[1]邓金城,彭应林,刘常春,等. 深度卷积神经网络在放射治疗计划图像分割中的应用[J].中国医学物理学杂志,2018,35(6):621-627.[doi:DOI:10.3969/j.issn.1005-202X.2018.06.001]
　DENG Jincheng,PENG Yinglin,LIU Changchun,et al. Application of deep convolution neural network in radiotherapy planning image segmentation[J].Chinese Journal of Medical Physics,2018,35(6):621-627.[doi:DOI:10.3969/j.issn.1005-202X.2018.06.001]
点击复制

 深度卷积神经网络在放射治疗计划图像分割中的应用()
分享到：

《中国医学物理学杂志》[ISSN:1005-202X/CN:44-1351/R]

卷:
35卷
期数:
2018年第6期
页码:
621-627
栏目:
医学放射物理
出版日期:
2018-06-22

文章信息/Info

Title:
 Application of deep convolution neural network in radiotherapy planning image segmentation
文章编号:
1005-202X（2018）06-0621-07
作者:
 邓金城1; 彭应林2; 刘常春1; 陈子杰1; 雷国胜1; 吴江华1; 张广顺2; 邓小武2
 1.深圳市医诺智能科技发展有限公司， 广东 深圳 518057； 2.中山大学肿瘤防治中心/华南肿瘤学国家重点实验室/肿瘤医学协同创新中心， 广东 广州 510060
Author(s):
 DENG Jincheng1;  PENG Yinglin2;  LIU Changchun1;  CHEN Zijie1;  LEI Guosheng1;  WU Jianghua1;  ZHANG Guangshun2;  DENG Xiaowu2
 1. Shenzhen Yino Intelligence Techonology Co, Ltd., Shenzhen 518057, China; 2. Collaborative Innovation Center for Cancer Medicine/State Key Laboratory of Oncology in South China/Sun Yat-Sen University Cancer Center, Guangzhou 510060, China
关键词:
 ; 深度学习; 卷积神经网络; 医学影像分割; 相似度系数; 放射治疗
Keywords:
 ; Keywords: deep learning;  convolution neural network;  medical image segmentation;  similarity coefficient;  radiotherapy
分类号:
R811.1
DOI:
DOI:10.3969/j.issn.1005-202X.2018.06.001
文献标志码:
A
摘要:
 目的：结合全卷积神经网络（Fully Convolutional Network, FCN）和多孔卷积（Atrous Convolution, AC）的深度学习方法，实现放射治疗计划图像的组织器官自动勾画。方法：选取122套已经由放疗医师勾画好正常器官结构轮廓的胸部患者CT图像，以其中71套图像（8 532张轴向切层图像）作为训练集，31套图像（5 559张轴向切层图像）作为验证集，20套图像（3 589张轴向切层图像）作为测试集。选取5种公开的FCN网络模型，并结合FCN和AC算法形成3种改进的深度卷积神经网络，即带孔全卷积神经网络（Dilation Fully Convolutional Network, D-FCN）。分别以训练集图像对上述8种网络进行调优训练，使用验证集图像在训练过程中对8种神经网络进行器官自动识别勾画验证，以获取各网络的最佳分割模型，最后使用测试集图像对充分训练后获取的最佳分割模型进行勾画测试，比较自动勾画与医师勾画的相似度系数（Dice）评价各模型的图像分割能力。 结果：使用训练图像集进行充分调优训练后，实验的各个神经网络均表现出较好的自动图像分割能力，其中改进的D-FCN 4s网络模型在测试实验中具有最佳的自动分割效果，其全局Dice为94.38%，左肺、右肺、心包、气管和食道等单个结构自动勾画的Dice分别为96.49%、96.75%、86.27%、61.51%和65.63%。 结论：提出了一种改进型全卷积神经网络D-FCN，实验测试表明该网络模型可以有效地提高胸部放疗计划图像的自动分割精度，并可同时进行多目标的自动分割。
Abstract:
 Objective To combine fully convolutional network (FCN) and atrous convolution (AC) for realizing automation segmentation of tissues and organs in radiotherapy planning image. Methods A total of 122 sets of chest CT images were selected in this study, in which the normal organ structures were sketched by radiotherapy physician, including 71 sets of CT images (8 532 axial slice images) as training set, 31 sets of CT images (5 559 axial slice images) as validation set, and 20 sets of CT images (3 589 axial slice images) as test set. Five kinds of published FCN models were selected and combined with AC algorithm to form 3 kinds of improved deep convolutional neural networks, namely dilation fully convolutional network (D-FCN). Training set was used for fully fine-tuning the above 8 kinds of network, and validation set was applied to validate the automatic segmentation results for obtaining the optimal model of each network, and finally, test set was used to perform segmentation test for the optimal models. The Dice similarity coefficients of automatic segmentation and physician sketching were compared for evaluating the performances of these image segmentation models. Results After being fully fine-tuned with the use of training set, each neural network model showed good performances in automatic image segmentation. The improved D-FCN 4s model achieved the best automatic segmentation results in validation test, with a global Dice of 94.38%. The Dice of automatic segmentations of left lung, right lung, pericardium, trachea and esophagus was 96.49%, 96.75%, 86.27%, 61.51% and 65.63%, respectively. Conclusion An improved D-FCN is put forward in this study and the verification test shows that the improved D-FCN can effectively improve the accuracy of automatic segmentation for radiotherapy planning image of chest, and segment multiple organs synchronously.

