分享:
分享到微信朋友圈
X
技术研究
均匀性校正在颅脑定量磁化率成像中的应用价值评估
甘凤玲 瞿筝 赵玮玮 钟昊东 李改英 李建奇

Cite this article as: Gan FL, Qu Z, Zhao WW, et al. Effect of signal intensity inhomogeneity correction on quantitative susceptibility mapping of brain[J]. Chin J Magn Reson Imaging, 2022, 13(4): 94-99.本文引用格式:甘凤玲, 瞿筝, 赵玮玮, 等. 均匀性校正在颅脑定量磁化率成像中的应用价值评估[J]. 磁共振成像, 2022, 13(4): 94-99. DOI:10.12015/issn.1674-8034.2022.04.017.


[摘要] 目的 评估颅脑定量磁化率成像(quantitative susceptibility mapping, QSM)中线圈选择和空间均匀性校正对深部核团磁化率测量值的影响。材料与方法 在2.89 T磁共振成像系统上采用20通道和64通道头颈联合线圈完成10名健康受试者扫描,得到分别施加和未施加空间均匀性校正的图像数据,再进行QSM重建。在重建得到的定量磁化率图上,手动勾画出六个双侧脑深部灰质核团,包括红核、黑质、苍白球、壳核、尾状核和齿状核,计算核团内磁化率平均值。采用配对样本t检验、线性相关分析和Bland-Altman分析比较均匀性校正前后、不同通道线圈采集获得的磁化率值的组间差异性和一致性。结果 用于QSM重建的图像数据施加空间均匀性校正后,磁化率图上大脑深部核团边界更加清晰,定量磁化率测量值均显著增加。均匀性校正前、后核团内磁化率平均值线性相关(20通道:斜率K=1.06,R2=0.96;64通道:斜率K=1.12,R2=0.95)。在20通道和64通道线圈中采集得到核团内磁化率平均值线性相关(校正前:斜率K=0.92,R2=0.96;校正后:斜率K=0.98,R2=0.96)。配对样本t检验结果显示均匀性校正后64通道和20通道线圈采集的核团磁化率平均值差异无统计学意义。Bland-Altman图显示均匀性校正后64通道和20通道线圈采集方法之间没有明显的偏差。结论 线圈选择和空间均匀性校正的使用对颅脑定量磁化率值有一定的影响,采用空间均匀性高的图像可提高QSM测量的准确性。
[Abstract] Objective To evaluate the effects of coil selection and signal intensity inhomogeneity on the susceptibility values of deep gray matter nuclei measured by quantitative susceptibility mapping (QSM).Materials and Methods Ten healthy subjects were scanned on a 2.89 T MRI system using 20- and 64-channel combined head-neck coils. QSM reconstructions were performed before and after correcting signal intensity inhomogeneity of the original gradient-echo images, respectively. Six bilateral deep cerebral gray matter nuclei were manually drawn on susceptibility maps, including red nucleus, substantia nigra, globus pallidus, putamen, caudate nucleus, and dentate nucleus. Paired sample t test, linear correlation analysis and Bland-Altman analysis were used to compare the differences and consistency of susceptibility values between groups with and without intensity inhomogeneity correction using different coils.Results The susceptibility maps with intensity inhomogeneity correction demonstrated improved boundaries of deep gray matter nuclei and susceptibility values increased significantly. The mean susceptibility values of deep gray matter nuclei after correction were highly correlated with those before correction (20-channel: slope K=1.06, R2=0.96; 64-channel: slope K=1.12, R2=0.95). The mean susceptibility values of deep gray matter nuclei from 20-channel coil were highly correlated with those from 64-channel coil (without correction: slope K=0.92, R2=0.96; with correction: slope K=0.98, R2=0.96). There was no statistically significant difference in the mean susceptibility values between the 20- and 64-channel coils acquisition after correcting the intensity inhomogeneity. Bland-Altman results showed no significant deviation between the 20- and 64-channel coils acquisition after the correction on the intensity inhomogeneity.Conclusions The signal intensity inhomogeneity affects the susceptibility values of the deep brain nuclei, and inhomogeneity correction can improve the accuracy of susceptibility values of deep gray matter nuclei.
[关键词] 定量磁化率成像;空间均匀性;线圈;深部灰质核团;磁共振成像
[Keywords] quantitative susceptibility mapping;intensity inhomogeneity correction;coil;deep gray matter nuclei;magnetic resonance imaging

