基于动态增强磁共振成像的影像组学和不同CNN的深度学习对乳腺良恶性病变的诊断价值

doi:10.3969/j.issn.2095-9400.2020.06.009

摘要
图/表
参考文献
相关文章 (15)

全文: PDF (1645 KB) HTML (1 KB)
输出: BibTeX | EndNote (RIS)

摘要目的：探讨基于动态增强磁共振成像（DCE-MRI）的影像组学和深度学习对乳腺良恶性病变的诊断价值。方法：选取2017年1月至2018年6月经手术或穿刺病理证实的乳腺恶性病变患者93例、良性59例。所有患者均术前行乳腺MRI平扫及增强检查。通过计算获得早期流入信号增强比、最大增强信号比和流出斜率相对应的参数图。采用基于纹理和强度直方图的影像组学分析和5个卷积神经元网络（CNN）模型（ResNet50、VGG16、VGG19、Xception和InceptionV3）的深度学习进行诊断分析，并采用包含不同量肿瘤周围组织的五种不同输入边界框分析。结果：影像组学的诊断准确度均值为80%。不同输入边界框的ResNet50模型中，含少量肿瘤周围组织的最小边界框诊断效能最高，高于1.5倍和2.0倍边界框分析（均P＜0.01）；ResNet50（93%）、Xcep-tion（94%）和InceptionV3（93%）的准确度高于VGG16（80%）和VGG19（79%）（均P＜0.01）。结论：深度学习对乳腺良恶性病灶的诊断效能优于影像组学，含少量肿瘤周围组织分析的诊断效能高于仅包含肿瘤组织和包含较多肿瘤周围组织。

	服务
	把本文推荐给朋友
	加入我的书架
	加入引用管理器
	E-mail Alert
	RSS
	作者相关文章
	周洁洁
	张洋
	苏敏莹
	何遐遐
	徐妮娜
	叶舒欣
	李建策
	王瓯晨
	王美豪

关键词 ：乳腺癌, 磁共振成像, 影像组学, 深度学习, 卷积神经元网络

Abstract：Objective: To evaluate and compare the performance of radiomics and deep learning in the diagnosis of benign and malignant breast lesions based on DCE-MRI. Methods: A total of 152 patients receiving breast MRI for diagnosis were analyzed, including 93 patients with malignant cancers, and 59 patients with benign lesions. Three DCE parametric maps corresponding to early wash-in signal enhancement (SE), maximum signal enhancement, and wash-out slope were generated. Radiomics analysis based on texture and intensity histogram, and deep learning using 5 networks (ResNet50, VGG16, VGG19, Xception, InceptionV3), were performed for differential diagnosis. Results: The accuracy of radiomics was 80%; the smallest bounding box obtained higher diagnostic accuracy than that with tumor only and 1.2 times box which was not significant (both P>0.05), and also higher than 1.5 times box and 2.0 times box (both P<0.01); the accuracy of ResNet50 (93%), Xception (94%), InceptionV3 (93%) was significantly higher than VGG16 (80%), VGG19 (79%) (both P<0.01). Conclusion: CNN achieved better diagnostic performance in the diagnosis of benign and malignant breast lesions. The smaller bounding box containing the tumor with small amount of per-tumor tissue had the higher diagnostic accuracy than that with the tumor only and than larger bounding box.

Key words： breast cancer magnetic resonance imaging radiomics deep learning convolutional neural network

收稿日期: 2019-09-05

基金资助:温州市科技计划项目（Y20190187）。

通讯作者: 王美豪，主任医师，Email：wzwmh@wmu.edu.cn。

作者简介: 周洁洁（1984-），女，浙江温州人，主治医师，硕士。

链接本文:

https://xb.wmu.edu.cn/CN/10.3969/j.issn.2095-9400.2020.06.009 或 https://xb.wmu.edu.cn/CN/Y2020/V50/I6/475

