基于HLSVD和AMARES磁共振波谱数据量化方法的比较

HLSVD versus AMARES for magnetic resonance spectroscopy data quantification

背景:磁共振波谱在临床医学诊断上正成为越来越重要的辅助诊断工具,目前应用最广泛的就是人脑的氢质子磁共振波谱,其在脑肿瘤的诊断上有重要的临床价值。目的:基于一定量数据的基础上,通过比较HLSVD和AMARES两种方法在量化结果的准确性和量化过程的时间性上的差异,找到更适用于临床辅助诊断的方法。方法:在分析软件JMRUI中分别运用两种量化方法,得到频率,幅值,相位等物理参数,这些参数即量化结果的表现形式,以此完成对数据信号的解读。结果及结论:HLSVD方法是根据波形的特征情况进行量化;而AMARES方法,由于使用了人工定义的先验知识,在图形的反应上就会出现偏差,导致量化的结果发生偏离。结果提示,HLSVD方法比AMARES方法在准确性和时间性上都更适合于应用到临床诊断上。

BACKGROUND：Magnetic resonance spectroscopy （MRS） has become important auxiliary diagnosis tool in clinical diagnosis.At present,1H MRS in human brain has been frequently used,which has important clinical value in diagnosis of brain tumor.OBJECTIVE：To quantify the human brain 1HMRS data based on HLSVD and AMARES methods separately,by comparing the two methods through a certain amount of data to find the method which is more suitable for the clinical auxiliary diagnosis.METHODS：We used the two methods separately in the JMRUI analysis software,and then the physical parameters such as frequency,amplitude and phrase,were obtained,which were results of the quantification in terms of these parameters,so we could deal with the interpretation of the signal.RESULTS AND CONCLUSION：HLSVD quantifies data according waveform features,while AMARES produced deviation in images.Therefore,HLSVD is more suitable than AMARES for the clinical diagnosis,both in accuracy and timeliness.