北京生物医学工程

基于稠残U-net神经网络在定位CT图像上自动分割甲状腺的研究

Study on automatic thyroid segmentation based on thick residue U-net neural network in localized CT images

作者：袁美芳杨毅赵彪文晓博易三莉

单位：昆明理工大学信息工程与自动化学院（云南昆明  650500）云南省肿瘤医院放射治疗科（云南昆明  650118） 通信作者：易三莉。E-mail：152514845@qq.com

分类号：R318.04

出版年·卷·期（页码）：2022·41·1（42-48）

摘要：

目的基于深度学习方法提出一种稠残U-net神经网络，探讨其在放疗定位CT上自动预测甲状腺轮廓的可行性，以减少放疗中甲状腺所受辐射剂量，降低甲减发生率。方法在U-net网络中引入残差机制和稠密连接机制建立一种稠残U-net网络。选取76名患者定位CT图像的甲状腺切片制作数据集，随机划分为训练集58例、验证集9例和测试集9例，对稠残U-net进行训练、验证和测试，得到稠残U-net自动预测甲状腺的结果。通过戴斯相似性系数（Dice）、杰卡德相似系数（Jaccard)）和豪斯多夫距离（HD）等评价指标来评估其分割性能。结果稠残U-net预测甲状腺的Dice值为0.86±0.09、Jaccard值为0.78±0.12、HD值为2.52±0.61，且预测的轮廓边界与专家勾画的标准边界非常接近。结论本文提出的稠残U-net能在定位CT图像上较为精准地预测甲状腺轮廓，且证明在卷积神经网络中引入残差机制和稠密连接机制能提高其分割性能。

Objective Based on the deep learning method, a dense-residual U-net neural network is proposed to explore the feasibility of automatically predicting the contour of the thyroid on the CT images of radiotherapy positioning, so as to reduce the radiation dose to the thyroid during radiotherapy and reduce the incidence of hypothyroidism .Methods Residual mechanism and dense connection mechanism were introduced into U-net to establish dense- residual U-net neural network. The thyroid tissue structures of 76 patients were selected from localized CT images to make a data set, which were randomly divided into a training set of 58 patients, a verification set of 9 patients and a test set of 9 patients.The dense-residual U-net neural network was trained, verified and tested, and the thyroid contour predicted by neural network were obtained. The segment performance of dense-residual U-net were evaluated by dice, jaccard and HD.Results The thyroid’ s dice was 0.86±0.09, the jaccard was 0.78±0.12, and the HD was 2.52±0.61 by dense-residual U-net, and the predicted contour boundary was very close to the standard drawn by experts.Conclusions The dense-residual U-net proposed in this study can accurately predict the thyroid contour on localized CT images, and it is proved that the residual mechanism and dense connection mechanism can improve performance of convolutional neural network in medical image segmentation.

参考文献：

[1] 中华医学会内分泌学分会.成人甲状腺功能减退症诊治指南[J].中华内分泌代谢杂志,2017,33(2):167-180.

[2] Ren L, Mei L, Zhang Y, et al. Nomogram for radiation-induced hypothyroidism prediction ?????in nasopharyngeal carcinoma after treatment[J]. The British Journal of Radiology,2017, 90(1070):20160686.

[3] R?njom M F, Brink C, Bentzen S M, et al. External validation of a normal tissue complication probability model for radiation-induced hypothyroidism in an independent cohort [J]. Acta Oncologica (Stockholm, Sweden), 2015, 54(9): 1301-1309.

[4] Hancock SL, Cox RS, Mcdougall IR. Thyroid diseases after treatment of Hodgkin's disease [J]. The New England journal of medicine, 1991, 325(9): 599-605.

[5] Diaz R, Jaboin JJ, Morales M, et al. Hypothyroidism as a consequence of intensity-modulated radiotherapy with concurrent taxane-based chemotherapy for locally advanced head-and-neck cancer [J]. International?Journal of Radiation Oncology, Biology, Physics, 2010, 77(2): 468-476.

[6] Dan CC, Giusti A, Gambardella LM, et al. Deep neural networks segment neuronal membranes in electron microscopy images[J]. Advances in Neural Information Processing Systems, 2012, 25:2852-2860.

[7] 赵飞,刘杰．基于卷积神经网络和图像显著性的心脏CT图像分割[J]．北京生物医学工程,2020,39(1):48-55

Zhao F, Liu J. Cardiac CT image segmentation based on convolutional neural network and image saliency[J].?Beijing Biomedical Engineering,2020, 39(1):48-55

[8] 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.

Deng JC, Peng YL, Liu CH?,et al.?Application of deep convolution neural network in radiotherapy planning image segmentation[J].?Chinese Journal of Medical Physics?,2018,35(6):621-627.

[9] Ronneberger O, Fischer P, Brox T. U-net: Convolutional networks for biomedical image segmentation [M]//Ronneberger O, Fischer P, Brox T, eds. Lecture notes in computer science. Cham: Springer International Publishing, 2015: 234- 241.

[10] He K, Zhang X, Ren S,?et al.?Identity mappings in deep residual networks[C]// European Conference on Computer Vision.??Computer Vision – ECCV 2016. London ：Springer，2016:630-645

[11] Huang G, Liu Z, Maaten L, et al. Densely connected convolutional networks[C]//?Conference?Proceedings.?Honolulu, HI, USA ：IEEE Conference on Computer Vision and Pattern Recognition , 2017: 243-256

[12]文晓博,袁美芳,赵彪,等.基于GDL损失函数U-net神经网络在放疗定位CT图像上对甲状腺分割的初步研究[J].山西医科大学学报,2021,52(3):350-355.

Wen XB,Yuan MF,Zhao B, et al.A preliminary study on thyroid segmentation by GDL-based U-net neural network in CT localization images for radiotherapy[J].?Journal of Shanxi Medical University,2021,52(3):350-355.

[13] 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):256-259.

Men K, Dai JR.?Automatic segmentation of organs at risk in radiotherapy using deep deconvolutional neural network[J]. Chinese Journal of Medical Physics,2018,35(3):256-259.

[14] 杨鑫,李学妍,张晓婷,等.基于自适应Unet网络的鼻咽癌放疗危及器官自动分割方法[J].南方医科大学学报,2020,40(11):1579-1586.?

Yang X, Li XY, Zhang XT, et al.?Segmentation of organs at risk in nasopharyngeal cancer for radiotherapy using a selfadaptive Unet network[J]. Journal of Southern Medical University,2020,40(11):1579-1586.

服务与反馈：

【文章下载】【加入收藏】

提示：您还未登录，请登录！点此登录