摘要
目的:本文基于影像组学方法,用常规灰阶超声(GS-US)图像对原发性肝细胞癌(HCC)的微血管侵犯(MVI)指标和肿瘤分化等级进行预测。方法:根据影像组学的基本流程,本研究通过四个步骤:图像分割、特征提取、特征筛选和分类判别,分别建立对经手术病理证实的HCC病人的两个指标的预测模型并进行回顾性预测。结果:结果表明,肝脏的二维图像和MVI以及分化等级之间存在相关性,通过留一法(LOOCV)使用支持向量机(SVM)进行预测,受试者操作特性曲线下的面积(ROC)分别达到0.76(MVI)、0.89(肿瘤分化)。结论:灰阶超声图像蕴含和MVI以及肿瘤分化相关的相关信息,对HCC的灰阶超声图像的影像组学研究有助于患者的术前诊断和预后预测。
Purpose: The purpose of this study is to predict micro vascular invasion(MVI) and differentiation grade of hepatocellular carcinoma(HCC) by gray scale ultrasound(GS-US) images based on a radiomics approach. Methods: According to the routine of radiomics, our study built two predicting models for MVI and differentiation grade of histopathologically proved HCC patients respectively by 4 steps: image segmentation, feature extraction, feature selection and classification. Results: The results showed that, there were correlative relationships between GS-US images of HCC patients and MVI or tumor differentiation grade. By SVM with LOOCV, models produced AUC of 0.76(MVI) and 0.89(tumor differentiation grade) respectively. Conclusion: GS-US images can reveal the information about MVI and tumor differentiation grade. The radiomics study of HCC might be helpful for clinical diagnosis and prognosis prediction.
引文
[1]刘晨,孙德光,高振明,等.原发性肝细胞性肝癌根治术后序贯性肝动脉化疗栓塞术的治疗进展.肝胆外科杂志,2012,20:308-310.
[2]Yang J D,Roberts L R.Hepatocellular carcinoma:a global view.Nature Reviews Gastroenterology&Hepatology,2010,7:448.
[3]Edmondson H A,Steiner P E.Primary carcinoma of the liver:a study of 100 cases among 48,900 necropsies.Cancer,1954,7:462.
[4]张梅,徐雪房,李艳明,等.CT联合血清指标对不同组织分化程度肝细胞肝癌患者T ACE术后的多元分析评价.滨州医学院学报,2014:327-330.
[5]中国抗癌协会肝癌专业委员会.原发性肝细胞肝癌规范化病理诊断指南(2015年版).解放军医学杂志,2015,21:833-839.
[6]金赟,李江涛.肝癌细胞侵犯微血管的临床相关因素及分子标志物的研究进展.临床肝胆病杂志,2013,29:550-553.
[7]刘臻玉,武丹,区锦玲,等.肝癌微血管侵犯的特点及相关临床指标.肝胆胰外科杂志,2017,29:107-111.
[8]Lambin P,Riosvelazquez E,Leijenaar R,et al.Radiomics:Extracting more information from medical images using advanced feature analysis.European Journal of Cancer,2012,48:441-446.
[9]Yu J,Shi Z,Lian Y,et al.Noninvasive IDH1 mutation estimation based on a quantitative radiomics approach for grade II glioma.European Radiology,2016,26:1-14.
[10]Aerts H J,Velazquez E R,Leijenaar R T,et al.Decoding tumour phenotype by noninvasive imaging using a quantitative radiomics approach.Nature Communications,2014,5:4006.
[11]Li Y,Namburi P,Yu Z,et al.Voxel selection in f MRI data analysis based on sparse representation.IEEE transactions on bio-medical engineering,2009,56:2439.
[12]Gouw A S,Balabaud C,Kusano H,et al.Markers for microvascular invasion in hepatocellular carcinoma:where do we stand?.Liver Transplantation,2011,17 Suppl 2(Supplement S2):S72.
[13]Chou C T,Chen R C,Lee C W,et al.Prediction of microvascular invasion of hepatocellular carcinoma by pre-operative CT imaging.British Journal of Radiology,2012,85:778-783.
[14]胡国祥,郑丽民.肝细胞性肝癌CT表现与肿瘤病理分化相关性研究.中国中西医结合消化杂志,2013,21:537-538.
[15]骆雁翎,熊文文,张迎春,等.不同分化程度肝癌患者的超声造影参数分析.医学临床研究,2017,34:381-383.