相似文献/References:

[1]王遥,霍万里,熊壮,等.TACE手术中不同站姿下铅眼镜和铅面罩对医生眼晶状体防护效果的蒙特卡洛模拟比较[J].中国医学物理学杂志,2016,33(6):553.[doi:DOI:10.3969/j.issn.1005-202X.2016.06.003]
　[J].Chinese Journal of Medical Physics,2016,33(6):553.[doi:DOI:10.3969/j.issn.1005-202X.2016.06.003]
[2]张新,谷晓芳,王培臣,等.轻离子束治疗设备注册检验关键技术问题[J].中国医学物理学杂志,2016,33(6):559.[doi:10.3969/j.issn.1005-202X.2016.06.004]
　[J].Chinese Journal of Medical Physics,2016,33(6):559.[doi:10.3969/j.issn.1005-202X.2016.06.004]
[3]江芬芬,王培,康盛伟,等. 热释光剂量片测量肺部肿瘤放疗剂量的方法[J].中国医学物理学杂志,2016,33(6):564.[doi:10.3969/j.issn.1005-202X.2016.06.005]
　[J].Chinese Journal of Medical Physics,2016,33(6):564.[doi:10.3969/j.issn.1005-202X.2016.06.005]
[4]刘洪源,彭威,杨锐,等. 锥形束CT离线校正肺癌摆位误差[J].中国医学物理学杂志,2016,33(6):573.[doi:10.3969/j.issn.1005-202X.2016.06.007]
　[J].Chinese Journal of Medical Physics,2016,33(6):573.[doi:10.3969/j.issn.1005-202X.2016.06.007]
[5]赵彪,潘香,杨毅,等. 右乳癌保乳术后瘤床同步X线和后程电子线补量调强放疗剂量学比较[J].中国医学物理学杂志,2016,33(6):576.[doi:10.3969/j.issn.1005-202X.2016.06.008]
　[J].Chinese Journal of Medical Physics,2016,33(6):576.[doi:10.3969/j.issn.1005-202X.2016.06.008]
[6]邓南,钱建庭,刁现芬,等. 基于宽带检测放疗X-光光声效应仿体实验[J].中国医学物理学杂志,2016,33(9):865.[doi:DOI:10.3969/j.issn.1005-202X.2016.09.001]
　[J].Chinese Journal of Medical Physics,2016,33(6):865.[doi:DOI:10.3969/j.issn.1005-202X.2016.09.001]
[7]张先稳,李军,张西志,等. 宫颈癌术后5野调强放疗4个变量组合的最佳治疗模式的剂量学[J].中国医学物理学杂志,2016,33(9):872.[doi:10.3969/j.issn.1005-202X.2016.09.002]
　[J].Chinese Journal of Medical Physics,2016,33(6):872.[doi:10.3969/j.issn.1005-202X.2016.09.002]
[8]胡健,李承军,徐利明,等. 床面倾斜对剂量验证通过率的影响[J].中国医学物理学杂志,2016,33(9):881.[doi:10.3969/j.issn.1005-202X.2016.09.003]
　[J].Chinese Journal of Medical Physics,2016,33(6):881.[doi:10.3969/j.issn.1005-202X.2016.09.003]
[9]陈亚正,肖明勇,孙春堂,等. 准直器角度对宫颈癌术后VMAT计划的影响[J].中国医学物理学杂志,2016,33(9):885.[doi:10.3969/j.issn.1005-202X.2016.09.004]
　[J].Chinese Journal of Medical Physics,2016,33(6):885.[doi:10.3969/j.issn.1005-202X.2016.09.004]
[10]李毅,唐丰文,张晓智. 基于四维CT和锥形束CT确定非小细胞肺癌放疗靶区外放边界[J].中国医学物理学杂志,2016,33(9):892.[doi:10.3969/j.issn.1005-202X.2016.09.005]
　[J].Chinese Journal of Medical Physics,2016,33(6):892.[doi:10.3969/j.issn.1005-202X.2016.09.005]
[11]门阔,戴建荣. 利用深度反卷积神经网络自动勾画放疗危及器官[J].中国医学物理学杂志,2018,35(3):256.[doi:DOI:10.3969/j.issn.1005-202X.2018.03.002]
　MEN Kuo,DAI Jianrong. Automatic segmentation of organs at risk in radiotherapy using deep deconvolutional neural network[J].Chinese Journal of Medical Physics,2018,35(6):256.[doi:DOI:10.3969/j.issn.1005-202X.2018.03.002]
[12]宫进昌,赵尚义,王远军. 基于深度学习的医学图像分割研究进展[J].中国医学物理学杂志,2019,36(4):420.[doi:DOI:10.3969/j.issn.1005-202X.2019.04.010]
　GONG Jinchang,ZHAO Shangyi,WANG Yuanjun.Research progress on deep learning-based medical image segmentation[J].Chinese Journal of Medical Physics,2019,36(6):420.[doi:DOI:10.3969/j.issn.1005-202X.2019.04.010]
[13]安莹,黄能军,杨荣,等. 基于深度学习的心血管疾病风险预测模型[J].中国医学物理学杂志,2019,36(9):1103.[doi:DOI:10.3969/j.issn.1005-202X.2019.09.021]
　AN Ying,HUANG Nengjun,YANG Rong,et al. Deep learning-based model for risk prediction of cardiovascular diseases[J].Chinese Journal of Medical Physics,2019,36(6):1103.[doi:DOI:10.3969/j.issn.1005-202X.2019.09.021]

备注/Memo

备注/Memo:
 【收稿日期】2018-03-19
【基金项目】国家重点研发计划（2017YFC0113200）；广东省科技计划项目（2015B020214002）；广州市科技计划项目（201508020105）；深圳市高技术产业化扶持计划项目（S2016I65100017）
【作者简介】邓金城，E-mail: djc@szyino.com；彭应林，E-mail: pengyl@sysucc.org.cn
【通信作者】邓小武，研究员，研究方向：肿瘤放射治疗物理学，E-mail: dengxw@sysucc.org.cn
更新日期/Last Update: 2018-06-22