Cargando…
An automatic interpretable deep learning pipeline for accurate Parkinson's disease diagnosis using quantitative susceptibility mapping and T1‐weighted images
Parkinson's disease (PD) diagnosis based on magnetic resonance imaging (MRI) is still challenging clinically. Quantitative susceptibility maps (QSM) can potentially provide underlying pathophysiological information by detecting the iron distribution in deep gray matter (DGM) nuclei. We hypothes...
Autores principales: | , , , , , , , , , , , , , , , , , |
---|---|
Formato: | Online Artículo Texto |
Lenguaje: | English |
Publicado: |
John Wiley & Sons, Inc.
2023
|
Materias: | |
Acceso en línea: | https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10365226/ https://www.ncbi.nlm.nih.gov/pubmed/37335041 http://dx.doi.org/10.1002/hbm.26399 |
_version_ | 1785076995583901696 |
---|---|
author | Wang, Yida He, Naying Zhang, Chunyan Zhang, Youmin Wang, Chenglong Huang, Pei Jin, Zhijia Li, Yan Cheng, Zenghui Liu, Yu Wang, Xinhui Chen, Chen Cheng, Jingliang Liu, Fangtao Haacke, Ewart Mark Chen, Shengdi Yang, Guang Yan, Fuhua |
author_facet | Wang, Yida He, Naying Zhang, Chunyan Zhang, Youmin Wang, Chenglong Huang, Pei Jin, Zhijia Li, Yan Cheng, Zenghui Liu, Yu Wang, Xinhui Chen, Chen Cheng, Jingliang Liu, Fangtao Haacke, Ewart Mark Chen, Shengdi Yang, Guang Yan, Fuhua |
author_sort | Wang, Yida |
collection | PubMed |
description | Parkinson's disease (PD) diagnosis based on magnetic resonance imaging (MRI) is still challenging clinically. Quantitative susceptibility maps (QSM) can potentially provide underlying pathophysiological information by detecting the iron distribution in deep gray matter (DGM) nuclei. We hypothesized that deep learning (DL) could be used to automatically segment all DGM nuclei and use relevant features for a better differentiation between PD and healthy controls (HC). In this study, we proposed a DL‐based pipeline for automatic PD diagnosis based on QSM and T1‐weighted (T1W) images. This consists of (1) a convolutional neural network model integrated with multiple attention mechanisms which simultaneously segments caudate nucleus, globus pallidus, putamen, red nucleus, and substantia nigra from QSM and T1W images, and (2) an SE‐ResNeXt50 model with an anatomical attention mechanism, which uses QSM data and the segmented nuclei to distinguish PD from HC. The mean dice values for segmentation of the five DGM nuclei are all >0.83 in the internal testing cohort, suggesting that the model could segment brain nuclei accurately. The proposed PD diagnosis model achieved area under the the receiver operating characteristic curve (AUCs) of 0.901 and 0.845 on independent internal and external testing cohorts, respectively. Gradient‐weighted class activation mapping (Grad‐CAM) heatmaps were used to identify contributing nuclei for PD diagnosis on patient level. In conclusion, the proposed approach can potentially be used as an automatic, explainable pipeline for PD diagnosis in a clinical setting. |
format | Online Article Text |
id | pubmed-10365226 |
institution | National Center for Biotechnology Information |
language | English |
publishDate | 2023 |
publisher | John Wiley & Sons, Inc. |
record_format | MEDLINE/PubMed |
spelling | pubmed-103652262023-07-25 An automatic interpretable deep learning pipeline for accurate Parkinson's disease diagnosis using quantitative susceptibility mapping and T1‐weighted images Wang, Yida He, Naying Zhang, Chunyan Zhang, Youmin Wang, Chenglong Huang, Pei Jin, Zhijia Li, Yan Cheng, Zenghui Liu, Yu Wang, Xinhui Chen, Chen Cheng, Jingliang Liu, Fangtao Haacke, Ewart Mark Chen, Shengdi Yang, Guang Yan, Fuhua Hum Brain Mapp Technical Report Parkinson's disease (PD) diagnosis based on magnetic resonance imaging (MRI) is still challenging clinically. Quantitative susceptibility maps (QSM) can potentially provide underlying pathophysiological information by detecting the iron distribution in deep gray matter (DGM) nuclei. We hypothesized that deep learning (DL) could be used to automatically segment all DGM nuclei and use relevant features for a better differentiation between PD and healthy controls (HC). In this study, we proposed a DL‐based pipeline for automatic PD diagnosis based on QSM and T1‐weighted (T1W) images. This consists of (1) a