-
肺癌是目前在世界范围内患病率(11.6%)和病死率(18.4%)最高的恶性肿瘤[1],其5年生存率极低,在我国仅为15.6%,超过75%的患者在确诊时已处于晚期[2]。因此,早发现、早诊断对于提高肺癌患者的生存率尤为关键。胸部X射线检查是最早应用于肺癌普查、筛查的影像学检查方法,但因其灵敏度、特异度较低,已被发展迅速的胸部CT扫描所替代。从最早的低分辨率CT到目前高分辨率薄层、低剂量螺旋、双源能谱CT,都包含了大量的影像学信息,放射科医师需要耗费大量的时间和精力阅片,并结合临床信息来诊断肺结节的良恶性。随着人工智能的发展,基于人工智能技术的计算机辅助诊断(computer aided diagnosis,CAD)系统可提高放射科医师的工作效率以及肺结节诊断结果的准确率。深度学习是目前人工智能技术的研究热点,基于深度学习的CAD系统在肺癌的早期诊断中的应用取得了突破性的进展[3]。
CAD系统应用于肺癌的早期诊断,通常包含以下几个步骤:数据预处理、肺区域分割、候选结节检测与分割以及结节诊断[4]。基于深度学习的CAD系统能有效解决肺癌早期诊断中的核心问题,包括特征提取、肺结节检测和假阳性率的降低3个方面[5]。深度学习模型通常分为监督学习和非监督学习两种形式,其中监督学习需要使用带有分类标签的数据,此类模型包括卷积神经网络(convolutional neural network,CNN)和大规模训练人工神经网络(massive-training artificial neural networks,MTANNs);非监督学习则使用无标签数据,此类模型包括自动编码器(autoencoder,AE)和深度置信网络(deep belief network,DBN)。
基于深度学习的计算机辅助诊断系统在肺癌早期诊断中的应用与进展
Application and development of computer-aided diagnosis systems based on deep learning for the early diagnosis of lung cancer
-
摘要: 胸部CT扫描是肺癌早期筛查和诊断的主要检查手段,应用于胸部影像诊断领域的基于深度学习的计算机辅助诊断(CAD)系统可对CT图像上的肺结节进行检测和分类。深度学习技术可提高CAD系统的性能,尤其是在提高肺结节检测的准确率和降低假阳性率方面。笔者就CAD系统中的深度学习模型在肺结节中的应用现状和研究进展作一综述。Abstract: Chest CT scan is the primary medical imaging method performed for the early screening and diagnosis of lung cancer. Deep-learning based computer aided diagnosis (CAD) system for chest CT imaging is helpful for detecting and classifying pulmonary nodules. Deep-learning techniques can improve the performance of CAD systems, especially in enhancing the accuracy of pulmonary nodule detection and reducing false-positive rates. This article reviewed the current application status of deep-learning models in CAD systems and the progress that has been achieved in using these systems for imaging pulmonary nodules.
-
[1] Bray F, Ferlay J, Soerjomataram I, et al. Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries[J]. CA Cancer J Clin, 2018, 68(6): 394−424. DOI: 10.3322/caac.21492. [2] Zeng HM, Zheng RS, Guo YM, et al. Cancer survival in China, 2003-2005: A population-based study[J]. Int J Cancer, 2015, 136(8): 1921−1930. DOI: 10.1002/ijc.29227. [3] Hinton GE, Salakhutdinov RR. Reducing the Dimensionality of Data with Neural Networks[J]. Science, 2006, 313(5786): 504−507. DOI: 10.1126/science.1127647. [4] El-Baz A, Beache GM, Gimel'Farb G, et al. Computer-aided diagnosis systems for lung cancer: challenges and methodologies[J]. Int J Biomed Imaging, 2013, 2013: 942353. DOI: 10.1155/2013/942353. [5] Greenspan H, van Ginneken B, Summers RM. Guest Editorial Deep Learning in Medical Imaging: Overview and Future Promise of an Exciting New Technique[J]. IEEE Trans Med Imaging, 2016, 35(5): 1153−1159. DOI: 10.1109/TMI.2016.2553401. [6] Setio AAA, Ciompi F, Litjens G, et al. Pulmonary Nodule Detection in CT Images: False Positive Reduction Using Multi-View Convolutional Networks[J]. IEEE Trans Med Imaging, 2016, 35(5): 1160−1169. DOI: 10.1109/TMI.2016.2536809. [7] Armato III SG, McLennan G, Bidaut L, et al. The Lung Image Database Consortium (LIDC) and Image Database Resource Initiative (IDRI): A Completed Reference Database of Lung Nodules on CT Scans[J]. Med Phys, 2011, 38(2): 915−931. DOI: 10.1118/1.3528204. [8] Xie HT, Yang DB, Sun NN, et al. Automated pulmonary nodule detection in CT images using deep convolutional neural networks[J]. Pattern Recognit, 2019, 85: 109−119. DOI: 10.1016/j.patcog.2018.07.031. [9] Jin HS, Li ZY, Tong RF, et al. A deep 3D residual CNN for false-positive reduction in pulmonary nodule detection[J]. Med Phys, 2018, 45(5): 2097−2107. DOI: 10.1002/mp.12846. [10] Tang H, Kim DR, Xie XH. Automated pulmonary nodule detection using 3D deep convolutional neural networks [C]//Proceedings of the 2018 IEEE 15th International Symposium on Biomedical Imaging. Washington: IEEE, 2018: 523−526. DOI: 10.1109/ISBI.2018.8363630. [11] Qin YL, Zheng H, Zhu YM, et al. Simultaneous Accurate Detection of Pulmonary Nodules and False Positive Reduction Using 3D CNNs[C]//Proceedings of 2018 IEEE International Conference on Acoustics, Speech and Signal Processing. Calgary: IEEE, 2018: 1005−1009. DOI: 10.1109/ICASSP.2018.8462546. [12] Zhu WT, Liu CC, Fan W, et