分享:
分享到微信朋友圈
X
技术研究
压缩感知联合并行采集技术的屏气 3D LAVA FLEX序列在肝脏磁共振快速成像中的应用
方子榕 陈秋雁 叶灵 余波 陈志健

Cite this article as: FANG Z R, CHEN Q Y, YE L, et al. Application of compressed sensing combined with parallel acqusition technique of breath-holding 3D LAVA FLEX sequence in rapid magnetic resonance imaging of liver[J]. Chin J Magn Reson Imaging, 2024, 15(2): 155-161.本文引用格式方子榕, 陈秋雁, 叶灵, 等. 压缩感知联合并行采集技术的屏气 3D LAVA FLEX序列在肝脏磁共振快速成像中的应用[J]. 磁共振成像, 2024, 15(2): 155-161. DOI:10.12015/issn.1674-8034.2024.02.023.


[摘要] 目的 探讨压缩感知(compressed sensing, CS)不同加速因子(acceleration factor, AF)联合并行采集技术(parallel acquisition technique, PAT)基于区域增长和相位校正水脂分离技术的三维肝脏容积加速采集(three-dimensional liver acqusition with volume acceleration flexible, 3D LAVA FLEX)屏气序列在肝脏MR快速成像中的应用。材料与方法 招募28名健康受试者,采用GE Architect 3.0 T MR仪行结合PAT AF 2以及不同CS AF的屏气3D LAVA FLEX序列的肝脏扫描。扫描序列分为五组,分别为PAT 2组和基于PAT 2的CS 1.2、CS 1.5、CS 2、CS 2.4组。由两位观察者分别对肝脏的水相、同相位(in-phase, IP)、反相位(opposed-phase, OP)图像质量进行5分制主观评分。取肝门水平的同一层面,放置感兴趣区域(region of interest, ROI)于肝右叶前、后段和肝左叶内、外段及其同一相位方向的右侧竖脊肌。记录各个ROI的平均信号强度(signal intensity, SI)及竖脊肌的噪声强度(standard deviationg, SD)值,分别计算图像信噪比(signal-to-noise ratio, SNR)以及对比噪声比(contrast-to-noise ratio, CNR)。用组内相关系数(intra-class correlation cofficient, ICC)分析两位观察者主观评分和测量数据一致性;用Kruskal-Wallis H检验分析主观评分间差异;使用单因素方差分析(analysis of variance, ANOVA)对比不同组间SNR、CNR的差异。结果 两位观察者对图像质量的主观评分及测量数据均为一致性良好(ICC>0.75)。五组水相、IP和OP图的主观评分组间差异有统计学意义(P<0.001,P<0.001,P<0.001)。五组图像肝右叶前、后段和肝左叶内、外段SNR的组间差异有统计学意义(P<0.001,P=0.004,P=0.002,P<0.001),CNR的组间差异无统计学意义(P=0.802,P=0.979,P=0.772,P=0.910)。当CS AF=2.4时,水相、IP、OP图像的主观评分与常规PAT 2组差异有统计学意义(P=0.009,P<0.001,P<0.001),肝右叶前、后段和肝左叶内、外段SNR与常规PAT 2组差异有统计学意义(P=0.010,P=0.002,P<0.001,P<0.001)。结论 随着CS AF的增加,扫描时间逐渐缩短,在保证图像质量的前提下,基于PAT AF 2的屏气3D LAVA FLEX序列临床推荐CS AF 2作为肝脏动态增强和IP、OP成像的扫描参数,并且比常规PAT AF 2的方案扫描时间下降44%,成像效率大幅提高。
[Abstract] Objective To investigate the application of compressed sensing (CS) with different acceleration factors (AF) and combined with parallel acqusition technique (PAT) of three-dimensional liver acqusition with volume acceleration flexible (3D LAVA FLEX) breath-holding sequence in rapid magnetic resonance imaging of liver.Materials and Methods A total of 28 healthy subjects were recruited for liver MRI scan using breath-holding 3D LAVA FLEX sequence combined with PAT AF 2 and CS with different AF at GE Architect 3.0 T MR scancer. The scanning sequences were divided into PAT 2 group and CS 1.2, CS 1.5, CS 2, CS 2.4 group based on PAT 2. Two observers were subjectively rated a 5-point scale on the image quality of water phase, in-phase (IP) and opposed-phase (OP) of liver. In the same slice of porta hepatis, the regions of interest (ROI) were placed on the anterior and posterior segments of the right hepatic lobe, the inner and outer segments of the left hepatic lobe, and the right erector spinae. The signal intensity (SI) value of each ROI and the standard deviationg (SD) value of erector spinae were recorded, and calculated the signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR). Inter-class correlation cofficient (ICC) was used to analyze the consistency of subjective scores and measurement data of the two observers. Kruskal-Wallis H test was used to analyze the differences among subjective scores. One-way analysis of variance (ANOVA) was used to compare the differences of SNR and CNR among different groups.Results There was good agreement between the two observers on the subjective score of image quality and the measured data (ICC>0.75). There was significant difference in subjective evaluation among the five groups of water phase, IP and OP images (P<0.001, P<0.001, P<0.001). There was significant difference in the SNR of the right anterior and posterior lobe and the inner and outer segments of the left lobe among the five groups (P<0.001, P=0.004, P=0.002, P<0.001), but there was no significant difference in the CNR among the five groups (P=0.802, P=0.979, P=0.772, P=0.910).When CS AF=2.4, the subjective scores of water phase, IP and OP images were significantly different from those in the regular PAT 2 group (P=0.009, P<0.001, P<0.001), and the SNR in the anterior and posterior segments of the right hepatic lobe,the inner and outer segments of the left hepatic lobe were significantly different from those in the conventional PAT 2 group (P=0.010, P=0.002, P<0.001, P<0.001).Conclusions With the increase of CS AF, the scanning time was gradually shortened. Under the premise of ensuring image quality, CS AF 2 was clinically recommended as the scanning parameter of liver dynamic enhancement and IP, OP imaging for breath-holding 3D LAVA FLEX sequence based on PAT AF 2, and the scanning time was reduced by 44% compared with the regular scanning scheme of PAT AF 2, and the imaging efficiency was greatly improved.
[关键词] 肝脏;图像质量;压缩感知;并行采集技术;加速因子;磁共振成像
[Keywords] liver;image quality;compressed sensing;parallel acqusition technique;acceleration factor;magnetic resonance imaging

