Temporally aware volumetric generative adversarial network-based MR image reconstruction with simultaneous respiratory motion compensation: Initial feasibility in 3D dynamic cine cardiac MRI.
Temporally aware volumetric generative adversarial network-based MR image reconstruction with simultaneous respiratory motion compensation: Initial feasibility in 3D dynamic cine cardiac MRI.
复制标题
DOI：
10.1002/mrm.28912
复制
发表时间：
2021-11
影响因子：
3.3
通讯作者：
Hu, Peng
中科院分区：
文献类型：
--
作者：
Ghodrati, Vahid;Bydder, Mark;Bedayat, Arash;Prosper, Ashley;Yoshida, Takegawa;Kim-Lien Nguyen;Finn, J. Paul;Hu, Peng
关键词：
开发了一种新的基于三维生成对抗网络(GAN)的4D MRI图像重建和呼吸运动补偿技术。我们的目标是实现高加速系数10.7x-15.8x，同时以较低的加速系数3.5x-7.9x保持优于最先进的自选通(SG)压缩传感小波(CS-WV)重建的健壮和诊断图像质量。我们的GAN基于像素级内容损失函数、对抗性损失函数和一种新的数据驱动的时间感知损失函数进行训练，以保持解剖的准确性和时间的一致性。除了图像重建，我们的网络还为自由呼吸扫描执行呼吸运动补偿。采用了一种新的基于渐进生长的策略，使所提出的GaN基结构的训练过程成为可能。该方法是基于42名患者的3D心脏电影数据开发并进行了全面的定性和定量评估。在10.7x-15.8x加速比3.5x-7.9x加速比SG CS-WV方法(4.53±0.540比3.13±0.681，4.12±0.429比2.97±0.434，P<；0.05)下，我们提出的方法在一般图像质量和图像伪影方面获得了明显更好的分数。在我们的图像中没有观察到虚假的解剖结构。该方法能够实现与传统SG CS-WV相似的心功能量化。与SG CS-WV(312秒/心脏时相)相比，该方法获得了更快的基于CPU的图像重建(6秒/心脏时相)。提出的方法在高分辨率(1mm3)自由呼吸4D磁共振数据采集中显示出良好的潜力，同时具有呼吸运动补偿和快速重建时间。
Develop a novel 3D generative adversarial network (GAN)-based technique for simultaneous image reconstruction and respiratory motion compensation of 4D MRI. Our goal was to enable high acceleration factors 10.7X-15.8X while maintaining robust and diagnostic image quality superior to state-of-the-art self-gating (SG) compressed sensing wavelet (CS-WV) reconstruction at lower acceleration factors 3.5X-7.9X. Our GAN was trained based on pixel-wise content loss functions, adversarial loss function, and a novel data-driven temporal aware loss function in order to maintain anatomical accuracy and temporal coherence. Besides image reconstruction, our network also performs respiratory motion compensation for free-breathing scans. A novel progressive growing based strategy was adapted to make the training process possible for the proposed GAN-based structure. The proposed method was developed and thoroughly evaluated qualitatively and quantitatively based on 3D cardiac cine data from 42 patients. Our proposed method achieved significantly better scores in general image quality and image artifacts at 10.7X-15.8X acceleration than the SG CS-WV approach at 3.5X-7.9X acceleration (4.53±0.540 vs. 3.13±0.681 for general image quality, 4.12±0.429 vs. 2.97±0.434 for image artifacts, p<0.05 for both). No spurious anatomical structures were observed in our images. The proposed method enabled similar cardiac function quantification as conventional SG CS-WV. The proposed method achieved faster CPU-based image reconstruction (6 sec/cardiac phase) than the SG CS-WV (312 sec/cardiac phase). The proposed method showed promising potential for high-resolution (1mm3) free-breathing 4D MR data acquisition with simultaneous respiratory motion compensation and fast reconstruction time.
登录
查看更多内容
影响因子：
3.3
作者：
Haskell, Melissa W.;Cauley, Stephen F.;Wald, Lawrence L.
通讯作者：
Wald, Lawrence L.
影响因子：
3.3
作者：
Hammernik K;Klatzer T;Kobler E;Recht MP;Sodickson DK;Pock T;Knoll F
通讯作者：
Knoll F
影响因子：
3.3
作者：
Eo, Taejoon;Jun, Yohan;Hwang, Dosik
通讯作者：
Hwang, Dosik
实时心血管MR具有时空伪影抑制，使用先天性心脏病中的概念深度学习。
影响因子：
3.3
作者：
Hauptmann A;Arridge S;Lucka F;Muthurangu V;Steeden JA
通讯作者：
Steeden JA
影响因子：
3.3
作者：
Griswold, MA;Jakob, PM;Haase, A
通讯作者：
Haase, A