convolutional neural network model integrated with multiple attention mechanisms which simultaneously segments caudate nucleus, globus pallidus, putamen, red nucleus, and substantia nigra from QSM and T1W images, and (2) an SE‐ResNeXt50 model with an anatomical attention mechanism, which uses QSM data and the segmented nuclei to distinguish PD from HC. The mean dice values for segmentation of the five DGM nuclei are all >0.83 in the internal testing cohort, suggesting that the model could segment brain nuclei accurately. The proposed PD diagnosis model achieved area under the the receiver operating characteristic curve (AUCs) of 0.901 and 0.845 on independent internal and external testing cohorts, respectively. Gradient‐weighted class activation mapping (Grad‐CAM) heatmaps were used to identify contributing nuclei for PD diagnosis on patient level. In conclusion, the proposed approach can potentially be used as an automatic, explainable pipeline for PD diagnosis in a clinical setting. John Wiley & Sons, Inc. 2023-06-19 /pmc/articles/PMC10365226/ /pubmed/37335041 http://dx.doi.org/10.1002/hbm.26399 Text en © 2023 The Authors. Human Brain Mapping published by Wiley Periodicals LLC. https://creativecommons.org/licenses/by-nc-nd/4.0/This is an open access article under the terms of the http://creativecommons.org/licenses/by-nc-nd/4.0/ (https://creativecommons.org/licenses/by-nc-nd/4.0/) License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non‐commercial and no modifications or adaptations are made. |
spellingShingle | Technical Report Wang, Yida He, Naying Zhang, Chunyan Zhang, Youmin Wang, Chenglong Huang, Pei Jin, Zhijia Li, Yan Cheng, Zenghui Liu, Yu Wang, Xinhui Chen, Chen Cheng, Jingliang Liu, Fangtao Haacke, Ewart Mark Chen, Shengdi Yang, Guang Yan, Fuhua An automatic interpretable deep learning pipeline for accurate Parkinson's disease diagnosis using quantitative susceptibility mapping and T1‐weighted images |
title | An automatic interpretable deep learning pipeline for accurate Parkinson's disease diagnosis using quantitative susceptibility mapping and T1‐weighted images
|
title_full | An automatic interpretable deep learning pipeline for accurate Parkinson's disease diagnosis using quantitative susceptibility mapping and T1‐weighted images
|
title_fullStr | An automatic interpretable deep learning pipeline for accurate Parkinson's disease diagnosis using quantitative susceptibility mapping and T1‐weighted images
|
title_full_unstemmed | An automatic interpretable deep learning pipeline for accurate Parkinson's disease diagnosis using quantitative susceptibility mapping and T1‐weighted images
|
title_short | An automatic interpretable deep learning pipeline for accurate Parkinson's disease diagnosis using quantitative susceptibility mapping and T1‐weighted images
|
title_sort | automatic interpretable deep learning pipeline for accurate parkinson's disease diagnosis using quantitative susceptibility mapping and t1‐weighted images |
topic | Technical Report |
url | https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10365226/ https://www.ncbi.nlm.nih.gov/pubmed/37335041 http://dx.doi.org/10.1002/hbm.26399 |
work_keys_str_mv | AT wangyida anautomaticinterpretabledeeplearningpipelineforaccurateparkinsonsdiseasediagnosisusingquantitativesusceptibilitymappingandt1weightedimages AT henaying anautomaticinterpretabledeeplearningpipelineforaccurateparkinsonsdiseasediagnosisusingquantitativesusceptibilitymappingandt1weightedimages AT zhangchunyan anautomaticinterpretabledeeplearningpipelineforaccurateparkinsonsdiseasediagnosisusingquantitativesusceptibilitymappingandt1weightedimages AT zhangyoumin anautomaticinterpretabledeeplearningpipelineforaccurateparkinsonsdiseasediagnosisusingquantitativesusceptibilitymappingandt1weightedimages AT wangchenglong anautomaticinterpretabledeeplearningpipelineforaccurateparkinsonsdiseasediagnosisusingquantitativesusceptibilitymappingandt1weightedimages AT huangpei anautomaticinterpretabledeeplearningpipelineforaccurateparkinsonsdiseasediagnosisusingquantitativesusceptibilitymappingandt1weightedimages AT jinzhijia anautomaticinterpretabledeeplearningpipelineforaccurateparkinsonsdiseasediagnosisusingquantitativesusceptibilitymappingandt1weightedimages AT liyan