Volume 46 Issue 6
Jun.  2022
Article Contents

Citation:

The new progress of 18F-FDG PET/CT in predicting the mutation status of EGFR gene in non-small-cell lung cancer

  • Corresponding author: Xiaonan Shao, scorey@sina.com
  • Received Date: 2021-07-12
  • Lung cancer is the second most common cancer in the world, among which non-small-cell lung cancer (NSCLC) is the predominant type. Epidermal growth factor receptor (EGFR) plays an important role in the development of NSCLC and has become an important target in the treatment of NSCLC. Epidermal growth factor receptor-tyrosine kinase inhibitors have been widely used in the targeted therapy of NSCLC and have been shown to effectively prolong the survival of patients with EGFR mutations, and their efficacy and prognosis are closely related to those of EGFR gene mutations. 18F-fluorodeoxyglucose (FDG) PET/CT can non-invasively evaluate NSCLC and is of great significance in predicting EGFR gene mutation status in NSCLC. The authors review the recent advances in 18F-FDG PET/CT metabolic parameters and radiomics related to EGFR gene mutation status.
  • 加载中
  • [1] Siegel RL, Miller KD, Fuchs HE, et al. Cancer statistics, 2021[J]. CA Cancer J Clin, 2021, 71(1): 7−33. DOI: 10.3322/caac.21654.
    [2] 中华医学会肿瘤学分会, 中华医学会杂志社. 中华医学会肿瘤学分会肺癌临床诊疗指南(2021版)[J]. 中华医学杂志, 2021, 101(23): 1725−1757. DOI: 10.3760/cma.j.cn112137-20210207-00377.Oncology Branch of Chinese Medical Association, Chinese Medical Association Pbulishing House. Clinical diagnosis and treatment guidelines for lung cancer of Oncology Branch of Chinese Medical Association (2021 Edition)[J]. Natl Med J China, 2021, 101(23): 1725−1757. DOI: 10.3760/cma.j.cn112137-20210207-00377.
    [3] Wu GY, Jochems A, Refaee T, et al. Structural and functional radiomics for lung cancer[J]. Eur J Nucl Med Mol Imaging, 2021, 48(12): 3961−3974. DOI: 10.1007/s00259-021-05242-1.
    [4] Lee SM, Bae SK, Jung SJ, et al. FDG uptake in non-small cell lung cancer is not an independent predictor of EGFR or KRAS mutation status: a retrospective analysis of 206 patients[J]. Clin Nucl Med, 2015, 40(12): 950−958. DOI: 10.1097/RLU.0000000000000975.
    [5] Kanmaz ZD, Aras G, Tuncay E, et al. Contribution of 18Fluorodeoxyglucose positron emission tomography uptake and TTF-1 expression in the evaluation of the EGFR mutation in patients with lung adenocarcinoma[J]. Cancer Biomark, 2016, 16(3): 489−498. DOI: 10.3233/CBM-160588.
    [6] 郭虹霞, 任筱璐, 张俊萍. 18F-FDG PET/CT对肺腺癌患者EGFR突变的预测价值[J]. 中华核医学与分子影像杂志, 2020, 40(8): 475−479. DOI: 10.3760/cma.j.cn321828-20200113-00017.Guo HX, Ren XL, Zhang JP, et al. Predictive value of 18F-FDG PET/CT for EGFR mutations in patients with lung adenocarcinoma[J]. Chin J Nucl Med Mol Imaging, 2020, 40(8): 475−479. DOI: 10.3760/cma.j.cn321828-20200113-00017.
