Alcicek Seyma, Ronellenfitsch Michael W, Steinbach Joachim P, Manzhurtsev Andrei, Thomas Dennis C, Weber Katharina J, Prinz Vincent, Forster Marie-Thérèse, Hattingen Elke, Pilatus Ulrich, Wenger Katharina J
Goethe University Frankfurt, University Hospital, Institute of Neuroradiology, Frankfurt am Main, Germany.
University Cancer Center Frankfurt (UCT), Frankfurt am Main, Germany.
J Magn Reson Imaging. 2025 Sep;62(3):890-901. doi: 10.1002/jmri.29787. Epub 2025 Apr 8.
Glutamate and glutamine are critical metabolites in gliomas, each serving distinct roles in tumor biology. Separate quantification of these metabolites using in vivo MR spectroscopy (MRS) at clinical field strengths (≤ 3T) is hindered by their molecular similarity, resulting in overlapping, hence indistinguishable, spectral peaks.
To develop an MRS imaging (MRSI) protocol to map glutamate and glutamine separately at 3T within clinically feasible time, using J-modulation to enhance spectral differentiation, demonstrate its reliability/reproducibility, and quantify the metabolites in glioma subregions.
Prospective.
Phantoms, 5 healthy subjects, and 30 patients with suspected glioma. IDH wild-type glioblastoma cases were evaluated to establish a uniform group.
FIELD STRENGTH/SEQUENCE: 3T, Research protocol: 2D H sLASER MRSI (40 and 120 ms TE) with water reference, 3D T1-weighted and 2D T2-weighted. Trial-screening process: T1-weighted, T1-weighted contrast-enhanced, T2-weighted, FLAIR.
Spectral simulations and phantom measurements were performed to design and validate the protocol. Spectral quality/fitting parameters for scan-rescan measurements were obtained using LCModel. The proposed long-TE data were compared with short-TE data. BraTS Toolkit was employed for fully automated tumor segmentation.
Scan-rescan comparison was performed using Bland-Altman analysis. LCModel coefficient of modeling covariance (CMC) between glutamate and glutamine was mapped to evaluate their model interactions for each spectral fitting. Metabolite levels in tumor subregions were compared using one-way ANOVA and Kruskal-Wallis. A p value < 0.05 was considered statistically significant.
Spectral quality/fitting parameters and metabolite levels were highly consistent between scan-rescan measurements. A negative association between glutamate and glutamine models at short TE (CMC = -0.16 ± 0.06) was eliminated at long TE (0.01 ± 0.05). Low glutamate in tumor subregions (non-enhancing-tumor-core: 5.35 ± 4.45 mM, surrounding-non-enhancing-FLAIR-hyperintensity: 7.39 ± 2.62 mM, and enhancing-tumor: 7.60 ± 4.16 mM) was found compared to contralateral (10.84 ± 2.94 mM), whereas glutamine was higher in surrounding-non-enhancing-FLAIR-hyperintensity (9.17 ± 6.84 mM) and enhancing-tumor (7.20 ± 4.42 mM), but not in non-enhancing-tumor-core (4.92 ± 3.38 mM, p = 0.18) compared to contralateral (2.94 ± 1.35 mM).
The proposed MRSI protocol (~12 min) enables separate mapping of glutamate and glutamine reliably along with other MRS-detectable standard metabolites in glioma subregions at 3T.
1 TECHNICAL EFFICACY: Stage 3.
谷氨酸和谷氨酰胺是神经胶质瘤中的关键代谢物，在肿瘤生物学中各自发挥着不同作用。在临床场强（≤3T）下使用体内磁共振波谱（MRS）对这些代谢物进行单独定量，会因它们的分子相似性而受阻，导致光谱峰重叠，因而无法区分。
开发一种MRS成像（MRSI）方案，在3T下于临床可行的时间内分别绘制谷氨酸和谷氨酰胺图谱，利用J调制增强光谱分辨能力，证明其可靠性/可重复性，并对神经胶质瘤亚区域中的代谢物进行定量分析。
前瞻性研究。
体模、5名健康受试者和30名疑似神经胶质瘤患者。对异柠檬酸脱氢酶（IDH）野生型胶质母细胞瘤病例进行评估，以建立一个统一的队列。
场强/序列：3T，研究方案：采用水作为参考的二维氢质子化学位移成像（2D H sLASER MRSI，回波时间分别为40和120毫秒）、三维T1加权成像和二维T2加权成像。试验筛查过程：T1加权成像、T1加权对比增强成像、T2加权成像、液体衰减反转恢复序列（FLAIR）成像。
进行光谱模拟和体模测量以设计和验证该方案。使用LCModel软件获得扫描-重复扫描测量的光谱质量/拟合参数。将所提出的长回波时间数据与短回波时间数据进行比较。采用BraTS Toolkit进行全自动肿瘤分割。
使用Bland-Altman分析进行扫描-重复扫描比较。绘制谷氨酸和谷氨酰胺之间的LCModel建模协方差系数（CMC），以评估每个光谱拟合中它们的模型相互作用。使用单因素方差分析和Kruskal-Wallis检验比较肿瘤亚区域中的代谢物水平。p值<0.05被认为具有统计学意义。
扫描-重复扫描测量之间的光谱质量/拟合参数和代谢物水平高度一致。短回波时间时谷氨酸和谷氨酰胺模型之间的负相关（CMC = -0.16±0.06）在长回波时间时消除（0.01±0.05）。与对侧（10.84±2.94毫摩尔）相比，肿瘤亚区域中谷氨酸水平较低（非增强肿瘤核心区：5.35±4.45毫摩尔，周围非增强FLAIR高信号区：7.39±2.62毫摩尔，增强肿瘤区：7.60±4.16毫摩尔），而谷氨酰胺在周围非增强FLAIR高信号区（9.17±6.84毫摩尔）和增强肿瘤区（7.20±4.42毫摩尔）较高，但在非增强肿瘤核心区（4.92±3.38毫摩尔，p = 0.18）与对侧（2.94±1.35毫摩尔）相比无差异。
所提出的MRSI方案（约12分钟）能够在3T下可靠地分别绘制谷氨酸和谷氨酰胺图谱，同时还能绘制神经胶质瘤亚区域中其他MRS可检测的标准代谢物图谱。
1级 技术效能：3级
