Rapid 4D-MRI reconstruction using a deep radial convolutional neural network: Dracula.
Rapid 4D-MRI reconstruction using a deep radial convolutional neural network: Dracula.
复制标题
使用深度径向卷积神经网络的快速4D-MRI重建：德古拉（此处“Dracula”可能是特定的项目名称、算法名称或有其他特定背景含义，如果有更多相关信息可能会翻译得更准确）
DOI：
10.1016/j.radonc.2021.03.034
复制
发表时间：
2021-06
期刊：
影响因子：
--
通讯作者：
Wetscherek A
中科院分区：
文献类型：
--
作者：
Freedman JN;Gurney-Champion OJ;Nill S;Shiarli AM;Bainbridge HE;Mandeville HC;Koh DM;McDonald F;Kachelrieß M;Oelfke U;Wetscherek A
深度学习加速用于在线自适应磁共振引导放射治疗的4D - MRI重建
28秒内首次重建16个时相的高分辨率全胸部4D - MRI
首次在28秒内使用深度卷积神经网络（dCNNs）从欠采样的4D - MRI进行中间位置重建
基于深度学习的肿瘤中间位置与参考位置高度吻合
4D和中间位置MRI可为肺部和腹部磁共振引导放射治疗中的计划调整提供信息。我们提出基于深度学习的解决方案，在保持高图像质量和短扫描时间的同时克服4D - MRI重建时间过长的问题。
训练了两个3D U - net深度卷积神经网络以加速4D联合运动校正 - 高维全变分（MoCo - HDTV）重建。对于第一个网络，网格化的和联合MoCo - HDTV重建的4D - MRI分别用作输入和目标数据，而第二个网络被训练用于直接计算中间位置图像。对于这两个网络，输入和目标数据的维度为256×256体素（二维）和16个呼吸时相。使用结构相似性指数（SSIM）和自然度图像质量评估器（NIQE），将基于深度学习的MRI与联合MoCo - HDTV重建的MRI进行验证。此外，两名经验丰富的观察者在一项盲法研究中勾勒出肿瘤大体体积并对图像进行评分。
对于12名网络之前未见过的受试者，仅在每个受试者28秒内从网格化图像中分别重建出高质量的4D和中间位置MRI（1.25×1.25×3.3 mm³）。基于深度学习的MRI与联合MoCo - HDTV重建的MRI之间高度吻合（平均SSIM≥0.96，NIQE评分分别为7.94和5.66）。基于深度学习的4D - MRI在临床靶区和危及器官勾画方面是可接受的。肿瘤在中间位置图像上的位置偏差在0.7 mm以内。
我们的结果表明，联合MoCo - HDTV和中间位置算法均可由深度卷积神经网络近似。4D和中间位置MRI的这种快速重建有助于胸部或腹部磁共振引导放射治疗中的在线治疗调整。
Deep learning accelerates 4D-MRI recon for online adaptive MR-guided radiotherapy. First reconstruction of high resolution whole thorax 4D-MRI for 16 phases in 28 s. First use of dCNNs for midposition reconstruction from undersampled 4D-MRI in 28 s. Excellent agreement between deep learning-based tumour midposition and reference. 4D and midposition MRI could inform plan adaptation in lung and abdominal MR-guided radiotherapy. We present deep learning-based solutions to overcome long 4D-MRI reconstruction times while maintaining high image quality and short scan times. Two 3D U-net deep convolutional neural networks were trained to accelerate the 4D joint MoCo-HDTV reconstruction. For the first network, gridded and joint MoCo-HDTV-reconstructed 4D-MRI were used as input and target data, respectively, whereas the second network was trained to directly calculate the midposition image. For both networks, input and target data had dimensions of 256 × 256 voxels (2D) and 16 respiratory phases. Deep learning-based MRI were verified against joint MoCo-HDTV-reconstructed MRI using the structural similarity index (SSIM) and the naturalness image quality evaluator (NIQE). Moreover, two experienced observers contoured the gross tumour volume and scored the images in a blinded study. For 12 subjects, previously unseen by the networks, high-quality 4D and midposition MRI (1.25 × 1.25 × 3.3 mm3) were each reconstructed from gridded images in only 28 seconds per subject. Excellent agreement was found between deep-learning-based and joint MoCo-HDTV-reconstructed MRI (average SSIM ≥ 0.96, NIQE scores 7.94 and 5.66). Deep-learning-based 4D-MRI were clinically acceptable for target and organ-at-risk delineation. Tumour positions agreed within 0.7 mm on midposition images. Our results suggest that the joint MoCo-HDTV and midposition algorithms can each be approximated by a deep convolutional neural network. This rapid reconstruction of 4D and midposition MRI facilitates online treatment adaptation in thoracic or abdominal MR-guided radiotherapy.
登录
查看更多内容
影响因子：
3.3
作者：
Feng, Li;Grimm, Robert;Block, Kai Tobias;Chandarana, Hersh;Kim, Sungheon;Xu, Jian;Axel, Leon;Sodickson, Daniel K.;Otazo, Ricardo
通讯作者：
Otazo, Ricardo
影响因子：
6.7
作者：
Freedman JN;Collins DJ;Bainbridge H;Rank CM;Nill S;Kachelrieß M;Oelfke U;Leach MO;Wetscherek A
通讯作者：
Wetscherek A
影响因子：
3.3
作者：
Cruz G;Atkinson D;Buerger C;Schaeffter T;Prieto C
通讯作者：
Prieto C
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：
3.5
作者：
Hyun, Chang Min;Kim, Hwa Pyung;Seo, Jin Keun
通讯作者：
Seo, Jin Keun