Alzheimer's disease classification method based on multi-modal medical images
HAN Kun1, PAN Haiwei1, ZHANG Wei2, BIAN Xiaofei1, CHEN Chunling1, HE Shuning1
1. College of Computer Science and Technology, Harbin Engineering University, Harbin 150001, China; 2. Modern Education Technology Center, Heilongjiang University, Harbin 150001, China
摘要多模态医学影像信息已经在计算机辅助检测和诊断中被广泛地应用。在对Alzheimer病（Alzheimer's disease，AD）的分类与诊断中，结合多个模态医学影像的特征信息能够更加准确且全面地对同一AD主题进行分类与诊断。该文提出了一种基于卷积神经网络的模型结构，分别对AD病患的磁共振图像（MRI）和正电子发射型计算机断层显像（PET）图像进行3D卷积操作来提取各自模态的特征信息，并应用模型融合方法对模态特征信息加以融合，从而得到包含更加丰富的多模态特征信息。最后用全连接神经网络将上述提取的多模态特征信息进行分类预测。通过在AD神经影像学倡议（Alzheimer's disease neuroimaging initiative，ADNI）公开数据集上的实验结果表明：该文所提出的模型在准确率（accuracy，ACC）和曲线下面积（area under the curve，AUC）的性能评价中都取得了更加优越的结果。
Abstract：Multi-modal medical image information has been widely used in computer-aided detection and computer-aided diagnosis in the medical community. Feature information from multi-modal medical images can be used to accurately classify and diagnosis Alzheimer's disease (AD) characteristics. This paper presents a convolutional neural network model for 3D convolution operations on magnetic resonance imaging (MRI) and positron emission computed tomography (PET) images of Alzheimer's subjects to extract the feature information for the various modalities. Then, models are used to fuse these modal information sets into a rich multi-modal feature information dataset. Finally, this dataset is classified and predicted using a fully connected neural network. Tests on the public data set of the AD neuroimaging initiative show that this model more accurately evaluates accuracy (ACC) and area under the curve (AUC) conditions.
[1] ZHAN L, ZHOU J Y, WANG Y L, et al. Comparison of nine tractography algorithms for detecting abnormal structural brain networks in Alzheimer's disease[J]. Frontiers in Aging Neuroscience, 2015, 7:48. [2] CHRISTINA P. World Alzheimer report 2018[R]. London:Alzheimer's Disease International, 2018. [3] PETERSEN R C. Mild cognitive impairment[J]. The Lancet, 2006, 367(9527):P1979. [4] LITJENS G, SÁNCHEZ C I, TIMOFEEVA N, et al. Deep learning as a tool for increased accuracy and efficiency of histopathological diagnosis[J]. Nature Scientific Reports, 2016, 6:26286. [5] GAO L L, PAN H W, LI Q, et al. Brain medical image diagnosis based on corners with importance-values[J]. BMC Bioinformatics, 2017, 18:505. [6] GAO L L, PAN H W, XIE X Q, et al. Graph modeling and mining methods for brain images[J]. Multimedia Tools and Applications, 2016, 75(15):9333-9369. [7] ZHANG D, SHEN D. Multi-modal multi-task learning for joint prediction of multiple regression and classification variables in Alzheimer's disease[J]. NeuroImage, 2012, 59(2):895-907. [8] KOROLEV S, SAFIULLIN A, BELYAEV M, et al. Residual and plain convolutional neural networks for 3D brain MRI classification[C]//2017 IEEE 14th International Symposium on Biomedical Imaging. Melbourne, Australia:IEEE, 2017, 4:18-21. [9] BRON E E, SMITS M, VAN DER FLIER W M, et al. Standardized evaluation of algorithms for computer-aided diagnosis of dementia based on structural MRI:The CAD dementia challenge[J]. NeuroImage, 2015, 111:562-579. [10] GRAY K R, WOLZ R, HECKEMANN R A, et al. Multi-region analysis of longitudinal FDG-PET for the classification of Alzheimer's disease[J]. NeuroImage, 2012, 60(1):221-229. [11] GRAY K R, WOLZ R, KEIHANINEJAD S, et al. Regional analysis of FDG-PET for use in the classification of Alzheimer's disease[C]//2011 IEEE International Symposium on Biomedical Imaging:From Nano to Macro. Chicago, USA:IEEE, 2011, 3:1082-1085. [12] SILVEIRA M, MARQUES J. Boosting Alzheimer disease diagnosis using PET images[C]//2010 20th International Conference on Pattern Recognition. Istanbul, Turkey:IEEE, 2015, 8:2556-2559. [13] LIU M H, CHENG D N, YAN W W. Classification of Alzheimer's disease by combination of convolutional and recurrent neural networks using FDG-PET images[J]. Frontiers in Neuroinformatics, 2018, 12:35. [14] ÇIÇEK Ö, ABDULKADIR A, LIENKAMP S S, et al. 3D U-net:Learning dense volumetric segmentation from sparse annotation[C]//Proceedings of the 19th International Conference on Medical Image Computing and Computer-Assisted Intervention. Athens, Greece:Springer, 2016:1-8. [15] DOLZ J, DESROSIERS C, AYED I B. 3D fully convolutional networks for subcortical segmentation in MRI:A large-scale study[J]. NeuroImage, 2018, 170:456-470. [16] SABUNCU M R, KONUKOGLU E. Clinical prediction from structural brain MRI Scans:A large-scale empirical study[J]. Neuroinformatics, 2015, 13(1):31-46. [17] LU S, XIA Y, CAI T W, et al. Semi-supervised manifold learning with affinity regularization for Alzheimer's disease identification using positron emission tomography imaging[C]//Proceedings of the 37th Annual International Conference of the IEEE Engineering in Medicine and Biology Society. Milan, Italy:IEEE, 2015:2251-2254. [18] SELVARAJU R R, COGSWELL M, DAS A, et al. Grad-CAM:Visual explanations from deep networks via gradient-based localization[C]//Proceedings of 2017 IEEE International Conference on Computer Vision. Venice, Italy:IEEE, 2017:618-626. [19] MU Y L, GAGE F H. Adult hippocampal neurogenesis and its role in Alzheimer's disease[J]. Molecular Neurodegeneration, 2011, 6:85. [20] OTT B R, COHEN R A, ASSAWIN G, et al. Brain ventricular volume and cerebrospinal fluid biomarkers of Alzheimer's disease[J]. Journal of Alzheimer's Disease, 2010, 20(2):647-657.
2020, Vol. 60 
