A generative whole-brain segmentation model for positron emission tomography images
Abstract Purpose Whole-brain segmentation via positron emission tomography (PET) imaging is crucial for advancing neuroscience research and clinical medicine, providing essential insights into biological metabolism and activity within different brain regions. However, the low resolution of PET image...
Saved in:
| Main Authors: | , , , , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
SpringerOpen
2025-02-01
|
| Series: | EJNMMI Physics |
| Subjects: | |
| Online Access: | https://doi.org/10.1186/s40658-025-00716-9 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1823861579944296448 |
|---|---|
| author | Wenbo Li Zhenxing Huang Hongyan Tang Yaping Wu Yunlong Gao Jing Qin Jianmin Yuan Yang Yang Yan Zhang Na Zhang Hairong Zheng Dong Liang Meiyun Wang Zhanli Hu |
| author_facet | Wenbo Li Zhenxing Huang Hongyan Tang Yaping Wu Yunlong Gao Jing Qin Jianmin Yuan Yang Yang Yan Zhang Na Zhang Hairong Zheng Dong Liang Meiyun Wang Zhanli Hu |
| author_sort | Wenbo Li |
| collection | DOAJ |
| description | Abstract Purpose Whole-brain segmentation via positron emission tomography (PET) imaging is crucial for advancing neuroscience research and clinical medicine, providing essential insights into biological metabolism and activity within different brain regions. However, the low resolution of PET images may have limited the segmentation accuracy of multiple brain structures. Therefore, we propose a generative multi-object segmentation model for brain PET images to achieve automatic and accurate segmentation. Methods In this study, we propose a generative multi-object segmentation model for brain PET images with two learning protocols. First, we pretrained a latent mapping model to learn the mapping relationship between PET and MR images so that we could extract anatomical information of the brain. A 3D multi-object segmentation model was subsequently proposed to apply whole-brain segmentation to MR images generated from integrated latent mapping models. Moreover, a custom cross-attention module based on a cross-attention mechanism was constructed to effectively fuse the functional information and structural information. The proposed method was compared with various deep learning-based approaches in terms of the Dice similarity coefficient, Jaccard index, precision, and recall serving as evaluation metrics. Results Experiments were conducted on real brain PET/MR images from 120 patients. Both visual and quantitative results indicate that our method outperforms the other comparison approaches, achieving 75.53% ± 4.26% Dice, 66.02% ± 4.55% Jaccard, 74.64% ± 4.15% recall and 81.40% ± 2.30% precision. Furthermore, the evaluation of the SUV distribution and correlation assessment in the regions of interest demonstrated consistency with the ground truth. Additionally, clinical tolerance rates, which are determined by the tumor background ratio, have confirmed the ability of the method to distinguish highly metabolic regions accurately from normal regions, reinforcing its clinical applicability. Conclusion For automatic and accurate whole-brain segmentation, we propose a novel 3D generative multi-object segmentation model for brain PET images, which achieves superior model performance compared with other deep learning methods. In the future, we will apply our whole-brain segmentation method to clinical practice and extend it to other multimodal tasks. |
| format | Article |
| id | doaj-art-9dd74093ffd24aa4ae26d01a4c12b268 |
| institution | Kabale University |
| issn | 2197-7364 |
| language | English |
| publishDate | 2025-02-01 |
| publisher | SpringerOpen |
| record_format | Article |
| series | EJNMMI Physics |
| spelling | doaj-art-9dd74093ffd24aa4ae26d01a4c12b2682025-02-09T12:54:48ZengSpringerOpenEJNMMI Physics2197-73642025-02-0112111610.1186/s40658-025-00716-9A generative whole-brain segmentation model for positron emission tomography imagesWenbo Li0Zhenxing Huang1Hongyan Tang2Yaping Wu3Yunlong Gao4Jing Qin5Jianmin Yuan6Yang Yang7Yan Zhang8Na Zhang9Hairong Zheng10Dong Liang11Meiyun Wang12Zhanli Hu13Research Center for Medical AI, Shenzhen Institute of Advanced Technology, Chinese Academy of SciencesResearch Center for Medical AI, Shenzhen Institute of Advanced Technology, Chinese Academy of SciencesResearch Center for Medical AI, Shenzhen Institute of Advanced Technology, Chinese Academy of SciencesDepartment of Medical Imaging, Henan Provincial People’s Hospital & People’s Hospital of Zhengzhou UniversityResearch Center for Medical AI, Shenzhen Institute of Advanced Technology, Chinese Academy of SciencesCentre for Smart Health, School of Nursing, The Hong Kong Polytechnic UniversityCentral Research Institute, United Imaging Healthcare GroupBeijing United Imaging Research Institute of Intelligent ImagingBeijing United Imaging Research Institute of Intelligent ImagingResearch Center for Medical AI, Shenzhen Institute of Advanced Technology, Chinese Academy of SciencesResearch Center for Medical AI, Shenzhen Institute of Advanced Technology, Chinese Academy of SciencesResearch Center for Medical AI, Shenzhen Institute of Advanced Technology, Chinese Academy of SciencesDepartment of Medical Imaging, Henan Provincial People’s Hospital & People’s Hospital of Zhengzhou UniversityResearch Center for Medical AI, Shenzhen Institute of Advanced Technology, Chinese Academy of SciencesAbstract Purpose Whole-brain segmentation via positron emission tomography (PET) imaging is crucial for advancing neuroscience research and clinical medicine, providing essential insights into biological metabolism and activity within different brain regions. However, the low resolution of PET images may have