• “中国科技期刊卓越行动计划”中文领军期刊
  • 百种中国杰出学术期刊
  • 中国百强报刊
  • RCCSE中国核心学术期刊(A+)
  • 中国自然科学类核心期刊
  • 中国高校百佳科技期刊
  • 中国精品科技期刊
  • 中国科技论文统计源期刊
  • 中华医学会优秀期刊
  • 中国精品科技期刊顶尖学术论文(5000)项目来源期刊
  • 入选中国高质量科技期刊分级目录(消化病学)T1级
  • 入选《中国学术期刊影响因子年报(自然科学与工程技术)》Q1区
  • 入选《科技期刊世界影响力指数(WJCI)报告(2022)》
  • “中国科技期刊卓越行动计划”中文领军期刊
  • 百种中国杰出学术期刊
  • 中国百强报刊
  • RCCSE中国核心学术期刊(A+)
  • 中国自然科学类核心期刊
  • 中国高校百佳科技期刊
  • 中国精品科技期刊
  • 中国科技论文统计源期刊
  • 中华医学会优秀期刊
  • 中国精品科技期刊顶尖学术论文(5000)项目来源期刊
  • 入选中国高质量科技期刊分级目录(消化病学)T1级
  • 入选《中国学术期刊影响因子年报(自然科学与工程技术)》Q1区
  • 入选《科技期刊世界影响力指数(WJCI)报告(2022)》

增强CT检查预测肝细胞癌微血管侵犯的影像基因组学研究

赵建新, 潘妮妮, 何迪梁, 施柳言, 何炫明, 熊恋秋, 马丽丽, 崔雅琼, 赵莲萍, 黄刚

赵建新, 潘妮妮, 何迪梁, 等. 增强CT检查预测肝细胞癌微血管侵犯的影像基因组学研究[J]. 中华消化外科杂志, 2023, 22(11): 1367-1377. DOI: 10.3760/cma.j.cn115610-20231013-00148
引用本文: 赵建新, 潘妮妮, 何迪梁, 等. 增强CT检查预测肝细胞癌微血管侵犯的影像基因组学研究[J]. 中华消化外科杂志, 2023, 22(11): 1367-1377. DOI: 10.3760/cma.j.cn115610-20231013-00148
Zhao Jianxin, Pan Nini, He Diliang, et al. Radiogenomics of enhanced CT imaging to predict microvascular invasion in hepatocellular carcinoma[J]. Chinese Journal of Digestive Surgery, 2023, 22(11): 1367-1377. DOI: 10.3760/cma.j.cn115610-20231013-00148
Citation: Zhao Jianxin, Pan Nini, He Diliang, et al. Radiogenomics of enhanced CT imaging to predict microvascular invasion in hepatocellular carcinoma[J]. Chinese Journal of Digestive Surgery, 2023, 22(11): 1367-1377. DOI: 10.3760/cma.j.cn115610-20231013-00148

增强CT检查预测肝细胞癌微血管侵犯的影像基因组学研究

基金项目: 

甘肃省人民医院院内科研基金 22GSSYD‑34

详细信息
    通讯作者:

    黄刚,Email:keen0999@163.com

Radiogenomics of enhanced CT imaging to predict microvascular invasion in hepatocellular carcinoma

Funds: 

Intramural Research Fund of Gansu Provincial Hospital 22GSSYD‑34

More Information
  • 摘要:
    目的 

    构建基于术前增强CT检查的联合影像组学模型,预测肝细胞癌微血管侵犯(MVI)状态,对影像组学模型进行生物学解释。

    方法 

    采用回顾性队列研究方法。收集癌症基因组图谱数据库建库至2023年1月纳入的424例肝细胞癌患者的mRNA数据,癌症图像档案馆数据库建库至2023年1月纳入的39例肝细胞癌患者和甘肃省人民医院2020年1月至2023年1月收治53例肝细胞癌患者的临床病理资料。92例肝细胞癌患者通过随机数字表法按7∶3分为训练集64例和测试集28例。分析动脉期及门静脉期CT检查图像及临床资料。使用3Dslicer软件(5.0.3版本)进行动脉期和门静脉期图像配准和三维感兴趣区勾画。使用开源软件FAE(0.5.5版本)对原始图像进行预处理并提取特征。通过最小绝对收缩和选择算子等方法筛选特征,构建影像组学模型并计算影像组学评分(R‑score),通过Logistic回归整合临床参数、影像学特征及R‑score构建列线图。通过加权基因共表达网络分析和相关性分析获取影像组学模型相关的基因模块并进行富集分析。观察指标:(1)不同MVI性质患者的临床特征比较。(2)MVI风险模型的建立。(3)MVI风险模型的评估。(4)基因模块聚类。(5)特征相关基因模块功能富集。正态分布的计量资料以x±s表示,组间比较采用独立样本t检验,偏态分布的计量资料以M(范围)表示,组间比较采用Mann⁃Whitney U检验,计数资料比较采用χ²检验。采用组内和组间相关系数(ICC)评估影像组学特征提取的观察者间的一致性。ICC>0.75表示特征提取的一致性良好。单因素和多因素分析采用Logistic回归模型。绘制受试者工作特征曲线,以曲线下面积(AUC)、决策曲线、校准曲线评估模型的诊断效能及临床实用性。

    结果 

    (1)不同MVI性质患者的临床特征比较。92例肝细胞癌患者中,MVI阳性47例,MVI阴性45例,两者肝炎、肿瘤长径、瘤周增强、瘤内动脉、假包膜及瘤周不光滑比较,差异均有统计学意义(χ²=5.308,9.977,47.370,32.368,21.105,31.711,P<0.05)。(2)MVI风险模型的建立。在动脉期及门静脉期的瘤内和瘤周分别提取了1 781个特征,经过特征降维后,从动脉期及门静脉期中确定8个影像组学特征构建联合模型。多因素分析结果显示:瘤周增强、瘤内动脉、假包膜、瘤周不光滑及R‑score是肝细胞癌患者MVI的独立危险因素[风险比=0.049,0.017,0.017,0.021,2.539,95%可信区间(CI)为0.005~0.446,0.001~0.435,0.001~0.518,0.001~0.473,1.220~3.283]。纳入瘤周增强、瘤内动脉、假包膜、瘤周不光滑及R‑score构建列线图模型。(3)MVI风险模型的评估。R‑score在训练集和测试集中AUC分别为0.923(95%CI为0.887~0.944)和0.918(95%CI为0.894~0.945);联合R‑score及影像学特征构建的列线图在训练集和测试集中AUC分别为0.973(95%CI为0.954~0.988)和0.962(95%CI为0.942~0.987)。决策曲线显示:列线图的临床效益优于R‑score。校准曲线显示:列线图和R‑score预测状态与实际观察结果间一致性良好。(4)基因模块聚类。经加权基因共表达网络分析后获取8个基因模块。(5)特征相关基因模块功能富集。4个基因模块与影像组学特征显著相关。预测MVI的影像组学特征可能与细胞周期、中性粒细胞外陷阱形成及PPAR信号通路有关。

    结论 

    基于术前增强CT检查的联合影像组学模型可以预测肝细胞癌MVI状态。通过获取影像组学特征相关的mRNA基因表达谱,为影像组学模型提供了生物学解释。

    Abstract:
    Objective 

    To construct a combined radiomics model based on preoperative enhanced computed tomography (CT) examination for predicting microvascular invasion (MVI) in hepatocellular carcinoma (HCC), and provide biological explanations for the radiomics model.