甘凤玲 1   瞿筝 2   赵玮玮 1   钟昊东 1   李改英 1   李建奇 1*  

1 华东师范大学物理与电子科学学院,上海市磁共振重点实验室,上海 200062

2 四川大学华西临床医学院/华西医院,成都 610041

李建奇,E-mail:jqli@phy.ecnu.edu.cn

作者利益冲突声明:全体作者均声明不存在利益冲突。


基金项目: 国家社会科学基金 15ZDB016
收稿日期:2021-11-11
接受日期:2022-03-25
中图分类号:R445.2 
文献标识码:A
DOI: 10.12015/issn.1674-8034.2022.04.017
本文引用格式:甘凤玲, 瞿筝, 赵玮玮, 等. 均匀性校正在颅脑定量磁化率成像中的应用价值评估[J]. 磁共振成像, 2022, 13(4): 94-99. DOI:10.12015/issn.1674-8034.2022.04.017

       磁共振定量成像能够量化组织某些特定生理参数,可提供形态学评估外的大量组织特定信息,在疾病早期诊断、疾病分期以及预后评估中具有重要价值。定量磁化率成像(quantitative susceptibility mapping,QSM)采用梯度回波类序列得到相位图,然后计算得到局部磁场分布图,再利用场图与磁化率之间的物理关系来反演出磁化率定量分布图[1, 2]。目前,QSM已广泛应用于临床诊断和科学研究[3],如评估大脑微出血或血肿大小[4]、研究神经退行性疾病大脑深部核团的铁沉积[5]、区分出血和钙化[6]、为脑深部刺激手术[7]提供准确和可靠的靶向核团成像等。

       在磁共振定量成像中,组织定量参数的准确性至关重要。QSM重建过程复杂,数据采集策略和QSM重建方案的变化会对磁化率图像产生不同的影响。先前的研究显示,不同场强磁共振扫描仪(1.5 T、3.0 T和7.0 T)[8, 9, 10, 11]和单双极采集[12, 13]得到的大脑深部核团磁化率重复性较好,而图像扫描分辨率[14, 15]、扫描覆盖范围[16]、不同回波时间[11,17]会对磁化率值有一定影响。而QSM重建的最大挑战是偶极核在54.7°附近的圆锥面区域存在零点而导致的不适定问题。为了解决这个难题,研究者提出了诸多不同的算法,包括k空间加权微分法[18]、k空间阈值相除法[19]、多方向采样磁化率计算法[20]、形态学偶极子反演法(morphology enabled dipole inversion,MEDI)[21]等,每种算法都各有优缺点。其中MEDI算法是目前大家广泛使用的方法,其利用模图作为先验信息,可有效抑制磁化率图中的条状伪影,并提高图像在组织边界处的清晰度,从而大幅提高磁化率图像质量,但是模图的信号强度在空间的不均匀性有可能影响磁化率计算的准确性。

       均匀性校正是在图像信号采集之前利用固定体线圈获得一个低分辨率大视野的图像,从而获得不同空间位置的信号强度分布信息,然后利用这个信息去校正图像以获得信号相对均匀的图像。空间均匀性校正会对模图产生影响,在使用MEDI算法时,这种影响可能会映射到QSM图像中。目前,还没有报道关注均匀性校正对QSM重建结果的影响。

       本研究中,采用20通道和64通道头颈联合线圈进行图像采集,并通过施加或未施加均匀性校正获得不同空间均匀性的颅脑图像,以评估均匀性校正在颅脑QSM的应用价值,提高临床上利用QSM进行疾病诊断和科学研究的准确性和横向对比的可靠性。