方子榕    陈秋雁 *   叶灵    余波    陈志健   

宁德师范学院附属宁德市医院放射科,宁德 352100

通信作者:陈秋雁,E-mail:lc7505723@163.com

作者贡献声明::陈秋雁设计本研究的方案,对稿件重要内容进行修改;方子榕实施研究,起草和撰写稿件,获取、分析和解释本研究的数据;叶灵、余波、陈志健获取、分析或解释本研究的数据,对稿件重要内容进行修改。全体作者都同意发表最后的修改稿,同意对本研究的所有方面负责,确保本研究的准确性和诚信。


收稿日期:2023-05-26
接受日期:2024-01-25
中图分类号:R445.2  R657.3 
文献标识码:A
DOI: 10.12015/issn.1674-8034.2024.02.023
本文引用格式方子榕, 陈秋雁, 叶灵, 等. 压缩感知联合并行采集技术的屏气 3D LAVA FLEX序列在肝脏磁共振快速成像中的应用[J]. 磁共振成像, 2024, 15(2): 155-161. DOI:10.12015/issn.1674-8034.2024.02.023.

0 引言

       MRI具有无电离辐射、非侵入性、软组织对比度高等优点[1, 2],在肝脏疾病如肝癌[3]、血管瘤[4]、转移瘤[5]等方面的诊断及鉴别诊断有着不可替代的优势。基于区域增长和相位校正水脂分离技术的三维肝脏容积加速采集(three-dimensional liver acqusition with volume acceleration flexible, 3D LAVA FLEX)屏气序列扫描效率高,一次扫描可获得四种对比图像:水相、脂相、同相位(in-phase, IP)、反相位(opposed-phase, OP)[6],其水相是临床动态增强扫描的主要影像,而IP、OP是临床诊断和鉴别诊断脂肪性病变的重要参照[7]。但是肝脏MRI检查对患者要求较高,在成像过程中需要受检者屏气,目前临床上常规的屏气扫描时间为15~20 s[8],有些心肺疾病或者年纪大的患者无法进行长时间的屏气,使得图像质量受到呼吸运动伪影的影响,对临床诊断造成不小的阻碍。因此如何提高屏气3D LAVA FLEX序列的成像速度至关重要。

       并行采集技术(parallel acquisition technique, PAT)是广泛使用的快速成像技术[9],其加速因子(acceleration factor, AF)的增加可使扫描时间下降,但过高的AF是以牺牲图像信噪比(signal-to-noise ratio, SNR)或空间分辨力为代价,已难以满足临床快速成像的要求。近年来随着压缩感知(compressed sensing, CS)的引入,使得进一步加快MRI的速度得以实现,同时能够减少伪影,提高成像质量[10]。CS利用了信号的稀疏特性,在远小于Nyquist采样率的条件下[11],通过随机采样来获得信号的离散样本,然后用非线性迭代算法进行图像重建,可以在仅仅采集少量数据的同时,保持良好的图像质量[12]。CS AF对MRI速度的影响也是近年来的研究热点[13, 14]。TRAN等[15]将CS与PAT结合应用于MRI的数学算法研究,并且通过计算颅脑影像的图像质量指数和SNR,得出CS与PAT的结合能更好地加速MRI采集,并且相比单纯应用PAT在图像质量方面均有所提高,为CS联合PAT的应用提供了理论依据。近年来,CS联合PAT的快速成像方法在临床上已有研究进展[16, 17],并且证实了联合成像方案在缩短成像时间和保证图像质量上的优越性。SUN等[18]将CS联合PAT的三维T1梯度回波序列应用于肝脏增强MRI的研究,得出CS联合PAT的成像方案相比常规方案在图像质量、病变对比方面更具优势。YOUNG等[19]将CS联合PAT的T1序列用于自由呼吸状态下的肝脏增强MRI,研究发现联合方案能够使呼吸运动伪影受到较好的控制,但存在容积伪影的影响,并不能很好地解决问题。以往将CS联合PAT应用于肝脏MRI的研究多数局限于普通T1序列的研究,也缺乏对CS AF的进一步探讨,目前国内尚未有将CS联合PAT应用于肝脏屏气3D LAVA FLEX序列的研究报道。本研究将CS联合PAT应用于对肝脏肿瘤病变的诊断及鉴别诊断更为敏感的屏气3D LAVA FLEX序列,在基于PAT AF 2的基础上,探讨不同CS AF对屏气3D LAVA FLEX序列中水相、IP、OP的图像质量的影响以及进一步提高成像效率的可行性。从而最大限度地解决屏气困难的患者难以进行肝脏屏气MRI的问题,对提高肝脏动态增强MRI和IP、OP成像的成功率,以及肝脏疾病的临床诊断效率有重要意义。

