1
|
Fan X, Li H, Liu L, Zhang K, Zhang Z, Chen Y, Wang Z, He X, Xu J, Hu Q. Early Diagnosing and Transformation Prediction of Alzheimer's Disease Using Multi-Scaled Self-Attention Network on Structural MRI Images with Occlusion Sensitivity Analysis. J Alzheimers Dis 2024; 97:909-926. [PMID: 38160355 DOI: 10.3233/jad-230705] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 01/03/2024]
Abstract
BACKGROUND Structural magnetic resonance imaging (sMRI) is vital for early Alzheimer's disease (AD) diagnosis, though confirming specific biomarkers remains challenging. Our proposed Multi-Scale Self-Attention Network (MUSAN) enhances classification of cognitively normal (CN) and AD individuals, distinguishing stable (sMCI) from progressive mild cognitive impairment (pMCI). OBJECTIVE This study leverages AD structural atrophy properties to achieve precise AD classification, combining different scales of brain region features. The ultimate goal is an interpretable algorithm for this method. METHODS The MUSAN takes whole-brain sMRI as input, enabling automatic extraction of brain region features and modeling of correlations between different scales of brain regions, and achieves personalized disease interpretation of brain regions. Furthermore, we also employed an occlusion sensitivity algorithm to localize and visualize brain regions sensitive to disease. RESULTS Our method is applied to ADNI-1, ADNI-2, and ADNI-3, and achieves high performance on the classification of CN from AD with accuracy (0.93), specificity (0.82), sensitivity (0.96), and area under curve (AUC) (0.95), as well as notable performance on the distinguish of sMCI from pMCI with accuracy (0.85), specificity (0.84), sensitivity (0.74), and AUC (0.86). Our sensitivity masking algorithm identified key regions in distinguishing CN from AD: hippocampus, amygdala, and vermis. Moreover, cingulum, pallidum, and inferior frontal gyrus are crucial for sMCI and pMCI discrimination. These discoveries align with existing literature, confirming the dependability of our model in AD research. CONCLUSION Our method provides an effective AD diagnostic and conversion prediction method. The occlusion sensitivity algorithm enhances deep learning interpretability, bolstering AD research reliability.
Collapse
Affiliation(s)
- Xinxin Fan
- Institute of Biomedical and Health Engineering, Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences, Shenzhen, China
- University of Chinese Academy of Sciences, Beijing, China
| | - Haining Li
- Department of Neurology, General Hospital of Ningxia Medical University, Yinchuan, China
| | - Lin Liu
- University of Chinese Academy of Sciences, Beijing, China
| | - Kai Zhang
- Institute of Biomedical and Health Engineering, Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences, Shenzhen, China
| | - Zhewei Zhang
- Institute of Biomedical and Health Engineering, Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences, Shenzhen, China
| | - Yi Chen
- Institute of Biomedical and Health Engineering, Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences, Shenzhen, China
| | - Zhen Wang
- Zhuhai Institute of Advanced Technology, Zhuhai, China
| | - Xiaoli He
- Department of Psychology, Ningxia University, Yinchuan, China
| | - Jinping Xu
- Institute of Biomedical and Health Engineering, Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences, Shenzhen, China
| | - Qingmao Hu
- Institute of Biomedical and Health Engineering, Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences, Shenzhen, China
| |
Collapse
|
2
|
Xu W, Sun X, Jiang H, Wang X, Wang B, Niu Q, Meng H, Du J, Yang G, Liu B, Zhang H, Tan Y. Diffusion Kurtosis Imaging in Evaluating the Mild Cognitive Impairment of Occupational Aluminum Workers. Acad Radiol 2023; 30:2225-2233. [PMID: 36690563 DOI: 10.1016/j.acra.2022.12.003] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 08/28/2022] [Revised: 12/02/2022] [Accepted: 12/02/2022] [Indexed: 01/23/2023]
Abstract
RATIONALE AND OBJECTIVES To investigate whether diffusion kurtosis imaging (DKI) can distinguish mild cognitive impairment (MCI) from normal controls (NC) in aluminum (Al)-exposed workers, and to explore the association of DKI with cognitive performance and plasma Al concentration. MATERIALS AND METHODS 28 patients with MCI and 25 NC at Al factory were enrolled in this study. All subjects underwent conventional MRI and DKI scans. The mean kurtosis (MK), axial kurtosis (Ka), radial kurtosis (Kr), mean diffusivity (MD) and fractional anisotropy (FA) parameters of the hippocampus, substantia nigra, red nucleus, thalamus, anterior cingulate gyrus, genu and crus of the corpus callosum, frontal, parietal and temporal lobe were measured. To compare the parameters between the two groups, the Mann-Whitney rank sum test was used. The correlation of parameter values with cognitive performance and plasma Al concentration was analyzed using Spearman correlation analysis. The receiver operating characteristic (ROC) curve and the Z-scores were used to evaluate the diagnostic efficacy of each parameter. RESULTS Compared with the NC group, the MK, Ka, Kr, and FA values in the MCI group were significantly decreased, and the MD values were significantly increased (p<0.05). For the diagnosis of MCI, MK in the right hippocampus showed the largest AUC (0.924). The MK, Kr, MD and FA values were correlated with the Montreal Cognitive Assessment (MoCA) scores, and MK values in the right hippocampus showed the greatest correlation with MoCA scores (r=0.744, p <0.001). Plasma Al in the MCI group was higher than that in the NC group, although there was no significant difference in plasma Al between the two groups (p=0.057). There was no correlation between DKI parameters and plasma Al. CONCLUSION The DKI method might be a sensitive imaging biomarker to discriminate MCI from NC, and could preliminarily assess the severity of cognitive impairment in Al-exposed workers. MK in the right hippocampus appeared to be the best independent predictor. The mechanism of cognitive decline is an important content of aluminum exposure research. This study indicates that the DKI technique could provide valuable information for the diagnosis of MCI.
