Integrating manual annotation with deep transfer learning and radiomics for vertebral fracture analysis
Abstract Background Vertebral compression fractures (VCFs) are prevalent in the elderly, often caused by osteoporosis or trauma. Differentiating acute from chronic VCFs is vital for treatment planning, but MRI, the gold standard, is inaccessible for some. However, CT, a more accessible alternative,...
Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
BMC
2025-02-01
|
| Series: | BMC Medical Imaging |
| Subjects: | |
| Online Access: | https://doi.org/10.1186/s12880-025-01573-9 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1823861513730916352 |
|---|---|
| author | Jing Wang Zhirui Dong Huanxin He Zhiyang Gao Yukai Huang Guangcheng Yuan Libo Jiang Mingdong Zhao |
| author_facet | Jing Wang Zhirui Dong Huanxin He Zhiyang Gao Yukai Huang Guangcheng Yuan Libo Jiang Mingdong Zhao |
| author_sort | Jing Wang |
| collection | DOAJ |
| description | Abstract Background Vertebral compression fractures (VCFs) are prevalent in the elderly, often caused by osteoporosis or trauma. Differentiating acute from chronic VCFs is vital for treatment planning, but MRI, the gold standard, is inaccessible for some. However, CT, a more accessible alternative, lacks precision. This study aimed to enhance CT’s diagnostic accuracy for VCFs using deep transfer learning (DTL) and radiomics. Methods We retrospectively analyzed 218 VCF patients scanned with CT and MRI within 3 days from Oct 2022 to Feb 2024. MRI categorized VCFs. 3D regions of interest (ROIs) from CT scans underwent feature extraction and DTL modeling. Receiver operating characteristic (ROC) analysis evaluated models, with the best fused with radiomic features via LASSO. AUCs compared via Delong test, and clinical utility assessed by decision curve analysis (DCA). Results Patients were split into training (175) and test (43) sets. Traditional radiomics with LR yielded AUCs of 0.973 (training) and 0.869 (test). Optimal DTL modeling improved to 0.992 (training) and 0.941 (test). Feature fusion further boosted AUCs to 1.000 (training) and 0.964 (test). DCA validated its clinical significance. Conclusion The feature fusion model enhances the differential diagnosis of acute and chronic VCFs, outperforming single-model approaches and offering a valuable decision-support tool for patients unable to undergo spinal MRI. |
| format | Article |
| id | doaj-art-e438a50f7bee4788907c6d58cb76d1fb |
| institution | Kabale University |
| issn | 1471-2342 |
| language | English |
| publishDate | 2025-02-01 |
| publisher | BMC |
| record_format | Article |
| series | BMC Medical Imaging |
| spelling | doaj-art-e438a50f7bee4788907c6d58cb76d1fb2025-02-09T12:59:58ZengBMCBMC Medical Imaging1471-23422025-02-0125111210.1186/s12880-025-01573-9Integrating manual annotation with deep transfer learning and radiomics for vertebral fracture analysisJing Wang0Zhirui Dong1Huanxin He2Zhiyang Gao3Yukai Huang4Guangcheng Yuan5Libo Jiang6Mingdong Zhao7Department of Orthopaedic Surgery, Jinshan Hospital, Fudan UniversityDepartment of Orthopaedic Surgery, Jinshan Hospital, Fudan UniversityDepartment of Orthopaedic Surgery, Jinshan Hospital, Fudan UniversityDepartment of Orthopaedic Surgery, Jinshan Hospital, Fudan UniversityDepartment of Orthopaedic Surgery, Jinshan Hospital, Fudan UniversityDepartment of Orthopaedic Surgery, Zhongshan Hospital, Fudan UniversityDepartment of Orthopaedic Surgery, Zhongshan Hospital, Fudan UniversityDepartment of Orthopaedic Surgery, Jinshan Hospital, Fudan UniversityAbstract Background Vertebral compression fractures (VCFs) are prevalent in the elderly, often caused by osteoporosis or trauma. Differentiating acute from chronic VCFs is vital for treatment planning, but MRI, the gold standard, is inaccessible for some. However, CT, a more accessible alternative, lacks precision. This study aimed to enhance CT’s diagnostic accuracy for VCFs using deep transfer learning (DTL) and radiomics. Methods We retrospectively analyzed 218 VCF patients scanned with CT and MRI within 3 days from Oct 2022 to Feb 2024. MRI categorized VCFs. 3D regions of interest (ROIs) from CT scans underwent feature extraction and DTL modeling. Receiver operating characteristic (ROC) analysis evaluated models, with the best fused with radiomic features via LASSO. AUCs compared via Delong test, and clinical utility assessed by decision curve analysis (DCA). Results Patients were split into training (175) and test (43) sets. Traditional radiomics with LR yielded AUCs of 0.973 (training) and 0.869 (test). Optimal DTL modeling improved to 0.992 (training) and 0.941 (test). Feature fusion further boosted AUCs to 1.000 (training) and 0.964 (test). DCA validated its clinical significance. Conclusion The feature fusion model enhances the differential diagnosis of acute and chronic VCFs, outperforming single-model approaches and offering a valuable decision-support tool for patients unable to undergo spinal MRI.https://doi.org/10.1186/s12880-025-01573-9Deep transfer learningVertebral compressionAutomatic segmentationFracturesRadiomics |
| spellingShingle | Jing Wang Zhirui Dong Huanxin He Zhiyang Gao Yukai Huang Guangcheng Yuan Libo Jiang Mingdong Zhao Integrating manual annotation with deep transfer learning and radiomics for vertebral fracture analysis BMC Medical Imaging Deep transfer learning Vertebral compression Automatic segmentation Fractures Radiomics |
| title | Integrating manual annotation with deep transfer learning and radiomics for vertebral fracture analysis |
| title_full | Integrating manual annotation with deep transfer learning and radiomics for vertebral fracture analysis |
| title_fullStr | Integrating manual annotation with deep transfer learning and radiomics for vertebral fracture analysis |
| title_full_unstemmed | Integrating manual annotation with deep transfer learning and radiomics for vertebral fracture analysis |
| title_short | Integrating manual annotation with deep transfer learning and radiomics for vertebral fracture analysis |
| title_sort | integrating manual annotation with deep transfer learning and radiomics for vertebral fracture analysis |
| topic | Deep transfer learning Vertebral compression Automatic segmentation Fractures Radiomics |
| url | https://doi.org/10.1186/s12880-025-01573-9 |
| work_keys_str_mv | AT jingwang integratingmanualannotationwithdeeptransferlearningandradiomicsforvertebralfractureanalysis AT zhiruidong integratingmanualannotationwithdeeptransferlearningandradiomicsforvertebralfractureanalysis AT huanxinhe integratingmanualannotationwithdeeptransferlearningandradiomicsforvertebralfractureanalysis AT zhiyanggao integratingmanualannotationwithdeeptransferlearningandradiomicsforvertebralfractureanalysis AT yukaihuang integratingmanualannotationwithdeeptransferlearningandradiomicsforvertebralfractureanalysis AT guangchengyuan integratingmanualannotationwithdeeptransferlearningandradiomicsforvertebralfractureanalysis AT libojiang integratingmanualannotationwithdeeptransferlearningandradiomicsforvertebralfractureanalysis AT mingdongzhao integratingmanualannotationwithdeeptransferlearningandradiomicsforvertebralfractureanalysis |