010 ), the Key R&D Program of Zhejiang Province ( 2019C03013 ) and the Beijing Nova Program ( Z181100006218008 ). The authors would like to thank Shukun (Beijing) Network Technology Co., Ltd. for providing additional computing resources and technical support.
Publisher Copyright:
© 2020
PY - 2021/4
Y1 - 2021/4
N2 - Aortic dissection (AD) is a life-threatening cardiovascular disease with a high mortality rate. The accurate and generalized 3-D reconstruction of AD from CT-angiography can effectively assist clinical procedures and surgery plans, however, is clinically unavaliable due to the lacking of efficient tools. In this study, we presented a novel multi-stage segmentation framework for type B AD to extract true lumen (TL), false lumen (FL) and all branches (BR) as different classes. Two cascaded neural networks were used to segment the aortic trunk and branches and to separate the dual lumen, respectively. An aortic straightening method was designed based on the prior vascular anatomy of AD, simplifying the curved aortic shape before the second network. The straightening-based method achieved the mean Dice scores of 0.96, 0.95 and 0.89 for TL, FL, and BR on a multi-center dataset involving 120 patients, outperforming the end-to-end multi-class methods and the multi-stage methods without straightening on the dual-lumen segmentation, even using different network architectures. Both the global volumetric features of the aorta and the local characteristics of the primary tear could be better identified and quantified based on the straightening. Comparing to previous deep learning methods dealing with AD segmentations, the proposed framework presented advantages in segmentation accuracy.
AB - Aortic dissection (AD) is a life-threatening cardiovascular disease with a high mortality rate. The accurate and generalized 3-D reconstruction of AD from CT-angiography can effectively assist clinical procedures and surgery plans, however, is clinically unavaliable due to the lacking of efficient tools. In this study, we presented a novel multi-stage segmentation framework for type B AD to extract true lumen (TL), false lumen (FL) and all branches (BR) as different classes. Two cascaded neural networks were used to segment the aortic trunk and branches and to separate the dual lumen, respectively. An aortic straightening method was designed based on the prior vascular anatomy of AD, simplifying the curved aortic shape before the second network. The straightening-based method achieved the mean Dice scores of 0.96, 0.95 and 0.89 for TL, FL, and BR on a multi-center dataset involving 120 patients, outperforming the end-to-end multi-class methods and the multi-stage methods without straightening on the dual-lumen segmentation, even using different network architectures. Both the global volumetric features of the aorta and the local characteristics of the primary tear could be better identified and quantified based on the straightening. Comparing to previous deep learning methods dealing with AD segmentations, the proposed framework presented advantages in segmentation accuracy.
KW - Aortic dissection
KW - CT-angiography
KW - Deep learning
KW - Prior anatomy simplification
KW - Segmentation
UR - http://www.scopus.com/inward/record.url?scp=85101164309&partnerID=8YFLogxK
U2 - 10.1016/j.media.2020.101931
DO - 10.1016/j.media.2020.101931
M3 - Article
C2 - 33618153
AN - SCOPUS:85101164309
SN - 1361-8415
VL - 69
JO - Medical Image Analysis
JF - Medical Image Analysis
M1 - 101931
ER -