    Methods 

    The retrospective cohort study was conducted. The messenger RNA (mRNA) of 424 HCC patients, the clinicopathological data of 39 HCC patients entered into the Cancer Genome Atlas database from its establishment until January 2023, and the clinicopathological data of 53 HCC patients who were admitted to the Gansu Provincial People′s Hospital from January 2020 to January 2023 were collected. The 92 HCC patients were randomly divided into a training dataset of 64 cases and a test dataset of 28 cases with a ratio of 7∶3 based on a random number table method. The CT images of patients in the arterial phase and portal venous phase as well as the corresponding clinical data were analyzed. The 3Dslicer software (version 5.0.3) was used to register the CT images in the arterial phase and portal venous phase and delineate the three‑dimensional regions of interest. The original images were preprocessed and the corresponding features were extracted by the open‑source software FAE (version 0.5.5). After selecting features using the Least Absolute Shrinkage and Selection Operator, the radiomics model was constructed and the radiomics score (R‑score) was calculated. The nomogram was constructed by integrating clinical parameters, imaging features and R‑score based on Logistic regression. The gene modules related to radiomics model were obtained and subjected to enrichment analysis by conducting weighted gene co‑expression network analysis and correlation analysis. Observation indicators: (1) comparison of clinical characteristics of patients with different MVI properties; (2) establishment of MVI risk model; (3) evaluation of MVI risk model; (4) clustering of gene modules; (5) functional enrichment of feature‑correlated gene modules. Measurement data with normal distribution were represented as Mean±SD, and comparison between groups was conducted using the independent sample t test. Measurement data with skewed distribution were represented as M(range), and comparison between groups was conducted using the Mann-Whitney U test. Comparison of count data was conducted using the chi‑square test. The intra‑/inter-class correlation coefficient (ICC) was used to assess the inter‑observer consistency of radiomics feature extracted by different observers. ICC >0.75 indicated a good consistency in feature extraction. The Logistic regression model was used for univariate and multivariate analyses. The receiver operating characteristic curve was drawn, and the area under curve (AUC), the decision curve and the calibration curve were used to evaluate the diagnostic efficacy and clinical practicality of the model.

    Results 

    (1) Comparison of clinical characteristics of patients with different MVI properties. Of 92 HCC patients, there were 47 cases with MVI‑positive and 45 cases with MVI‑negative, and there were significant differences in hepatitis, tumor diameter, peritumoral enhancement, intratumoral arteries, pseudocapsule and smoothness of tumor margin between them (χ²=5.308, 9.977, 47.370, 32.368, 21.105, 31.711, P<0.05). (2) Establishment of MVI risk model. A total of 1 781 features were extrac-ted from arterial and portal venous phases of the intratumoral and peritumoral regions. After feature dimension reduction, 8 radiomics features were selected from arterial and portal venous phases to construct the combined model. Results of multivariate analysis showed that peritumoral enhancement, intratumoral arteries, pseudocapsule, smoothness of tumor margins, and R‑score were independent risk factors for MVI in patients with HCC [hazard ratio=0.049, 0.017, 0.017, 0.021, 2.539, 95% confidence interval (CI) as 0.005-0.446, 0.001-0.435, 0.001-0.518, 0.001-0.473, 1.220-5.283, P<0.05]. A nomogram model was constructed incorporating peritumoral enhancement, intratumoral arteries, pseudocapsule, smoothness of tumor margins, and R‑score. (3) Evaluation of the MVI risk model. The AUC of radiomics model was 0.923 (95%CI as 0.887-0.944) and 0.918 (95%CI as 0.894-0.945) in the training dataset and test dataset, respectively. The AUC of nomogram model, incorpora-ting both the R‑score and radiomics features, was 0.973 (95%CI as 0.954-0.988) and 0.962 (95%CI as 0.942-0.987) in the training dataset and test dataset, respectively. Results of decision curve showed that the nomogram had better clinical utility compared to the R‑score. Results of calibration curve showed good consistency between the actual observed outcomes and the nomogram or the R‑score. (4) Clustering of gene module. Results of weighted gene co‑expression network analysis showed that 8 gene modules were obtained. (5) Functional enrichment of feature‑related gene modules. Results of correlation analysis showed 4 gene modules were significantly associated with radiomics features. The radiomics features predicting of MVI may be related to pathways such as the cell cycle, neutrophil extracellular trap formation, and PPAR signaling pathway.

    Conclusions 

    The combined radiomics model based on preoperative enhanced CT imaging can predict the MVI status of HCC. By obtaining mRNA gene expression profiles associated with radiomics features, a biological interpretation of the radiomics model is provided.

  • 肝细胞癌具有高度侵袭性,肝切除术后70%的患者复发,肝移植后25%患者复发,5年总生存率约为10%~20%[15]。微血管侵犯(microvascular invasion,MVI)是肝细胞癌术后复发的危险因素[611]。术后病理学检查是MVI诊断的金标准,若在术前诊断MVI可以指导治疗方案的选择,进而改善患者预后[1215]。术前CT检查中部分影像学特征具有预测MVI的价值,但严重依赖观察者的经验判断[3,1618]。影像组学通过计算机高通量提取定量成像特征,已经在术前MVI和预后预测中展现出广泛的应用前景[8,16,1921]。本研究回顾性分析癌症基因组图谱(The Cancer Genome Atlas,TCGA)数据库建库至2023年1月纳入的424例肝细胞癌患者的生物学信息,癌症图像档案馆(The Cancer Imaging Archive,TCIA)数据库建库至2023年1月纳入的39例肝细胞癌患者和甘肃省人民医院2020年1月至2023年1月收治53例肝细胞癌患者的临床病理资料,构建基于术前增强CT检查的联合影像组学模型,预测肝细胞癌MVI状态,对影像组学模型进行生物学解释。

    采用回顾性队列研究方法。收集TCGA数据库424例肝细胞癌患者的mRNA数据,收集TCIA数据库纳入的73例及甘肃省人民医院收治的103例肝细胞癌患者的临床病理资料。TCIA数据库和甘肃省人民医院经遴选最终分别纳入39例和53例患者。92例肝细胞癌患者通过随机数字表法按7∶3分为训练集64例和测试集28例,训练集用于构建预测模型,测试集用于测试预测模型效能。两组患者年龄、AFP、肝炎、肿瘤长径、瘤周增强、瘤内动脉、假包膜及瘤周不光滑比较,差异均无统计学意义(P>0.05)。见表1。本研究通过甘肃省人民医院医学伦理委员会审批,批号为2022⁃432,患者及家属均签署知情同意书。

    表  1  92例训练集和测试集肝细胞癌患者的临床及影像学特征比较
    Table  1.  Comparison of clinical and radiological characteristics of 92 patients with hepatocellular carcinoma in the training dataset and the test dataset
    临床及影像学特征赋值例数训练集(64例)测试集(28例)统计量值P
    性别(例)
    0664224χ²=3.8770.049
    126224
    年龄(x±s,岁)--57±1260±13t=-1.0290.306
    甲胎蛋白(例)
    ≤400 ng/L0594019χ²=0.2430.622
    >400 ng/L133249
    肝炎(例)
    036279χ²=0.8250.364
    有(HBV/HCV)1563719
    肿瘤长径(例)
    ≤5 cm046297χ²=3.3740.066
    >5 cm1463521
    瘤周增强(例)
    0463115χ²=0.2050.650
    1463313
    瘤内动脉(例)
    026206χ²=0.9270.336
    1664422
    假包膜(例)
    0473611χ²=2.2430.134
    1452817
    瘤周不光滑(例)
    0463115χ²=0.2050.650
    1463313
    注:HBV为乙型病毒肝炎;HCV为丙型病毒肝炎;“-”表示此项无
    下载: 导出CSV 
    | 显示表格