1 材料与方法

1.1 磁共振扫描

       根据预试验结果,基于PASS 15.0软件的Test for Paired Means算法进行样本量估算,最终纳入10名健康受试者(女6名,男4名,年龄20~24岁)参与试验。所有受试者均无磁共振成像扫描禁忌证,也无神经学、心血管或其他严重躯体疾病史。本前瞻性研究经华东师范大学人体试验伦理委员会批准(批准文号:HR 442-2019),全体受试者均签署了知情同意书。所有扫描均在2.89 T磁共振成像系统(Magnetom Prisma Fit;西门子医疗,埃朗根,德国)上采用20通道和64通道头颈联合线圈完成。QSM扫描采用三维多回波梯度回波序列,具体参数为:TR=31 ms,TE1 (第一个回波时间)=4.07 ms,ΔTE (回波间隔时间)=4.35 ms,回波数=6,FA=12°,FOV=240 mm×200 mm,采集矩阵=288×240,体素大小=0.83 mm×0.83 mm×0.80 mm,层数=192,并行成像加速因子为2。通过扫描参数设置可同时得到施加和未施加空间均匀性校正的复数图像数据。

1.2 QSM重建

       QSM重建具体步骤如下:首先,采用BET (brain extraction tool)方法去除颅骨[22],并对相位图中每个体素的相位进行一维时间域解缠绕,接着对每个体素不同TE时间的相位进行加权最小二乘法拟合来估计场图[23];然后,采用基于快速傅里叶变换的拉普拉斯法进行空间域相位解缠绕[24];再者,使用拉普拉斯边界值法去除背景场[25];最后,将剩余的组织场使用MEDI+0算法反演计算得到磁化率分布图。MEDI+0方法是MEDI算法中添加正则项,以自动将脑室内脑脊液作为组织磁化率参考值[26],该方法可以抑制常规MEDI方法中脑脊液磁化率不均匀的问题[27]。QSM重建采用MEDI toolbox 2020程序包完成(http://pre.weill.cornell.edu/mri/pages/qsm.html),软件运行平台为MATLAB R2016b (MathWorks,MA,USA)。

1.3 感兴趣区勾画

       一名研究人员(两年磁共振神经影像学研究经验)基于均匀性校正后的磁化率图,使用ITK-SNAP图像处理软件(http://www.itk-snap.org)手动勾画感兴趣区域(region of interest,ROI),包括六个双侧脑深部灰质核团:红核、黑质、苍白球、尾状核、壳核和齿状核。所得到的ROI要求覆盖双侧核团所有可见区域。为了确保定量评估的准确性和可比性,本研究使用同一个ROI对均匀性校正前后的磁化率图进行分析。

1.4 统计学分析

       采用组间相关系数(intraclass correlation coefficient,ICC)评估20通道和64通道线圈采集图像的ROI勾画的一致性,ICC>0.75为一致性良好。采用配对样本t检验比较均匀性校正前后两种线圈采集所得磁化率值的组间差异性,P<0.05表示差异具有统计学意义。采用线性相关分析和Bland-Altman分析方法评价均匀性校正前后两种线圈采集获得的磁化率值的一致性和可重复性。本研究使用 IBM SPSS 23.0软件进行统计分析。

2 结果

2.1 图像结果

       10例受试者均顺利完成磁共振成像检查,图像质量良好。结果显示,均匀性校正会校正模图信号强度(图1A1D),未对相位图(图2A2D)产生影响。未施加均匀性校正的模图(图1A1C)靠近线圈处的颅脑外部区域信号强度大,脑深部区域信号强度明显偏弱;相比20通道线圈采集得到的模图(图1A),64通道线圈采集得到的模图的信号空间分布更加不均匀(图1C)。而采用均匀性校正后,模图(图1B1D)上信号趋于均匀,由相应模图生成的梯度掩模图(图3B3D)中组织边界更加清晰合理。相较于均匀性校正后数据重建得到的磁化率图像(图4B4D),由均匀性校正前数据重建所得磁化率图像(图4A4C)更加平滑。均匀性校正的应用为QSM提供了更为清晰的核团边界和更好的组织对比度,改善了图像的可视化。均匀性校正后,两种线圈采集得到的磁化率图像基本相同。