anautomaticinterpretabledeeplearningpipelineforaccurateparkinsonsdiseasediagnosisusingquantitativesusceptibilitymappingandt1weightedimages AT chengzenghui anautomaticinterpretabledeeplearningpipelineforaccurateparkinsonsdiseasediagnosisusingquantitativesusceptibilitymappingandt1weightedimages AT liuyu anautomaticinterpretabledeeplearningpipelineforaccurateparkinsonsdiseasediagnosisusingquantitativesusceptibilitymappingandt1weightedimages AT wangxinhui anautomaticinterpretabledeeplearningpipelineforaccurateparkinsonsdiseasediagnosisusingquantitativesusceptibilitymappingandt1weightedimages AT chenchen anautomaticinterpretabledeeplearningpipelineforaccurateparkinsonsdiseasediagnosisusingquantitativesusceptibilitymappingandt1weightedimages AT chengjingliang anautomaticinterpretabledeeplearningpipelineforaccurateparkinsonsdiseasediagnosisusingquantitativesusceptibilitymappingandt1weightedimages AT liufangtao anautomaticinterpretabledeeplearningpipelineforaccurateparkinsonsdiseasediagnosisusingquantitativesusceptibilitymappingandt1weightedimages AT haackeewartmark anautomaticinterpretabledeeplearningpipelineforaccurateparkinsonsdiseasediagnosisusingquantitativesusceptibilitymappingandt1weightedimages AT chenshengdi anautomaticinterpretabledeeplearningpipelineforaccurateparkinsonsdiseasediagnosisusingquantitativesusceptibilitymappingandt1weightedimages AT yangguang anautomaticinterpretabledeeplearningpipelineforaccurateparkinsonsdiseasediagnosisusingquantitativesusceptibilitymappingandt1weightedimages AT yanfuhua anautomaticinterpretabledeeplearningpipelineforaccurateparkinsonsdiseasediagnosisusingquantitativesusceptibilitymappingandt1weightedimages AT wangyida automaticinterpretabledeeplearningpipelineforaccurateparkinsonsdiseasediagnosisusingquantitativesusceptibilitymappingandt1weightedimages AT henaying automaticinterpretabledeeplearningpipelineforaccurateparkinsonsdiseasediagnosisusingquantitativesusceptibilitymappingandt1weightedimages AT zhangchunyan automaticinterpretabledeeplearningpipelineforaccurateparkinsonsdiseasediagnosisusingquantitativesusceptibilitymappingandt1weightedimages AT zhangyoumin automaticinterpretabledeeplearningpipelineforaccurateparkinsonsdiseasediagnosisusingquantitativesusceptibilitymappingandt1weightedimages AT wangchenglong automaticinterpretabledeeplearningpipelineforaccurateparkinsonsdiseasediagnosisusingquantitativesusceptibilitymappingandt1weightedimages AT huangpei automaticinterpretabledeeplearningpipelineforaccurateparkinsonsdiseasediagnosisusingquantitativesusceptibilitymappingandt1weightedimages AT jinzhijia automaticinterpretabledeeplearningpipelineforaccurateparkinsonsdiseasediagnosisusingquantitativesusceptibilitymappingandt1weightedimages AT liyan automaticinterpretabledeeplearningpipelineforaccurateparkinsonsdiseasediagnosisusingquantitativesusceptibilitymappingandt1weightedimages AT chengzenghui automaticinterpretabledeeplearningpipelineforaccurateparkinsonsdiseasediagnosisusingquantitativesusceptibilitymappingandt1weightedimages AT liuyu automaticinterpretabledeeplearningpipelineforaccurateparkinsonsdiseasediagnosisusingquantitativesusceptibilitymappingandt1weightedimages AT wangxinhui automaticinterpretabledeeplearningpipelineforaccurateparkinsonsdiseasediagnosisusingquantitativesusceptibilitymappingandt1weightedimages AT chenchen automaticinterpretabledeeplearningpipelineforaccurateparkinsonsdiseasediagnosisusingquantitativesusceptibilitymappingandt1weightedimages AT chengjingliang automaticinterpretabledeeplearningpipelineforaccurateparkinsonsdiseasediagnosisusingquantitativesusceptibilitymappingandt1weightedimages AT liufangtao automaticinterpretabledeeplearningpipelineforaccurateparkinsonsdiseasediagnosisusingquantitativesusceptibilitymappingandt1weightedimages AT haackeewartmark automaticinterpretabledeeplearningpipelineforaccurateparkinsonsdiseasediagnosisusingquantitativesusceptibilitymappingandt1weightedimages AT chenshengdi automaticinterpretabledeeplearningpipelineforaccurateparkinsonsdiseasediagnosisusingquantitativesusceptibilitymappingandt1weightedimages AT yangguang automaticinterpretabledeeplearningpipelineforaccurateparkinsonsdiseasediagnosisusingquantitativesusceptibilitymappingandt1weightedimages AT yanfuhua automaticinterpretabledeeplearningpipelineforaccurateparkinsonsdiseasediagnosisusingquantitativesusceptibilitymappingandt1weightedimages |