    纳入标准:(1)组织病理学检查确诊为肝细胞癌。(2)既往无干预治疗或肝部分切除史。(3)术前行增强CT检查。(4)影像学检查无肉眼可见门静脉、肝静脉、肝动脉及胆管癌栓。(5)术前临床资料完整。

    排除标准:(1)增强CT检查图像质量不佳、存在伪影等无法进行分析。(2)术后组织病理学检查结果无MVI诊断。(3)合并其他系统恶性肿瘤。(4)术前临床资料缺失。

    收集临床参数为性别、年龄、肝炎病毒感染史(0,阴性;1,HBV史、HCV史或HBV+HCV史)、AFP(0,≤400 ng/L;1,>400 ng/L)。影像学特征包括由放射科医师识别的5个特征:肿瘤长径、瘤周增强、瘤内动脉、假包膜、瘤周不光滑。

    将CT检查图像导入3Dslicer软件(5.0.3版本;https://www.slicer.org)进行动脉期和门静脉期图像配准和三维感兴趣区勾画。由1位具有7年诊断经验的放射科医师在门静脉期图像沿肿瘤边界进行全肿瘤分割,然后对三维感兴趣区进行10 mm全自动外扩并手动擦除肝外区域。勾画过程如图1。间隔3个月后,由同一医师随机选择30例图像进行三维感兴趣区分割。所有医师均不知晓病理学检查结果。使用基于pyradiomics的开源软件FAE(0.5.5版本;https://github.com/salan668/FAE)对原始图像进行预处理,体素重采样为1 mm×1 mm×1 mm,灰度级别离散体素强度设置为16,然后提取一阶、形态、纹理及高阶特征。

    图  1  肝细胞癌患者三维感兴趣区勾画与病理学检查结果 1A:微血管侵犯阴性患者增强CT检查图像;1B:微血管侵犯阴性患者沿横轴位逐层勾画示意图;1C:微血管侵犯阴性患者三维感兴趣区示意图;1D:微血管侵犯阴性患者病理学检查结果 HE染色 低倍放大;1E:微血管侵犯阳性患者增强CT检查图像;1F:微血管侵犯阳性患者沿横轴位逐层勾画示意图;1G:微血管侵犯阳性患者三维感兴趣区示意图;1H:微血管侵犯阳性患者病理学检查结果 HE染色 低倍放大
    Figure  1.  Delineation of three‑dimensional regions of interest and pathological examination in patient with hepatocellular carcinoma 1A: Enhanced computed tomography (CT) image of patient with negative microvascular invasion (MVI); 1B: Schematic diagram of sequential delineation of regions of interest along the transverse axis of patient with negative MVI; 1C: Three‑dimensional representa⁃tion of regions of interest of patient with negative MVI; 1D: Pathological examination of patient with negative MVI HE staining Low magnification; 1E: Enhanced CT image of patient with positive MVI; 1F: Schematic diagram of sequential delineation of regions of interest along the transverse axis of patient with positive MVI; 1G: Three‑dimensional representation of regions of interest of patient with positive MVI; 1H: Pathological examination of patient with positive MVI HE staining Low magnification

    使用Z‑score算法对所有特征进行标准化,通过Mann‑Whitney U检验或χ²检验进行特征筛选,选择皮尔森相关系数<0.8的特征,使用最大相关最小冗余算法筛选前20个特征,通过最小绝对收缩和选择算子(least absolute shrinkage and selection opera⁃tor,LASSO)进行特征降维。

    应用Logistic回归构建影像组学模型:(1)基于动脉期和门静脉期的瘤内和瘤周特征分别构建模型,并比较各期的模型效能。(2)基于动脉期及门静脉期特征构建联合模型。根据影像组学特征及系数计算影像组学评分(radiomics score,R‑score)。对临床变量、影像学特征及R‑score进行单因素和多因素分析,确定MVI危险因素并构建列线图。使用受试者工作特征(receiver operating characteristic curve,ROC)曲线的曲线下面积(area under the curve,AUC)量化模型的预测性能,进行决策曲线分析评估模型的临床实用性,通过校准曲线观察预测概率和实际结果的一致性。

    使用R软件(4.3.0版本)的limma软件包处理TCGA数据库中肝细胞癌患者的mRNA数据以鉴定差异表达基因,选取错误发现率<0.05且倍数变化>1的差异表达基因,随后将差异表达基因用于加权基因共表达网络分析进行下一步分析。

    使用R软件(4.3.0版本)的加权基因共表达网络分析软件包进行加权基因共表达网络分析。选择软合适的阈值,使得基因表达关系符合无尺度网络,基于软阈值构建邻近矩阵及拓扑重叠矩阵。根据相异度对基因进行聚类,使用动态剪切法将聚类基因分为不同模块。将模块内的基因数设置为≥30个,基因模块合并的高度阈值设置为0.25。

    将TCGA‑TCIA数据库中同时拥有mRNA及影像学数据的患者进行Pearson相关性分析,构建影像组学特征和基因模块的影像基因组图谱。使用R软件(4.3.0版本)中的clusterProfiler软件包进行基因本体论(Gene Ontology,GO)富集分析、京都基因和基因组百科全书(Kyoto Encyclopedia of Genes and Gnomes,KEGG)通路分析,注释影像组学特征相关模块的生物学功能,为影像组学模型提供可能的生物学解释(错误发现率<0.05)。

    (1)不同MVI性质患者的临床特征比较:性别、年龄、AFP、肝炎、肿瘤长径、瘤周增强、瘤内动脉、假包膜及瘤周不光滑。(2)MVI风险模型的建立:影像组学特征提取数量、最终筛选数量及MVI相关危险因素。(3)MVI风险模型的评估:动脉期、门静脉期及列线图模型在训练集和测试集中的效能。(4)基因模块聚类:聚类模块数量。(5)特征相关基因模块功能富集:影像组学特征相关模块数量及基因富集通路。

    应用R软件(4.3.0版本)和SPSS 26.0统计软件进行分析。正态分布的计量资料以x±s表示,组间比较采用独立样本t检验,偏态分布的计量资料以M(范围)表示,组间比较采用Mann‑Whitney U检验。计数资料比较采用χ²检验。采用组内和组间相关系数(intra‑/inter‑class correlation coefficients,ICC)评估影像组学特征提取的观察者间的一致性。ICC>0.75表示特征提取的一致性良好。单因素和多因素分析采用Logistic回归模型。绘制ROC曲线,以AUC、决策曲线、校准曲线评估模型的诊断效能及临床实用性。P<0.05为差异有统计学意义。

    92例肝细胞癌患者中,MVI阳性47例,MVI阴性45例,两者肝炎、肿瘤长径、瘤周增强、瘤内动脉、假包膜及瘤周不光滑比较,差异均有统计学意义(P<0.05);而两者性别、年龄、AFP比较,差异均无统计学意义(P>0.05)。见表2

