[1]梁翠霞,李明强,边兆英,等.基于深度学习特征的乳腺肿瘤分类模型评估[J].南方医科大学学报,2019,(01):88.[doi:10.12122/j.issn.1673-4254.2019.01.14]
点击复制

基于深度学习特征的乳腺肿瘤分类模型评估()
分享到:

《南方医科大学学报》[ISSN:1673-4254/CN:44-1627/R]

卷:
期数:
2019年01期
页码:
88
栏目:
出版日期:
2019-01-26

文章信息/Info

Title:
Establishment of a deep feature-based classification model for distinguishing benign and malignant breast tumors on full-filed digital mammography
作者:
梁翠霞李明强边兆英吕闻冰曾栋马建华
关键词:
乳腺肿瘤全数字乳腺成像计算机辅助诊断深度学习放射组学
Keywords:
breast tumors full-filed digital mammography computer-aided diagnosis deep learning radiomics
DOI:
10.12122/j.issn.1673-4254.2019.01.14
摘要:
目的本文结合深度学习特征(DF)和传统图像特征(HCF)特点,利用多分类器融合的方法建立一个乳腺肿瘤分类模型,并 深入评估和分析不同深度学习网络特征的肿瘤分类性能。方法回顾性分析106例乳腺肿瘤患者的头尾位和内外倾斜位投影 的全数字乳腺成像数据。首先从肿瘤区域提取23维HCF(12维形态及11维纹理特征),用t检验进行显著性特征选择;然后分别 从3 个卷积神经网络模型提取不同维度DF,在实验中,3 个不同深度学习网络产生了相应DF,分别是AlexNet,VGG16 和 GoogLeNet;最后结合2个投影数据的DF和HCF,采用多分类器的融合模型对特征进行训练和测试,实验重点分析不同DF在 肿瘤分类上的性能。结果结合DF和HCF建立的分类模型比使用单独HCF的分类模型表现出更好的性能;相比于其它网络框 架,DFAlexNet和HCF的结合表现出更高精度的分类结果。结论结合DF和HCF的特征方法建立一个分类模型,对于良恶性乳腺 肿瘤具有优秀的鉴别能力,且泛化能力更强,能作为临床辅助诊断工具。
Abstract:
Objective To develop a deep features-based model to classify benign and malignant breast lesions on full- filed digital mammography. Method The data of full-filed digital mammography in both craniocaudal view and mediolateral oblique view from 106 patients with breast neoplasms were analyzed. Twenty-three handcrafted features (HCF) were extracted from the images of the breast tumors and a suitable feature set of HCF was selected using t-test. The deep features (DF) were extracted from the 3 pre-trained deep learning models, namely AlexNet, VGG16 and GoogLeNet. With abundant breast tumor information from the craniocaudal view and mediolateral oblique view, we combined the two extracted features (DF and HCF) as the two-view features. A multi-classifier model was finally constructed based on the combined HCF and DF sets. The classification ability of different deep learning networks was evaluated. Results Quantitative evaluation results showed that the proposed HCF+DF model outperformed HCF model, and AlexNet produced the best performances among the 3 deep learning models. Conclusion The proposed model that combines DF and HCF sets of breast tumors can effectively distinguish benign and malignant breast lesions on full-filed digital mammography.

相似文献/References:

[1]徐蕾,白中红,许软成,等.Her-2、EGFR、PS-2、ER在乳腺癌中的表达及其临床意义[J].南方医科大学学报,2006,(02):231.
 XU Lei,BAI Zhong-hong,XU Ruan-cheng,et al.Expressions of Her-2, EGFR, PS-2 and ER in breast cancer and their clinical implications[J].Journal of Southern Medical University,2006,(01):231.
[2]王晓玉,林绍强,李君武,等.细胞表面唾液酸及其连接方式对乳腺癌细胞黏附的影响[J].南方医科大学学报,2006,(06):742.
 WANG Xiao-yu,LIN Shao-qiang,LI Jun-wu,et al.Effect of cell surface sialic acid and their linkages on adhesion of mammary carcinoma cells[J].Journal of Southern Medical University,2006,(01):742.
[3]徐学虎,苏榕,伍尚标,等.早期乳腺癌保乳手术36例疗效评价[J].南方医科大学学报,2006,(09):1378.
[4]吴武军,曾健,潘承恩,等.活性炭-表阿霉素混悬液对乳腺癌腋窝淋巴结转移的治疗作用[J].南方医科大学学报,2006,(12):1812.
 WU Wu-jun,ZENG Jian,PAN Cheng-en Department of General Surgery,et al.Efficacy of activated charcoal-epirubicin suspension for treatment of breast cancer with axillary metastasis[J].Journal of Southern Medical University,2006,(01):1812.
[5]康华峰,王西京,刘小旭,等.塞来昔布预防大鼠乳腺癌发生及其机制[J].南方医科大学学报,2006,(11):1599.
 KANG Hua-feng,WANG Xi-jing,LIU Xiao-xu,et al.Chemopreventive effect of celecoxib against DMBA-induced breast cancer and its mechanism[J].Journal of Southern Medical University,2006,(01):1599.
[6]王小拍,吴兴平,颜黎栩,等.血清miR-103可作为乳腺癌潜在的诊断标记物[J].南方医科大学学报,2012,(05):631.
[7]李荣,郑航,罗荣城.HER2基因的克隆及其在MCF-7细胞中的表达[J].南方医科大学学报,2005,(10):1264.
 LI Rong,ZHENG Hang,LUO Rong-cheng.Cloning and expression of the HER2 gene in MCF-7 cells[J].Journal of Southern Medical University,2005,(01):1264.
[8]郑航,罗荣城.TPS、CA-153和CEA联合检测对乳腺癌的诊断价值[J].南方医科大学学报,2005,(10):1293.
 ZHENG Hang,LUO Rong-cheng.Diagnostic value of combined detection of TPS, CA153 and CEA in breast cancer[J].Journal of Southern Medical University,2005,(01):1293.
[9]苏国强,黄宗海,刘志锋,等.重组腺病毒驱动KDR-CDglyTK融合基因系统对MCF-7细胞及血管内皮细胞的靶向杀伤作用[J].南方医科大学学报,2004,(12):1346.
 SU Guo-qiang,HUANG Zong-hai,LIU Zhi-feng,et al.Adenovirus-mediated CDglyTK fusion gene system driven by KDR promoter selectively kills MCF-7 breast cancer cells and vascular endothelial cells[J].Journal of Southern Medical University,2004,(01):1346.
[10]黄宗海,苏国强,毛乾国,等.血管内皮生长因子与nm23-H1基因在青年女性乳腺癌组织中的表达及其与预后之间的关系[J].南方医科大学学报,2004,(12):1398.
 HUANG Zong-hai,SU Guo-qiang,MAO Qian-guo,et al.Expressions of vascular endothelial growth factor and nm23-H1 gene and their relation to the prognosis of breast cancer in young women[J].Journal of Southern Medical University,2004,(01):1398.

更新日期/Last Update: 1900-01-01