Deep learning-based identification and localization of intracranial hemorrhage in patients using a large annotated head computed tomography dataset: A retrospective multicenter study
Background: Accurately identifying and localizing the five subtypes of intracranial hemorrhage (ICH) are crucial steps for subsequent clinical treatment; however, the lack of a large computed tomography (CT) dataset with annotations of the categorization and localization of ICH considerably limits t...
Saved in:
| Main Authors: | , , , , , , , , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-02-01
|
| Series: | Intelligent Medicine |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2667102624000895 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1850026420124778496 |
|---|---|
| author | Jingjing Liu Weijie Fan Yi Yang Qi Peng Bingjun Ji Luxing He Yang Li Jing Yuan Wei Li Xianqi Wang Yi Wu Chen Liu Qingfang Gong Mi He Yeqin Fu Dong Zhang Si Zhang Yongjian Nian |
| author_facet | Jingjing Liu Weijie Fan Yi Yang Qi Peng Bingjun Ji Luxing He Yang Li Jing Yuan Wei Li Xianqi Wang Yi Wu Chen Liu Qingfang Gong Mi He Yeqin Fu Dong Zhang Si Zhang Yongjian Nian |
| author_sort | Jingjing Liu |
| collection | DOAJ |
| description | Background: Accurately identifying and localizing the five subtypes of intracranial hemorrhage (ICH) are crucial steps for subsequent clinical treatment; however, the lack of a large computed tomography (CT) dataset with annotations of the categorization and localization of ICH considerably limits the development of deep learning-based identification and localization methods. We aimed to construct this large dataset and develop a deep learning-based model to identify and localize the five ICH subtypes, including intraventricular hemorrhage (IVH), intraparenchymal hemorrhage (IPH), subdural hemorrhage (SDH), subarachnoid hemorrhage (SAH), and epidural hemorrhage (EDH), in non-contrast head CT scans. Methods: Based on the public Radiological Society of North America (RSNA) 2019 dataset, we constructed a large CT dataset named RSNA 2019+ that was annotated for bleeding localization of the five ICH subtypes by three radiologists. An improved YOLOv8 architecture with the bidirectional feature pyramid network was proposed and trained using the RSNA 2019+ training dataset and evaluated on the RSNA 2019+ test dataset. The public CQ500, and two private datasets collected from the Xinqiao and Sunshine Union Hospitals, respectively, were also annotated to perform multicenter validation. Furthermore, the performance of the deep learning model was compared with that of four radiologists. Multiple performance metrics, including the average precision (AP), precision, recall and F1-score, were used for performance evaluation. The McNemar and chi-squared tests were performed, and the 95% Wilson confidence intervals were given for the precision and recall. Results: There were 175,125; 4,707; 8,259; and 3,104 bounding boxes after annotation on the RSNA 2019+; CQ500+; and the PD 1 and PD 2 datasets, respectively. With an intersection-over-union threshold of 0.5, the APs of IVH, IPH, SAH, SDH and EDH are 0.852, 0.820, 0.574, 0.639, and 0.558, respectively, yielding a mean average precision (mAP) of 0.688 for our proposed deep learning model on the RSNA 2019+ test dataset. For the multicenter validation involving the three external datasets, the mAPs for CQ500, PD1, and PD2 were 0.594, 0.734, and 0.66, respectively, which is comparable to those of radiologist with eight years of experience in head CT interpretation. Conclusion: The deep learning model developed from the constructed RSNA 2019+ dataset exhibited good potential in identifying and localizing the five ICH subtypes in CT slices and has the potential to assist in the clinical diagnosis. |
| format | Article |
| id | doaj-art-ac13984e104c48988c20c43a0bc4d456 |
| institution | DOAJ |
