AFEM-Transformer:基于自适应特征提取的Transformer阿尔兹海默症早期诊断研究

doi:10.12052/gdutxb.230194

摘要/Abstract

摘要： 目前诊断阿尔兹海默症(Alzheimer's Disease, AD)的主要手段是通过结构磁共振成像(structural Magnetic Resonance Imaging, sMRI) 实现的，现有基于深度学习的AD诊断主要利用2D(将3D sMRI转2D切片) 或者3D的卷积神经网络，不能很好地捕获3D sMRI全局特征，为此本文提出改进Swin Transformer实现3D块划分提取全局特征，构建Transformer预测分类模型。又根据现有阿尔兹海默患者发生萎缩的区域对尺寸的变换很敏感，而且现有深度学习模型在病灶区域定位方面能力不强的问题，提出自适应特征提取模块(Adaptive Feature Extraction Module, AFEM)，实现了可变形自适应特征提取，对基础3D Transformer模型进行扩展，构建AFEM-Transformer深度学习模型，进一步提升模型特征学习能力，实现自适应的定位病理区域的具体位置，用以辅助临床诊断，实现阿尔兹海默病情分类和预测。本文选择阿尔茨海默症神经影像学计划(Alzheimer's Disease Neuroimaging Initiative，ADNI)提供的2248名受试者sMRI影像作为实验数据集。在阿尔兹海默诊断任务和轻度认知障碍(Mild Cognitive Impairment, MCI) 进展预测任务中将提出的AFEM-Transformer模型与现有基于卷积神经网络的模型和基础Transformer模型进行分析比较，评估该模型的价值。结果表明本文提出的AFEM-Transformer模型在两个任务上的准确率、敏感度、特异性、AUC(Area Under Curve) 值的实验结果相较于基础和卷积神经网络模型和Transformer模型性能提升明显，证实了所提出的AFEM模块的有效性。本文提出的AFEM-Transformer深度学习模型能够准确诊断阿尔兹海默症和预测MCI的进展，并可自动定位出病灶区域，可以作为阿尔兹海默症临床诊断中一个有效的计算机辅助方法。

关键词: 深度学习, 人工智能, 阿尔兹海默症, 轻度认知障碍, 核磁共振成像

Abstract: Currently, the main approach of Alzheimer's Disease (AD) diagnosis is realized by structural magnetic resonance imaging (sMRI) , and the existing deep learning-based AD diagnosis is mainly based on 2D (converting 3D sMRI to 2D slices) or 3D convolutional neural networks, which cannot effectively capture 3D sMRI global features. To address this, this paper improves Swin Transformer to realize 3D block division for global features extraction and constructs a Transformer predictive classification model. Due to the sensitivity of existing Alzheimer's patients' regions with atrophy to the transformation of dimensions, the existing deep learning models are not capable of localizing the lesion regions. To overcome this, we propose an adaptive feature extraction module (AFEM) to realize the deformable adaptive feature extraction, and extend the basic 3D Transformer model to construct the AFEM-Transformer deep learning model to further enhance the feature learning ability of the model and realize adaptive localization of the specific location of the pathological region, which can be used to assist clinical diagnosis and realize the classification and prediction of Alzheimer's disease. In this study, sMRI images of 2248 subjects provided by the Alzheimer's Disease Neuroimaging Initiative (ADNI) were selected as the experimental dataset. The proposed AFEM-Transformer model for Alzheimer's disease diagnosis and mild cognitive impairment (MCI) progression prediction tasks will be evaluated and compared with existing convolutional neural network-based models and basic Transformer models. The results show that the experimental results of accuracy, sensitivity, specificity, and area under curve (AUC) value of the proposed AFEM-Transformer model on the two tasks show significant performance improvement compared to the convolutional neural network-based models and basic Transformer model, demonstrating the effectiveness of the proposed AFEM module. The proposed AFEM-Transformer deep learning model is able to accurately diagnose Alzheimer's disease and predict the progression of MCI, and can automatically localize the lesion area, which can be used as an effective computer-aided method in the clinical diagnosis of Alzheimer's disease.

