Abstract：Objective: To evaluate the atrophy pattern of subcortical gray matter structures in patients with mild Alzheimer’s disease(AD) by using an automatic segmentation tool. These structures include hippocampi, amygdalae, accumbens, caudates, putamen, pallidum, thalami. Methods: Twenty-nine cases of mild AD and 30 normal aging volunteers as control subjects were scanned with 3D FSPGRIR sequence using 1.5T MR system． The subcortical gray matter structures were segmented using FIRST tool in FSL. The mean values of each structure’s relative volume(gray matter volume/intracranial volume×100) were compared with 2-sample t test. The correlation between the relative volume of each structure and MMSE scores were evaluated using partial correlation analysis. Binary logistic regression analysis and ROC curve were applied to assess the diagnostic significance of the relative volume of each structure. Results: Compared with normal control group, the relative volume of left hippocampus, right amygdala, bilateral thalami, right caudate, left putamen and bilateral pallidum in mild AD group was significantly smaller than that in the control group(P<0.05). The relative volume of left hippocampus, right amygdala, left thalamus and left putamen was correlated with MMSE scores. It was the relative volume of left hippocampus that entered the final model of binary logistic regression analysis. The accuracy was 86.3%. The area under the curve of ROC curve was 0.897. In the cut-off point, i.e. maximal value of Yuden index, the sensitivity was 83.3% and the specificity was 96.3%. Conclusion: In addition to hippocampus, volumetric atrophy was found in other subcortical gray matter structures in mild AD. The atrophy of right amygdala, left putamen and left thalamus may take part in the pathological mechanism of cognitive impairment in AD patients. Among these structures, hippocampus is of the most diagnostic significance.
2012, Vol. 23 