图1  两种线圈采集得到的均匀性校正前、后的模图。1A:20通道线圈采集,均匀性校正前;1B:20通道线圈采集,均匀性校正后;1C:64通道线圈采集,均匀性校正前;1D:64通道线圈采集,均匀性校正后。
Fig. 1  The magnitude images acquired with 20-channel (1A&1B) and 64-channel (1C&1D) coils, respectively. 1A&1C: the signal intensity inhomogeneity correction was not performed; 1B&1D: the signal intensity inhomogeneity correction was performed.
图2  两种线圈采集得到的均匀性校正前、后的相位图。2A:20通道线圈采集,均匀性校正前;2B:20通道线圈采集,均匀性校正后;2C:64通道线圈采集,均匀性校正前;2D:64通道线圈采集,均匀性校正后。
Fig. 2  The Phase images acquired with 20-channel (2A&2B) and 64-channel (2C&2D) coils, respectively. 2A&2C: the signal intensity inhomogeneity correction was not performed; 2B&2D: the signal intensity inhomogeneity correction was performed.
图3  两种线圈采集得到的均匀性校正前、后的幅值梯度掩模图。3A:20通道线圈采集,均匀性校正前;3B:20通道线圈采集,均匀性校正后;3C:64通道线圈采集,均匀性校正前;3D:64通道线圈采集,均匀性校正后。
Fig. 3  The amplitude gradient masks calculated from the magnitude images acquired with 20-channel (3A&3B) and 64-channel (3C&3D) coils, respectively. 3A&3C: the signal intensity inhomogeneity correction was not performed; 3B&3D: the signal intensity inhomogeneity correction was performed.
图4  两种线圈采集得到的均匀性校正前、后数据重建得到的磁化率图。4A:20通道线圈采集,均匀性校正前;4B:20通道线圈采集,均匀性校正后;4C:64通道线圈采集,均匀性校正前;4D:64通道线圈采集,均匀性校正后。
Fig. 4  The susceptibility maps reconstructed from the data acquired with 20-channel (4A&4B) and 64-channel (4C&4D) coils, respectively. 4A&4C: the signal intensity inhomogeneity correction was not performed; 4B&4D: the signal intensity inhomogeneity correction was performed.

2.2 统计结果

       在20通道和64通道线圈采集得到的磁化率图上勾画的ROI表现出良好的体积一致性(ICC>0.91)。

       表1为两种线圈均匀性校正前后数据得到的核团磁化率值对比。无论是使用20通道线圈还是64通道线圈采集,均匀性校正后图像重建得到的核团内磁化率平均值较之校正前均有显著性提高(P均<0.001)。均匀性校正前,20通道线圈采集获得的核团内磁化率平均值较之64通道采集有显著性提高(P均<0.05)。均匀性校正后,两组通道线圈采集得到的核团磁化平均值基本相同,差异无统计学意义(P均>0.05)。

       obtained from two head-neck coils and before or after correcting signal intensity inhomogeneity. 5A: Comparison between before and after correcting signal intensity inhomogeneity using 20-channel coil acquisition; 5B: Comparison between before and after correcting signal intensity inhomogeneity using 64-channel coil acquisition; 5C: Comparison between 20- and 64-channel coils acquisition before correcting signal intensity inhomogeneity; The solid and dotted lines are the trend line of the linear regression and the line of equality, respectively. RN: the red nucleus, SN: substantia nigra, GP: globus pallidus, PU: putamen, CN: caudate nucleus, DN: dentate nucleus.

       均匀性校正前、后核团内磁化率平均值线性相关(20通道:斜率K=1.06,R2=0.96;64通道:斜率K=1.12,R2=0.95) (图5A5B)。均匀性校正前,在20通道和64通道线圈中采集得到核团内磁化率平均值线性相关(斜率K=0.92,R2=0.96) (图5C)。

       均匀性校正后,20通道和64通道线圈测量结果具有良好一致性,核团内磁化率平均值线性相关(图6A),两种采集方法之间的线性回归斜率(K=0.98)接近于1,截距(y0=-1.8 ppb)接近于0,相关系数(R2=0.96)也接近于1。Bland-Altman图显示均匀性校正后20通道和64通道线圈采集方法之间没有明显的偏差,95%置信区间的范围为(-3.1±13.3) ppb (图6B)。

