Deep adaptive learning predicts and diagnoses CSVD-related cognitive decline using radiomics from T2-FLAIR: a multi-centre study
Abstract Early identification of cerebral small vessel disease related cognitive impairment (CSVD-CI) is crucial for timely clinical intervention. We developed a Transformer-based deep learning model using white matter hyperintensity (WMH) radiomics features from T2-fluid-attenuated inversion recove...
Saved in:
| Main Authors: | , , , , , , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2025-07-01
|
| Series: | npj Digital Medicine |
| Online Access: | https://doi.org/10.1038/s41746-025-01813-w |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1849331572392591360 |
|---|---|
| author | Lili Huang Zhuoyuan Li Xiaolei Zhu Hui Zhao Chenglu Mao Zhihong Ke Yuting Mo Dan Yang Yue Cheng Ruomeng Qin Zheqi Hu Pengfei Shao Ying Chen Min Lou Kelei He Yun Xu |
| author_facet | Lili Huang Zhuoyuan Li Xiaolei Zhu Hui Zhao Chenglu Mao Zhihong Ke Yuting Mo Dan Yang Yue Cheng Ruomeng Qin Zheqi Hu Pengfei Shao Ying Chen Min Lou Kelei He Yun Xu |
| author_sort | Lili Huang |
| collection | DOAJ |
| description | Abstract Early identification of cerebral small vessel disease related cognitive impairment (CSVD-CI) is crucial for timely clinical intervention. We developed a Transformer-based deep learning model using white matter hyperintensity (WMH) radiomics features from T2-fluid-attenuated inversion recovery images to detect CSVD-CI. A total of 783 subjects (161 longitudinally followed) were enrolled from three centres for model development and external validation, using a domain adaptation strategy. The model achieved AUCs of 0.841 (training) and 0.859/0.749 (validation cohorts), outperforming conventional machine learning models. The gradient-weighted class activation mapping approach highlighted WMH textural features, particularly the logarithm-transformed gray level size zone matrix features, as key contributors. These features were significantly correlated with CSVD macro- and microstructural changes, mediated age-cognition relationships and predicted longitudinal cognitive decline. Our findings indicate that WMH radiomics features, reflecting CI-related biological changes in CSVD, combined with a Transformer-based deep learning model, constitute a feasible, automated, and non-invasive tool for CSVD-CI detection. |
| format | Article |
| id | doaj-art-acf1bd5562d94e35a4e9e308941c272b |
| institution | Kabale University |
| issn | 2398-6352 |
| language | English |
| publishDate | 2025-07-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | npj Digital Medicine |
| spelling | doaj-art-acf1bd5562d94e35a4e9e308941c272b2025-08-20T03:46:29ZengNature Portfolionpj Digital Medicine2398-63522025-07-018111310.1038/s41746-025-01813-wDeep adaptive learning predicts and diagnoses CSVD-related cognitive decline using radiomics from T2-FLAIR: a multi-centre studyLili Huang0Zhuoyuan Li1Xiaolei Zhu2Hui Zhao3Chenglu Mao4Zhihong Ke5Yuting Mo6Dan Yang7Yue Cheng8Ruomeng Qin9Zheqi Hu10Pengfei Shao11Ying Chen12Min Lou13Kelei He14Yun Xu15Department of Neurology, Nanjing Drum Tower Hospital, Affiliated Hospital of Medical School, Nanjing UniversityMedical School of Nanjing UniversityDepartment of Neurology, Nanjing Drum Tower Hospital, Affiliated Hospital of Medical School, Nanjing UniversityDepartment of Neurology, Nanjing Drum Tower Hospital, Affiliated Hospital of Medical School, Nanjing UniversityDepartment of Neurology, Nanjing Drum Tower Hospital, Affiliated Hospital of Medical School, Nanjing UniversityDepartment of Neurology, Nanjing Drum Tower Hospital, Affiliated Hospital of Medical School, Nanjing UniversityDepartment of Neurology, Nanjing Drum Tower Hospital, Affiliated Hospital of Medical School, Nanjing UniversityDepartment of Neurology, Nanjing Drum Tower Hospital, Affiliated Hospital of Medical School, Nanjing UniversityDepartment of Neurology, Nanjing Drum Tower Hospital, Affiliated Hospital of Medical School, Nanjing UniversityDepartment of Neurology, Nanjing Drum Tower Hospital, Affiliated Hospital of Medical School, Nanjing UniversityDepartment of Neurology, Nanjing Drum Tower Hospital, Affiliated Hospital of Medical School, Nanjing UniversityDepartment of Neurology, Nanjing Drum Tower Hospital, Affiliated Hospital of Medical School, Nanjing UniversityDepartment of Neurology, Nanjing Drum Tower Hospital, Affiliated Hospital of Medical School, Nanjing UniversityDepartment of Neurology, the Second Affiliated Hospital of Zhejiang University, School of MedicineJiangsu Key Laboratory for Molecular Medicine, Medical School of Nanjing