Deep-learning-based automatic segmentation and classification for craniopharyngiomas
ObjectiveNeuronavigation and classification of craniopharyngiomas can guide surgical approaches and prognostic information. The QST classification has been developed according to the origin of craniopharyngiomas; however, accurate preoperative automatic segmentation and the QST classification remain...
| Main Authors: | , , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2023-05-01
|
| Series: | Frontiers in Oncology |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/fonc.2023.1048841/full |
| _version_ | 1827953733700419584 |
|---|---|
| author | Xiaorong Yan Bingquan Lin Jun Fu Shuo Li He Wang He Wang Wenjian Fan Yanghua Fan Ming Feng Renzhi Wang Jun Fan Songtao Qi Changzhen Jiang |
| author_facet | Xiaorong Yan Bingquan Lin Jun Fu Shuo Li He Wang He Wang Wenjian Fan Yanghua Fan Ming Feng Renzhi Wang Jun Fan Songtao Qi Changzhen Jiang |
| author_sort | Xiaorong Yan |
| collection | DOAJ |
| description | ObjectiveNeuronavigation and classification of craniopharyngiomas can guide surgical approaches and prognostic information. The QST classification has been developed according to the origin of craniopharyngiomas; however, accurate preoperative automatic segmentation and the QST classification remain challenging. This study aimed to establish a method to automatically segment multiple structures in MRIs, detect craniopharyngiomas, and design a deep learning model and a diagnostic scale for automatic QST preoperative classification.MethodsWe trained a deep learning network based on sagittal MRI to automatically segment six tissues, including tumors, pituitary gland, sphenoid sinus, brain, superior saddle cistern, and lateral ventricle. A deep learning model with multiple inputs was designed to perform preoperative QST classification. A scale was constructed by screening the images.ResultsThe results were calculated based on the fivefold cross-validation method. A total of 133 patients with craniopharyngioma were included, of whom 29 (21.8%) were diagnosed with type Q, 22 (16.5%) with type S and 82 (61.7%) with type T. The automatic segmentation model achieved a tumor segmentation Dice coefficient of 0.951 and a mean tissue segmentation Dice coefficient of 0.8668 for all classes. The automatic classification model and clinical scale achieved accuracies of 0.9098 and 0.8647, respectively, in predicting the QST classification.ConclusionsThe automatic segmentation model can perform accurate multi-structure segmentation based on MRI, which is conducive to clearing tumor location and initiating intraoperative neuronavigation. The proposed automatic classification model and clinical scale based on automatic segmentation results achieve high accuracy in the QST classification, which is conducive to developing surgical plans and predicting patient prognosis. |
| first_indexed | 2024-04-09T14:18:12Z |
| format | Article |
| id | doaj.art-24dfe04316914679a8a7258ba3aa7791 |
| institution | Directory Open Access Journal |
| issn | 2234-943X |
| language | English |
| last_indexed | 2024-04-09T14:18:12Z |
| publishDate | 2023-05-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Oncology |
| spelling | doaj.art-24dfe04316914679a8a7258ba3aa77912023-05-05T05:15:28ZengFrontiers Media S.A.Frontiers in Oncology2234-943X2023-05-011310.3389/fonc.2023.10488411048841Deep-learning-based automatic segmentation and classification for craniopharyngiomasXiaorong Yan0Bingquan Lin1Jun Fu2Shuo Li3He Wang4He Wang5Wenjian Fan6Yanghua Fan7Ming Feng8Renzhi Wang9Jun Fan10Songtao Qi11Changzhen Jiang12Department of Neurosurgery, First affiliated Hospital of Fujian Medical University, Fuzhou, Fujian, ChinaDepartment of Medical Image Center, Southern Medical University, Nanfang Hospital, Guangzhou, ChinaDepartment of Neurosurgery, First affiliated Hospital of Fujian Medical University, Fuzhou, Fujian, ChinaDepartment of Plastic Surgery, Peking Union Medical College Hospital, Beijing, ChinaDepartment of Neurosurgery, Peking Union Medical College Hospital, Beijing, ChinaDepartment of Neurosurgery, Xuanwu Hospital, Capital Medical University, China International Neuroscience Institute, Beijing, ChinaDepartment of Neurosurgery, First affiliated Hospital of Fujian Medical University, Fuzhou, Fujian, ChinaDepartment of Neurosurgery, Beijing Tiantan Hospital, Beijing Neurosurgical Institute, Capital Medical University, Beijing, ChinaDepartment of Neurosurgery, Peking Union Medical College Hospital, Beijing, ChinaDepartment of Neurosurgery, Peking Union Medical College Hospital, Beijing, ChinaDepartment