使用点弥散函数模型法校正扩散频谱影像內磁化率差异引致影像扭曲的效力评估(英文)

Evaluation of Diffusion Spectrum Imaging Tractography after Susceptibility-Induced Distortion Correction by Model-Based Point Spread Function Method

平面回波影像(EPI)常常在额叶、颞叶或脑干处出现严重的影像扭曲.目前已有数种方法可校正这种由磁化率差异造成的影像扭曲,该文提出一种简单有效的方法叫点弥散函数模型法.由纤维追踪的角度来探讨影像扭曲校正的有效性.评估的方法是将点弥散函数模型法应用于扩散频谱影像,并比较由扭曲影像与校正影像所得之扣带回纤维数目的多寡.定义了一个指标用来量化影像扭曲校正的有效性:r=( Nc-Nd)/( Nc+Nd),此值越大代表校正影像的纤维追踪结果越好.对于左扣带回与右扣带回,其r值分别为0 .424±0 .452(平均±标准差)及0 .343±0 .452 ,且在统计上皆显著大于零.结果显示,就左右扣带回而言,校正后的影像其纤维追踪的结果较扭曲影像更佳.因此,将点弥散函数模型法应用于校正扩散频谱影像的扭曲可增进左右扣带回的纤维追踪的结果.

Echo-planar imaging suffers from susceptibility-induced distortion in the frontal lobe,temporal lobe and brain stem regions.To correct for the distortion,recently we proposed an easy-implemented and effective method called model-based point spread function （PSF） method.This study aimed to investigate the effectiveness of distortion correction from the perspective of fiber tracking.The evaluation was carried out by applying the model-based PSF method on diffusion spectrum imaging （DSI） dataset and comparing the number of cingulum tracts generated from the distorted DSI dataset and the corrected DSI dataset.An index representing the effectiveness of distortion correction was defined as：r=（Nc-Nd）/（Nc＋Nd）,where Nc and Nd were,respectively,the number of tracts generated from DSI dataset after distortion correction and that without correction.For the left and right cingulum bundles,the ratios of the effectiveness of distortion correction were 0.424±0.452 （mean±SD） and 0.343±0.452,respectively.In addition,the ratios of both cingulum bundles were significantly greater than zero,indicating that the tracking results after distortion correction were better than those without correction.In conclusion,our results suggest that application of distortion correction on DSI dataset using the model-based PSF method can significantly improve the fiber tracking results.