欢迎来到《四川大学学报(医学版)》
谢薇, 孙怀强, 陈嘉伟, 等. 基于几何深度学习的脑形态学研究在阿尔茨海默病诊断中的初步应用[J]. 四川大学学报(医学版), 2021, 52(2): 300-305. DOI: 10.12182/20210360103
引用本文: 谢薇, 孙怀强, 陈嘉伟, 等. 基于几何深度学习的脑形态学研究在阿尔茨海默病诊断中的初步应用[J]. 四川大学学报(医学版), 2021, 52(2): 300-305. DOI: 10.12182/20210360103
XIE Wei, SUN Huai-qiang, CHEN Jia-wei, et al. A Preliminary Study of Applying Geometric Deep Learning in Brain Morphometry for Diagnosis of Alzheimer’s Disease[J]. Journal of Sichuan University (Medical Sciences), 2021, 52(2): 300-305. DOI: 10.12182/20210360103
Citation: XIE Wei, SUN Huai-qiang, CHEN Jia-wei, et al. A Preliminary Study of Applying Geometric Deep Learning in Brain Morphometry for Diagnosis of Alzheimer’s Disease[J]. Journal of Sichuan University (Medical Sciences), 2021, 52(2): 300-305. DOI: 10.12182/20210360103

基于几何深度学习的脑形态学研究在阿尔茨海默病诊断中的初步应用

A Preliminary Study of Applying Geometric Deep Learning in Brain Morphometry for Diagnosis of Alzheimer’s Disease

  • 摘要:
      目的  基于脑表面图形和几何深度学习建立阿尔茨海默病(Alzheimer’s disease,AD)的分类预测模型,并评估其性能。
      方法  纳入临床确诊AD患者76例,健康老年人83例,并按4∶1的比例随机划分为训练集和测试集。从受试者的MR成像中三维T1加权高分辨率结构像中构建脑表面图形,进行一系列图形简化操作后将训练集输入几何深度神经网络进行训练,用测试集对训练产生的预测模型进行性能评估,评估参数包括准确率、敏感性和特异性。
      结果  在右脑面数为6 000的脑表面图形上训练得到的预测模型取得最佳性能(准确性93.8%,敏感性91.7%,特异性94.1%)。脑表面图形在卷积与池化操作过程中的变化揭示AD患者相较健康老年人存在全脑弥漫分布的脑组织损失。
      结论  基于图形数据和几何深度学习的脑形态学分析方法在AD的诊断和鉴别诊断中有较大的发展潜力。

     

    Abstract:
      Objective  A predictive model of Alzheimer’s disease (AD) was established based on brain surface meshes and geometric deep learning, and its performance was evaluated.
      Methods  Seventy-six clinically diagnosed AD patients and 83 healthy older adults were enrolled and randomly assigned to the training set and the test set according to a 4-to-1 ratio. Brain surface mesh was constructed from 3-D T1-weighted high-resolution structural MR volumes of each participant. After applying a series of simplification to the surface meshes, the training set was fed into the geometric deep neural network for training. The performance of the prediction model was evaluated with the test set, and the evaluation metrics included accuracy, sensitivity and specificity.
      Results  The prediction model trained on the right brain surface meshes with 6000 faces achieved the best performance, with accuracy reaching 93.8%, sensitivity, 91.7%, and specificity, 94.1%. The evolution of the brain surface meshes during convolution and pooling revealed that AD patients had diffuse brain tissue loss compared with healthy older adults.
      Conclusion  Morphological brain analysis based on mesh data and geometric deep learning has great potential in the differential diagnosis of AD.

     

/

返回文章
返回