1 材料与方法

1.1 研究对象

       于2022年11月至2023年4月在我院招募28名健康受试者进行肝脏MRI扫描。其中男16名,女12名,年龄27~49(35.64±5.44)岁,纳入标准:(1)无MRI检查禁忌证;(2)无肝脏肿物或弥漫性病灶;(3)无肝脏手术史。排除标准:(1)体质量指数高于24 kg/m2或体质量指数低于18.5 kg/m2者;(2)呼吸配合不佳者。本研究遵守《赫尔辛基宣言》,经宁德师范学院附属宁德市医院伦理委员会批准(批准文号:NSRKYLL-2023-024),所有受试者均签署知情同意书。

1.2 设备与方法

       采用美国GE Architect 3.0 T 超导MR系统,30通道腹部相控阵线圈。嘱受检者检查前4 h禁食,尽可能避免肝左叶信号受到过多胃容物的影响。检查前对受试者进行屏气训练,于呼气末屏气。检查取仰卧位、头先进,中心线定于剑突。分别用PAT AF 2以及基于PAT AF 2,CS AF分别为1.2、1.5、2和2.4的屏气3D LAVA FLEX序列行肝脏MRI。各组除CS AF外其他参数的设置保持一致,扫描参数:TR 4.9 ms,TE 2 ms,翻转角12°,视野400 mm×320 mm,矩阵320×256,层数52,层厚4.4 mm,层间距0,体素大小1.2 mm×1.6 mm×4.4 mm。

1.3 图像分析和数据测量

1.3.1 图像主观评分

       由观察者1(具有9年MRI工作经验的主管技师)和观察者2(具有6年MRI工作经验的主管技师)采用双盲法于肝门水平对图像质量进行独立分析。水相图从脂肪抑制效果、噪声和伪影、肝内解剖结构方面[20],IP、OP图从噪声和伪影、肝内解剖结构、诊断价值对五组图像进行5分制主观评分,大于等于3分为满足诊断要求,见表1图12

图1  男,32岁,肝门水平层面。1A~1E分别为并行采集技术(PAT)加速因子(AF)为2和基于PAT AF 2,压缩感知(CS)AF分别为1.2、1.5、2、2.4的横断位基于区域增长和相位校正水脂分离技术的三维肝脏容积加速采集(3D LAVA FLEX)序列肝脏同层面水相图。可见随着CS AF的增加,PAT AF 2与基于PAT AF 2,CS AF分别为1.2、1.5、2时图像无明显噪声和伪影,CS AF为2.4时噪声和伪影较为明显。1F为在肝右叶前、后段和肝左叶内、外段以及右侧竖脊肌放置感兴趣区(ROI)测量平均信号强度及竖脊肌的噪声强度值,圆形蓝色区域为ROI。
Fig. 1  Male, 32 years old, slice of hilar level. 1A-1E show water phase images of transsectional three-dimensional liver acqusition with volume acceleration flexible (3D LAVA FLEX) sequence of liver at the sames level with parallel acquisition technique (PAT) acceleration factor (AF) 2 and based on PAT AF 2, compressed sensing (CS) AF are 1.2, 1.5, 2 and 2.4, respectively. There is no obvious noise and artifact with PAT AF 2 and based on PAT AF 2, CS AF were 1.2, 1.5 and 2. When CS AF is 2.4, noise and artifacts are more obvious. 1F shows the region of interest (ROIs) are placed on anterior and posterior segments of the right hepatic lobe, the inner and outer segments of the left hepatic lobe and the right erector spinae to measure signal intensity (SI) and standard deviationg (SD) values, and the circular blue areas are ROIs.
图2  男,32岁,肝门层面图像质量较差。2A:基于并行采集技术(PAT)加速因子(AF)为2,压缩感知(CS)AF为2.4的同相位(IP)图;2B:基于PAT AF为2,CS AF为 2.4的反相位(OP)图。当CS AF为2.4时,IP和OP图像的噪声和伪影较为明显。
Fig. 2  Male, 32 years old, the image quality of hilar level is poor. 2A: The in-phase (IP) image of compressed sensing (CS) acceleration factor (AF) 2.4 based on parallel acquisition technique (PAT) AF 2. 2B: The opposed-phase (OP) images of CS AF 2.4 based on PAT AF 2. When CS AF is 2.4, the noise and artifacts of IP and OP images are obvious.
表1  主观评分标准表
Tab. 1  Table of subjective scoring standards