图5  核团磁化率值线性回归分析的散点图。5A:20通道线圈采集得到均匀性校正前、后的核团磁化率值线性回归;5B:64通道线圈采集得到均匀性校正前、后的核团磁化率值线性回归;5C:均匀性校正前,20通道和64通道线圈采集得到的核团磁化率值线性回归。图中的实线和虚线分别是线性回归的趋势线和等式线。RN:红核,SN:黑质,GP:苍白球,PUT:壳核,CN:尾状核,DN:齿状核。
Fig. 5  Scattered plots of the linear regression analysis of susceptibility values obtained from two head-neck coils and before or after correcting signal intensity inhomogeneity. 5A: Comparison between before and after correcting signal intensity inhomogeneity using 20-channel coil acquisition; 5B: Comparison between before and after correcting signal intensity inhomogeneity using 64-channel coil acquisition; 5C: Comparison between 20- and 64-channel coils acquisition before correcting signal intensity inhomogeneity; The solid and dotted lines are the trend line of the linear regression and the line of equality, respectively. RN: the red nucleus, SN: substantia nigra, GP: globus pallidus, PU: putamen, CN: caudate nucleus, DN: dentate nucleus.
图6  20通道和64通道线圈采集的图像均匀性校正后重建得到的核团磁化率值定量对比。6A:核团磁化率值线性回归图;6B:Bland-Altman图。散点图中的实线和虚线分别是线性回归的趋势线和等式线,Bland-Altman图中的实线和虚线分别表示平均值和1.96倍标准差的平均差异。RN:红核, SN:黑质,GP:苍白球,PUT:壳核,CN:尾状核,DN:齿状核。
Fig. 6  Quantitative comparison of the susceptibility values between 20- and 64-channel coils after correcting the intensity inhomogeneity. 6A: Scattered plots of the Linear regression analysis of susceptibility values; 6B: Bland-Altman plots. The solid and dotted lines in A are the trend line of the linear regression and the line of equality, respectively. The solid and dotted lines in B indicate the mean difference ± 1.96 times the standard deviation of the difference, respectively. RN: red nucleus, SN: substantia nigra, GP: globus pallidus, PU: putamen, CN: caudate nucleus, CN: caudate nucleus, DN: dentate nucleus.
表1  两种线圈均匀性校正前后数据得到的核团磁化率平均值对比
Tab. 1  Mean susceptibility values of deep gray matter nuclei obtained from the two head-neck coils with and without signal intensity inhomogeneity correction

3 讨论

       本研究首次通过对比20通道和64通道线圈采集图像在均匀性校正前后的磁化率测量值,评估均匀性校正在颅脑QSM中的应用价值。定性分析结果显示,图像经过空间均匀性校正后重建得到的磁化率图像具有更清晰的核团边界和更好的组织对比度。统计分析结果显示,经过均匀性校正后图像得到的脑深部核团磁化率值显著升高。均匀性校正后,20通道与64通道线圈采集得到的核团磁化率值表现出良好的定量一致性。因此,我们发现提高图像的空间均匀性可以获得更准确的磁化率测量值,这一研究结果或将有助于提高临床上利用QSM进行疾病诊断和科学研究的准确性和横向对比的可靠性。

3.1 空间均匀性对磁化率值的影响

       MEDI算法通过模图提供额外的结构信息来约束反演过程中梯度的稀疏性。由于模图和磁化率图具有共同的组织解剖结构,模图上的组织边界也就对应于磁化率图上的组织边界,两者具有一致的空间变化趋势。通过引入平滑性约束限制,不仅可以抑制磁化率图上的条状伪影,也提高了QSM图像在组织边界处的清晰程度,大幅提升QSM图像质量。但MEDI算法在计算组织边界的掩模时采用了单一阈值[28],空间不均匀的模图可能会导致部分不正确的组织边界信息,从而影响磁化率计算的准确性。

       均匀性校正前,相较于64通道线圈采集得到的数据,20通道线圈采集得到的模图均匀性更好,组织边界的信息相对更准确,因此得到的核团磁化率值也更准确。而均匀性校正后,两种线圈得到的模图空间信号强度都比较均匀,组织边界更加清晰合理,为QSM提供一个合理准确的梯度掩模,最终得到磁化率值一致性较好,差异无统计学意义。因此,在采用MEDI算法时,空间均匀性校正非常重要。

3.2 临床价值

       本研究为QSM技术应用于临床进行疾病诊断和科学研究提供了一个更为准确稳健的采集方案。QSM的优势是能够定量测量组织磁化率的变化,被广泛地应用到与铁代谢相关的退行性疾病的诊断与疗效评估、脑出血的诊断和颅内出血与钙化的鉴别。其中,在帕金森病[29, 30, 31]、阿尔茨海默病[32, 33, 34]、多发性硬化症[35]和威尔逊病[36]等神经退行性疾病中,通过QSM对铁浓度的量化,可以早期检测神经退行性疾病,并判定疾病的发展程度及预后情况,磁化率测量准确性对于QSM在应用研究中的可行性具有关键性作用。均匀性校正后,64通道和20通道线圈采集得到的磁化率一致性非常好,都更加准确。因此均匀性校正将有利于QSM在脑疾病的横向和纵向研究中的应用推广。

