自动分割技术联合神经突方向离散度及密度成像观察海马硬化型内侧颞叶癫痫患者海马亚区体积与微结构改变

目的 观察自动分割技术联合神经突方向离散度和密度成像(NODDI)显示海马硬化型内侧颞叶癫痫(mTLE-HS)患者海马亚区体积与微结构改变的价值。方法 回顾性分析33例左侧mTLE-HS患者(mTLE-HS组)及35名健康成人(对照组)的头颅MRI,以FreeSurfer软件自动分割海马亚区,测量阿蒙角(CA)1、CA2~3、CA4、颗粒细胞-齿状回(GC-DG)及下托体积,经后处理获取各亚区NODDI参数;比较组内及组间海马各亚区体积及NODDI参数,分析其间差异有统计学意义者与发病年龄及病程的相关性。结果 mTLE-HS组左侧海马各亚区体积均低于对照组(P均<0.05)。mTLE-HS组左侧CA1及CA4亚区神经突密度指数(NDI)低于、而CA1亚区各向同性体积分数(fiso)均高于右侧(P均<0.05)。mTLE-HS组左侧CA1、CA2~3及CA4亚区ODI,CA1、CA4及GC-DG亚区NDI均低于对照组(P均<0.05),而左侧CA1、GC-DG及下托亚区fiso均高于对照组(P均<0.05)。mTLE-HS左侧海马各亚区体积均与发病年龄呈中度正相关(r=0.540~0.667,P均<0.001)而与病程呈低度负相关(r=－0.492~－0.386,P均<0.05);左侧海马CA4及GC-DG亚区NDI均与病程呈低度负相关(r=－0.418、－0.388,P均<0.05)。结论 自动分割技术联合NODDI可显示mTLE-HS患者海马体积及微结构改变;NDI可能是评估mTLE-HS的较敏感的神经元进行性损伤生物学标志物。

Automatic segmentation technique combined with neurite orientation dispersion and density imaging for observing volume and microstructure changes of hippocampal subregion in patients with hippocampal sclerosis medial temporal lobe epilepsy

Objective To explore the value of automatic segmentation technique combined with neurite dispersion and density imaging (NODDI) for displaying volume and microstructure changes of hippocampal subregion in patients with hippocampal sclerosis medial temporal lobe epilepsy (mTLE-HS). Methods MRI data of 33 patients with left mTLE-HS (mTLE-HS group) and 35 healthy adults (control group) were retrospectively analyzed. The hippocampal subregions were automatically segmented using FreeSurfer software, the volume of cornu Ammonis (CA) 1, CA2—3, CA4, granulose cell-dentate gyrus (GC-DG) and subiculum were measured, then the NODDI parameters of each subregion were obtained through post-processing. The intra- and inter-groups hippocampal subregion volumes and NODDI parameters were compared, and the correlations of parameters being significantly different with the onset age and disease courses were analyzed. Results The volume of hippocampal subregions in mTLE-HS group were all lower than those in control group (all P<0.05). In mTLE-HS group, the neurite density index (NDI) of left CA1 and CA4 subregions were both lower, while the free-water isotropic volume fraction (fiso) of the left CA1 subregion was higher than those of the right side (all P<0.05). The orientation dispersion index (ODI) of left CA1, CA2—3 and CA4 subregions, as well as NDI of left CA1, CA4 and GC-DG subregions in mTLE-HS group were all lower than those in control group (all P<0.05), while fiso of left CA1, GC-DG and subiculum subregions in mTLE-HS group were all higher than those in control group (all P<0.05). The volume of left hippocampal subregions in patients with mTLE-HS were all moderately positively correlated with the onset age (r=0.540—0.667, all P<0.001) but weakly negatively correlated with disease courses (r=－0.492—－0.386, all P<0.05). NDI of left CA4 and GC-DG subregions in patients with mTLE-HS were both weakly negatively correlated with disease courses (r=－0.418, －0.388, both P<0.05). Conclusion Automatic segmentation technique combined with NODDI could be used to display the volume and microstructure changes of mTLE-HS. NDI might be a biomarker of mTLE-HS being sensitive to progressive neuronal damage.