    表  2  92例微血管侵犯阴性和阳性肝细胞癌患者的临床及影像学特征比较
    Table  2.  Comparison of clinical and radiological characteristics of 92 hepatocellular carcinoma patients with negative and positive microvascular invasion
    临床及影像学特征赋值例数微血管侵犯阴性(45例)微血管侵犯阳性(47例)统计量值P
    性别(例)
    0663036χ²=1.1180.290
    1261511
    年龄(x±s,岁)--59±1457±11t=0.9170.362
    甲胎蛋白(例)
    ≤400 ng/L0593326χ²=3.2430.072
    >400 ng/L1331221
    肝炎(例)
    0362313χ²=5.3080.021
    有(HBV/HCV)1562234
    肿瘤长径(例)
    ≤5 cm0362511χ²=9.9770.002
    >5 cm1562036
    瘤周增强(例)
    046397χ²=47.370<0.001
    146640
    瘤内动脉(例)
    026251χ²=32.368<0.001
    1662046
    假包膜(例)
    0473413χ²=21.105<0.001
    1451134
    瘤周不光滑(例)
    0463610χ²=31.711<0.001
    146937
    注:HBV为乙型肝炎病毒;HCV为丙型肝炎病毒;“-”表示此项无
    下载: 导出CSV 
    | 显示表格

    在动脉期及门静脉期的瘤内和瘤周分别提取了1 781个特征,经过特征降维后,从动脉期及门静脉期中确定8个影像组学特征构建联合模型。

    单因素分析结果显示:肝炎、肿瘤长径、瘤周增强、瘤内动脉、假包膜、瘤周不光滑、R‑score是肝细胞癌患者MVI的影响因素(P<0.05),见表3。多因素分析结果显示:瘤周增强、瘤内动脉、假包膜、瘤周不光滑及R‑score是肝细胞癌患者MVI的独立危险因素(P<0.05),见表4。纳入瘤周增强、瘤内动脉、假包膜、瘤周不光滑及R‑score构建列线图模型(图2)。

    表  3  影响64例训练集肝细胞癌患者微血管侵犯的单因素分析
    Table  3.  Univariate analysis of microvascular invasion in 64 patients with hepatocellular carcinoma of the training dataset
    临床及影像学因素b标准误Wald优势比95%可信区间P值
    甲胎蛋白(≤400 ng/L比>400 ng/L)-0.7980.4473.1890.4500.188~1.0810.074
    肝炎(无比有)-1.0060.4425.1810.3660.154~0.8700.023
    肿瘤长径(≤5 cm比>5 cm)-1.4090.4579.5100.2440.100~0.5980.002
    瘤周增强(无比有)-3.6150.60036.2790.0270.008~0.087<0.001
    瘤内动脉(无比有)-4.0521.05414.7670.0170.002~0.137<0.001
    假包膜(无比有)-2.0900.47619.2700.1240.049~0.315<0.001
    瘤周不光滑(无比有)-2.6950.51627.3060.0680.025~0.186<0.001
    影像组学评分0.6130.13620.1831.8451.413~2.411<0.001
    下载: 导出CSV 
    | 显示表格
    表  4  影响64例训练集肝细胞癌患者微血管侵犯的多因素分析
    Table  4.  Multivariate analysis of microvascular invasion in 64 patients with hepatocellular carcinoma of the training dataset
    临床及影像学因素b标准误Wald风险比95%可信区间P
    瘤周增强(无比有)-3.0071.1227.1780.0490.005~0.4460.007
    瘤内动脉(无比有)-4.0861.6606.0590.0170.001~0.4350.014
    假包膜(无比有)-4.0941.7545.4520.0170.001~0.5180.020
    瘤周不光滑(无比有)-3.8701.5935.9030.0210.001~0.4730.015
    影像组学评分0.9320.3746.2092.5391.220~5.2830.013
    下载: 导出CSV 
    | 显示表格
    图  2  肝细胞癌患者微血管侵犯的列线图预测模型
    Figure  2.  The nomogram predictive model for microvascular invasion in patients with hepatocellular carcinoma

    在瘤内和瘤周中,与动脉期影像组学模型比较,门静脉期影像组学模型具有更好的MVI预测效能。将门静脉期瘤内与瘤周影像组学特征相结合后,预测效能显著提高,训练集的AUC为0.914(95%CI为0.865~0.934),测试集的AUC为0.895(95%CI为0.845~0.933);通过联合模型(动脉期和门静脉期瘤内+瘤周)的影像组学特征计算R‑score,AUC在训练集及测试集分别为0.923(95%CI为0.887~0.944)和0.918(95%CI为0.894~0.945);列线图模型的AUC在训练集及测试集分别为0.973(95%CI为0.954~0.988)和0.962(95%CI为0.942~0.987)。见表5,图3

    表  5  肝细胞癌微血管侵犯预测的多期增强CT影像组学模型
    Table  5.  The radiomics model in predicting of microvascular invasion of hepatocellular carcinoma based on multi‑phase enhanced computed tomography
    参数训练集测试集
    曲线下面积95%可信区间准确度(%)特异度(%)灵敏度(%)曲线下面积95%可信区间准确度(%)特异度(%)灵敏度(%)
    A0.8180.776~0.84570.7770.9770.590.8120.768~0.85675.8675.0076.92
    A100.7740.723~0.80572.3173.5370.970.7740.743~0.82268.9769.2368.75
    V0.8540.821~0.88776.9278.7975.000.8760.805~0.88982.7680.0085.71
    V100.8410.795~0.87375.3866.6784.380.8330.798~0.87175.8678.5773.33
    A+A100.9040.864~0.92176.9278.1275.760.8760.835~0.92282.7680.0085.71
    V+V100.9140.865~0.93484.6281.2587.880.8950.845~0.93375.8680.0071.43
    R‑score0.9230.887~0.94483.0874.1991.180.9180.894~0.94586.2187.5084.62
    Nomogram0.9730.954~0.98892.3196.1593.550.9620.942~0.98786.1286.9293.75
    注:A为动脉期瘤内影像组学模型;A10为动脉期瘤周10 mm影像组学模型;V为门静脉期瘤内影像组学模型;V10为门静脉期瘤周10 mm影像组学模型;A+A10为动脉期瘤内+瘤周影像组学模型;V+V10为门静脉期瘤内+瘤周影像组学模型;R‑score为影像组学评分;Nomogram为列线图
    下载: 导出CSV 
    | 显示表格
    图  3  测试集的受试者工作特征(ROC)曲线 3A:动脉期模型ROC曲线;3B:门静脉期模型ROC曲线;3C:联合模型ROC曲线;3D:影像组学评分及列线图模型ROC曲线
    注:A为动脉期瘤内影像组学模型;A10为动脉期瘤周10 mm影像组学模型;V为门静脉期瘤内影像组学模型;V10为门静脉期瘤周10 mm影像组学模型;A+A10为动脉期瘤内+瘤周影像组学模型;V+V10为门静脉期瘤内+瘤周影像组学模型;A+A10+V+V10为动脉期及门静脉期瘤内+瘤周影像组学模型
    Figure  3.  The receiver operating characteristic (ROC) curve of test dataset 3A: ROC curve of arterial phase model; 3B: ROC curve of venous phase; 3C: ROC curve of combined model; 3D: ROC curve of R‑score and nomogram model

    决策曲线显示:列线图的临床效益优于R‑score;校准曲线显示:列线图和R‑score预测状态与实际观察结果间一致性良好。见图4,5

    图  4  影像组学评分与列线图预测肝细胞癌微血管侵犯的决策曲线 4A:训练集决策曲线;4B:测试集决策曲线
    Figure  4.  Decision curve of R⁃socre and nomogram in predicting microvascular invasion of hepatocellular carcinoma 4A: Decision curve in the training dataset; 4B: Decision curve in the test dataset
    图  5  影像组学评分与列线图预测肝细胞癌微血管侵犯的校准曲线 5A:训练集校准曲线;5B:测试集校准曲线
    Figure  5.  Calibration curve of R‑score and nomogram in predicting microvascular invasion hepatocellular carcinoma 5A: Calibration curve in the training dataset; 5B: Calibration curve in the test dataset