3.3 局限性及今后研究方向

       本研究也存在一定的局限性。首先,QSM重建方法众多,我们尚未研究空间均匀性对其他重建算法是否会产生跟MEDI算法相同的影响。从理论上来说,对于未采用模图得到的组织边界正则化化反演计算的算法,均匀性校正的使用不会对磁化率图像产生影响,在之后的研究中我们将进行更为全面的分析。另外,还可以通过回顾性后处理方法进行图像空间均匀性校正[37, 38],常规的方法包括滤波法、曲面拟合法、分割法、直方图法等,未来我们也将对这些方法进行评估以探求一种最为鲁棒性的均匀性校正方法,并将均匀性校正方法集成到QSM重建流程中。

       综上所述,均匀性校正的使用可以提高QSM图像质量,且通道数不同的线圈采集得到的定量磁化率图像具有良好的定量一致性。因此,采用空间均匀性高的图像测量得到的磁化率值将更加准确,有助于提高临床上利用QSM进行疾病诊断的准确性和可靠性。

[1]
Wang Y, Liu T. Quantitative susceptibility mapping (QSM): Decoding MRI data for a tissue magnetic biomarker[J]. Magn Reson Med, 2015, 73(1): 82-101. DOI: 10.1002/mrm.25358.
[2]
郑志伟, 蔡聪波, 蔡淑惠, 等. 定量磁化率成像的基本原理及方法概述[J]. 磁共振成像, 2016, 7(6): 454-460. DOI: 10.12015/issn.1674-8034.2016.06.011.
Zheng ZW, Cai CB, Cai SH, et al. Brief overview of principles and methods of quantitative susceptibility mapping[J]. Chin J Magn Reson Imaging, 2016, 7(6): 454-460. DOI: 10.12015/issn.1674-8034.2016.06.011.
[3]
孙彤彤, 吴巧玲, 倪红艳. 定量磁化率成像方法及临床应用进展[J]. 中华放射学杂志, 2020, 54(9): 912-916. DOI: 10.3760/cma.j.cn112149-20190924-00513.
Sun TT, Wu QL, Ni HY. Imaging method and clinical application of quantitative susceptibility mapping[J]. Chin J Radiol, 2020, 54(9): 912-916. DOI: 10.3760/cma.j.cn112149-20190924-00513.
[4]
Liu T, Surapaneni K, Lou M, et al. Cerebral microbleeds: burden assessment by using quantitative susceptibility mapping[J]. Radiology, 2012, 262(1): 269-278. DOI: 10.1148/radiol.11110251.
[5]
Li GY, Zhai GQ, Zhao XX, et al. 3D texture analyses within the substantia nigra of Parkinson's disease patients on quantitative susceptibility maps and R2 maps[J]. Neuroimage, 2019, 188: 465-472. DOI: 10.1016/j.neuroimage.2018.12.041.
[6]
Chen WW, Zhu WZ, Kovanlikaya I, et al. Intracranial calcifications and hemorrhages: characterization with quantitative susceptibility mapping[J]. Radiology, 2014, 270(2): 496-505. DOI: 10.1148/radiol.13122640.
[7]
Dimov AV, Gupta A, Kopell BH, et al. High-resolution QSM for functional and structural depiction of subthalamic nuclei in DBS presurgical mapping[J]. J Neurosurg, 2018, 131(2): 360-367. DOI: 10.3171/2018.3.JNS172145.
[8]
Hinoda T, Fushimi Y, Okada T, et al. Quantitative susceptibility mapping at 3 T and 1.5 T: evaluation of consistency and reproducibility[J]. Invest Radiol, 2015, 50(8): 522-530. DOI: 10.1097/RLI.0000000000000159.
[9]
Ippoliti M, Adams LC, Winfried B, et al. Quantitative susceptibility mapping across two clinical field strengths: contrast-to-noise ratio enhancement at 1.5T[J]. J Magn Reson Imaging, 2018, 48(5): 1410-1420. DOI: 10.1002/jmri.26045.
[10]
Isaacs BR, Mulder MJ, Groot JM, et al. 3 versus 7 Tesla magnetic resonance imaging for parcellations of subcortical brain structures in clinical settings[J]. PLoS One, 2020, 15(11): e0236208. DOI: 10.1371/journal.pone.0236208.
[11]
Spincemaille P, Anderson J, Wu GH, et al. Quantitative susceptibility mapping: MRI at 7T versus 3T[J]. J Neuroimaging, 2020, 30(1): 65-75. DOI: 10.1111/jon.12669.
[12]
