MR 脑图像海马自动分割法在AD 早期诊断中的应用研究

目的:研究磁共振(Magnetic resonance,MR)脑图像中海马的自动分割方法及海马的形态学分析方法,为阿尔茨海默病(Alzheimer’s disease,AD)的早期诊断提供依据。方法:对20例AD患者和60名正常对照者行MRI T1 WI 3D容积扫描,建立海马的三维主动表观模型,并以此模型对每个个体脑部磁共振图像上的海马进行自动识别和三维分割,分别建立正常对照组和AD组的海马统计形状模型,比较AD组与正常对照组间海马形状的差异性。结果:海马三维分割方法与手动分割方法在海马体积测量上无统计学差别(P>0.05);AD患者海马头部发生萎缩(P<0.05)。结论:基于主动表观模型的MR脑图像海马自动识别和三维分割法是准确可靠的;海马头部萎缩可作为AD诊断的依据之一。

Hippocampal Automatic Recognition and 3D Segmentation Based on ActiveAppearance Model in Brain MR Images for Early Diagnosis ofAlzheimer's Disease

Objective: To investigate the three-dimensional segmentation method and the differences of regional pattern between AD and normal aging based on the MRI hippocampal shape analysis to provide effective evidence to assist the early diagnosis of AD.Methods: 20 AD patients and 60 health persons were included in this study.3D structure images were obtained on a 3.0 T high-resolution MR imaging system.Data were processed to create three-dimensional active appearance model of hippocampus.Three-dimensional segmentation and automatic identification were carried out in the hippocampus for each individual brain MR images with this model,and the hippocampal statistical shape model respectively for control group and AD group was established,which could compare the difference of hippocampal shape between AD group and control group.Results: There was no significant difference between conventional hand-drawing ROIs and 3D segmentation and auto-detected method in the measurement of hippocampal volume(P>0.05).Hippocampal head atrophy was found in AD patients(P<0.05).Conclusions: Hippocampal three-dimensional segmentation and automatic identification method based on active appearance model in brain MR image is accurate and reliable;the feature of hippocampal head atrophy can be used as a basis for diagnosis of AD.