    使用LogFC为1的阈值进行mRNA差异分析,424例有基因表达数据的肝细胞癌患者获取5 350个差异基因。加权基因共表达网络分析结果显示:选择β=3作为构建无标度网络的合适软阈值,将差异基因聚类为8个基因模块。见图6A。

    图  6  生物信息学分析 6A:加权基因共表达网络分析获取8个基因模块;6B:基因模块与影像组学特征相关性分析;6C:显著相关基因的基因本体论富集分析;6D:显著相关基因的京都基因和基因组百科全书通路分析
    注:F1为A_exponential_gldm_DependenceVariance;F2为A_wavelet‑LHL_glcm_ClusterShade;F3为A_exponential_firstorder_Maximum;F4为A10_lbp‑2D_glrlm_LongRunEmphasis;F5为A10_wavelet‑LHH glszm ZoneEntropy;F6为V_lbp‑3D‑m2_glszm_GrayLevelVariance;F7为V_wavelet‑HLH_firstorder_Median;F8为V_wavelet‑LHH_glcm_Correlation
    Figure  6.  Bioinformatics analysis 6A: Eight gene modules obtaining by weighted gene co‑expression network analysis; 6B: Correlation analysis between gene modules and radiomics features; 6C: Gene ontology analysis of significantly correlated genes; 6D: Kyoto encyclopedia of genes and gnomes enrichment analysis of significantly correlated genes

    含有3 162个基因的4个基因模块(棕色、黄色、绿松石色、品红色)与影像组学特征显著相关(r>0.3,P<0.05)。见图6B。上述4个基因模块的GO和KEGG富集分析结果显示:GO主要富集在细胞器裂变、DNA复制、小分子代谢过程等通路;KEGG通路中,包括细胞周期、中性粒细胞外陷阱形成及PPAR信号通路等。见图6C~6D。

    AFP是肝细胞癌的肿瘤标志物,也是影响MVI的独立危险因素[2223]。已有研究结果显示:肝细胞癌患者血清AFP与MVI无相关性[19]。本研究多因素分析结果显示:肝细胞癌患者血清AFP不是影响MVI的独立因素,因此未被纳入列线图预测模型。笔者认为:AFP与MVI的相关性可能受医疗机构的血清学检测标准影响。

    影像学特征中,肿瘤长径、瘤周增强、瘤内动脉、假包膜、瘤周不光滑预测MVI的状态均有临床价值[3,2425]。本研究结果显示:肿瘤长径不是影响MVI的独立因素。其原因可能是本研究样本量较小造成影像数据选择偏差,也可能是较大的肿瘤易导致病理学检查取材偏差。

    MVI更易发生在肿瘤边缘,瘤周取材对病理学检查结果非常重要。本研究结果显示:与MVI病理改变区域不同,以肝细胞癌患者瘤周影像组学特征构建预测模型,其预测MVI的效能却低于基于瘤内影像组学模型,与已有文献报道结果类似[8,16,26]。目前对造成这种矛盾现象的原因缺乏合理解释。笔者团队认为:原因可能为病理学检查是直接观察MVI改变情况,而影像组学特征除反映MVI病理学改变之外,还反映肿瘤MVI本身的部分特征,这类影像组学特征比瘤周MVI病理改变区域的影像组学特征更能反映整个肿瘤的MVI状态。本研究结果显示:联合瘤内和瘤周影像组学特征构建的模型能够更好地预测MVI状态。在不同时相的增强CT检查图像中,门静脉期影像组学模型预测MVI效能优于动脉期[2729]。这可能与CT检查门静脉期扫描时间长于动脉期,提取的影像组学特征相对稳定有关。

    影像组学模型虽然对肝细胞癌患者术前MVI状态有较高预测效能,但是联合临床资料、影像学特征构建的列线图具有更高预测效能[16,30]。在肺癌、乳腺癌、宫颈癌等肿瘤的预测模型中也有类似结果[3133]。因此,对于MVI的预测模型而言,纳入更多的相关影响因素可建立更高效能的预测模型。

    影像组学特征相关基因在细胞周期、中性粒细胞外陷阱形成及PPAR信号通路中显著富集。细胞周期与肝细胞癌的生长及增殖相关,细胞周期调控相关的信号通路,如PI3K/Akt信号通路和Wnt/β‑catenin信号通路等也可以调节肝细胞癌的血管生成[3436]。在肝细胞癌的肿瘤微环境中,中性粒细胞可以被激活并释放中性粒细胞外陷阱形成[37]。而中性粒细胞外陷阱形成可使促血管生成因子增加促进血管生成,为肝细胞癌细胞提供侵入周围血管和转移的通道[38]。此外,中性粒细胞外陷阱形成还可以通过激活PLT和血管内皮细胞,促进微血管生成和血管壁通透性的增加,从而进一步促进MVI发生[3940]。PPAR信号通路在调节脂质代谢、炎症反应以及细胞增殖和分化等生理过程中起着重要作用[4143]。激活PPAR信号通路可以下调细胞增殖和侵袭相关基因的表达,从而降低肿瘤细胞对血管的侵袭[44]。因此,MVI发生可能与炎细胞浸润、细胞增殖及血管生成有关。已有研究结果显示:Ras信号通路可以调控MVI的发生发展,这条通路与细胞增殖和血管生成有关[45]。由此可见,肝细胞癌MVI的发生可能与多种途径相关。在影像基因组学分析中,基因可以映射到影像组学特征中为影像组学特征提供生物学解释。

    (1)本研究纳入2家医疗中心数据,但样本数量仍较小。(2)虽然对所有图像进行预处理,但是从TCIA下载的影像数据集在扫描仪模式和采集协议方面存在差异,这可能对结果存在影响。(3)由于实验条件有限,影像组学特征与基因之间的相关性无法得到验证。

    综上,基于术前增强CT检查的联合影像组学模型可以准确预测肝细胞癌患者MVI状态。同时通过获取影像组学特征相关的mRNA基因表达谱,为影像组学模型提供了可能的生物学解释。

    赵建新:酝酿设计实验,实施研究,采集数据,分析和解释数据,起草文章,统计分析;潘妮妮、何迪梁、施柳言:分析和解释数据,对文章的知识性内容作批评性审阅,支持性贡献;何炫明、熊恋秋、马丽丽、崔雅琼、赵莲萍:指导,支持性贡献;黄刚:酝酿和设计实验,分析并解释数据,对文章的知识性内容作批评性审阅,获取研究经费,指导,支持性贡献
    所有作者均声明不存在利益冲突
    赵建新, 潘妮妮, 何迪梁, 等. 增强CT检查预测肝细胞癌微血管侵犯的影像基因组学研究[J]. 中华消化外科杂志, 2023, 22(11): 1367-1377. DOI: 10.3760/cma.j.cn115610-20231013-00148.

    http://journal.yiigle.com/LinkIn.do?linkin_type=cma&DOI=10.3760/cma.j.cn115610-20231013-23148

  • 图  1   肝细胞癌患者三维感兴趣区勾画与病理学检查结果 1A:微血管侵犯阴性患者增强CT检查图像;1B:微血管侵犯阴性患者沿横轴位逐层勾画示意图;1C:微血管侵犯阴性患者三维感兴趣区示意图;1D:微血管侵犯阴性患者病理学检查结果 HE染色 低倍放大;1E:微血管侵犯阳性患者增强CT检查图像;1F:微血管侵犯阳性患者沿横轴位逐层勾画示意图;1G:微血管侵犯阳性患者三维感兴趣区示意图;1H:微血管侵犯阳性患者病理学检查结果 HE染色 低倍放大