Li JQ, Chang SX, Liu T, et al. Phase-corrected bipolar gradients in multi-echo gradient-echo sequences for quantitative susceptibility mapping[J]. MAGMA, 2015, 28(4): 347-355. DOI: 10.1007/s10334-014-0470-3.
[13]
Lauzon ML, McCreary CR, McLean DA, et al. Quantitative susceptibility mapping at 3 T: comparison of acquisition methodologies[J]. NMR Biomed, 2017, 30(4): e3492. DOI: 10.1002/nbm.3492.
[14]
Cong F, Liu XR, Liu CJ, et al. Improved depiction of subthalamic nucleus and globus pallidus internus with optimized high-resolution quantitative susceptibility mapping at 7 T[J]. NMR Biomed, 2020, 33(11): e4382. DOI: 10.1002/nbm.4382.
[15]
Zhou D, Cho J, Zhang JW, et al. Susceptibility underestimation in a high-susceptibility phantom: dependence on imaging resolution, magnitude contrast, and other parameters[J]. Magn Reson Med, 2017, 78(3): 1080-1086. DOI: 10.1002/mrm.26475.
[16]
Karsa A, Punwani S, Shmueli K. The effect of low resolution and coverage on the accuracy of susceptibility mapping[J]. Magn Reson Med, 2019, 81(3): 1833-1848. DOI: 10.1002/mrm.27542.
[17]
Lancione M, Cencini M, Costagli M, et al. Diagnostic accuracy of quantitative susceptibility mapping in multiple system atrophy: the impact of echo time and the potential of histogram analysis[J]. Neuroimage Clin, 2022, 34: 102989. DOI: 10.1016/j.nicl.2022.102989.
[18]
Li W, Wu B, Liu CL. Quantitative susceptibility mapping of human brain reflects spatial variation in tissue composition[J]. NeuroImage, 2011, 55(4): 1645-1656. DOI: 10.1016/j.neuroimage.2010.11.088.
[19]
Wharton S, Schäfer A, Bowtell R. Susceptibility mapping in the human brain using threshold-based k-space division[J]. Magn Reson Med, 2010, 63(5): 1292-1304. DOI: 10.1002/mrm.22334.
[20]
Liu T, Spincemaille P, de Rochefort L, et al. Calculation of susceptibility through multiple orientation sampling (COSMOS): a method for conditioning the inverse problem from measured magnetic field map to susceptibility source image in MRI[J]. Magn Reson Med, 2009, 61(1): 196-204. DOI: 10.1002/mrm.21828.
[21]
Liu J, Liu T, de Rochefort L, et al. Morphology enabled dipole inversion for quantitative susceptibility mapping using structural consistency between the magnitude image and the susceptibility map[J]. Neuroimage, 2012, 59(3): 2560-2568. DOI: 10.1016/j.neuroimage.2011.08.082.
[22]
Smith SM. Fast robust automated brain extraction[J]. Hum Brain Mapp, 2002, 17(3): 143-155. DOI: 10.1002/hbm.10062.
[23]
Liu T, Wisnieff C, Lou M, et al. Nonlinear formulation of the magnetic field to source relationship for robust quantitative susceptibility mapping[J]. Magn Reson Med, 2013, 69(2): 467-476. DOI: 10.1002/mrm.24272.
[24]
Abdul-Rahman HS, Gdeisat MA, Burton DR, et al. Fast and robust three-dimensional best path phase unwrapping algorithm[J]. Appl Opt, 2007, 46(26): 6623-6635. DOI: 10.1364/ao.46.006623.
[25]
Zhou D, Liu T, Spincemaille P, et al. Background field removal by solving the Laplacian boundary value problem[J]. NMR Biomed, 2014, 27(3): 312-319. DOI: 10.1002/nbm.3064.
[26]
