Relaxation Parameter Mapping Adapted for 7T and Validation against Optimized Single Voxel MRS
Relaxation Parameter Mapping Adapted for 7T and Validation against Optimized Single Voxel MRS
复制标题
适用于 7T 的松弛参数映射以及针对优化的单体素 MRS 的验证
DOI：
--
复制
发表时间：
2013
期刊：
影响因子：
--
通讯作者：
A. Henning
中科院分区：
文献类型：
--
作者：
M. Wyss;T. Kirchner;A. Ringenbach;K. Prüssmann;A. Henning
先前发表的脑组织在7T时的T1松弛时间在结果和方法上差异很大[1-7]，而只有两篇出版物[8,9]评估了人脑在7T时的体内T2松弛时间。这两篇文章发表了来自视觉皮层的T2值，我们知道T2松弛与大脑其他区域不同。因此，本研究的目的是开发和验证可靠的、高分辨率的T1和T2定位序列，适用于存在相关传输B1不均匀性的7T，并确定多脑区水的T1和T2弛豫时间。对于T1弛豫时间映射，采用了具有绝热反演预脉冲和改进拟合程序的Look-Locker序列[10]，并根据光谱数据进行了验证。对于T2弛豫时间映射，供应商预先实现的混合成像序列[11]通过MRS数据进行了验证，并被证明是一个可靠的T2映射序列，适用于超高场系统。为了交叉验证目的，我们还通过优化的单体素MRS参数序列扫描方案，在7特斯拉下人脑白质（wm）、前扣带回灰质（gm PAC）、尾状核和脑脊液（CSF）等解剖位置测量了T1和T2松弛时间。材料和方法所有测量数据均在7T MR系统（Philips Healthcare, Cleveland， USA）上获得，使用正交发射头线圈和32通道接收阵列（NOVA Medical, Wilmington， USA）。在每次测量之前，自动音量基于三阶FASTERMAP振荡[12]和手动F0调整。对9名健康志愿者（中位年龄26岁，3名女性，6名男性）进行测量，并根据当地伦理法规给予知情同意。SV-MRS：使用非常小（0.343cm）体素尺寸的未抑制单体素水光谱，结合STEAM定位和翻转角优化的外体积抑制。T1测量记录了绝热反演脉冲（TR=10-20s， 12次红外，NSA= 2-4， TA=6 ~ 10分钟）的反演恢复序列，T2测量记录了回波时间序列（TR=6-10s， 8次回波，NSA=4-12， TA=5-9分钟）。基于体素的翻转角度优化应用于最小化B1效果。在水峰区域拟合三参数模型Mz(t)=M0-(M0-Mz(0))*exp（-t/T1）确定T1，在相应区域拟合两参数模型Mxy(TE)=Mxy(0)*exp（-TE/T2）确定T2。为了解释部分体积效应，在PAC区域同时激发gm和CSF的两种情况下都使用了双指数拟合。成像：对于Look-Locker序列，使用以下参数：TR= 10s， TR=8.2ms， TE=4.9ms，读出翻转角= 7°，5mm切片厚度，30张图像，TA= 4:50 min，平面分辨率为1mm。采用高绝热反演脉冲（超割脉冲，持续时间22ms，振幅15μT），通过修正拟合模型，修正了由于B1不均匀性（尤其是顶叶区域）引起的残余反演缺陷：M(t)=M0-(-cos(β)+ M∞)*exp(-τ/T1*)，其中1/T1*=1/T1 - ln (cos(α)/τ)， β为反演预脉冲，α为TFE读取翻转角。已经预先实现的混合序列同时提供T1、T2和rho映射。由于序列中B1敏感的90°和180°脉冲，混合序列的T1值（722ms（±94））被低估，因此本研究仅用MRS验证T2图谱。分辨率参数与Look-Locker保持一致。
INTRODUCTION The previously published T1 relaxation times for brain tissue at 7T vary greatly in results and in methods [1-7] while only two publications [8, 9] assessed in vivo T2 relaxation times in the human brain at 7T. These two articles published T2 values from the visual cortex where it is known that T2 relaxation is different than in other brain regions. The aim of this study was hence the development and validation of reliable, high resolution T1 and T2 mapping sequences applicable to 7T in the presence of related transmit B1 inhomogeneity and the determination of T1 and T2 relaxation times of water in multiple brain regions. For T1 relaxation time mapping, a Look-Locker sequence [10] with an adiabatic inversion prepulse and a modified fitting routine was implemented and validated against spectroscopic data. For T2 relaxation time mapping the vendor pre-implemented mixed imaging sequence [11] was validated with MRS data and proofed to be a reliable T2 mapping sequence for application at ultra-high field systems. For cross validation purpose T1 and T2 relaxation times were also measured in the human brain at 7 Tesla at selected anatomical locations like white matter (wm), grey matter in the perigenual anterior cingulate gyrus (gm PAC), nucleus caudate and in cerebrospinal fluid (CSF) by optimized single voxel MRS parameter series scan protocols. MATERIAL AND METHODS All measurements were acquired on a 7T MR system (Philips Healthcare, Cleveland, USA) using a quadrature transmit head coil together with a 32-channel receive array (NOVA Medical, Wilmington, USA). Automatic volume based third order FASTERMAP shimming [12] and manual F0 adjustment was performed prior to each measurement. Nine healthy volunteers (median age 26 years, 3 female, 6 male) were measured and gave informed consent in line with local ethics regulations. SV-MRS: Unsuppressed single voxel water spectra using very small (0.343cm) voxel sizes with STEAM localization combined with flip angle-optimized outer volume suppression has been used. For T1 measurement an inversion recovery series with an adiabatic inversion pulse (TR=10-20s, 12 IR times, NSA =2-4, TA=6 to 10 minutes) and for T2 measurements an echo time series (TR=6-10s, 8 echo times, NSA=4-12, TA=5-9 minutes) were recorded. Voxel based flip angle optimization was applied to minimize B1 effects. The three parameter model Mz(t)=M0-(M0-Mz(0))*exp(-t/T1) was fitted to the areas of the water peak to determine T1, and the two parameter model Mxy(TE)=Mxy(0)*exp(-TE/T2) to determine T2 in the respective region. A biexponential fit was used in both cases in the PAC region where gm and CSF was excited simultaneously in order to account for partial volume effects. Imaging: For the Look-Locker sequence the following parameters were used: TR for Inversion=10s, TR=8.2ms, TE=4.9ms, readout flipangle = 7°, 5mm slice thickness, 30 images, TA= 4:50 minutes with an in plane resolution of 1 mm. A highly adiabatic inversion pulse (hypersecant, duration 22ms, amplitude 15μT) was used and the residual inversion imperfections due to B1 inhomogeneity (especially in parietal regions) were corrected by modifying the fitting model as follows: M(t)=M0-(-cos(β)+m∞)*exp(-τ/T1*), where 1/T1*=1/T1 – ln (cos(α)/τ), β is the inversion prepulse and α the TFE readout flip angle. The already pre-implemented mixed sequence provides T1, T2 and rho maps simultaneously. Only the T2 maps are being validated with MRS in this study because T1 values from the mixed sequence (722ms (±94)) are underestimated due to the B1 sensitive 90° and 180° pulses in the sequence. Resolution parameters were kept identical as the Look-Locker.