    Figure  1.   Delineation of three‑dimensional regions of interest and pathological examination in patient with hepatocellular carcinoma 1A: Enhanced computed tomography (CT) image of patient with negative microvascular invasion (MVI); 1B: Schematic diagram of sequential delineation of regions of interest along the transverse axis of patient with negative MVI; 1C: Three‑dimensional representa⁃tion of regions of interest of patient with negative MVI; 1D: Pathological examination of patient with negative MVI HE staining Low magnification; 1E: Enhanced CT image of patient with positive MVI; 1F: Schematic diagram of sequential delineation of regions of interest along the transverse axis of patient with positive MVI; 1G: Three‑dimensional representation of regions of interest of patient with positive MVI; 1H: Pathological examination of patient with positive MVI HE staining Low magnification

    图  2   肝细胞癌患者微血管侵犯的列线图预测模型

    Figure  2.   The nomogram predictive model for microvascular invasion in patients with hepatocellular carcinoma

    图  3   测试集的受试者工作特征(ROC)曲线 3A:动脉期模型ROC曲线;3B:门静脉期模型ROC曲线;3C:联合模型ROC曲线;3D:影像组学评分及列线图模型ROC曲线

    注:A为动脉期瘤内影像组学模型;A10为动脉期瘤周10 mm影像组学模型;V为门静脉期瘤内影像组学模型;V10为门静脉期瘤周10 mm影像组学模型;A+A10为动脉期瘤内+瘤周影像组学模型;V+V10为门静脉期瘤内+瘤周影像组学模型;A+A10+V+V10为动脉期及门静脉期瘤内+瘤周影像组学模型

    Figure  3.   The receiver operating characteristic (ROC) curve of test dataset 3A: ROC curve of arterial phase model; 3B: ROC curve of venous phase; 3C: ROC curve of combined model; 3D: ROC curve of R‑score and nomogram model

    图  4   影像组学评分与列线图预测肝细胞癌微血管侵犯的决策曲线 4A:训练集决策曲线;4B:测试集决策曲线

    Figure  4.   Decision curve of R⁃socre and nomogram in predicting microvascular invasion of hepatocellular carcinoma 4A: Decision curve in the training dataset; 4B: Decision curve in the test dataset

    图  5   影像组学评分与列线图预测肝细胞癌微血管侵犯的校准曲线 5A:训练集校准曲线;5B:测试集校准曲线

    Figure  5.   Calibration curve of R‑score and nomogram in predicting microvascular invasion hepatocellular carcinoma 5A: Calibration curve in the training dataset; 5B: Calibration curve in the test dataset

    图  6   生物信息学分析 6A:加权基因共表达网络分析获取8个基因模块;6B:基因模块与影像组学特征相关性分析;6C:显著相关基因的基因本体论富集分析;6D:显著相关基因的京都基因和基因组百科全书通路分析

    注:F1为A_exponential_gldm_DependenceVariance;F2为A_wavelet‑LHL_glcm_ClusterShade;F3为A_exponential_firstorder_Maximum;F4为A10_lbp‑2D_glrlm_LongRunEmphasis;F5为A10_wavelet‑LHH glszm ZoneEntropy;F6为V_lbp‑3D‑m2_glszm_GrayLevelVariance;F7为V_wavelet‑HLH_firstorder_Median;F8为V_wavelet‑LHH_glcm_Correlation

    Figure  6.   Bioinformatics analysis 6A: Eight gene modules obtaining by weighted gene co‑expression network analysis; 6B: Correlation analysis between gene modules and radiomics features; 6C: Gene ontology analysis of significantly correlated genes; 6D: Kyoto encyclopedia of genes and gnomes enrichment analysis of significantly correlated genes

    表  1   92例训练集和测试集肝细胞癌患者的临床及影像学特征比较

    Table  1   Comparison of clinical and radiological characteristics of 92 patients with hepatocellular carcinoma in the training dataset and the test dataset

    临床及影像学特征赋值例数训练集(64例)测试集(28例)统计量值P
    性别(例)
    0664224χ²=3.8770.049
    126224
    年龄(x±s,岁)--57±1260±13t=-1.0290.306
    甲胎蛋白(例)
    ≤400 ng/L0594019χ²=0.2430.622
    >400 ng/L133249
    肝炎(例)
    036279χ²=0.8250.364
    有(HBV/HCV)1563719
    肿瘤长径(例)
    ≤5 cm046297χ²=3.3740.066
    >5 cm1463521
    瘤周增强(例)
    0463115χ²=0.2050.650
    1463313
    瘤内动脉(例)
    026206χ²=0.9270.336
    1664422
    假包膜(例)
    0473611χ²=2.2430.134
    1452817
    瘤周不光滑(例)
    0463115χ²=0.2050.650
    1463313
    注:HBV为乙型病毒肝炎;HCV为丙型病毒肝炎;“-”表示此项无
    下载: 导出CSV

    表  2   92例微血管侵犯阴性和阳性肝细胞癌患者的临床及影像学特征比较

    Table  2   Comparison of clinical and radiological characteristics of 92 hepatocellular carcinoma patients with negative and positive microvascular invasion

    临床及影像学特征赋值例数微血管侵犯阴性(45例)微血管侵犯阳性(47例)统计量值P
    性别(例)
    0663036χ²=1.1180.290
    1261511
    年龄(x±s,岁)--59±1457±11t=0.9170.362
    甲胎蛋白(例)
    ≤400 ng/L0593326χ²=3.2430.072
    >400 ng/L1331221
    肝炎(例)
    0362313χ²=5.3080.021
    有(HBV/HCV)1562234
    肿瘤长径(例)
    ≤5 cm0362511χ²=9.9770.002
    >5 cm1562036
    瘤周增强(例)
    046397χ²=47.370<0.001
    146640
    瘤内动脉(例)
    026251χ²=32.368<0.001
    1662046
    假包膜(例)
    0473413χ²=21.105<0.001
    1451134
    瘤周不光滑(例)
    0463610χ²=31.711<0.001
    146937
    注:HBV为乙型肝炎病毒;HCV为丙型肝炎病毒;“-”表示此项无
    下载: 导出CSV

    表  3   影响64例训练集肝细胞癌患者微血管侵犯的单因素分析

    Table  3   Univariate analysis of microvascular invasion in 64 patients with hepatocellular carcinoma of the training dataset

    临床及影像学因素b标准误Wald优势比95%可信区间P值
    甲胎蛋白(≤400 ng/L比>400 ng/L)-0.7980.4473.1890.4500.188~1.0810.074
    肝炎(无比有)-1.0060.4425.1810.3660.154~0.8700.023
    肿瘤长径(≤5 cm比>5 cm)-1.4090.4579.5100.2440.100~0.5980.002
    瘤周增强(无比有)-3.6150.60036.2790.0270.008~0.087<0.001
    瘤内动脉(无比有)-4.0521.05414.7670.0170.002~0.137<0.001
    假包膜(无比有)-2.0900.47619.2700.1240.049~0.315<0.001
    瘤周不光滑(无比有)-2.6950.51627.3060.0680.025~0.186<0.001
    影像组学评分0.6130.13620.1831.8451.413~2.411<0.001
    下载: 导出CSV