1.3.2 数据测量

       将图像传至GE后处理工作站(AW Volumeshare 7),取水相图,由观察者1、观察者2独立进行测量,分别于肝门水平层面的肝右叶前段、右叶后段、左叶内段、左叶外段以及右侧竖脊肌内的信号均匀处勾画圆形感兴趣区(region of interest, ROI),肝内ROI大小约为200 mm2,距肝包膜约3 cm,应尽量避开胆管和血管;右侧竖脊肌内ROI大小约为100 mm2,避开脂肪间隙。分别测量肝脏、竖脊肌的平均信号强度(signal intnsity, SI)和竖脊肌的噪声强度(standard deviationg, SD),并计算肝脏各叶的SNR和对比噪声比(contrast-to-noise ratio, CNR),计算公式见式(1)、(2)。

1.4 统计学分析

       采用SPSS 27.0统计学软件处理数据。通过计算组内相关系数(intra-class correlation cofficient, ICC),验证两位观察者主观评分和客观测量的一致性,ICC≤0.40表示一致性较差,0.40<ICC≤0.75为一致性一般,0.75<ICC≤1.00为一致性良好。根据Kolmogorov-Smirnov检验数据是否服从正态分布,采用Kruskal-Wallis H检验分析主观评分间差异,采用单因素方差分析(analysis of variance, ANOVA)及最小显著性差异法(least-significant difference, LSD)进行组间SNR、CNR总体差异的分析及组间两两比较。以P<0.05为差异具有统计学意义。

2 结果

2.1 五组扫描时间

       当PAT AF=2时,扫描时间为18 s,基于PAT AF 2,随着CS AF的增加,扫描时间可进一步缩短,当CS AF分别为1.2、1.5、2、2.4时,扫描时间分别为15、12、10、9 s。

2.2 一致性分析

       两位观察者根据评分标准评价五组图像质量,水相图主观评分一致性分析ICC值分别为0.818、0.878、0.918、0.768、0.794,IP图主观评分一致性分析ICC值分别为0.788、0.813、0.825、0.796、0.787,OP图主观评分一致性分析ICC值分别为0.786、0.781、0.764、0.804、0.802,一致性均为良好(ICC>0.75)。两位观察者对肝脏、右侧竖脊肌SI和右侧竖脊肌SD的测量的一致性良好(ICC>0.75),结果见表2

       取观察者1(高年资,具有9年MRI工作经验的主管技师)的数据进行图像质量分析,通过Kruskal-Wallis H检验水相、IP和OP图像质量主观评分,各组间差异有统计学意义(P<0.001,P<0.001,P<0.001)。采用ANOVA分析五组组间肝脏SNR及CNR,结果显示在肝右叶前、后段和肝左叶内、外段的五组图像的SNR的组间差异有统计学意义(P<0.001,P=0.004,P=0.002,P<0.001),CNR组间差异无统计学意义(P=0.802,P=0.979,P=0.772,P=0.910),结果见表3

       主观评分两两比较:当CS=2.4时,水相、IP、OP图的主观评分与常规PAT 2组差异有统计学意义(P=0.009,P<0.001,P<0.001),结果见表4。采用LSD进一步对SNR进行组间两两比较,结果显示当CS=2.4时,肝右叶前、后段和肝左叶内、外段SNR与常规PAT 2组差异均有统计学意义(P=0.010,P=0.002,P<0.001,P<0.001),结果见图3

图3  各个肝脏区域在并行采集技术(PAT)加速因子(AF)为2和基于PAT AF 2,压缩感知(CS)AF分别为1.2、1.5、2、2.4扫描参数下的信噪比(SNR)箱式图。3A:肝右叶前段SNR;3B:肝右叶后段SNR;3C:肝左叶内段SNR;3D:肝左叶外段SNR。当CS=2.4时,在肝右叶前、后段和肝左叶内、外段的SNR与常规PAT 2组的差异均有统计学意义。CS 2.4组较PAT 2、CS 1.2、CS 1.5、CS 2组的肝右叶前、后段和肝左叶内、外段的SNR均明显下降。
Fig. 3  The box diagrams of signal noise ratio (SNR) of each liver region with parallel acquisition technique (PAT) acceleration factor (AF) 2 and based on PAT 2, compressed sensing (CS) AF are 1.2, 1.5, 2 and 2.4, respectively. 3A: SNR of anterior segment of the right hepatic lobe; 3B: SNR of posterior segment of the right hepatic lobe; 3C: SNR of inner segment of the left hepatic lobe; 3D: SNR of outer segment of the left hepatic lobe. When CS=2.4, SNRs in anterior and posterior segments of the right hepatic lobe and the inner and outer segments of the left hepatic lobe are significanly different from that in regular PAT 2 group. It can be seen that SNRs in anterior and posterior segments of the right hepatic lobe, the inner and outer segments of the left hepatic lobe in group CS 2.4 are significantly decreased compared with those in PAT 2, CS 1.2, CS 1.5 and CS 2 groups.
表2  两位观察者测量数据及主观评价一致性检验
Tab. 2  Consistency test of measurement data and subjective evaluation between the two observers
表3  SNR、CNR以及主观评分的差异性比较
Tab. 3  Comparison of SNR, CNR and subjective score differences
表4  常规PAT AF 2组与不同CS AF组的主观评分两两比较结果
Tab. 4  Results of pairwise comparison of subjective ratings between regular PAT AF 2 group and different CS AF group