| issn | 2667-1026 |
| language | English |
| publishDate | 2025-02-01 |
| publisher | Elsevier |
| record_format | Article |
| series | Intelligent Medicine |
| spelling | doaj-art-ac13984e104c48988c20c43a0bc4d4562025-08-20T03:00:32ZengElsevierIntelligent Medicine2667-10262025-02-0151142210.1016/j.imed.2024.11.002Deep learning-based identification and localization of intracranial hemorrhage in patients using a large annotated head computed tomography dataset: A retrospective multicenter studyJingjing Liu0Weijie Fan1Yi Yang2Qi Peng3Bingjun Ji4Luxing He5Yang Li6Jing Yuan7Wei Li8Xianqi Wang9Yi Wu10Chen Liu11Qingfang Gong12Mi He13Yeqin Fu14Dong Zhang15Si Zhang16Yongjian Nian17Department of Digital Medicine, School of Biomedical Engineering and Imaging Medicine, Army Medical University, Chongqing 400038, ChinaDepartment of Radiology, Second Affiliated Hospital, Army Medical University, Chongqing 400037, ChinaDepartment of Digital Medicine, School of Biomedical Engineering and Imaging Medicine, Army Medical University, Chongqing 400038, ChinaDepartment of Digital Medicine, School of Biomedical Engineering and Imaging Medicine, Army Medical University, Chongqing 400038, ChinaRadiotherapy Center, Sunshine Union Hospital, Weifang, Shandong 261061, ChinaDepartment of Digital Medicine, School of Biomedical Engineering and Imaging Medicine, Army Medical University, Chongqing 400038, ChinaDepartment of Radiology, Air Force Hospital of Western Theater Command, Chengdu, Sichuan 610011, ChinaDepartment of Radiology, Army Medical Center, Army Medical University, Chongqing 400037, ChinaDepartment of Radiology, Second Affiliated Hospital, Army Medical University, Chongqing 400037, ChinaDepartment of Radiology, First Affiliated Hospital, Army Medical University, Chongqing 400037, ChinaDepartment of Digital Medicine, School of Biomedical Engineering and Imaging Medicine, Army Medical University, Chongqing 400038, ChinaDepartment of Radiology, First Affiliated Hospital, Army Medical University, Chongqing 400037, ChinaDepartment of Digital Medicine, School of Biomedical Engineering and Imaging Medicine, Army Medical University, Chongqing 400038, ChinaDepartment of Medical Instruments and Metrology, School of Biomedical Engineering and Imaging Medicine, Army Medical University, Chongqing 400038, ChinaDepartment of Digital Medicine, School of Biomedical Engineering and Imaging Medicine, Army Medical University, Chongqing 400038, ChinaDepartment of Radiology, Second Affiliated Hospital, Army Medical University, Chongqing 400037, China; Corresponding authors: Yongjian Nian, Department of Digital Medicine, School of Biomedical Engineering and Imaging Medicine, Army Medical University, Chongqing 400038, China; Si Zhang, Department of Radiology, Second Affiliated Hospital, Army Medical University, Chongqing 400037, China; Dong Zhang, Department of Radiology, Second Affiliated Hospital, Army Medical University, Chongqing 400037, China.Department of Radiology, Second Affiliated Hospital, Army Medical University, Chongqing 400037, China; Corresponding authors: Yongjian Nian, Department of Digital Medicine, School of Biomedical Engineering and Imaging Medicine, Army Medical University, Chongqing 400038, China; Si Zhang, Department of Radiology, Second Affiliated Hospital, Army Medical University, Chongqing 400037, China; Dong Zhang, Department of Radiology, Second Affiliated Hospital, Army Medical University, Chongqing 400037, China.Department of Digital Medicine, School of Biomedical Engineering and Imaging Medicine, Army Medical University, Chongqing 400038, China; Corresponding authors: Yongjian Nian, Department of Digital Medicine, School of Biomedical Engineering and Imaging Medicine, Army Medical University, Chongqing 400038, China; Si Zhang, Department of Radiology, Second Affiliated Hospital, Army Medical University, Chongqing 400037, China; Dong Zhang, Department