Deep learning-based thigh muscle segmentation for reproducible fat fraction quantification using fat–water decomposition MRI
Abstract Background Time-efficient and accurate whole volume thigh muscle segmentation is a major challenge in moving from qualitative assessment of thigh muscle MRI to more quantitative methods. This study developed an automated whole thigh muscle segmentation method using deep learning for reprodu...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
SpringerOpen
2020-11-01
|
| Series: | Insights into Imaging |
| Subjects: | |
| Online Access: | https://doi.org/10.1186/s13244-020-00946-8 |
| _version_ | 1818854796923240448 |
|---|---|
| author | Jie Ding Peng Cao Hing-Chiu Chang Yuan Gao Sophelia Hoi Shan Chan Varut Vardhanabhuti |
| author_facet | Jie Ding Peng Cao Hing-Chiu Chang Yuan Gao Sophelia Hoi Shan Chan Varut Vardhanabhuti |
| author_sort | Jie Ding |
| collection | DOAJ |
| description | Abstract Background Time-efficient and accurate whole volume thigh muscle segmentation is a major challenge in moving from qualitative assessment of thigh muscle MRI to more quantitative methods. This study developed an automated whole thigh muscle segmentation method using deep learning for reproducible fat fraction quantification on fat–water decomposition MRI. Results This study was performed using a public reference database (Dataset 1, 25 scans) and a local clinical dataset (Dataset 2, 21 scans). A U-net was trained using 23 scans (16 from Dataset 1, seven from Dataset 2) to automatically segment four functional muscle groups: quadriceps femoris, sartorius, gracilis and hamstring. The segmentation accuracy was evaluated on an independent testing set (3 × 3 repeated scans in Dataset 1 and four scans in Dataset 2). The average Dice coefficients between manual and automated segmentation were > 0.85. The average percent difference (absolute) in volume was 7.57%, and the average difference (absolute) in mean fat fraction (meanFF) was 0.17%. The reproducibility in meanFF was calculated using intraclass correlation coefficients (ICCs) for the repeated scans, and automated segmentation produced overall higher ICCs than manual segmentation (0.921 vs. 0.902). A preliminary quantitative analysis was performed using two-sample t test to detect possible differences in meanFF between 14 normal and 14 abnormal (with fat infiltration) thighs in Dataset 2 using automated segmentation, and significantly higher meanFF was detected in abnormal thighs. Conclusions This automated thigh muscle segmentation exhibits excellent accuracy and higher reproducibility in fat fraction estimation compared to manual segmentation, which can be further used for quantifying fat infiltration in thigh muscles. |
| first_indexed | 2024-12-19T07:58:25Z |
| format | Article |
| id | doaj.art-f36f086572894eb49caec55a6bca659e |
| institution | Directory Open Access Journal |
| issn | 1869-4101 |
| language | English |
| last_indexed | 2024-12-19T07:58:25Z |
| publishDate | 2020-11-01 |
| publisher | SpringerOpen |
| record_format | Article |
| series | Insights into Imaging |
| spelling | doaj.art-f36f086572894eb49caec55a6bca659e2022-12-21T20:29:56ZengSpringerOpenInsights into Imaging1869-41012020-11-0111111110.1186/s13244-020-00946-8Deep learning-based thigh muscle segmentation for reproducible fat fraction quantification using fat–water decomposition MRIJie Ding0Peng Cao1Hing-Chiu Chang2Yuan Gao3Sophelia Hoi Shan Chan4Varut Vardhanabhuti5Department of Diagnostic Radiology, Li Ka Shing Faculty of Medicine, The University of Hong KongDepartment of Diagnostic Radiology, Li Ka Shing Faculty of Medicine, The University of Hong KongDepartment of Diagnostic Radiology, Li Ka Shing Faculty of Medicine, The University of Hong KongDivision of Neurology, Department of Medicine, Queen Mary Hospital, The University of Hong KongDivision of Paediatric Neurology, Department of Paediatrics and Adolescent Medicine, Li Ka Shing Faculty of Medicine, The University of Hong KongDepartment of Diagnostic Radiology, Li Ka Shing Faculty of Medicine, The University