Key words: deep learning, artificial intelligence, Alzheimer's disease, mild cognitive impairment, magnetic resonance imaging

中图分类号:

TP391

徐萍萍, 黄国恒, 赵钦, 陈一嘉. AFEM-Transformer:基于自适应特征提取的Transformer阿尔兹海默症早期诊断研究[J]. 广东工业大学学报, 2025, 42(1): 51-59.doi: 10.12052/gdutxb.230194

Xu Pingping, Huang Guoheng, Zhao Qin, Chen Yijia. AFEM-Transformer: Early Diagnosis of Alzheimer's Disease Based on Adaptive Feature Extraction with Transformer[J]. Journal of Guangdong University of Technology, 2025, 42(1): 51-59.doi: 10.12052/gdutxb.230194

参考文献

[1] SCHELTENS P, DE STROOPER B, KIVIPELTO M, et al. Alzheimer's disease [J]. Lancet, 2021, 397(10284): 1577-1590.
[2] LEMPRIÈRE S. New tools could improve Alzheimer disease diagnosis from structural MRI [J]. Nature Reviews Neurology, 2020, 16(6): 297-297.
[3] ANTONAKAKIS M, DIMITRIADIS S I, ZERVAKIS M, et al. Aberrant whole-brain transitions and dynamics of spontaneous network microstates in mild traumatic brain injury [J]. Frontiers in Computational Neuroscience, 2019, 13: 90-90.
[4] SUDHARSAN M, THAILAMBAL G. Alzheimer's disease prediction using machine learning techniques and principal component analysis (PCA) [J]. Materials Today: Proceedings, 2023, 81: 182-190.
[5] DOU X, YAO H, FENG F, et al. Characterizing white matter connectivity in Alzheimer's disease and mild cognitive impairment: an automated fiber quantification analysis with two independent datasets [J]. Cortex, 2020, 129: 390-405.
[6] 罗刚, 泮思林, 乔思波, 等. 深度学习技术在胎儿超声心动图图像自动识别中的应用[J]. 实用医学杂志, 2022, 38(14): 1830-1833.
LUO G, PAN S L, QIAO S B, et al. Deep learning technology for automatic recognition of fetal echocardiography images [J]. The Journal of Practical Medicine, 2022, 38(14): 1830-1833.
[7] 杨磊, 唐灿. 人工智能在乳腺癌超声诊断的应用价值[J]. 实用医学杂志, 2022, 38(1): 106-110.
YANG L, TANG C. CAD machine diagnosis system in ultrasonic diagnosis of breast cancer [J]. The Journal of Practical Medicine, 2022, 38(1): 106-110.
[8] 廖淑婷, 于向荣. 能谱CT和人工智能在甲状腺癌诊断中的应用[J]. 实用医学杂志, 2022, 38(2): 129-133.
LIAO S T, YU X R. Application of spectral CT and artificial intelligence in the diagnosis of thyroid cancer [J]. The Journal of Practical Medicine, 2022, 38(2): 129-133.
[9] LIU M, ZHANG J, ADELI E, et al. Landmark-based deep multi-instance learning for brain disease diagnosis [J]. Medical Image Analysis, 2018, 43: 157-168.
[10] LIAN C, LIU M, ZHANG J, et al. Hierarchical fully convolutional network for joint atrophy localization and Alzheimer’s disease diagnosis using structural mri [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2018, 42(4): 880-893.
[11] ZHU W, SUN L, HUANG J, et al. Dual attention multi-instance deep learning for Alzheimer’s disease diagnosis with structural MRI [J]. IEEE Transactions on Medical Imaging, 2021, 40(9): 2354-2366.
[12] 曾安, 黄殷, 潘丹, 等. 基于卷积循环神经网络的阿尔茨海默症早期诊断[J]. 生物医学工程研究, 2020, 39(3): 249-255.
ZENG A, HUANG Y, PAN D, et al. Early diagnosis of Alzheimer's disease based on convolutional recurrent neural network [J]. Journal of Biomedical Engineering Research, 2020, 39(3): 249-255.
[13] 王聪, 袁榕澳, 李川. 基于3D卷积神经网络的阿尔茨海默症及轻度认知功能障碍诊断[J]. 现代计算机, 2021(11): 120-123.
WANG C, YUAN R A, LI C. Analysis of Alzheimer’s disease medical images based on deep machine learning [J]. Modern Computer, 2021(11): 120-123.
[14] QIN Z, LIU Z, GUO Q, et al. 3D convolutional neural networks with hybrid attention mechanism for early diagnosis of Alzheimer’s disease [J]. Biomedical Signal Processing and Control, 2022, 77: 103828.
