The purpose of this study was to analyze the lesion brightness (image contrast) in multiple MRI sequences in patients with relapsing-remitting MS (RRMS), secondary progressive MS (SPMS), primary progressive MS (PPMS),...The purpose of this study was to analyze the lesion brightness (image contrast) in multiple MRI sequences in patients with relapsing-remitting MS (RRMS), secondary progressive MS (SPMS), primary progressive MS (PPMS), and clinically isolated syndrome (CIS);and to correlate the lesion contrast with lesion volumes and neurological disability. MRI ex- amination at 1.5 T was performed on 80 patients with RRMS, SPMS, PPMS, or CIS. The protocol included T1- and T2-weighted spin echo (SE), fluid attenuated inversion recovery (FLAIR), T1-weighted SE with magnetization transfer preparation, and diffusion weighted imaging (DWI). Contrast was measured between MS lesions and normal appearing white matter. Lesion volume was calculated in T1-weighted- and FLAIR-images. All patients were examined neurologically including evaluation of expanded disability status scale (EDSS) score. Lesion contrast correlated with total brain lesion volume (p = 0.000 - 0.040). In patients with low EDSS, three sequences were able to differentiate between CIS and RRMS. SPMS and PPMS were separated by DWI. Lesion contrast correlated with EDSS score on T1-weighted imaging, with or without magnetization transfer preparation. Patient age correlated with lesion contrasts. Contrast measurements seem limited in radiological and clinical diagnosis of MS in reference to disease course, its activity and progression. The differentiation between MS subgroups might improve at 3 T and could help in leading to earlier treatment of the disease.展开更多
为了提高融合多序列MR图像应用于脑肿瘤提取时分割区域的准确性,基于核稀疏表示分类方法,联合多序列MR图像中的空间结构和灰度特征信息,提出一种空间特征联合的脑肿瘤核稀疏表示分类方法.首先构建各个类别的子字典,再用邻域滤波核稀疏...为了提高融合多序列MR图像应用于脑肿瘤提取时分割区域的准确性,基于核稀疏表示分类方法,联合多序列MR图像中的空间结构和灰度特征信息,提出一种空间特征联合的脑肿瘤核稀疏表示分类方法.首先构建各个类别的子字典,再用邻域滤波核稀疏表示方法对多序列脑MR图像进行分类,该邻域滤波核可以有效地将灰度特征与空间结构结合起来提高脑肿瘤提取的准确性.对国际数据库MICCAI Bra TS提供的临床和仿真数据进行分割.结果表明:与稀疏表示分类方法相比,所提出的基于空间特征联合核稀疏表示的脑肿瘤提取方法由于增加了空间结构信息,所得的提取准确率提高了5%~6%.展开更多
文摘The purpose of this study was to analyze the lesion brightness (image contrast) in multiple MRI sequences in patients with relapsing-remitting MS (RRMS), secondary progressive MS (SPMS), primary progressive MS (PPMS), and clinically isolated syndrome (CIS);and to correlate the lesion contrast with lesion volumes and neurological disability. MRI ex- amination at 1.5 T was performed on 80 patients with RRMS, SPMS, PPMS, or CIS. The protocol included T1- and T2-weighted spin echo (SE), fluid attenuated inversion recovery (FLAIR), T1-weighted SE with magnetization transfer preparation, and diffusion weighted imaging (DWI). Contrast was measured between MS lesions and normal appearing white matter. Lesion volume was calculated in T1-weighted- and FLAIR-images. All patients were examined neurologically including evaluation of expanded disability status scale (EDSS) score. Lesion contrast correlated with total brain lesion volume (p = 0.000 - 0.040). In patients with low EDSS, three sequences were able to differentiate between CIS and RRMS. SPMS and PPMS were separated by DWI. Lesion contrast correlated with EDSS score on T1-weighted imaging, with or without magnetization transfer preparation. Patient age correlated with lesion contrasts. Contrast measurements seem limited in radiological and clinical diagnosis of MS in reference to disease course, its activity and progression. The differentiation between MS subgroups might improve at 3 T and could help in leading to earlier treatment of the disease.
文摘为了提高融合多序列MR图像应用于脑肿瘤提取时分割区域的准确性,基于核稀疏表示分类方法,联合多序列MR图像中的空间结构和灰度特征信息,提出一种空间特征联合的脑肿瘤核稀疏表示分类方法.首先构建各个类别的子字典,再用邻域滤波核稀疏表示方法对多序列脑MR图像进行分类,该邻域滤波核可以有效地将灰度特征与空间结构结合起来提高脑肿瘤提取的准确性.对国际数据库MICCAI Bra TS提供的临床和仿真数据进行分割.结果表明:与稀疏表示分类方法相比,所提出的基于空间特征联合核稀疏表示的脑肿瘤提取方法由于增加了空间结构信息,所得的提取准确率提高了5%~6%.
