Institutional Repository of Chinese Acad Sci, Inst Automat, Natl Lab Pattern Recognit, Beijing 100190, Peoples R China
| A multi-task two-path deep learning system for predicting the invasiveness of craniopharyngioma | |
Zhu, Lin1,2 ; Zhang, Lingling3; Hu, Wenxing4; Chen, Haixu5,6; Li, Han2,7; Wei, Shoushui1; Chen, Xuzhu3; Ma, Xibo2,8
| |
| Source Publication | COMPUTER METHODS AND PROGRAMS IN BIOMEDICINE
 |
| ISSN | 0169-2607 |
| 2022-04-01 | |
| Volume | 216Pages:13 |
| Corresponding Author | Wei, Shoushui(sswei@sdu.edu.cn) ; Chen, Xuzhu(radiology888@aliyun.com) ; Ma, Xibo(xibo.ma@ia.ac.cn) |
| Abstract | Background and Objective: Craniopharyngioma is a kind of benign brain tumor in histography. However, it might be clinically aggressive and have severe manifestations, such as increased intracranial pressure, hypothalamic-pituitary dysfunction, and visual impairment. It is considered challenging for radiologists to predict the invasiveness of craniopharyngioma through MRI images. Therefore, developing a non-invasive method that can predict the invasiveness and boundary of CP as a reference before surgery is of clinical value for making more appropriate and individualized treatment decisions and reducing the occurrence of inappropriate surgical plan choices. Methods: The MT-Brain system has consisted of two pathways, a sub-path based on 2D CNN for capturing the features from each slice of MRI images, and a 3D sub-network for capturing additional context information between slices. By introducing the two-path architecture, our system can make full use of the fusion of the above 2D and 3D features for classification. Furthermore, position encoding and mask-guided attention also have been introduced to improve the segmentation and diagnosis performance. To verify the performance of the MT-Brain system, we have enrolled 1032 patients with craniopharyngioma (302 invasion and 730 non-invasion patients), segmented the tumors on postcontrast coronal T1WI and randomized them into a training dataset and a testing dataset at a ratio of 8:2. Results: The MT-Brain system achieved a remarkable performance in diagnosing the invasiveness of craniopharyngioma with the AUC of 83.84%, the accuracy of 77.94%, the sensitivity of 70.97%, and the specificity of 80.99%. In the lesion segmentation task, the predicted boundaries of lesions were similar to those labeled by radiologists with the dice of 66.36%. In addition, some explorations also have been made on the interpretability of deep learning models, illustrating the reliability of the model. Conclusions: To the best of our knowledge, this study is the first to develop an integrated deep learning model to predict the invasiveness of craniopharyngioma preoperatively and locate the lesion boundary synchronously on MRI. The excellent performances indicate that the MT-Brain system has great potential in real-world clinical applications. (C) 2022 Published by Elsevier B.V. |
| Keyword | Craniopharyngioma MRI Imaging Deep learning Invasiveness diagnosis Lesion segmentation |
| DOI | 10.1016/j.cmpb.2022.106651 |
| Indexed By | SCI |
| Language | 英语 |
| Funding Project | National Key Research Program of China[2016YFA0100902] ; National Natural Science Foundation Projects of China[82072014] ; National Natural Science Foundation Projects of China[82090051] ; National Natural Science Foundation Projects of China[81871442] ; National Natural Science Foundation Projects of China[817720 05] ; Shandong Province Natural Science Foundation[ZR2020MF028] ; Youth Innovation Promotion Association CAS[Y201930] ; Beijing Municipal Science & Technology Commission[Z191100006619088] |
| Funding Organization | National Key Research Program of China ; National Natural Science Foundation Projects of China ; Shandong Province Natural Science Foundation ; Youth Innovation Promotion Association CAS ; Beijing Municipal Science & Technology Commission |
| WOS Research Area | Computer Science ; Engineering ; Medical Informatics |
| WOS Subject | Computer Science, Interdisciplinary Applications ; Computer Science, Theory & Methods ; Engineering, Biomedical ; Medical Informatics |
| WOS ID | WOS:000766134000003 |
| Publisher | ELSEVIER IRELAND LTD |
| Sub direction classification | 人工智能+医疗 |
| Citation statistics | |
| Document Type | 期刊论文 |
| Identifier | http://ir.ia.ac.cn/handle/173211/48018 |
| Collection | 模式识别国家重点实验室_生物识别与安全技术 |
| Corresponding Author | Wei, Shoushui; Chen, Xuzhu; Ma, Xibo |
| Affiliation | 1.Shandong Univ, Sch Control Sci & Engn, Jinan 250061, Shandong, Peoples R China 2.Chinese Acad Sci, Inst Automat, CBSR & NLPR, Beijing 100190, Peoples R China 3.Capital Med Univ, Beijing Tiantan Hosp, Dept Radiol, Beijing, Peoples R China 4.Univ New South Wales, Sydney, NSW, Australia 5.Chinese Peoples Liberat Army Gen Hosp, Inst Geriatr, Med Ctr 2, Beijing 100853, Peoples R China 6.Chinese Peoples Liberat Army Gen Hosp, Natl Clin Res Ctr Geriatr Dis, Med Ctr 2, Beijing 100853, Peoples R China 7.Guangdong Acad Med Sci, Dept Orthoped, Guangdong Prov Peoples Hosp, Guangzhou, Peoples R China 8.Univ Chinese Acad Sci, Sch Artificial Intelligence, Beijing 100049, Peoples R China |
| First Author Affilication | Chinese Acad Sci, Inst Automat, Natl Lab Pattern Recognit, Beijing 100190, Peoples R China |
| Corresponding Author Affilication | Chinese Acad Sci, Inst Automat, Natl Lab Pattern Recognit, Beijing 100190, Peoples R China |
| Recommended Citation GB/T 7714 | Zhu, Lin,Zhang, Lingling,Hu, Wenxing,et al. A multi-task two-path deep learning system for predicting the invasiveness of craniopharyngioma[J]. COMPUTER METHODS AND PROGRAMS IN BIOMEDICINE,2022,216:13. |
| APA | Zhu, Lin.,Zhang, Lingling.,Hu, Wenxing.,Chen, Haixu.,Li, Han.,...&Ma, Xibo.(2022).A multi-task two-path deep learning system for predicting the invasiveness of craniopharyngioma.COMPUTER METHODS AND PROGRAMS IN BIOMEDICINE,216,13. |
| MLA | Zhu, Lin,et al."A multi-task two-path deep learning system for predicting the invasiveness of craniopharyngioma".COMPUTER METHODS AND PROGRAMS IN BIOMEDICINE 216(2022):13. |
| Files in This Item: | There are no files associated with this item. | |||||
Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.
Edit Comment