[15] ZHANG J, ZHOU L, WANG L, et al. Diffusion kernel attention network for brain disorder classification [J]. IEEE Transactions on Medical Imaging, 2022, 41(10): 2814-2827.
[16] 石磊, 彭少康, 张亚萌, 等. 基于特征增强金字塔网络的阿尔茨海默症早期诊断研究[J]. 数据采集与处理, 2022, 37(4): 727-735.
SHI L, PENG S K, ZHANG Y M, et al. Early diagnosis of Alzheimer’s disease based on feature enhanced pyramid network [J]. Journal of Data Acquisition and Processing, 2022, 37(4): 727-735.
[17] ZHANG X, HAN L, ZHU W, et al. An explainable 3D residual self-attention deep neural network for joint atrophy localization and Alzheimer’s disease diagnosis using structural MRI [J]. IEEE Journal of Biomedical and Health Informatics, 2021, 26(11): 5289-5297.
[18] ZHANG X, HAN L, HAN L, et al. sMRI-PatchNet: A novel efficient explainable patch-based deep learning network for Alzheimer’s disease diagnosis with structural MRI [J]. IEEE Access, 2023, 11: 108603-108616.
[19] VASWANI A, SHAZEER N, PARMAR N, et al. Attention is all you need[C]//Proceedings of the 31st International Conference on Neural Information Processing Systems (NeurIPS). New York: ACM, 2017: 6000-6010.
[20] DOSOVITSKIY A, BEYER L, KOLESNIKOV A, et al. An image is worth 16×16 words: transformers for image recognition at scale[EB/OL]. arXiv: 2010.11929 (2021-01-03) [2024-01-25]. https://doi.org/10.48550/arXiv.2010.11929
[21] LIU Z, LIN Y, CAO Y, et al. Swin transformer: hierarchical vision transformer using shifted windows[C]//Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) . New York: IEEE, 2021: 9992-10002.
[22] ZHU X, HU H, LIN S, et al. Deformable ConvNets V2: more deformable, better results[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) . New York: IEEE, 2019: 9300-9308.
[23] ZHU X, SU W, LU L, et al. Deformable detr: deformable transformers for end-to-end object detection[EB/OL]. arXiv: 2010.04159 (2021-03-28) [2024-01-25]. https://doi.org/10.48550/arXiv.2010. 04159
[24] CHEN Z, ZHU Y, ZHAO C, et al. DPT: deformable patch-based transformer for visual recognition[C]//Proceedings of the 29th ACM International Conference on Multimedia. New York: Association for Computing Machinery, 2021: 2899-2907.
[25] XIA Z, PAN X, SONG S, et al. Vision transformer with deformable attention[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Los Alamitos: IEEE, 2022: 4784-4793.
[26] JACK J C R, BERNSTEIN M A, FOX N C, et al. The Alzheimer's disease neuroimaging initiative (ADNI) : MRI methods [J]. Journal of Magnetic Resonance Imaging, 2008, 27(4): 685-691.
[27] PENNY W D, FRISTON K J, ASHBURNER J T, et al. Statistical parametric mapping: the analysis of functional brain images[M]. Amsterdam: Academic Press, 2007: 1-647.
[28] DING X, ZHANG X, MA N, et al. RepVGG: Making VGG-style convnets great again[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Nashville: IEEE, 2021: 13728-13737.
[29] PAN X, PHAN T L, ADEL M, et al. Multi-view separable pyramid network for AD prediction at MCI stage by 18 F-FDG brain PET imaging [J]. IEEE Transactions on Medical Imaging, 2020, 40(1): 81-92.
[30] HUANG J, ZHOU L, WANG L, et al. Attention-diffusion-bilinear neural network for brain network analysis [J]. IEEE Transactions on Medical Imaging, 2020, 39(7): 2541-2552.

Metrics

Viewed

Full text

230

HTML			PDF

Just accepted	Online first	Issue	Just accepted	Online first	Issue
0	0	0	0	9	221

From	Others	local

Times	25	205
Rate	11%	89%

Abstract

123

Just accepted	Online first	Issue

0	10	113

Cited

Shared

Discussed