3 讨论

       本研究将CS联合PAT的屏气3D LAVA FLEX序列运用于肝脏MRI,优化成像时间较短,图像质量较高的CS AF。结果显示基于PAT AF 2,当CS AF为2时,肝脏的SNR、CNR以及主观评分与常规PAT AF 2序列差异无统计学意义,能够在保证图像质量前提下使扫描时间缩短44%,成像效率大幅提高,为国内首次提出。不仅可以最大限度地解决心肺疾病或者年纪大的患者行肝脏动态增强扫描和IP、OP成像时无法进行长时间屏气的问题,还可提高临床对肝脏疾病诊断效率,具有良好的临床应用前景。

3.1 CS联合PAT的3D LAVA FLEX序列应用于肝脏MRI的优势和可行性

       CS可通过随机稀疏采样的方式获得离散样本信号,然后采用非线性迭代重建算法获得稀疏信号,实现部分K空间采集,从而明显缩短MRI时间[21]。魏强等[22]的研究得出:在单纯应用PAT AF 2时,在肝脏三维T1水脂分离屏气序列的成像时间为15.1 s,而使用结合CS AF 2的序列能在保证图像质量的前提下使扫描时间缩短至13.1 s,印证了结合CS的屏气MRI序列在肝脏MRI中应用的可行性。本研究在基于PAT AF 2的基础上展开对CS AF的探讨,并优化最佳的CS AF。IHARA等[23]在肝脏MRI应用结合CS的T1梯度回波增强扫描序列与常规PAT序列进行对比,结果表明应用CS的成像序列能够获得肝脏解剖结构、病灶对比与常规PAT序列无差异的图像,而成像时间降低至17 s。本研究以CS联合PAT的3D LAVA FLEX序列在CS AF为2时能使成像时间缩短至10 s,与IHARA等的研究结果相比,本研究在各个肝区的图像质量分析与其结果一致,但在成像时间上,本研究优化后序列大幅缩短。RAJLAWOT等[24]将LAVA FLEX序列用于肝细胞癌内脂肪的评估,结果得出LAVA FLEX序列的IP、OP像在测定肝细胞癌的脂肪分数方面准确性高,但成像时间需19 s,而本研究的成像方案在CS AF为2时,成像时间比前者缩短47%,而且图像主观评分与常规序列相当,极大地提高了成像效率。本研究还注意到当3D LAVA FLEX序列CS AF为1.2、1.5、2时,图像的主观评分、SNR并没有随着CS AF的增加而出现明显下降,而CS AF为2.4时在肝左外叶、右后叶、尾状叶处容易出现噪声和伪影。分析发现,本研究CS AF为1.2、1.5、2时,虽然获取离散样本数减少,但随机丢失的信号不足以影响图像的重建,因此噪声和伪影不明显。当选择CS AF 2.4时,获取的离散样本数进一步减少,使得图像重建步骤频繁发生错误,K空间欠采样加重,从而产生较为明显的混叠样噪声和伪影[25]。潘珂等[26]在膝关节MRI的研究发现,CS AF大于2时,随着AF的增加图像噪声增多,主观评分下降,与本研究结果相符。而TOSHIMORI等[27]在结合CS与PAT的肝脏MRI研究中得出,当CS大于1.4时图像的主观评分降低,肝右叶图像的模糊效应增加,与本研究结果不一致,经分析,可能为3D LAVA FLEX序列有着独特的K空间数据填充技术和水脂分离技术[28],使其在图像分辨率上比常规肝脏MRI序列更具优势,因此,能够在较大的CS AF下,仍能保证图像质量。

3.2 CS联合PAT的屏气3D LAVA FLEX序列优化肝脏MRI扫描方案

       PAT虽然能提高MRI的时间分辨率,提高MRI检查效率[29],但随着AF的增加,PAT的快速采集是以K空间的欠采样为基础,以牺牲图像SNR或空间分辨力为代价[30]。早期YOON等[31]将PAT AF 4的三维T1加权梯度回波用于肝脏MRI,能在18.5 s时间获得单次屏气的图像,但是图像噪声相比PAT AF 2.6的成像序列明显增加。本研究在PAT AF 2的基础上结合了CS,随着CS AF的增加,成像时间进一步缩短,并且能够在一定程度上保证了图像质量,对比前者研究中增加PAT AF所导致的图像质量问题,本研究利用了CS使用非线性重建来避免欠采样伪影,从而快速获取高质量的图像的特点[32],证实了本研究比传统单纯基于PAT的肝脏MRI更具优越性,同时也进一步实现了肝脏快速MRI。此外,有研究通过自适应的呼吸训练来改善肝脏MRI的呼吸配合不耐受问题[33],在一定程度上改善了图像质量,但其局限性是在检查前期需要耗费一定的训练时间,并且对于心肺疾病或者年长者仍是挑战。本研究结果显示CS AF为2时,能够在保证图像质量前提下使扫描时间缩短44%,相比前者研究,本研究能够在较短的时间内完成扫描,提高了屏气困难的受检者行肝脏MRI的成功率,进一步扩大肝脏MRI的受检群体。PAN等[34]通过调整序列参数,应用连续采集的肝脏动态增强MRI,对难以长时间屏气的受检者进行自由呼吸状态下的扫描,在减少呼吸运动伪影和提高图像清晰度方面有所提升,但其局限性是存在容积伪影,容易对微小病变漏诊。本研究应用CS AF 2联合PAT AF 2的屏气3D LAVA FLEX序列可使屏气时间大于等于10 s的受检者完成扫描,成像时间大幅缩短,可较大程度地减少以往研究中无法长时间屏气的受检者选择自由呼吸状态下行肝脏MRI的可能性,从而尽可能避免了容积伪影及漏诊问题。此外,3D LAVA FLEX序列具有独特的K空间数据填充技术以及水脂分离技术,使其在时间、空间分辨率和脂肪抑制效果上相比常规的二维和三维扰相梯度回波序列更具优势[35],更有利于临床对肝脏病变的诊断。

