Explainable graph neural network based on metabolic brain imaging for differential diagnosis of parkinsonism
BackgroundAccurate differentiation of parkinsonian syndromes remains challenging due to overlapping clinical manifestations and subtle neuroimaging variations. This study introduces an explainable graph neural network (GNN) framework integrating a Regional Radiomics Similarity Network (R2SN) and Tra...
Saved in:
| Main Authors: | , , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2025-04-01
|
| Series: | Frontiers in Aging Neuroscience |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/fnagi.2025.1580910/full |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1850186232704794624 |
|---|---|
| author | Ronghua Ling Ronghua Ling Xingxing Cen Shaoyou Wu Min Wang Ying Zhang Juanjuan Jiang Jiaying Lu Yingqian Liu Chuantao Zuo Jiehui Jiang Yinghui Yang Zhuangzhi Yan Zhuangzhi Yan |
| author_facet | Ronghua Ling Ronghua Ling Xingxing Cen Shaoyou Wu Min Wang Ying Zhang Juanjuan Jiang Jiaying Lu Yingqian Liu Chuantao Zuo Jiehui Jiang Yinghui Yang Zhuangzhi Yan Zhuangzhi Yan |
| author_sort | Ronghua Ling |
| collection | DOAJ |
| description | BackgroundAccurate differentiation of parkinsonian syndromes remains challenging due to overlapping clinical manifestations and subtle neuroimaging variations. This study introduces an explainable graph neural network (GNN) framework integrating a Regional Radiomics Similarity Network (R2SN) and Transformer-based attention mechanisms to address this diagnostic dilemma.MethodsOur study prospectively enrolled 1,495 participants, including 220 healthy controls and 1,275 patients diagnosed with idiopathic Parkinson’s disease (IPD), multiple system atrophy (MSA), or progressive supranuclear palsy (PSP), all undergoing standardized 18F-fluorodeoxyglucose positron emission tomography imaging. Metabolic networks were constructed by encoding edge weights derived from radiomic feature similarity matrices, enabling simultaneous quantification of microscopic metabolic heterogeneity and macroscale network reorganization.ResultsThe proposed framework achieved superior classification performance with F1-scores of 92.5% (MSA), 96.3% (IPD), and 86.7% (PSP), significantly outperforming comparators by 5.5–8.3%. Multiscale interpretability analysis revealed: (1) Regional hypometabolism in pathognomonic nodes (putamen in IPD, midbrain tegmentum in PSP); (2) Disease-specific connectivity disruptions (midbrain-prefrontal disconnection in PSP, cerebellar-pontine decoupling in MSA). The substructure attention mechanism reduced computational complexity by 41% while enhancing diagnostic specificity (PSP precision +5.2%).ConclusionThe proposed R2SN-based explainable GNN framework for parkinsonian syndrome differentiation establishes a new paradigm for precision subtyping of neurodegenerative disorders, with methodological extensibility to other network-driven neurological conditions. |
| format | Article |
| id | doaj-art-3e455ebffbf24952acf19da7cfa01c24 |
| institution | OA Journals |
| issn | 1663-4365 |
| language | English |
| publishDate | 2025-04-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Aging Neuroscience |
| spelling | doaj-art-3e455ebffbf24952acf19da7cfa01c242025-08-20T02:16:23ZengFrontiers Media S.A.Frontiers in Aging Neuroscience1663-43652025-04-011710.3389/fnagi.2025.15809101580910Explainable graph neural network based on metabolic brain imaging for differential diagnosis of parkinsonismRonghua Ling0Ronghua Ling1Xingxing Cen2Shaoyou Wu3Min Wang4Ying Zhang5Juanjuan Jiang6Jiaying Lu7Yingqian Liu8Chuantao Zuo9Jiehui Jiang10Yinghui Yang11Zhuangzhi Yan12Zhuangzhi Yan13School of Communication and Information Engineering, Shanghai University, Shanghai, ChinaSchool of Medical Imaging, Shanghai University of Medicine and Health Science, Shanghai, ChinaGlorious Sun School of Business and Management, Donghua University, Shanghai, ChinaSchool of Life Sciences, Shanghai University, Shanghai, ChinaSchool of Life Sciences, Shanghai University, Shanghai, ChinaSchool of Life Sciences, Shanghai University, Shanghai, ChinaSchool of Medical Imaging, Shanghai University of Medicine and Health Science, Shanghai, ChinaDepartment of Nuclear Medicine and PET Center, National Clinical Research Center for Aging and Medicine, National Center for Neurological Disorders, Huashan Hospital, Fudan University, Shanghai, ChinaSchool of Electrical Engineering, Shandong University of Aeronautics, Binzhou, ChinaDepartment of Nuclear Medicine and PET Center, National Clinical Research Center for Aging and Medicine, National Center for Neurological Disorders, Huashan Hospital, Fudan University, Shanghai, ChinaSchool of Life Sciences, Shanghai University, Shanghai, ChinaCollege of Health Management, Shanghai Jian Qiao University, Shanghai, ChinaSchool of Communication and Information Engineering, Shanghai University, Shanghai, ChinaSchool