机构：[1]Department of Biomedical Engineering, Southern Medical University, Guangzhou 510515, Guangdong, China[2]Sun Yat-sen University Cancer Center, State Key Laboratory of Oncology in South China, Collaborative Innovation Center for Cancer Medicine, Guangdong Key Laboratory of Nasopharyngeal Carcinoma Diagnosis and Therapy, Guangzhou 510060, Guangdong, China[3]Department of Biomedical Engineering, Southern Medical University, Guangzhou 510515, Guangdong, China[4]Sun Yat-sen University Cancer Center, State Key Laboratory of Oncology in South China, Collaborative Innovation Center for Cancer Medicine, Guangdong Key Laboratory of Nasopharyngeal Carcinoma Diagnosis and Therapy, Guangzhou 510060, Guangdong, China[5]Department of Radiation Oncology, Guangdong Province Traditional Medical Hospital, Guangzhou 510000, Guangdong, China[6]Department of Biomedical Engineering, Southern Medical University, Guangzhou 510515, Guangdong, China[7]Sun Yat-sen University Cancer Center, State Key Laboratory of Oncology in South China, Collaborative Innovation Center for Cancer Medicine, Guangdong Key Laboratory of Nasopharyngeal Carcinoma Diagnosis and Therapy, Guangzhou 510060, Guangdong, China[8]Department of Biomedical Engineering, Southern Medical University, Guangzhou 510515, Guangdong, China
Purpose The purpose of this study is to investigate the effect of different magnetic resonance (MR) sequences on the accuracy of deep learning-based synthetic computed tomography (sCT) generation in the complex head and neck region. Methods Four sequences of MR images (T1, T2, T1C, and T1DixonC-water) were collected from 45 patients with nasopharyngeal carcinoma. Seven conditional generative adversarial network (cGAN) models were trained with different sequences (single channel) and different combinations (multi-channel) as inputs. To further verify the cGAN performance, we also used a U-net network as a comparison. Mean absolute error, structural similarity index, peak signal-to-noise ratio, dice similarity coefficient, and dose distribution were evaluated between the actual CTs and sCTs generated from different models. Results The results show that the cGAN model with multi-channel (i.e., T1 + T2 + T1C + T1DixonC-water) as input to predict sCT achieves higher accuracy than any single MR sequence model. The T1-weighted MR model achieves better results than T2, T1C, and T1DixonC-water models. The comparison between cGAN and U-net shows that the sCTs predicted by cGAN retains additional image details are less blurred and more similar to the actual CT. Conclusions Conditional generative adversarial network with multiple MR sequences as model input shows the best accuracy. The T1-weighted MR images provide sufficient image information and are suitable for sCT prediction in clinical scenarios with limited acquisition sequences or limited acquisition time.
基金：
National Key R&D Program of China [2017YFC0113203]; National Natural Science Foundation of ChinaNational Natural Science Foundation of China [11805292, 81601577, 81571771]; Natural Science Foundation of Guangdong, ChinaNational Natural Science Foundation of Guangdong Province [2018A0303100020]
第一作者机构：[1]Department of Biomedical Engineering, Southern Medical University, Guangzhou 510515, Guangdong, China
共同第一作者：
通讯作者：
推荐引用方式(GB/T 7714)：
Qi Mengke,Li Yongbao,Wu Aiqian,et al.Multi-sequence MR image-based synthetic CT generation using a generative adversarial network for head and neck MRI-only radiotherapy[J].MEDICAL PHYSICS.2020,47(4):1880-1894.doi:10.1002/mp.14075.
APA：
Qi, Mengke,Li, Yongbao,Wu, Aiqian,Jia, Qiyuan,Li, Bin...&Song, Ting.(2020).Multi-sequence MR image-based synthetic CT generation using a generative adversarial network for head and neck MRI-only radiotherapy.MEDICAL PHYSICS,47,(4)
MLA：
Qi, Mengke,et al."Multi-sequence MR image-based synthetic CT generation using a generative adversarial network for head and neck MRI-only radiotherapy".MEDICAL PHYSICS 47..4(2020):1880-1894