3.3 局限性

       本研究的局限性:(1)本研究样本量偏少,不能完全评估大样本的实际情况,有待进一步扩大;(2)因实验组序列有多组,成像时间较长,本研究只针对健康且会配合屏气人群,对肝脏疾病患者(如肿瘤)未触及,但在后续的临床应用中显示,结合了CS和PAT的屏气3D LAVA FLEX序列能够提高屏气困难患者的检查成功率,且在动态增强扫描和IP、OP成像上均能良好地显示及鉴别病变;(3)3D LAVA FLEX序列应用于肝脏动态增强扫描时需对患者注射对比剂,本研究受试者均未注射对比剂,但在后期的临床应用上,结合了CS和PAT的肝脏3D LAVA FLEX序列在各个增强期上对病灶的对比具有良好的表现。

4 结论

       综上所述,CS联合PAT的3D LAVA FLEX屏气序列能在保证图像质量的前提下,使成像时间进一步缩短。基于PAT AF 2的基础上,在CS AF为2时能在保证图像SNR、CNR以及良好的主观评分的同时缩短成像时间至10 s,相比常规PAT AF 2的扫描方案减少了约44%的成像时间,提高了屏气困难患者行肝脏动态增强MRI和IP、OP成像的成功率,以及临床对肝脏疾病诊断的效率,具有良好的临床应用及推广价值。