of Radiology, Second Affiliated Hospital, Army Medical University, Chongqing 400037, China.Background: Accurately identifying and localizing the five subtypes of intracranial hemorrhage (ICH) are crucial steps for subsequent clinical treatment; however, the lack of a large computed tomography (CT) dataset with annotations of the categorization and localization of ICH considerably limits the development of deep learning-based identification and localization methods. We aimed to construct this large dataset and develop a deep learning-based model to identify and localize the five ICH subtypes, including intraventricular hemorrhage (IVH), intraparenchymal hemorrhage (IPH), subdural hemorrhage (SDH), subarachnoid hemorrhage (SAH), and epidural hemorrhage (EDH), in non-contrast head CT scans. Methods: Based on the public Radiological Society of North America (RSNA) 2019 dataset, we constructed a large CT dataset named RSNA 2019+ that was annotated for bleeding localization of the five ICH subtypes by three radiologists. An improved YOLOv8 architecture with the bidirectional feature pyramid network was proposed and trained using the RSNA 2019+ training dataset and evaluated on the RSNA 2019+ test dataset. The public CQ500, and two private datasets collected from the Xinqiao and Sunshine Union Hospitals, respectively, were also annotated to perform multicenter validation. Furthermore, the performance of the deep learning model was compared with that of four radiologists. Multiple performance metrics, including the average precision (AP), precision, recall and F1-score, were used for performance evaluation. The McNemar and chi-squared tests were performed, and the 95% Wilson confidence intervals were given for the precision and recall. Results: There were 175,125; 4,707; 8,259; and 3,104 bounding boxes after annotation on the RSNA 2019+; CQ500+; and the PD 1 and PD 2 datasets, respectively. With an intersection-over-union threshold of 0.5, the APs of IVH, IPH, SAH, SDH and EDH are 0.852, 0.820, 0.574, 0.639, and 0.558, respectively, yielding a mean average precision (mAP) of 0.688 for our proposed deep learning model on the RSNA 2019+ test dataset. For the multicenter validation involving the three external datasets, the mAPs for CQ500, PD1, and PD2 were 0.594, 0.734, and 0.66, respectively, which is comparable to those of radiologist with eight years of experience in head CT interpretation. Conclusion: The deep learning model developed from the constructed RSNA 2019+ dataset exhibited good potential in identifying and localizing the five ICH subtypes in CT slices and has the potential to assist in the clinical diagnosis.http://www.sciencedirect.com/science/article/pii/S2667102624000895Intracranial hemorrhageIdentification and localizationDeep learning modelBounding box |
| spellingShingle | Jingjing Liu Weijie Fan Yi Yang Qi Peng Bingjun Ji Luxing He Yang Li Jing Yuan Wei Li Xianqi Wang Yi Wu Chen Liu Qingfang Gong Mi He Yeqin Fu Dong Zhang Si Zhang Yongjian Nian Deep learning-based identification and localization of intracranial hemorrhage in patients using a large annotated head computed tomography dataset: A retrospective multicenter study Intelligent Medicine Intracranial hemorrhage Identification and localization Deep learning model Bounding box |
| title | Deep learning-based identification and localization of intracranial hemorrhage in patients using a large annotated head computed tomography dataset: A retrospective multicenter study |
| title_full | Deep learning-based identification and localization of intracranial hemorrhage in patients using a large annotated head computed tomography dataset: A retrospective multicenter study |
| title_fullStr | Deep learning-based identification and localization of intracranial hemorrhage in patients using a large annotated head computed tomography dataset: A retrospective multicenter study |