UniversityDepartment of Neurology, Nanjing Drum Tower Hospital, Affiliated Hospital of Medical School, Nanjing UniversityAbstract Early identification of cerebral small vessel disease related cognitive impairment (CSVD-CI) is crucial for timely clinical intervention. We developed a Transformer-based deep learning model using white matter hyperintensity (WMH) radiomics features from T2-fluid-attenuated inversion recovery images to detect CSVD-CI. A total of 783 subjects (161 longitudinally followed) were enrolled from three centres for model development and external validation, using a domain adaptation strategy. The model achieved AUCs of 0.841 (training) and 0.859/0.749 (validation cohorts), outperforming conventional machine learning models. The gradient-weighted class activation mapping approach highlighted WMH textural features, particularly the logarithm-transformed gray level size zone matrix features, as key contributors. These features were significantly correlated with CSVD macro- and microstructural changes, mediated age-cognition relationships and predicted longitudinal cognitive decline. Our findings indicate that WMH radiomics features, reflecting CI-related biological changes in CSVD, combined with a Transformer-based deep learning model, constitute a feasible, automated, and non-invasive tool for CSVD-CI detection.https://doi.org/10.1038/s41746-025-01813-w |
| spellingShingle | Lili Huang Zhuoyuan Li Xiaolei Zhu Hui Zhao Chenglu Mao Zhihong Ke Yuting Mo Dan Yang Yue Cheng Ruomeng Qin Zheqi Hu Pengfei Shao Ying Chen Min Lou Kelei He Yun Xu Deep adaptive learning predicts and diagnoses CSVD-related cognitive decline using radiomics from T2-FLAIR: a multi-centre study npj Digital Medicine |
| title | Deep adaptive learning predicts and diagnoses CSVD-related cognitive decline using radiomics from T2-FLAIR: a multi-centre study |
| title_full | Deep adaptive learning predicts and diagnoses CSVD-related cognitive decline using radiomics from T2-FLAIR: a multi-centre study |
| title_fullStr | Deep adaptive learning predicts and diagnoses CSVD-related cognitive decline using radiomics from T2-FLAIR: a multi-centre study |
| title_full_unstemmed | Deep adaptive learning predicts and diagnoses CSVD-related cognitive decline using radiomics from T2-FLAIR: a multi-centre study |
| title_short | Deep adaptive learning predicts and diagnoses CSVD-related cognitive decline using radiomics from T2-FLAIR: a multi-centre study |
| title_sort | deep adaptive learning predicts and diagnoses csvd related cognitive decline using radiomics from t2 flair a multi centre study |
| url | https://doi.org/10.1038/s41746-025-01813-w |
| work_keys_str_mv | AT lilihuang deepadaptivelearningpredictsanddiagnosescsvdrelatedcognitivedeclineusingradiomicsfromt2flairamulticentrestudy AT zhuoyuanli deepadaptivelearningpredictsanddiagnosescsvdrelatedcognitivedeclineusingradiomicsfromt2flairamulticentrestudy AT xiaoleizhu deepadaptivelearningpredictsanddiagnosescsvdrelatedcognitivedeclineusingradiomicsfromt2flairamulticentrestudy AT huizhao deepadaptivelearningpredictsanddiagnosescsvdrelatedcognitivedeclineusingradiomicsfromt2flairamulticentrestudy AT chenglumao deepadaptivelearningpredictsanddiagnosescsvdrelatedcognitivedeclineusingradiomicsfromt2flairamulticentrestudy AT zhihongke deepadaptivelearningpredictsanddiagnosescsvdrelatedcognitivedeclineusingradiomicsfromt2flairamulticentrestudy AT yutingmo deepadaptivelearningpredictsanddiagnosescsvdrelatedcognitivedeclineusingradiomicsfromt2flairamulticentrestudy AT danyang deepadaptivelearningpredictsanddiagnosescsvdrelatedcognitivedeclineusingradiomicsfromt2flairamulticentrestudy AT yuecheng deepadaptivelearningpredictsanddiagnosescsvdrelatedcognitivedeclineusingradiomicsfromt2flairamulticentrestudy AT ruomengqin deepadaptivelearningpredictsanddiagnosescsvdrelatedcognitivedeclineusingradiomicsfromt2flairamulticentrestudy AT zheqihu deepadaptivelearningpredictsanddiagnosescsvdrelatedcognitivedeclineusingradiomicsfromt2flairamulticentrestudy AT pengfeishao deepadaptivelearningpredictsanddiagnosescsvdrelatedcognitivedeclineusingradiomicsfromt2flairamulticentrestudy AT yingchen deepadaptivelearningpredictsanddiagnosescsvdrelatedcognitivedeclineusingradiomicsfromt2flairamulticentrestudy AT minlou deepadaptivelearningpredictsanddiagnosescsvdrelatedcognitivedeclineusingradiomicsfromt2flairamulticentrestudy AT keleihe deepadaptivelearningpredictsanddiagnosescsvdrelatedcognitivedeclineusingradiomicsfromt2flairamulticentrestudy AT yunxu deepadaptivelearningpredictsanddiagnosescsvdrelatedcognitivedeclineusingradiomicsfromt2flairamulticentrestudy |