[1]
GATTI M, MAINO C, TORE D, et al. Benign focal liver lesions: the role of magnetic resonance imaging[J]. World J Hepatol, 2022, 14(5): 923-943. DOI: 10.3934/mbe.2021116.
[2]
BELLO R DAL, LAPAEVA M, LA GRECA SAINT-ESTEVEN A, et al. Patient-specific quality assurance strategies for synthetic computed tomography in magnetic resonance-only radiotherapy of the abdomen[J/OL]. Phys Imaging Radiat Oncol, 2023, 27: 100464 [2023-05-25]. https://pubmed.ncbi.nlm.nih.gov/37497188/. DOI: 10.1016/j.phro.2023.100464.
[3]
XIAO Y Y, ZHOU C W, NI X Y, et al. Preoperative subcategorization based on magnetic resonance imaging in intrahepatic cholangiocarcinoma[J/OL]. Cancer Imaging, 2023, 23(1): 15 [2023-05-25]. https://pubmed.ncbi.nlm.nih.gov/36782276/. DOI: 10.1186/s40644-023-00533-2.
[4]
LIU D, QIAO Z G, XU L H, et al. Gd-EOB-DTPA enhanced MRI features of liver hemangiomatosis coexistent with GCH[J]. Curr Med Imaging, 2022, 18(13): 1390-1395. DOI: 10.2174/1573405618666220602113223.
[5]
ICHIKAWA S, GOSHIMA S. Clinical significance of liver MR imaging[J]. Magn Reson Med Sci, 2023, 22(2): 157-175. DOI: 10.2463/mrms.rev.2022-0100.
[6]
SAMJI K, ALRASHED A, SHABANA W M, et al. Comparison of high-resolution T1W 3D GRE (LAVA) with 2-point Dixon fat/water separation (FLEX) to T1W fast spin echo (FSE) in prostate cancer (PCa)[J]. Clin Imaging, 2016, 40(3): 407-413. DOI: 10.1016/j.clinimag.2015.11.023.
[7]
AYDIN H, AYDIN H, KARAIBRAHIMOĞLU A, et al. Evaluation of the parenchymal distribution of renal steatosis in chronic kidney disease using chemical shift magnetic resonance imaging[J/OL]. Adv Clin Exp Med, 2023 [2023-05-25]. https://pubmed.ncbi.nlm.nih.gov/37341174/. DOI: 10.17219/acem/166512.
[8]
邓和平, 张宇, 杨学刚, 等. 光梭径向采集技术在屏气配合欠佳状态下肝脏动态增强扫描中的应用价值[J]. 放射学实践, 2023, 38(5): 631-635. DOI: 10.13609/j.cnki.1000-0313.2023.05.018.
DENG H P, ZHANG Y, YANG X G, et al. Application of united compressed sensing with radial acquisition technology in liver dynamic contrast-en-hanced scanning in patients with poor breath-holding[J]. Radiol Pract, 2023, 38(5): 631-635. DOI: 10.13609/j.cnki.1000-0313.2023.05.018.
[9]
DESHMANE A, GULANI V, GRISWOLD M A, et al. Parallel MR imaging[J]. J Magn Reson Imaging, 2012, 36(1): 55-72. DOI: 10.1002/jmri.23639.
[10]
KIM J, SEO N, BAE H, et al. Comparison of sensitivity encoding (SENSE) and compressed sensing-SENSE for contrast-enhanced T1-weighted imaging in patients with crohn disease undergoing MR enterography[J]. AJR Am J Roentgenol, 2022, 218(4): 678-686. DOI: 10.2214/AJR.21.26733.
[11]
VORNETTI G, BARTIROMO F, TONI F, et al. Follow-up assessment of intracranial aneurysms treated with endovascular coiling: comparison of compressed sensing and parallel imaging time-of-flight magnetic resonance angiography[J]. Tomography, 2022, 8(3): 1608-1617. DOI: 10.3390/tomography8030133.
[12]
MEISTER R L, GROTH M, JÜRGENS J H W, et al. Compressed SENSE in pediatric brain tumor MR imaging: assessment of image quality, examination time and energy release[J]. Clin Neuroradiol, 2022, 32(3): 725-733. DOI: 10.1007/s00062-021-01112-3.
[13]
CHOI E S, KIM J S, NICKEL M D, et al. Free-breathing contrast-enhanced multiphase MRI of the liver in patients with a high risk of breath-holding failure: comparison of compressed sensing-accelerated radial and Cartesian acquisition techniques[J]. Acta Radiol, 2022, 63(11): 1453-1462. DOI: 10.1177/02841851211052988.
[14]
杨婧, 刘杨颖秋, 刘娜, 等. 不同压缩感知因子对脑SWI "燕尾征" 分析的影响[J]. 磁共振成像, 2022, 13(5): 89-93. DOI: 10.12015/issn.1674-8034.2022.05.016.
YANG J, LIU Y Y Q, LIU N, et al. The effect of different compressed sensing factors on analysis of\ "swallow tail sign\" in cerebral SWI[J]. Chin J Magn Reson Imag, 2022, 13(5): 89-93. DOI: 10.12015/issn.1674-8034.2022.05.016.
[15]
TRAN A Q, NGUYEN T A, DOAN P T, et al. Parallel magnetic resonance imaging acceleration with a hybrid sensing approach[J]. Math Biosci Eng, 2021, 18(3): 2288-2302. DOI: 10.3934/mbe.2021116.
[16]
AOIKE T, FUJIMA N, YONEYAMA M, et al. Development of three-dimensional MR neurography using an optimized combination of compressed sensing and parallel imaging[J/OL]. Magn Reson Imaging, 2022, 87: 32-37 [2023-05-25]. https://pubmed.ncbi.nlm.nih.gov/34968698/. DOI: 10.1016/j.mri.2021.12.002.
[17]
FUJIWARA H, YABUUCHI H, WADA T, et al. High-resolution magnetic resonance imaging of the triangular fibrocartilage complex using compressed sensing sensitivity encoding (SENSE)[J/OL]. Eur J Radiol, 2022, 149: 110191 [2023-05-25]. https://pubmed.ncbi.nlm.nih.gov/35149336/. DOI: 10.1016/j.ejrad.2022.110191.
[18]
SUN W, WANG W T, ZHU K, et al. Feasibility of compressed sensing technique for isotropic dynamic contrast-enhanced liver magnetic resonance imaging[J/OL]. Eur J Radiol, 2021, 139: 109729 [2023-05-25]. https://pubmed.ncbi.nlm.nih.gov/33905976/. DOI: 10.1016/j.ejrad.2021.109729.
[19]
YOUNG PARK J, SANG M, LEE J SUB, et al. Free-breathing dynamic T1WI using compressed sensing-golden angle radial sparse parallel imaging for liver MRI in patients with limited breath-holding capability[J/OL]. Eur J Radiol, 2022, 152: 110342 [2023-05-25]. https://pubmed.ncbi.nlm.nih.gov/35597070/. DOI: 10.1016/j.ejrad.2022.110342.