of Hong KongAbstract Background Time-efficient and accurate whole volume thigh muscle segmentation is a major challenge in moving from qualitative assessment of thigh muscle MRI to more quantitative methods. This study developed an automated whole thigh muscle segmentation method using deep learning for reproducible fat fraction quantification on fat–water decomposition MRI. Results This study was performed using a public reference database (Dataset 1, 25 scans) and a local clinical dataset (Dataset 2, 21 scans). A U-net was trained using 23 scans (16 from Dataset 1, seven from Dataset 2) to automatically segment four functional muscle groups: quadriceps femoris, sartorius, gracilis and hamstring. The segmentation accuracy was evaluated on an independent testing set (3 × 3 repeated scans in Dataset 1 and four scans in Dataset 2). The average Dice coefficients between manual and automated segmentation were > 0.85. The average percent difference (absolute) in volume was 7.57%, and the average difference (absolute) in mean fat fraction (meanFF) was 0.17%. The reproducibility in meanFF was calculated using intraclass correlation coefficients (ICCs) for the repeated scans, and automated segmentation produced overall higher ICCs than manual segmentation (0.921 vs. 0.902). A preliminary quantitative analysis was performed using two-sample t test to detect possible differences in meanFF between 14 normal and 14 abnormal (with fat infiltration) thighs in Dataset 2 using automated segmentation, and significantly higher meanFF was detected in abnormal thighs. Conclusions This automated thigh muscle segmentation exhibits excellent accuracy and higher reproducibility in fat fraction estimation compared to manual segmentation, which can be further used for quantifying fat infiltration in thigh muscles.https://doi.org/10.1186/s13244-020-00946-8Thigh muscle segmentationDeep learningFat–water decomposition MRIQuantitative MRI analysis |
| spellingShingle | Jie Ding Peng Cao Hing-Chiu Chang Yuan Gao Sophelia Hoi Shan Chan Varut Vardhanabhuti Deep learning-based thigh muscle segmentation for reproducible fat fraction quantification using fat–water decomposition MRI Insights into Imaging Thigh muscle segmentation Deep learning Fat–water decomposition MRI Quantitative MRI analysis |
| title | Deep learning-based thigh muscle segmentation for reproducible fat fraction quantification using fat–water decomposition MRI |
| title_full | Deep learning-based thigh muscle segmentation for reproducible fat fraction quantification using fat–water decomposition MRI |
| title_fullStr | Deep learning-based thigh muscle segmentation for reproducible fat fraction quantification using fat–water decomposition MRI |
| title_full_unstemmed | Deep learning-based thigh muscle segmentation for reproducible fat fraction quantification using fat–water decomposition MRI |
| title_short | Deep learning-based thigh muscle segmentation for reproducible fat fraction quantification using fat–water decomposition MRI |
| title_sort | deep learning based thigh muscle segmentation for reproducible fat fraction quantification using fat water decomposition mri |
| topic | Thigh muscle segmentation Deep learning Fat–water decomposition MRI Quantitative MRI analysis |
| url | https://doi.org/10.1186/s13244-020-00946-8 |
| work_keys_str_mv | AT jieding deeplearningbasedthighmusclesegmentationforreproduciblefatfractionquantificationusingfatwaterdecompositionmri AT pengcao deeplearningbasedthighmusclesegmentationforreproduciblefatfractionquantificationusingfatwaterdecompositionmri AT hingchiuchang deeplearningbasedthighmusclesegmentationforreproduciblefatfractionquantificationusingfatwaterdecompositionmri AT yuangao deeplearningbasedthighmusclesegmentationforreproduciblefatfractionquantificationusingfatwaterdecompositionmri AT sopheliahoishanchan deeplearningbasedthighmusclesegmentationforreproduciblefatfractionquantificationusingfatwaterdecompositionmri AT varutvardhanabhuti deeplearningbasedthighmusclesegmentationforreproduciblefatfractionquantificationusingfatwaterdecompositionmri |