[1] MANN R M, KUHL C K, MOY L. Contrast-enhanced MRI for breast cancer screening[J]. J Magn Reson Imaging, 2019, 50(2): 377-390.
[2] LEITHNER D, WENGET G J, HELBICH T H, et al. Clinical role of breast MRI and going forward[J]. Clin Radiol, 2018, 73(8): 700-714.
[3] LAMBIN P, RIOS-VELAZQUEZ E, LEIJENAAR R, et al. Radiomics: extracting more information from medical images using advanced feature analysis[J]. Eur J Cancer, 2012, 48(4): 441-446.
[4] HOSNY A, PARMAR C, QUACKENBUSH J, et al. Artificial intelligence in radiology[J]. Nat Rev Cancer, 2018, 18(8): 500-510.
[5] ANWAR S M, MAJID M, QAYYUM A,et al. Medical image analysis using convolutional neural networks: A review [J]. J Med Syst, 2018, 42(11): 226.
[6] NIE K, CHEN J H, YU H J, et al. Quantitative analysis of lesion morphology and texture features for diagnostic prediction in breast MRI[J]. Acad Radiol, 2008, 15(12): 1513-1525.
[7] LANGN, SU M, YU H, et al. Differentiation of myeloma and metastatic cancer in the spine using dynamic contrast-enhanced MRI[J]. Magn Reson Imaging, 2013, 31(8): 1285-1291.
[8] CHITALIA R D, KONTOS D. Role of texture analysis in breast MRI as a cancer biomarker: A review[J]. J Magn Reson Imaging, 2018, 49(4): 927-938.
[9] HARALICK R M, SHANMUGAM K, DINSTEIN I. Textural features for image classification[J]. IEEE Trans Syst Man Cybern, 1973, 3(6): 610-621.
[10] LANG N, ZHANG Y, ZHANG E, et al. Differentiation of spinal metastases originated from lung and other cancers using radiomics and deep learning based on DCE-MRI[J]. Magn Reson Imaging, 2019, 64: 4-12.
[11] HE K, ZHANG X, REN S, et al. Deep residual learning for image recognition[EB/OL].(2015-12-10)[2019-09-05]. https://arxiv.org/pdf/1512.03385.pdf.
[12] SIMONYAN K, ZISSERMAN A. Very deep convolutional networks for large-scale image recognition[EB/OL].(2015-04-10)[2019-09-05]. https://arxiv.org/pdf/1409.1556.pdf.
[13] Chollet F. Xception: Deep learning with depthwise separable convolutions[EB/OL].(2017-04-04)[2019-09-05]. https://arxiv.org/pdf/1610.02357.pdf.
[14] SZEGEDY C, VANHOUCKE V, IOFFE S, et al. Rethinking the inception architecture for computer vision[EB/OL].(2015-12-11)[2019-09-05].https://arxiv.org/pdf/1512.00567.pdf.
[15] KINGMA D, BA J. Adam: A method for stochastic optimization[EB/OL]. (2017-01-30)[2019-09-05]. https://arxiv.org/pdf/1412.6980.pdf.
[16] DENG J, DONG W, SOCHER R, et al. Imagenet: A large-
scale hierarchical image database[EB/OL].(2009-08-18)[2019-09-05]. https://ieeexplore.ieee.org/document/ 5206848.
[17] BURNSIDE E S, DRUKKER K, LI H, et al. Using computer-extracted image phenotypes from tumors on breast magnetic resonance imaging to predict breast cancer pathologic stage[J]. Cancer, 2016, 122(5): 748-757.
[18] LI H, ZHU Y, BURNSIDE E S, et al. Quantitative MRI radiomics in the prediction of molecular classifications of breast cancer subtypes in the TCGA/TCIA data set[J]. NPJ Breast Cancer, 2016, 2: 16012.
[19] BECKER A S, MARCON M, GHAFOOR S, et al. Deep learning in mammography diagnostic accuracy of a multipurpose image analysis software in the detection of breast cancer[J]. Invest Radiol, 2017, 52(7): 434-440.
[20] TRUHN D, SCHRADING S, HAARBURGERR C, et al. Radiomic versus convolutional neural networks analysis for classification of contrast-enhancing lesions at multiparametric breast MRI[J]. Radiology, 2018, 290(2): 181352.
[21] WU J S, SHENG S R, LIANG X H, et al. The role of tumor microenvironment in collective tumor cell invasion[J]. Futur Oncol, 2017, 13(11): 991-1002.