| 摘 要: 针对桥小脑角区听神经瘤和脑膜瘤在临床诊断中不易区分的问题,提出了一种基于深度学习的辅助诊断模型。首先,采集肿瘤的T1WI(T1 Weighted Imaging)增强图像和T2WI(T2 Weighted Imaging)图像,构建基于VGG-net改进的s-VGG网络对两组图像分别进行训练,得到s-VGG-T1和s-VGG-T2两个分类模型。其次,集合放射科与放疗科的临床诊断结果,建立深度学习辅助诊断模型,将分类模型结果与临床诊断结果加权平均得到诊断模型结果。相比单独的诊断结果,诊断模型对10例肿瘤的诊断准确率有所提升,表明基于深度学习的辅助诊断模型具有良好的性能,可降低误诊率,提升诊断的准确性和临床工作的效率。 |
| 关键词: 深度学习;辅助诊断;听神经瘤;脑膜瘤 |
|
中图分类号: TP391
文献标识码: A
|
|
| An Aided Diagnosis Model of Acoustic Neuroma and Meningioma in Cerebellopontine Angle Based on Deep Learning |
|
HU Xiaoyang1,2, LIU Ying1, NI Chunxia2, CHEN Shu2, DONG Binbin3
|
(1.School of Health Science and Engineering, University of Shanghai f or Science and Technology, Shanghai 200093, China;
2.Department of Radiotherapy, Shanghai Gamma Hospital, Shanghai 200235, China;
3.Department of Radiology, Shanghai Gamma Hospital, Shanghai 200235, China)
hxy2006x@126.com; miusst@163.com; 564056708@qq.com; moonshushu@126.com; bdon@foxmail.com
|
| Abstract: Aiming at the problem that acoustic neuroma and meningioma in cerebellopontine angle are not easy to distinguish in clinical diagnosis, this paper proposes an aided diagnosis model based on deep learning. Firstly, T1WI (T1 Weighted Imaging) enhancement images and T2WI (T2 Weighted Imaging) images of tumors are collected, and the improved s-VGG network based on VGG-net is constructed to train the two groups of images respectively. Thus, the s-VGG-T1 and s-VGG-T2 classification models are obtained. Then a deep learning aided diagnosis model is established based on the clinical diagnosis results of the radiology and the radiotherapy department. And the diagnosis model results are calculated by weighted average of the classification model results and the clinical diagnosis results. Compared to the individual diagnostic results, the diagnostic accuracy of diagnosis model for 10 tumors is improved, which shows that the aided diagnosis model based on deep learning has good performance, and it can reduce the misdiagnosis rate and improve the diagnostic accuracy and the efficiency of clinical work. |
| Keywords: deep learning; aided diagnosis; acoustic neuroma; meningioma |