of Life Sciences, Shanghai University, Shanghai, ChinaBackgroundAccurate differentiation of parkinsonian syndromes remains challenging due to overlapping clinical manifestations and subtle neuroimaging variations. This study introduces an explainable graph neural network (GNN) framework integrating a Regional Radiomics Similarity Network (R2SN) and Transformer-based attention mechanisms to address this diagnostic dilemma.MethodsOur study prospectively enrolled 1,495 participants, including 220 healthy controls and 1,275 patients diagnosed with idiopathic Parkinson’s disease (IPD), multiple system atrophy (MSA), or progressive supranuclear palsy (PSP), all undergoing standardized 18F-fluorodeoxyglucose positron emission tomography imaging. Metabolic networks were constructed by encoding edge weights derived from radiomic feature similarity matrices, enabling simultaneous quantification of microscopic metabolic heterogeneity and macroscale network reorganization.ResultsThe proposed framework achieved superior classification performance with F1-scores of 92.5% (MSA), 96.3% (IPD), and 86.7% (PSP), significantly outperforming comparators by 5.5–8.3%. Multiscale interpretability analysis revealed: (1) Regional hypometabolism in pathognomonic nodes (putamen in IPD, midbrain tegmentum in PSP); (2) Disease-specific connectivity disruptions (midbrain-prefrontal disconnection in PSP, cerebellar-pontine decoupling in MSA). The substructure attention mechanism reduced computational complexity by 41% while enhancing diagnostic specificity (PSP precision +5.2%).ConclusionThe proposed R2SN-based explainable GNN framework for parkinsonian syndrome differentiation establishes a new paradigm for precision subtyping of neurodegenerative disorders, with methodological extensibility to other network-driven neurological conditions.https://www.frontiersin.org/articles/10.3389/fnagi.2025.1580910/fullparkinsonismglucose metabolismgraph neural networkbrain imagingclassification |
| spellingShingle | Ronghua Ling Ronghua Ling Xingxing Cen Shaoyou Wu Min Wang Ying Zhang Juanjuan Jiang Jiaying Lu Yingqian Liu Chuantao Zuo Jiehui Jiang Yinghui Yang Zhuangzhi Yan Zhuangzhi Yan Explainable graph neural network based on metabolic brain imaging for differential diagnosis of parkinsonism Frontiers in Aging Neuroscience parkinsonism glucose metabolism graph neural network brain imaging classification |
| title | Explainable graph neural network based on metabolic brain imaging for differential diagnosis of parkinsonism |
| title_full | Explainable graph neural network based on metabolic brain imaging for differential diagnosis of parkinsonism |
| title_fullStr | Explainable graph neural network based on metabolic brain imaging for differential diagnosis of parkinsonism |
| title_full_unstemmed | Explainable graph neural network based on metabolic brain imaging for differential diagnosis of parkinsonism |
| title_short | Explainable graph neural network based on metabolic brain imaging for differential diagnosis of parkinsonism |
| title_sort | explainable graph neural network based on metabolic brain imaging for differential diagnosis of parkinsonism |
| topic | parkinsonism glucose metabolism graph neural network brain imaging classification |
| url | https://www.frontiersin.org/articles/10.3389/fnagi.2025.1580910/full |
| work_keys_str_mv | AT ronghualing explainablegraphneuralnetworkbasedonmetabolicbrainimagingfordifferentialdiagnosisofparkinsonism AT ronghualing explainablegraphneuralnetworkbasedonmetabolicbrainimagingfordifferentialdiagnosisofparkinsonism AT xingxingcen explainablegraphneuralnetworkbasedonmetabolicbrainimagingfordifferentialdiagnosisofparkinsonism AT shaoyouwu explainablegraphneuralnetworkbasedonmetabolicbrainimagingfordifferentialdiagnosisofparkinsonism AT minwang explainablegraphneuralnetworkbasedonmetabolicbrainimagingfordifferentialdiagnosisofparkinsonism AT yingzhang explainablegraphneuralnetworkbasedonmetabolicbrainimagingfordifferentialdiagnosisofparkinsonism AT juanjuanjiang explainablegraphneuralnetworkbasedonmetabolicbrainimagingfordifferentialdiagnosisofparkinsonism AT jiayinglu explainablegraphneuralnetworkbasedonmetabolicbrainimagingfordifferentialdiagnosisofparkinsonism AT yingqianliu explainablegraphneuralnetworkbasedonmetabolicbrainimagingfordifferentialdiagnosisofparkinsonism AT chuantaozuo explainablegraphneuralnetworkbasedonmetabolicbrainimagingfordifferentialdiagnosisofparkinsonism AT jiehuijiang explainablegraphneuralnetworkbasedonmetabolicbrainimagingfordifferentialdiagnosisofparkinsonism AT yinghuiyang explainablegraphneuralnetworkbasedonmetabolicbrainimagingfordifferentialdiagnosisofparkinsonism AT zhuangzhiyan explainablegraphneuralnetworkbasedonmetabolicbrainimagingfordifferentialdiagnosisofparkinsonism AT zhuangzhiyan explainablegraphneuralnetworkbasedonmetabolicbrainimagingfordifferentialdiagnosisofparkinsonism |