关键词: 多发性硬化, 阿尔茨海默病, MR图像, 纹理分析, 感兴趣区

Abstract: Objective Multiple sclerosis(MS) is a chronic inflammatory disease of the central nervous system, which leads to severe progressive neurological dysfunction. Alzheimer’s disease(AD) is a severe neurodegenerative disease of the central nervous system. The previous researches shown that there were significant differences between MRI texture characteristics of MS and normal control(NC). To compare texture characteristics of MS and other cerebral parenchyma diseases, the differences of cerebral parenchyma texture characteristics on MR images between patients with Multiple Sclerosis(MS) and Alzheimer’s disease(AD) were analyzed. Methods In this research, the total sample set was 39 with MR images of group AD, group MS and group NC 13 for each. Texture analysis of gray level co-occurrence matrix was performed on MR T1-weighted images, and the region of interest(ROI) was cerebral parenchyma. Texture characteristics, such as energy, contrast, entropy and inverse different moment were calculated in order to test if there was a significant difference between these three groups. Results Monotonic trend of texture values existed in AD, MS and NC group the mean values were 0.002 2 vs 0.003 8 vs 0.008 3 for energy, 2.863 3 vs 2.678 9 vs 1.317 8 for entropy, 47.524 1 vs 44.396 1 vs 36.809 0 for contrast, 0.221 7 vs 0.249 8 vs 0.338 3 for inverse different moment. There were significant differences for energy, entropy and inverse different moment between every two groups of AD, MS and NC. Concerning contrast, there was a significant difference between groups AD and NC, groups MS and NC, but not difference between AD and MS. Conclusion The texture differences between MR image of MS and NC could reveal the pathological changes of brain substance for patients with MS, which may be helpful to aid diagnose of MS. At the same time, the texture difference may reveal pathological differences between AD and MS, which may be helpful to differentially diagnose AD, MS and other pathological changes. The data of our subjects is limited, so a larger sample set maybe require to confirm the above conclusion in further studies.

Key words: multiple sclerosis, Alzheimer’s disease, MR imaging, texture analysis, region of interest