limited the segmentation accuracy of multiple brain structures. Therefore, we propose a generative multi-object segmentation model for brain PET images to achieve automatic and accurate segmentation. Methods In this study, we propose a generative multi-object segmentation model for brain PET images with two learning protocols. First, we pretrained a latent mapping model to learn the mapping relationship between PET and MR images so that we could extract anatomical information of the brain. A 3D multi-object segmentation model was subsequently proposed to apply whole-brain segmentation to MR images generated from integrated latent mapping models. Moreover, a custom cross-attention module based on a cross-attention mechanism was constructed to effectively fuse the functional information and structural information. The proposed method was compared with various deep learning-based approaches in terms of the Dice similarity coefficient, Jaccard index, precision, and recall serving as evaluation metrics. Results Experiments were conducted on real brain PET/MR images from 120 patients. Both visual and quantitative results indicate that our method outperforms the other comparison approaches, achieving 75.53% ± 4.26% Dice, 66.02% ± 4.55% Jaccard, 74.64% ± 4.15% recall and 81.40% ± 2.30% precision. Furthermore, the evaluation of the SUV distribution and correlation assessment in the regions of interest demonstrated consistency with the ground truth. Additionally, clinical tolerance rates, which are determined by the tumor background ratio, have confirmed the ability of the method to distinguish highly metabolic regions accurately from normal regions, reinforcing its clinical applicability. Conclusion For automatic and accurate whole-brain segmentation, we propose a novel 3D generative multi-object segmentation model for brain PET images, which achieves superior model performance compared with other deep learning methods. In the future, we will apply our whole-brain segmentation method to clinical practice and extend it to other multimodal tasks.https://doi.org/10.1186/s40658-025-00716-9Positron emission tomographyBrainGenerative medical segmentationCross-attention |
| spellingShingle | Wenbo Li Zhenxing Huang Hongyan Tang Yaping Wu Yunlong Gao Jing Qin Jianmin Yuan Yang Yang Yan Zhang Na Zhang Hairong Zheng Dong Liang Meiyun Wang Zhanli Hu A generative whole-brain segmentation model for positron emission tomography images EJNMMI Physics Positron emission tomography Brain Generative medical segmentation Cross-attention |
| title | A generative whole-brain segmentation model for positron emission tomography images |
| title_full | A generative whole-brain segmentation model for positron emission tomography images |
| title_fullStr | A generative whole-brain segmentation model for positron emission tomography images |
| title_full_unstemmed | A generative whole-brain segmentation model for positron emission tomography images |
| title_short | A generative whole-brain segmentation model for positron emission tomography images |
| title_sort | generative whole brain segmentation model for positron emission tomography images |
| topic | Positron emission tomography Brain Generative medical segmentation Cross-attention |
| url | https://doi.org/10.1186/s40658-025-00716-9 |
| work_keys_str_mv | AT wenboli agenerativewholebrainsegmentationmodelforpositronemissiontomographyimages AT zhenxinghuang agenerativewholebrainsegmentationmodelforpositronemissiontomographyimages AT hongyantang agenerativewholebrainsegmentationmodelforpositronemissiontomographyimages AT yapingwu agenerativewholebrainsegmentationmodelforpositronemissiontomographyimages AT yunlonggao agenerativewholebrainsegmentationmodelforpositronemissiontomographyimages AT jingqin agenerativewholebrainsegmentationmodelforpositronemissiontomographyimages AT jianminyuan agenerativewholebrainsegmentationmodelforpositronemissiontomographyimages AT yangyang agenerativewholebrainsegmentationmodelforpositronemissiontomographyimages AT yanzhang agenerativewholebrainsegmentationmodelforpositronemissiontomographyimages AT nazhang agenerativewholebrainsegmentationmodelforpositronemissiontomographyimages AT hairongzheng agenerativewholebrainsegmentationmodelforpositronemissiontomographyimages AT dongliang agenerativewholebrainsegmentationmodelforpositronemissiontomographyimages AT meiyunwang agenerativewholebrainsegmentationmodelforpositronemissiontomographyimages AT zhanlihu agenerativewholebrainsegmentationmodelforpositronemissiontomographyimages AT wenboli generativewholebrainsegmentationmodelforpositronemissiontomographyimages AT zhenxinghuang generativewholebrainsegmentationmodelforpositronemissiontomographyimages AT hongyantang generativewholebrainsegmentationmodelforpositronemissiontomographyimages AT yapingwu generativewholebrainsegmentationmodelforpositronemissiontomographyimages AT yunlonggao generativewholebrainsegmentationmodelforpositronemissiontomographyimages AT jingqin generativewholebrainsegmentationmodelforpositronemissiontomographyimages AT jianminyuan generativewholebrainsegmentationmodelforpositronemissiontomographyimages AT yangyang generativewholebrainsegmentationmodelforpositronemissiontomographyimages AT yanzhang generativewholebrainsegmentationmodelforpositronemissiontomographyimages AT nazhang generativewholebrainsegmentationmodelforpositronemissiontomographyimages AT hairongzheng generativewholebrainsegmentationmodelforpositronemissiontomographyimages AT dongliang generativewholebrainsegmentationmodelforpositronemissiontomographyimages AT meiyunwang generativewholebrainsegmentationmodelforpositronemissiontomographyimages AT zhanlihu generativewholebrainsegmentationmodelforpositronemissiontomographyimages |