[20]
王诗瑜, 刘爱连, 王家正, 等. 压缩感知不同加速因子对肝脏3D-mDixon序列成像质量的影响[J]. 放射学实践, 2020, 35(12): 1605-1609. DOI: 10.13609/j.cnki.1000-0313.2020.12.021.
WANG S Y, LIU A L, WANG J Z, et al. Study on the influence of compressed sense technique with different acceleration factors on the image quality of 3D modified-Dixon sequence of liver[J]. Radiol Pract, 2020, 35(12): 1605-1609. DOI: 10.13609/j.cnki.1000-0313.2020.12.021.
[21]
张浩南, 宋清伟, 张楠, 等. 压缩感知技术在腰椎磁共振快速成像中的应用[J]. 磁共振成像, 2023, 14(2): 132-137, 144. DOI: 10.12015/issn.1674-8034.2023.02.022.
ZHANG H N, SONG Q W, ZHANG N, et al. Application of compressed sensing technology in rapid lumbar magnetic resonance imaging[J]. Chin J Magn Reson Imag, 2023, 14(2): 132-137, 144. DOI: 10.12015/issn.1674-8034.2023.02.022.
[22]
魏强, 王家正, 伊东娜, 等. 结合压缩感知技术与敏感度编码的3D mDixon序列以及3D Vane序列对肝脏成像的影响的对比研究[J]. 磁共振成像, 2020, 11(9): 781-785. DOI: 10.12015/issn.1674-8034.2020.09.012.
WEI Q, WANG J Z, YI D N, et al. A comparative study on the effects of 3D mDixon sequence and 3D Vane sequence on liver imaging by combining the Compressed SENSE technology and SENSE[J]. Chin J Magn Reson Imag, 2020, 11(9): 781-785. DOI: 10.12015/issn.1674-8034.2020.09.012.
[23]
IHARA K, ONODA H, TANABE M, et al. Breath-hold high-resolution T1-weighted gradient echo liver MR imaging with compressed sensing obtained during the gadoxetic acid-enhanced hepatobiliary phase: image quality and lesion visibility compared with a standard T1-weighted sequence[J/OL]. Magn Reson Med Sci, 2023 [2023-05-25]. https://pubmed.ncbi.nlm.nih.gov/36740257/. DOI: 10.2463/mrms.mp.2022-0137.
[24]
RAJLAWOT K, JIANG T, ZHOU J, et al. Accuracies of chemical shift In/opposed phase and chemical shift encoded magnetic resonance imaging to detect intratumoral fat in hepatocellular carcinoma[J]. J Magn Reson Imaging, 2021, 53(6): 1791-1802. DOI: 10.1002/jmri.27539.
[25]
GAO Z F, GUO Y F, ZHANG J J, et al. Hierarchical perception adversarial learning framework for compressed sensing MRI[J]. IEEE Trans Med Imaging, 2023, 42(6): 1859-1874. DOI: 10.1109/TMI.2023.3240862.
[26]
潘珂, 刘倩倩, 唐玲玲, 等. 人工智能-压缩感知与压缩感知在膝关节MRI中的加速效率及图像质量评价研究[J]. 磁共振成像, 2022, 13(5): 94-98. DOI: 10.12015/issn.1674-8034.2022.05.017.
PAN K, LIU Q Q, TANG L L, et al. Study on acceleration efficiency and image quality of artificial intelligence compressed sensing and compressed sensing in knee MRI[J]. Chin J Magn Reson Imag, 2022, 13(5): 94-98. DOI: 10.12015/issn.1674-8034.2022.05.017.
[27]
TOSHIMORI W, MATSUDA M, TSUDA T, et al. Comparisons of hepatobiliary phase imaging using combinations of parallel imaging and variable degrees of compressed sensing with use of parallel imaging alone[J]. J Comput Assist Tomogr, 2023, 47(4): 524-529. DOI: 10.1097/rct.0000000000001451.
[28]
SAITO S, TANAKA K, HASHIDO T. Liver acquisition with volume acceleration flex on 70-cm wide-bore and 60-cm conventional-bore 3.0-T MRI[J]. Radiol Phys Technol, 2016, 9(2): 154-160. DOI: 10.1007/s12194-015-0344-z.
[29]
LEE S M, KIM M, PARK C, et al. Deep learning-reconstructed parallel accelerated imaging for knee MRI[J/OL]. Curr Med Imaging, 2024, 20: e240523217293 [2023-05-25]. https://pubmed.ncbi.nlm.nih.gov/37226797/. DOI: 10.2174/1573405620666230524151816.
[30]
TAO H, ZHANG W, WANG H F, et al. Multi-weight respecification of scan-specific learning for parallel imaging[J/OL]. Magn Reson Imaging, 2023, 97: 1-12 [2023-05-25]. https://pubmed.ncbi.nlm.nih.gov/36567001/. DOI: 10.1016/j.mri.2022.12.009.
[31]
YOON J H, LEE J M, YU M H, et al. High-resolution T1-weighted gradient echo imaging for liver MRI using parallel imaging at high-acceleration factors[J]. Abdom Imaging, 2014, 39(4): 711-721. DOI: 10.1007/s00261-014-0099-8.
[32]
张小艳, 马培旗, 袁玉山, 等. 压缩感知技术三维MRI矢状位T2加权频率衰减反转恢复序列显示距腓前韧带的价值探索[J]. 磁共振成像, 2023, 14(6): 71-74, 81. DOI: 10.12015/issn.1674-8034.2023.06.011.
ZHANG X Y, MA P Q, YUAN Y S, et al. The Value of compressed sensing 3D MRI sagittal T2 weighted imaging-spectral attenuated in-version recovery to display the anterior talofibular ligament[J]. Chin J Magn Reson Imag, 2023, 14(6): 71-74, 81. DOI: 10.12015/issn.1674-8034.2023.06.011.
[33]
FU K, RAN Q S, GUO G K, et al. Adaptive training: a new method to improve the image quality of abdominal magnetic resonance imaging[J]. Curr Med Imaging, 2023, 19(9): 1090-1095. DOI: 10.2174/1573405619666221130114646.
[34]
PAN J Y, SHAO X, LIU H, et al. Image quality optimization: dynamic contrast-enhanced MRI of the abdomen at 3T using a continuously acquired radial golden-angle compressed sensing acquisition[J/OL]. Abdom Radiol (NY), 2023 [2023-05-25]. https://pubmed.ncbi.nlm.nih.gov/37792056/. DOI: 10.1007/s00261-023-04035-4.
[35]
LIU T F, WANG Y H, XU Z Y, et al. Application study of 3D LAVA-Flex on lumbar intervertebral disc degeneration[J]. Eur J Med Res, 2021, 26(1): 43 [2023-05-25]. https://pubmed.ncbi.nlm.nih.gov/33962698/. DOI: 10.1186/s40001-021-00512-y.

上一篇 不同翻转角和延迟时间对Gd-BOPTA增强肝胆期肝脏和胆道系统图像质量的影响
下一篇 脾脏高分化神经内分泌肿瘤MRI表现一例
  
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