al. DeepLung: Deep 3D Dual Path Nets for Automated Pulmonary Nodule Detection and Classification[C]//Proceedings of 2018 IEEE Winter Conference on Applications of Computer Vision. Lake Tahoe: IEEE, 2018: 673−681. DOI: 10.1109/WACV.2018.00079. [13] Ciompi F, Chung K, van Riel SJ, et al. Towards automatic pulmonary nodule management in lung cancer screening with deep learning[J/OL]. Sci Rep, 2017, 7: 46479[2019-11-12]. https://www.nature.com/articles/srep46479. DOI: 10.1038/srep46479. [14] Liu XL, Hou F, Qin H, et al. Multi-view multi-scale CNNs for lung nodule type classification from CT images[J]. Pattern Recognit, 2018, 77: 262−275. DOI: 10.1016/j.patcog.2017.12.022. [15] Yuan JJ, Liu XL, Hou F, et al. Hybrid-feature-guided lung nodule type classification on CT images[J]. Comput Graph, 2018, 70: 288−299. DOI: 10.1016/j.cag.2017.07.020. [16] Wu BT, Zhou Z, Wang JW, et al. Joint learning for pulmonary nodule segmentation, attributes and malignancy prediction [C]//Proceedings of the 2018 IEEE 15th International Symposium on Biomedical Imaging. Washington: IEEE, 2018: 1109−1113. DOI: 10.1109/ISBI.2018.8363765. [17] Suzuki K, Doi K. How Can a Massive Training Artificial Neural Network (MTANN) be Trained With a Small Number of Cases in the Distinction Between Nodules and Vessels in Thoracic CT?[J]. Acad Radiol, 2005, 12(10): 1333−1341. DOI: 10.1016/j.acra.2005.06.017. [18] Tajbakhsh N, Suzuki K. Comparing two classes of end-to-end machine-learning models in lung nodule detection and classification: MTANNs vs. CNNs[J]. Pattern Recognit, 2017, 63: 476−486. DOI: 10.1016/j.patcog.2016.09.029. [19] Song Q, Zhao L, Luo X, et al. Using Deep Learning for Classification of Lung Nodules on Computed Tomography Images[J]. J Healthc Eng, 2017, 2017: 8314740. DOI: 10.1155/2017/8314740. [20] Vincent P, Larochelle H, Bengio Y, et al. Extracting and composing robust features with denoising autoencoders [C]//Proceedings of the 25th International Conference on Machine Learning. New York: ACM, 2008: 1096−1103. DOI: 10.1145/1390156.1390294. [21] Vincent P, Larochelle H, Lajoie I, et al. Stacked Denoising Autoencoders: Learning Useful Representations in a Deep Network with a Local Denoising Criterion[J]. J Mach Learn Res, 2010, 11(12): 3371−3408. [22] Kumar D, Wong A, Clausi DA. Lung Nodule Classification Using Deep Features in CT Images[C]//Proceedings of the 2015 12th Conference on Computer and Robot Vision. Halifax: IEEE, 2015: 133−138. DOI: 10.1109/CRV.2015.25. [23] Kim BC, Sung YS, Suk HI. Deep feature learning for pulmonary nodule classification in a lung CT[C]//Proceedings of the 2016 4th International Winter Conference on Brain-Computer Interface. Yongpyong: IEEE, 2016: 1−3. DOI: 10.1109/IWW-BCI.2016.7457462. [24] Mao KM, Tang RJ, Wang XQ, et al. Feature Representation Using Deep Autoencoder for Lung Nodule Image Classification[J]. Complexity, 2018, 2018: 3078374. DOI: 10.1155/2018/3078374. [25] 高琰, 陈白帆, 晁绪耀, 等. 基于对比散度-受限玻尔兹曼机深度学习的产品评论情感分析[J]. 计算机应用, 2016, 36(4): 1045−1049. DOI: 10.11772/j.issn.1001−9081.2016.04.1045.
Gao Y, Chen BF, Chao XY, et al. Sentiment analysis of product reviews based on contrastive divergence-restricted Boltzmann machine deep learning[J]. J Comput Appl, 2016, 36(4): 1045−1049. DOI: 10.11772/j.issn.1001−9081.2016.04.1045.[26] Hua KL, Hsu CH, Hidayati SC, et al. Computer-aided classification of lung nodules on computed tomography images via deep learning technique[J/OL]. Onco Targets Ther, 2015, 8: 2015−2022[2019-11-12]. https://www.dovepress.com/computer-aided-classification-of-lung-nodules-on-computed-tomography-i-peer-reviewed-article-OTT. DOI: 10.2147/OTT.S80733. [27] Nibali A, He Z, Wollersheim D. Pulmonary nodule classification with deep residual networks[J]. Int J Comput Assist Radiol Surg, 2017, 12(10): 1799−1808. DOI: 10.1007/s11548−017−1605−6. [28] Zhang BH, Qi SL, Monkam P, et al. Ensemble Learners of Multiple Deep CNNs for Pulmonary Nodules Classification Using CT Images[J]. IEEE Access, 2019, 7: 110358−110371. DOI: 10.1109/ACCESS.2019.2933670.
计量
- 文章访问数: 12161
- HTML全文浏览量: 11670
- PDF下载量: 25