中文摘要：

目的探讨在体肝脏3.0T质子磁共振波谱（MRS）谱线用SAGE软件Spectrum后处理方法处理评估者间和评估者自身可重复性。资料与方法19例自愿者,使用GESigna Excite HD3.0T超导磁共振扫描仪,8通道相控阵线圈,单体素点分辨波谱分析法（PRESS）序列。采集参数为TR1500ms,TE30ms,信号采集次数（NSA）64次,感兴趣容积（VOI）2cm×2cm×2cm,共获得19条谱线。采用SAGE软件包,Spectrum后处理方法,测量者1测量2次和测量者2测量1次,经过5Hz高斯窗函数截趾（Apodization）处理,傅立叶变换,谱线自动基线校正,谱线自动相位校正,手动确定4.7ppm水和1.3ppm脂肪峰亚甲基峰位置,拟合积分共得到57组水峰和脂肪峰峰高及峰下面积。（1）计算评估者间和评估者自身等级相关系数。（2）采用多组配对样本秩和检验（FriedmanTest）比较三组的水峰和脂肪峰峰高及峰下面积差异是否存在统计学意义。结果（1）测量者1前后2次测量结果脂质峰峰高及峰下面积相关系数分别为0.981（P〈0.000）、0.989（P〈0.000）,水峰峰高及峰下面积相关系数分别为1.000、0.989（P〈0.000）,测量者1第1次与测量者2测量结果脂质峰峰高及峰下面积相关系数分别为0.993（P〈0.000）、0.988（P〈0.000）,水峰峰高及峰下面积相关系数分别为0.921（P〈0.000）、1.000,测量者1第2次与测量者2测量结果脂质峰峰高及峰下面积相关系数分别为0.977（P〈0.000）、0.988（P〈0.000）,水峰峰高及峰下面积相关系数分别为0.921（P〈0.000）、0.989（P〈0.000）。（2）三组水峰峰高（χ2=3.500,P=0.174）及峰下面积（χ2=1.500,P=0.472）和脂肪峰峰高（χ2=2.333,P=0.311）及峰下面积（χ2=1.600,P=0.449）差异均无统计学意义。结论采用SAGE后处理软件Spectrum后处理方法,在体肝脏3.0T质子MRS谱线测量者内及测量者间可重复性高,测量结果高度相关。

英文摘要：

Objective To assess the Intrarater and interrater reliability of hepatic MRS at 3.0 T using spectrum methods with SAGE software. Materials and Methods The prospective study of hepatic proton MRS was performed using GE Signa Excite HD 3.0 T system and eight-channel torso phased-array coils. 19 healthy individuals were enrolled in this study. Liver spectra were collected using PRESS sequence with a TR of 1500 ms,TE of 30 ms,VOI of 2 cm×2 cm×2 cm,NSA of 64 times. 19 scans were analyzed twice by one reader and once by another reader to calculated the intra-and interrater reliability.The results were statistically analyzed by Friedman test and Spearman correlation test. Results The interater and intrarater reliability were excellent for fat and water. There were no significant differences between three groups. Conclusion The Intrarater and interrater reliability were good for liver proton MRS at 3.0 T in vivo.