| title_full_unstemmed | Deep learning-based identification and localization of intracranial hemorrhage in patients using a large annotated head computed tomography dataset: A retrospective multicenter study |
| title_short | Deep learning-based identification and localization of intracranial hemorrhage in patients using a large annotated head computed tomography dataset: A retrospective multicenter study |
| title_sort | deep learning based identification and localization of intracranial hemorrhage in patients using a large annotated head computed tomography dataset a retrospective multicenter study |
| topic | Intracranial hemorrhage Identification and localization Deep learning model Bounding box |
| url | http://www.sciencedirect.com/science/article/pii/S2667102624000895 |
| work_keys_str_mv | AT jingjingliu deeplearningbasedidentificationandlocalizationofintracranialhemorrhageinpatientsusingalargeannotatedheadcomputedtomographydatasetaretrospectivemulticenterstudy AT weijiefan deeplearningbasedidentificationandlocalizationofintracranialhemorrhageinpatientsusingalargeannotatedheadcomputedtomographydatasetaretrospectivemulticenterstudy AT yiyang deeplearningbasedidentificationandlocalizationofintracranialhemorrhageinpatientsusingalargeannotatedheadcomputedtomographydatasetaretrospectivemulticenterstudy AT qipeng deeplearningbasedidentificationandlocalizationofintracranialhemorrhageinpatientsusingalargeannotatedheadcomputedtomographydatasetaretrospectivemulticenterstudy AT bingjunji deeplearningbasedidentificationandlocalizationofintracranialhemorrhageinpatientsusingalargeannotatedheadcomputedtomographydatasetaretrospectivemulticenterstudy AT luxinghe deeplearningbasedidentificationandlocalizationofintracranialhemorrhageinpatientsusingalargeannotatedheadcomputedtomographydatasetaretrospectivemulticenterstudy AT yangli deeplearningbasedidentificationandlocalizationofintracranialhemorrhageinpatientsusingalargeannotatedheadcomputedtomographydatasetaretrospectivemulticenterstudy AT jingyuan deeplearningbasedidentificationandlocalizationofintracranialhemorrhageinpatientsusingalargeannotatedheadcomputedtomographydatasetaretrospectivemulticenterstudy AT weili deeplearningbasedidentificationandlocalizationofintracranialhemorrhageinpatientsusingalargeannotatedheadcomputedtomographydatasetaretrospectivemulticenterstudy AT xianqiwang deeplearningbasedidentificationandlocalizationofintracranialhemorrhageinpatientsusingalargeannotatedheadcomputedtomographydatasetaretrospectivemulticenterstudy AT yiwu deeplearningbasedidentificationandlocalizationofintracranialhemorrhageinpatientsusingalargeannotatedheadcomputedtomographydatasetaretrospectivemulticenterstudy AT chenliu deeplearningbasedidentificationandlocalizationofintracranialhemorrhageinpatientsusingalargeannotatedheadcomputedtomographydatasetaretrospectivemulticenterstudy AT qingfanggong deeplearningbasedidentificationandlocalizationofintracranialhemorrhageinpatientsusingalargeannotatedheadcomputedtomographydatasetaretrospectivemulticenterstudy AT mihe deeplearningbasedidentificationandlocalizationofintracranialhemorrhageinpatientsusingalargeannotatedheadcomputedtomographydatasetaretrospectivemulticenterstudy AT yeqinfu deeplearningbasedidentificationandlocalizationofintracranialhemorrhageinpatientsusingalargeannotatedheadcomputedtomographydatasetaretrospectivemulticenterstudy AT dongzhang deeplearningbasedidentificationandlocalizationofintracranialhemorrhageinpatientsusingalargeannotatedheadcomputedtomographydatasetaretrospectivemulticenterstudy AT sizhang deeplearningbasedidentificationandlocalizationofintracranialhemorrhageinpatientsusingalargeannotatedheadcomputedtomographydatasetaretrospectivemulticenterstudy AT yongjiannian deeplearningbasedidentificationandlocalizationofintracranialhemorrhageinpatientsusingalargeannotatedheadcomputedtomographydatasetaretrospectivemulticenterstudy |