| 齐志刚,钱天翼,安彦红,张默,陈楠,李坤成.联机测量灰质皮层体积在阿尔茨海默病诊断中的应用[J].中国医学影像技术,2016,32(8):1165~1168 |
| 联机测量灰质皮层体积在阿尔茨海默病诊断中的应用 |
| Application of inline measurement of cerebral grey matter in diagnosis of Alzheimer disease |
| 投稿时间：2016-04-01 修订日期：2016-06-07 |
| DOI：10.13929/j.1003-3289.2016.08.004 |
| 中文关键词: 轻度认知障碍 阿尔茨海默病 磁共振成像 测量 |
| 英文关键词:Mild cognitive impairment Alzheimer disease Magnetic resonance imaging Measurement |
| 基金项目:国家自然科学基金（81471649、81271556）、国家“十二五”科技支撑计划课题项目（2012BAI10B04）、北京市医院管理局重点医学专业发展计划（ZYLX201609）。 |
| 作者 | 单位 | E-mail | | 齐志刚 | 首都医科大学宣武医院放射科, 北京 100053
磁共振成像脑信息学北京市重点实验室, 北京 100053 | | | 钱天翼 | 西门子医疗MRI东北亚合作中心, 北京 100102 | | | 安彦红 | 首都医科大学宣武医院放射科, 北京 100053
磁共振成像脑信息学北京市重点实验室, 北京 100053 | | | 张默 | 首都医科大学宣武医院放射科, 北京 100053
磁共振成像脑信息学北京市重点实验室, 北京 100053 | | | 陈楠 | 首都医科大学宣武医院放射科, 北京 100053
磁共振成像脑信息学北京市重点实验室, 北京 100053 | | | 李坤成 | 首都医科大学宣武医院放射科, 北京 100053
磁共振成像脑信息学北京市重点实验室, 北京 100053 | cjr.likuncheng@vip.163.com |
|
| 摘要点击次数: 1758 |
| 全文下载次数: 818 |
| 中文摘要: |
| 目的 探讨采用联机软件测量脑灰质体积对阿尔茨海默病（AD）患者的诊断价值。方法 对36例轻度认知障碍（MCI）患者（MCI组）、29例AD患者（AD组）进行MR扫描，正常对照（NC）组为28名认知正常的老年人。获得三维脑结构数据，通过联机软件计算各脑区体积的相对定量值。采用单因素方差分析和非参数秩和检验获得组间差异脑区，并利用ROC曲线和支持向量机（SVM）分析各脑区在组间鉴别诊断中的效能。结果 3组双侧海马体积、左侧和右侧海马体积及双侧扣带回、岛叶、额叶、顶叶、颞叶体积差异均有统计学意义（P均<0.001），且AD组与MCI组、AD组与NC组间差异均有统计学意义（P均<0.001）。ROC分析显示，鉴别诊断AD组与NC组、MCI组与NC组、AD组与MCI组、（AD组+MCI组）与（NC组）时，最大AUC值分别位于左侧颞叶（0.95）、左侧岛叶（0.69）、右侧扣带回（0.85）、左侧颞叶（0.80）。SVM分析结果提示，AD组与NC组、MCI组与NC组、AD组与MCI组、（AD组+MCI组）与（NC组）中分类准确率最高的区域分别位于双侧海马（89.09%）、左侧岛叶（64.52%）、右侧海马（77.78%）和左侧海马（71.11%）；综合海马、颞叶与岛叶体积鉴别AD/NC的准确率高达94.55%。结论 联机测量反映的脑区改变符合AD病理改变及其发展过程，可以用于临床诊断。 |
| 英文摘要: |
| Objective To investigate the value of inline morphometry analysis package in Alzheimer disease (AD) diagnosis. Methods Three groups were enrolled, including 28 normal control (NC), 36 cases in mild cognitive impairment (MCI) group, 29 cases in AD group and 28 healthy olders in normal control (NC) group. Three dimensional MRI structural data was obtained. The relative quantity volume of brain regions were calculated through the online software. The MRI data were transfered to the inline software. And volume of each brain area were normalized to an index number of percentage using the whole brain volume. Brain areas with statistical volumetric differences were obtained with one-way ANOVA and non-parametric rank sum test. The diagnosis efficacy of different brain regions was evaluated with ROC curve and support vector machine (SVM) among three groups. Results The volume of bilateral hippocampus, left and right cingulated cortex, insula, bilateral frontal, bilateral temporal, and bilateral parietal lobe showed significant differences among NC, MCI and AD groups (P<0.001). And the difference were also found between AD group and MCI group, AD group and NC group. ROC analysis showed that the highest the highest area under the curve (AUC) value between AD group and NC group, MCI group and NC group, AD group and MCI group, AD+MCI groups and NC group were observed in the left temporal lobe (0.95), left insula (0.69), right cingulated lobe (0.85) and left temporal lobe (0.80). SVM analysis showed that the highest classification accuracy rates between AD group and NC group, MCI group and NC group, AD group and MCI group, AD+MCI groups and NC group were observed in bilateral hippocampus (89.09%), left insula (64.52%), right hippocampus (77.78%) and left hippocampus (71.11%), respectively. Accuracy rate as high as 94.55% was obtained with combination analysis between AD group and NC group. Conclusion The individual-level inline morphometry analysis is consistent with the Alzheimer pathology, and the software can be applied in the clinics. |
| 查看全文 查看/发表评论 下载PDF阅读器 |
|
|
|
