活体磁共振波谱中定位方法的研究

发布时间：2018-11-10 15:43

【摘要】：磁共振波谱作为对个体进行纵向研究的无创性方法,被广泛应用于临床,作为评估疾病进展和反馈治疗效果的手段。在这种纵向研究中,需要对同一个被试在不同的时间进行多次重复扫描。为了保证研究结果的准确性,提高扫描的可重复性和降低扫描误差成为必须。影响磁共振波谱数据可重复性的因素有很多,例如扫描过程中随机产生的系统误差、多次扫描被试的生理变化以及重定位的影响等等。其中一个主要干扰因素由不准确的感兴趣体素重定位产生。传统解决重定位问题的方法主要有两种。一种是手动重定位(主要靠眼睛定位)方法,在多次重复扫描前,调整被试平躺的位置,保证其与基线扫描(第一次扫描)时所处位置完全一致,得到相同的结构图像并根据解剖结构进行定位。这种方法主要适用于连续扫描或者扫描间隔很短的研究,对于扫描间隔长的研究则结果较差。另一种方法是通过配准调整高分辨结构像的位置,使其与基线扫描结构像一致,再根据解剖结构进行定位。这两种方法不仅耗费时间,而且操作员作为主观因素对结果的影响很大。本研究采用高效率的配准算法,用于对感兴趣体素进行重定位以获取磁共振波谱数据。这个自动重定位(追踪定位)过程以磁共振结构图像的配准为基础,通过直接对感兴趣体素进行位置变动(平动和转动)操作,保证后续扫描时感兴趣体素覆盖脑区与基线扫描时基本一致。为验证本方法的效果,本研究对多次扫描感兴趣体素内各组织体素的重合率、化合物含量以及含量比进行评估计算。同时,通过分析重定位后的扫描结果与基线扫描结果的差异性(用差异性系数表示)和可重复性(用可重复性系数表示),我们对追踪定位方法与手动定位方法进行比较。
[Abstract]:As a noninvasive method for longitudinal study of individuals, magnetic resonance spectroscopy (MRS) is widely used in clinical practice as a means of evaluating disease progression and feedback treatment. In this longitudinal study, multiple repeated scans were performed on the same subject at different times. In order to ensure the accuracy of the research results, it is necessary to improve the repeatability of scanning and reduce the scanning error. There are many factors that influence the repeatability of magnetic resonance spectrum data, such as the random systematic error in the scanning process, the physiological changes of the multiple scanning subjects, and the influence of the repositioning, and so on. One of the main interference factors is caused by inaccurate voxel relocalization. There are two main methods to solve the problem of repositioning. One is manual repositioning (mainly eye positioning), which adjusts the position of the subjects lying flat before repeated scans for many times to ensure that they are exactly the same as those in the baseline scan (the first scan). The same structure image was obtained and located according to the anatomical structure. This method is mainly suitable for continuous scanning or short scan interval, but the results for long scan interval are poor. Another method is to adjust the position of the high resolution structure image by registration to make it consistent with the baseline scan structure image, and then to locate the image according to the anatomical structure. These two methods not only cost time, but also have a great influence on the result as subjective factors. In this study, an efficient registration algorithm is used to relocate voxels to obtain magnetic resonance spectrum data. This automatic resetting (tracking location) process is based on the registration of magnetic resonance structural images and is performed directly on the position changes (translation and rotation) of the voxels of interest. Make sure that the area covered by voxel is consistent with baseline scan. In order to verify the effectiveness of this method, the coincidence rate, compound content and content ratio of various tissue voxels in voxels were evaluated and calculated. At the same time, by analyzing the difference between the scanning result and the baseline scan result (expressed by the coefficient of difference) and the repeatability (expressed by the repeatability coefficient), we compare the tracking localization method with the manual location method.
【学位授予单位】：华东师范大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：R310;O482.531

【相似文献】