    表  4   影响64例训练集肝细胞癌患者微血管侵犯的多因素分析

    Table  4   Multivariate analysis of microvascular invasion in 64 patients with hepatocellular carcinoma of the training dataset

    临床及影像学因素b标准误Wald风险比95%可信区间P
    瘤周增强(无比有)-3.0071.1227.1780.0490.005~0.4460.007
    瘤内动脉(无比有)-4.0861.6606.0590.0170.001~0.4350.014
    假包膜(无比有)-4.0941.7545.4520.0170.001~0.5180.020
    瘤周不光滑(无比有)-3.8701.5935.9030.0210.001~0.4730.015
    影像组学评分0.9320.3746.2092.5391.220~5.2830.013
    下载: 导出CSV

    表  5   肝细胞癌微血管侵犯预测的多期增强CT影像组学模型

    Table  5   The radiomics model in predicting of microvascular invasion of hepatocellular carcinoma based on multi‑phase enhanced computed tomography

    参数训练集测试集
    曲线下面积95%可信区间准确度(%)特异度(%)灵敏度(%)曲线下面积95%可信区间准确度(%)特异度(%)灵敏度(%)
    A0.8180.776~0.84570.7770.9770.590.8120.768~0.85675.8675.0076.92
    A100.7740.723~0.80572.3173.5370.970.7740.743~0.82268.9769.2368.75
    V0.8540.821~0.88776.9278.7975.000.8760.805~0.88982.7680.0085.71
    V100.8410.795~0.87375.3866.6784.380.8330.798~0.87175.8678.5773.33
    A+A100.9040.864~0.92176.9278.1275.760.8760.835~0.92282.7680.0085.71
    V+V100.9140.865~0.93484.6281.2587.880.8950.845~0.93375.8680.0071.43
    R‑score0.9230.887~0.94483.0874.1991.180.9180.894~0.94586.2187.5084.62
    Nomogram0.9730.954~0.98892.3196.1593.550.9620.942~0.98786.1286.9293.75
    注:A为动脉期瘤内影像组学模型;A10为动脉期瘤周10 mm影像组学模型;V为门静脉期瘤内影像组学模型;V10为门静脉期瘤周10 mm影像组学模型;A+A10为动脉期瘤内+瘤周影像组学模型;V+V10为门静脉期瘤内+瘤周影像组学模型;R‑score为影像组学评分;Nomogram为列线图
    下载: 导出CSV
  • [1]

    FornerA, ReigM, BruixJ. Hepatocellular carcinoma[J]. Lancet,2018,391(10127):1301‑1314. DOI:10.1016/S014 0-6736(18)30010-2.

    [2]

    FujiwaraN, FriedmanSL, GoossensN, et al. Risk factors and prevention of hepatocellular carcinoma in the era of precision medicine[J]. J Hepatol,2018,68(3):526‑549. DOI: 10.1016/j.jhep.2017.09.016.

    [3]

    RenzulliM, BrocchiS, CucchettiA, et al. Can current preoperative imaging be used to detect microvascular invasion of hepatocellular carcinoma?[J]. Radiology,2016,279(2):432‑442. DOI: 10.1148/radiol.2015150998.

    [4] 秦小琰,刘彦尧,康权.肝癌肝移植术后复发转移的防治策略和靶向免疫治疗[J].器官移植,2022,13(2):271-276. DOI: 10.3969/j.issn.1674-7445.2022.02.018.
    [5] 刘少儒,许磊波.肝癌肝移植术后复发转移的免疫治疗[J].器官移植,2021,12(3):272-279. DOI: 10.3969/j.issn.1674-7445.2021.03.004.
    [6]

    GouwAS, BalabaudC, KusanoH, et al. Markers for microvascular invasion in hepatocellular carcinoma: where do we stand?[J]. Liver Transpl,2011,17Suppl 2:S72‑S80. DOI: 10.1002/lt.22368.

    [7]

    LeiZ, LiJ, WuD, et al. Nomogram for preoperative estimation of microvascular invasion risk in hepatitis B virus-related hepatocellular carcinoma within the Milan criteria[J]. JAMA Surg,2016,151(4):356‑363. DOI: 10.1001/jamasurg.2015.4257.

    [8]

    ChongHH, YangL, ShengRF, et al. Multi‑scale and multi-parametric radiomics of gadoxetate disodium‑enhanced MRI predicts microvascular invasion and outcome in pati-ents with solitary hepatocellular carcinoma ≤5 cm[J]. Eur Radiol,2021,31(7):4824‑4838. DOI: 10.1007/s00330-020-07601-2.

    [9] 石鑫林,王继洲,姜洪池,等.影响肝癌患者肝移植术后复发的血管因素研究进展[J].器官移植,2019,10(6):727-730. DOI: 10.3969/j.issn.1674-7445.2019.06.017.
    [10] 何涛,邹婕妤,雷婷婷,等.肝细胞癌微血管侵犯的研究进展[J].中华肝脏病杂志,2022,30(8):899-904. DOI:10.3760/cma. j.cn501113-20210428-00211.
    [11] 金涛,李涛,孙青,等.肝细胞癌微血管侵犯与临床病理指标的相关性[J].中国现代普通外科进展,2018,21(7):520-523. DOI: 10.3969/j.issn.1009-9905.2018.07.006.
    [12]

    ChenQ, XiaoH, GuY, et al. Deep learning for evaluation of microvascular invasion in hepatocellular carcinoma from tumor areas of histology images[J]. Hepatol Int,2022,16(3):590‑602. DOI: 10.1007/s12072-022-10323-w.

    [13]

    LinWD, YeLN, SongZS, et al. Wide surgical margins im-prove prognosis for HCC with microvascular invasion[J]. Eur Rev Med Pharmacol Sci,2023,27(5):2052‑2059. DOI: 10.26355/eurrev_202303_31576.

    [14]

    HwangS, LeeYJ, KimKH, et al. The impact of tumor size on long‑term survival outcomes after resection of solitary hepatocellular carcinoma: single‑institution experience with 2558 patients[J]. J Gastrointest Surg,2015,19(7):1281-1290. DOI: 10.1007/s11605-015-2849-5.

    [15] 向德森,徐雪松,龚建平,等.肝细胞癌诊断技术的新进展[J]. 中国现代普通外科进展,2017,20(5):417-420. DOI: 10.3969/j.issn.1009-9905.2017.05.027.
    [16]

    XuX, ZhangHL, LiuQP, et al. Radiomic analysis of contrast-enhanced CT predicts microvascular invasion and outcome in hepatocellular carcinoma[J]. J Hepatol,2019, 70(6):1133‑1144. DOI: 10.1016/j.jhep.2019.02.023.

    [17] 任新平,李若坤.肝细胞癌规范化筛查及早期影像诊断策略[J].中华肝脏病杂志,2021,29(4):304-307. DOI:10.3760/cma. j.cn501113-20210406-00164.
    [18]

    SartorisR, GregoryJ, Dioguardi BurgioM, et al. HCC advan-ces in diagnosis and prognosis: digital and Imaging[J]. Liver Int,2021,41Suppl 1:73‑77. DOI: 10.1111/liv.14865.

    [19]

    XiaTY, ZhouZH, MengXP, et al. Predicting microvascular invasion in hepatocellular carcinoma using CT‑based radio-mics model[J]. Radiology,2023,307(4):e222729. DOI:10. 1148/radiol.222729.

    [20]

    LvK, CaoX, DuP, et al. Radiomics for the detection of microvascular invasion in hepatocellular carcinoma[J]. World J Gastroenterol,2022,28(20):2176‑2183. DOI: 10.3748/wjg.v28.i20.2176.