of Neurosurgery, Southern Medical University, Nanfang Hospital, Fuzhou, Fujian, ChinaDepartment of Neurosurgery, Southern Medical University, Nanfang Hospital, Fuzhou, Fujian, ChinaDepartment of Neurosurgery, First affiliated Hospital of Fujian Medical University, Fuzhou, Fujian, ChinaObjectiveNeuronavigation and classification of craniopharyngiomas can guide surgical approaches and prognostic information. The QST classification has been developed according to the origin of craniopharyngiomas; however, accurate preoperative automatic segmentation and the QST classification remain challenging. This study aimed to establish a method to automatically segment multiple structures in MRIs, detect craniopharyngiomas, and design a deep learning model and a diagnostic scale for automatic QST preoperative classification.MethodsWe trained a deep learning network based on sagittal MRI to automatically segment six tissues, including tumors, pituitary gland, sphenoid sinus, brain, superior saddle cistern, and lateral ventricle. A deep learning model with multiple inputs was designed to perform preoperative QST classification. A scale was constructed by screening the images.ResultsThe results were calculated based on the fivefold cross-validation method. A total of 133 patients with craniopharyngioma were included, of whom 29 (21.8%) were diagnosed with type Q, 22 (16.5%) with type S and 82 (61.7%) with type T. The automatic segmentation model achieved a tumor segmentation Dice coefficient of 0.951 and a mean tissue segmentation Dice coefficient of 0.8668 for all classes. The automatic classification model and clinical scale achieved accuracies of 0.9098 and 0.8647, respectively, in predicting the QST classification.ConclusionsThe automatic segmentation model can perform accurate multi-structure segmentation based on MRI, which is conducive to clearing tumor location and initiating intraoperative neuronavigation. The proposed automatic classification model and clinical scale based on automatic segmentation results achieve high accuracy in the QST classification, which is conducive to developing surgical plans and predicting patient prognosis.https://www.frontiersin.org/articles/10.3389/fonc.2023.1048841/fullcraniopharyngiomasQST typing systemdeep learningsegmentationclassification |
| spellingShingle | Xiaorong Yan Bingquan Lin Jun Fu Shuo Li He Wang He Wang Wenjian Fan Yanghua Fan Ming Feng Renzhi Wang Jun Fan Songtao Qi Changzhen Jiang Deep-learning-based automatic segmentation and classification for craniopharyngiomas Frontiers in Oncology craniopharyngiomas QST typing system deep learning segmentation classification |
| title | Deep-learning-based automatic segmentation and classification for craniopharyngiomas |
| title_full | Deep-learning-based automatic segmentation and classification for craniopharyngiomas |
| title_fullStr | Deep-learning-based automatic segmentation and classification for craniopharyngiomas |
| title_full_unstemmed | Deep-learning-based automatic segmentation and classification for craniopharyngiomas |
| title_short | Deep-learning-based automatic segmentation and classification for craniopharyngiomas |
| title_sort | deep learning based automatic segmentation and classification for craniopharyngiomas |
| topic | craniopharyngiomas QST typing system deep learning segmentation classification |
| url | https://www.frontiersin.org/articles/10.3389/fonc.2023.1048841/full |
| work_keys_str_mv | AT xiaorongyan deeplearningbasedautomaticsegmentationandclassificationforcraniopharyngiomas AT bingquanlin deeplearningbasedautomaticsegmentationandclassificationforcraniopharyngiomas AT junfu deeplearningbasedautomaticsegmentationandclassificationforcraniopharyngiomas AT shuoli deeplearningbasedautomaticsegmentationandclassificationforcraniopharyngiomas AT hewang deeplearningbasedautomaticsegmentationandclassificationforcraniopharyngiomas AT hewang deeplearningbasedautomaticsegmentationandclassificationforcraniopharyngiomas AT wenjianfan deeplearningbasedautomaticsegmentationandclassificationforcraniopharyngiomas AT yanghuafan deeplearningbasedautomaticsegmentationandclassificationforcraniopharyngiomas AT mingfeng deeplearningbasedautomaticsegmentationandclassificationforcraniopharyngiomas AT renzhiwang deeplearningbasedautomaticsegmentationandclassificationforcraniopharyngiomas AT junfan deeplearningbasedautomaticsegmentationandclassificationforcraniopharyngiomas AT songtaoqi deeplearningbasedautomaticsegmentationandclassificationforcraniopharyngiomas AT changzhenjiang deeplearningbasedautomaticsegmentationandclassificationforcraniopharyngiomas |