Collapse
Affiliation(s)
- Wenji Xu
- College of Medical Imaging, Shanxi Medical University, Taiyuan, Shanxi Province, China
| | - Xiangru Sun
- College of Medical Imaging, Shanxi Medical University, Taiyuan, Shanxi Province, China
| | - Haoru Jiang
- College of Medical Imaging, Shanxi Medical University, Taiyuan, Shanxi Province, China
| | - Xiaochun Wang
- Department of Radiology, First Hospital of Shanxi Medical University, Taiyuan, Shanxi Province, China.; Shanxi Key Laboratory of Intelligent Imaging and Nanomedicine, First Hospital of Shanxi Medical University, Taiyuan, Shanxi Province, China.; Intelligent Imaging Big Data and Functional Nano-imaging Engineering Research Center of Shanxi Province, First Hospital of Shanxi Medical University, Taiyuan, Shanxi Province, China
| | - Bin Wang
- Department of Radiology, First Hospital of Shanxi Medical University, Taiyuan, Shanxi Province, China.; Shanxi Key Laboratory of Intelligent Imaging and Nanomedicine, First Hospital of Shanxi Medical University, Taiyuan, Shanxi Province, China.; Intelligent Imaging Big Data and Functional Nano-imaging Engineering Research Center of Shanxi Province, First Hospital of Shanxi Medical University, Taiyuan, Shanxi Province, China
| | - Qiao Niu
- School of Public Health, Shanxi Medical University, Taiyuan, Shanxi Province, China
| | - Huaxing Meng
- School of Public Health, Shanxi Medical University, Taiyuan, Shanxi Province, China
| | - Jiangfeng Du
- Department of Radiology, First Hospital of Shanxi Medical University, Taiyuan, Shanxi Province, China.; Shanxi Key Laboratory of Intelligent Imaging and Nanomedicine, First Hospital of Shanxi Medical University, Taiyuan, Shanxi Province, China.; Intelligent Imaging Big Data and Functional Nano-imaging Engineering Research Center of Shanxi Province, First Hospital of Shanxi Medical University, Taiyuan, Shanxi Province, China
| | - Guoqiang Yang
- Department of Radiology, First Hospital of Shanxi Medical University, Taiyuan, Shanxi Province, China.; Shanxi Key Laboratory of Intelligent Imaging and Nanomedicine, First Hospital of Shanxi Medical University, Taiyuan, Shanxi Province, China.; Intelligent Imaging Big Data and Functional Nano-imaging Engineering Research Center of Shanxi Province, First Hospital of Shanxi Medical University, Taiyuan, Shanxi Province, China
| | - Bo Liu
- Department of Radiology, First Hospital of Shanxi Medical University, Taiyuan, Shanxi Province, China.; Shanxi Key Laboratory of Intelligent Imaging and Nanomedicine, First Hospital of Shanxi Medical University, Taiyuan, Shanxi Province, China.; Intelligent Imaging Big Data and Functional Nano-imaging Engineering Research Center of Shanxi Province, First Hospital of Shanxi Medical University, Taiyuan, Shanxi Province, China
| | - Hui Zhang
- Department of Radiology, First Hospital of Shanxi Medical University, Taiyuan, Shanxi Province, China.; Shanxi Key Laboratory of Intelligent Imaging and Nanomedicine, First Hospital of Shanxi Medical University, Taiyuan, Shanxi Province, China.; Intelligent Imaging Big Data and Functional Nano-imaging Engineering Research Center of Shanxi Province, First Hospital of Shanxi Medical University, Taiyuan, Shanxi Province, China
| | - Yan Tan
- Department of Radiology, First Hospital of Shanxi Medical University, Taiyuan, Shanxi Province, China.; Shanxi Key Laboratory of Intelligent Imaging and Nanomedicine, First Hospital of Shanxi Medical University, Taiyuan, Shanxi Province, China.; Intelligent Imaging Big Data and Functional Nano-imaging Engineering Research Center of Shanxi Province, First Hospital of Shanxi Medical University, Taiyuan, Shanxi Province, China..
| |
Collapse
|