    [21] 王莉莉,李俊峰,雷军强,等.影像组学方法评价肝细胞癌微血管侵犯的研究进展[J].中华肝脏病杂志,2023,31(3):327-331. DOI: 10.3760/cma.j.cn501113-20230312-00107.
    [22]

    ChenY, XiaY, TolatPP, et al. Comparison of conventional gadoxetate disodium‑enhanced MRI features and radio-mics signatures with machine learning for diagnosing microvascular invasion[J]. AJR Am J Roentgenol,2021,216(6):1510‑1520. DOI: 10.2214/AJR.20.23255.

    [23] 俞超,胡晓欣,张汐,等.基于CT影像组学及临床资料预测肝细胞癌微血管侵犯的研究[J]复旦学报:医学版,2023,50(1):15‑25. DOI: 10.3969/j.issn.1672-8467.2023.01.003
    [24]

    LiuHF, ZhangYZ, WangQ, et al. A nomogram model integrating LI‑RADS features and radiomics based on contrast-enhanced magnetic resonance imaging for predicting microvascular invasion in hepatocellular carcinoma falling the Milan criteria[J]. Transl Oncol,2023,27:101597. DOI: 10.1016/j.tranon.2022.101597.

    [25]

    LeeS, KimSH, LeeJE, et al. Preoperative gadoxetic acid-enhanced MRI for predicting microvascular invasion in patients with single hepatocellular carcinoma[J]. J Hepatol,2017, 67(3):526‑534. DOI: 10.1016/j.jhep.2017.04.024.

    [26]

    FengST, JiaY, LiaoB, et al. Preoperative prediction of microvascular invasion in hepatocellular cancer: a radio-mics model using Gd‑EOB‑DTPA‑enhanced MRI[J]. Eur Radiol,2019,29(9):4648‑4659. DOI: 10.1007/s00330-018-5935-8.

    [27]

    MaX, WeiJ, GuD, et al. Preoperative radiomics nomogram for microvascular invasion prediction in hepatocellular carcinoma using contrast‑enhanced CT[J]. Eur Radiol,2019,29(7):3595‑3605. DOI: 10.1007/s00330-018-5985-y.

    [28]

    ZhangW, YangR, LiangF, et al. Prediction of microvascular invasion in hepatocellular carcinoma with a multi-disciplinary team‑like radiomics fusion model on dynamic contrast-enhanced computed tomography[J]. Front Oncol,2021,11:660629. DOI: 10.3389/fonc.2021.660629.

    [29]

    YaoW, YangS, GeY, et al. Computed tomography radio-mics‑based prediction of microvascular invasion in hepatocellular carcinoma[J]. Front Med (Lausanne),2022,9:819670. DOI: 10.3389/fmed.2022.819670.

    [30] 刘燕娜,许乙凯,邓淑云,等.基于CT和MRI影像组学预测肝细胞癌微血管侵犯研究进展[J].中华肝脏病杂志,2022,30(8):809-813. DOI: 10.3760/cma.j.cn501113-20220628-00359.
    [31] 李阳,杨昭,李淑华,等.基于非小细胞肺癌双能CT表现及影像组学列线图模型预测其血管生成拟态[J].中国医学影像技术,2023,39(5):684‑689. DOI:10.13929/j.issn.1003-3289. 2023.05.009
    [32]

    ContiA, DuggentoA, IndovinaI, et al. Radiomics in breast cancer classification and prediction[J]. Semin Cancer Biol,2021,72:238‑250. DOI: 10.1016/j.semcancer.2020.04.002.

    [33] 崔雅琼,黄刚,王莉莉,等.基于临床‑多参数磁共振影像组学特征预测宫颈癌脉管浸润和预后的研究[J].磁共振成像,2023,14(2):73‑82. DOI:10.12015/issn.1674-8034.2023. 02.013.
    [34]

    ZhouT, QinR, ShiS, et al. DTYMK promote hepatocellular carcinoma proliferation by regulating cell cycle[J]. Cell Cycle,2021,20(17):1681‑1691. DOI:10.1080/15384101.2021. 1958502.

    [35]

    SunEJ, WankellM, PalamuthusingamP, et al. Targeting the PI3K/Akt/mTOR pathway in hepatocellular carcinoma[J]. Biomedicines,2021,9(11):1639. DOI:10.3390/biomedi cines 9111639.

    [36]

    LeungR, LeeT. Wnt/β‑catenin signaling as a driver of stemness and metabolic reprogramming in hepatocellular carcinoma[J]. Cancers (Basel),2022,14(21):5468. DOI:10. 3390/cancers14215468.

    [37]

    van der WindtDJ, SudV, ZhangH, et al. Neutrophil extracellular traps promote inflammation and development of hepatocellular carcinoma in nonalcoholic steatohepatitis[J]. Hepatology,2018,68(4):1347‑1360. DOI:10.1002/hep. 29914.

    [38]

    ZhanX, WuR, KongXH, et al. Elevated neutrophil extrace-llular traps by HBV‑mediated S100A9‑TLR4/RAGE‑ROS cascade facilitate the growth and metastasis of hepatoce-llular carcinoma[J]. Cancer Commun (Lond),2023,43(2):225‑245. DOI: 10.1002/cac2.12388.

    [39]

    JiangZZ, PengZP, LiuXC, et al. Neutrophil extracellular traps induce tumor metastasis through dual effects on cancer and endothelial cells[J]. Oncoimmunology,2022,11(1):2052418. DOI: 10.1080/2162402X.2022.2052418.

    [40]

    ZenlanderR, HavervallS, MagnussonM, et al. Neutrophil extracellular traps in patients with liver cirrhosis and hepatocellular carcinoma[J]. Sci Rep,2021,11(1):18025. DOI: 10.1038/s41598-021-97233-3.

    [41]

    IshtiaqSM, ArshadMI, KhanJA. PPARγ signaling in hepatocarcinogenesis: mechanistic insights for cellular reprogramming and therapeutic implications[J]. Pharmacol Ther,2022,240:108298. DOI: 10.1016/j.pharmthera.2022.108298.

    [42] 马炳奇,孟慧娟,董小锋,等.炎症和肿瘤指标预后评分在肝内胆管癌中的预测价值[J].中华肝脏病杂志,2022,30(7):777-783. DOI: 10.3760/cma.j.cn501113-20211108-00539.
    [43] 张向颖,徐玲,田原,等.CHOP信号分子在PPARα激活抑制小鼠急性肝衰竭炎症反应中的作用研究[J].中华肝脏病杂志,2020,28(7):613-618. DOI:10.3760/cma.j.cn501113-2020 0608-00298.
    [44]

    HanW, WangN, KongR, et al. Ligand‑activated PPARδ expression promotes hepatocellular carcinoma progression by regulating the PI3K‑AKT signaling pathway[J]. J Transl Med,2022,20(1):86. DOI:10.1186/s12967-022-03 288-9.

    [45]

    XinZ, LiJ, ZhangH, et al. Cancer genomic alterations can be potential biomarkers predicting microvascular invasion and early recurrence of hepatocellular carcinoma[J]. Front Oncol,2022,12:783109. DOI:10.3389/fonc.2022.78 3109.

图(6)  /  表(5)
计量
  • 文章访问数:  32
  • HTML全文浏览量:  0
  • PDF下载量:  0
  • 被引次数: 0
出版历程
  • 收稿日期:  2023-10-12
  • 网络出版日期:  2024-06-24
  • 刊出日期:  2023-11-19

目录

/

返回文章
返回