Liu Z, Spincemaille P, Yao YH, et al. MEDI+0: Morphology enabled dipole inversion with automatic uniform cerebrospinal fluid zero reference for quantitative susceptibility mapping[J]. Magn Reson Med, 2018, 79(5): 2795-2803. DOI: 10.1002/mrm.26946.
[27]
Liu T, Xu WY, Spincemaille P, et al. Accuracy of the morphology enabled dipole inversion (MEDI) algorithm for quantitative susceptibility mapping in MRI[J]. IEEE Trans Med Imaging, 2012, 31(3): 816-824. DOI: 10.1109/TMI.2011.2182523.
[28]
Liu T, Liu J, de Rochefort L, et al. Morphology enabled dipole inversion (MEDI) from a single-angle acquisition: comparison with COSMOS in human brain imaging[J]. Magn Reson Med, 2011, 66(3): 777-783. DOI: 10.1002/mrm.22816.
[29]
Li KR, Avecillas-Chasin J, Nguyen TD, et al. Quantitative evaluation of brain iron accumulation in different stages of Parkinson's disease[J]. J Neuroimaging, 2022, 32(2): 363-371. DOI: 10.1111/jon.12957.
[30]
He NY, Ghassaban K, Huang P, et al. Imaging iron and neuromelanin simultaneously using a single 3D gradient echo magnetization transfer sequence: combining neuromelanin, iron and the nigrosome-1 sign as complementary imaging biomarkers in early stage Parkinson's disease[J]. Neuroimage, 2021, 230: 117810. DOI: 10.1016/j.neuroimage.2021.117810.
[31]
Uchida Y, Kan H, Sakurai K, et al. Magnetic susceptibility associates with dopaminergic deficits and cognition in Parkinson's disease[J]. Mov Disord, 2020, 35(8): 1396-1405. DOI: 10.1002/mds.28077.
[32]
Au CKF, Abrigo J, Liu CL, et al. Quantitative susceptibility mapping of the hippocampal Fimbria in Alzheimer's disease[J]. J Magn Reson Imaging, 2021, 53(6): 1823-1832. DOI: 10.1002/jmri.27464.
[33]
Gong NJ, Dibb R, Bulk M, et al. Imaging beta amyloid aggregation and iron accumulation in Alzheimer's disease using quantitative susceptibility mapping MRI[J]. Neuroimage, 2019, 191: 176-185. DOI: 10.1016/j.neuroimage.2019.02.019.
[34]
Bulk M, Abdelmoula WM, Geut H, et al. Quantitative MRI and laser ablation-inductively coupled plasma-mass spectrometry imaging of iron in the frontal cortex of healthy controls and Alzheimer's disease patients[J]. Neuroimage, 2020, 215: 116808. DOI: 10.1016/j.neuroimage.2020.116808.
[35]
Huang WY, Sweeney EM, Kaunzner UW, et al. Quantitative susceptibility mapping versus phase imaging to identify multiple sclerosis iron rim lesions with demyelination[J]. J Neuroimaging, 2022. DOI: 10.1111/jon.12987.
[36]
Li GY, Wu R, Tong R, et al. Quantitative measurement of metal accumulation in brain of patients with Wilson's disease[J]. Mov Disord, 2020, 35(10): 1787-1795. DOI: 10.1002/mds.28141.
[37]
Song S, Zheng YJ, He YL. A review of methods for bias correction in medical images[J]. Biomed Eng Rev, 2017, 3(1). DOI: 10.18103/bme.v3i1.1550.
[38]
Vovk U, Pernus F, Likar B. A review of methods for correction of intensity inhomogeneity in MRI[J]. IEEE Trans Med Imaging, 2007, 26(3): 405-421. DOI: 10.1109/TMI.2006.891486.

上一篇 合成MRI与传统定量方法对T1、T2弛豫值测定的模体验证对比研究
下一篇 基于压缩感知3D-FLAIR序列与常规2D-FLAIR序列在脑白质病变成像中的对比研究
  
诚聘英才 | 广告合作 | 免责声明 | 版权声明
联系电话:010-67113815
京ICP备19028836号-2