Zhang Zhe, Zhang Bing, Li Ming, Liang Xue, Chen Xiaodong, Liu Renyuan, Zhang Xin, Guo Hua
Center for Biomedical Imaging Research, Department of Biomedical Engineering, Tsinghua University, Beijing, China.
Department of Radiology, Affiliated Drum Tower Hospital of Nanjing University Medical School, Nanjing, China.
J Magn Reson Imaging. 2017 Jul;46(1):167-174. doi: 10.1002/jmri.25522. Epub 2016 Oct 20.
To report a diffusion imaging technique insensitive to off-resonance artifacts and motion-induced ghost artifacts using multishot Cartesian turbo spin-echo (TSE) acquisition and iterative POCS-based reconstruction of multiplexed sensitivity encoded magnetic resonance imaging (MRI) (POCSMUSE) for phase correction.
Phase insensitive diffusion preparation was used to deal with the violation of the Carr-Purcell-Meiboom-Gill (CPMG) conditions of TSE diffusion-weighted imaging (DWI), followed by a multishot Cartesian TSE readout for data acquisition. An iterative diffusion phase correction method, iterative POCSMUSE, was developed and implemented to eliminate the ghost artifacts in multishot TSE DWI. The in vivo human brain diffusion images (from one healthy volunteer and 10 patients) using multishot Cartesian TSE were acquired at 3T and reconstructed using iterative POCSMUSE, and compared with single-shot and multishot echo-planar imaging (EPI) results. These images were evaluated by two radiologists using visual scores (considering both image quality and distortion levels) from 1 to 5.
The proposed iterative POCSMUSE reconstruction was able to correct the ghost artifacts in multishot DWI. The ghost-to-signal ratio of TSE DWI using iterative POCSMUSE (0.0174 ± 0.0024) was significantly (P < 0.0005) smaller than using POCSMUSE (0.0253 ± 0.0040). The image scores of multishot TSE DWI were significantly higher than single-shot (P = 0.004 and 0.006 from two reviewers) and multishot (P = 0.008 and 0.004 from two reviewers) EPI-based methods.
The proposed multishot Cartesian TSE DWI using iterative POCSMUSE reconstruction can provide high-quality diffusion images insensitive to motion-induced ghost artifacts and off-resonance related artifacts such as chemical shifts and susceptibility-induced image distortions.
1 Technical Efficacy: Stage 1 J. MAGN. RESON. IMAGING 2017;46:167-174.
报告一种扩散成像技术，该技术使用多激发笛卡尔涡轮自旋回波（TSE）采集以及基于迭代投影-优化-共轭梯度-敏感性编码磁共振成像（MRI）（POCSMUSE）的相位校正方法，对失谐伪影和运动诱导的鬼影伪影不敏感。
采用相位不敏感扩散准备来处理TSE扩散加权成像（DWI）中违反Carr-Purcell-Meiboom-Gill（CPMG）条件的情况，随后进行多激发笛卡尔TSE读出以采集数据。开发并实施了一种迭代扩散相位校正方法——迭代POCSMUSE，以消除多激发TSE DWI中的鬼影伪影。使用多激发笛卡尔TSE在3T下采集了体内人脑扩散图像（来自1名健康志愿者和10名患者），并使用迭代POCSMUSE进行重建，然后与单次激发和多激发回波平面成像（EPI）结果进行比较。两名放射科医生使用1至5的视觉评分（同时考虑图像质量和失真程度）对这些图像进行评估。
所提出的迭代POCSMUSE重建能够校正多激发DWI中的鬼影伪影。使用迭代POCSMUSE的TSE DWI的鬼影与信号比（0.0174±0.0024）显著（P<0.0005）低于使用POCSMUSE时（0.0253±0.0040）。多激发TSE DWI的图像评分显著高于基于单次激发（两位审阅者分别为P = 0.004和0.006）和多激发（两位审阅者分别为P = 0.008和0.004）EPI的方法。
所提出的使用迭代POCSMUSE重建的多激发笛卡尔TSE DWI能够提供对运动诱导的鬼影伪影以及失谐相关伪影（如化学位移和磁化率诱导的图像失真）不敏感的高质量扩散图像。
1技术效能：1期 J.MAGN.RESON.IMAGING 2017;46:167 - 174。