    [7] Gu JC, Xu SQ, Huang LX, et al. Value of combining serum carcinoembryonic antigen and PET/CT in predicting EGFR mutation in non-small cell lung cancer[J]. J Thorac Dis, 2018, 10(2): 723−731. DOI: 10.21037/jtd.2017.12.143.
    [8] 丁重阳, 李天女, 孙晋, 等. 18F-FDG摄取与甲状腺转录因子-1表达预测肺腺癌患者表皮生长因子受体突变的价值[J]. 中华核医学与分子影像杂志, 2018, 38(2): 92−96. DOI: 10.3760/cma.j.issn.2095-2848.2018.02.004.Ding CY, Li TN, Sun J, et al. Values of 18F-FDG uptake and thyroid transcription factor-1 expression to predict the mutations of epidermal growth factor receptor in lung adenocarcinoma[J]. Chin J Nucl Med Mol Imaging, 2018, 38(2): 92−96. DOI: 10.3760/cma.j.issn.2095-2848.2018.02.004.
    [9] 姜阳, 马晓伟, 董楚宁, 等. 基于18F-FDG PET-CT代谢参数的风险模型对非小细胞肺癌EGFR基因突变的预测价值及效能评价[J]. 中华放射学杂志, 2020, 54(7): 688−693. DOI: 10.3760/cma.j.cn112149-20190731-00652.Jiang Y, Ma XW, Dong CN, et al. The predictive value and efficacy of the risk model based on the metabolic parameters of 18F-FDG PET-CT for EGFR gene mutations in non-small cell lung cancer[J]. Chin J Radiol, 2020, 54(7): 688−693. DOI: 10.3760/cma.j.cn112149-20190731-00652.
    [10] 赵承勇, 邓小毅, 王洪松, 等. 18F-脱氧葡萄糖正电子发射计算机断层扫描显像在预测肺鳞癌表皮生长因子受体突变中的价值[J]. 中华肿瘤杂志, 2021, 43(7): 795−800. DOI: 10.3760/cma.j.cn112152-20190111-00006.Zhao CY, Deng XY, Wang HS, et al. Value of 18F-FDG PET-CT imaging to predict epidermal growth factor receptor mutations in patients with lung squamous cell carcinoma[J]. Chin J Oncol, 2021, 43(7): 795−800. DOI: 10.3760/cma.j.cn112152-20190111-00006.
    [11] Lee EY, Khong PL, Lee VHF, et al. Metabolic phenotype of stage IV lung adenocarcinoma: relationship with epidermal growth factor receptor mutation[J]. Clin Nucl Med, 2015, 40(3): e190−e195. DOI: 10.1097/RLU.0000000000000684.
    [12] Mayerhoefer ME, Materka A, Langs G, et al. Introduction to radiomics[J]. J Nucl Med, 2020, 61(4): 488−495. DOI: 10.2967/jnumed.118.222893.
    [13] Avanzo M, Stancanello J, Pirrone G, et al. Radiomics and deep learning in lung cancer[J]. Strahlenther Onkol, 2020, 196(10): 879−887. DOI: 10.1007/s00066-020-01625-9.
    [14] Yip SSF, Kim J, Coroller TP, et al. Associations between somatic mutations and metabolic imaging phenotypes in non-small cell lung cancer[J]. J Nucl Med, 2017, 58(4): 569−576. DOI: 10.2967/jnumed.116.181826.
    [15] Zhang JY, Zhao XM, Zhao Y, et al. Value of pre-therapy 18F-FDG PET/CT radiomics in predicting EGFR mutation status in patients with non-small cell lung cancer[J]. Eur J Nucl Med Mol Imaging, 2020, 47(5): 1137−1146. DOI: 10.1007/s00259-019-04592-1.
    [16] Jiang MM, Zhang YQ, Xu JS, et al. Assessing EGFR gene mutation status in non-small cell lung cancer with imaging features from PET/CT[J]. Nucl Med Commun, 2019, 40(8): 842−849. DOI: 10.1097/MNM.0000000000001043.
    [17] 杨天红, 张胤, 李淑仪, 等. 18F-FDG PET/CT影像组学在预测肺腺癌患者EGFR突变中的价值[J]. 中华核医学与分子影像杂志, 2021, 41(2): 65−70. DOI: 10.3760/cma.j.cn321828-20191108-00255.Yang TH, Zhang Y, Li SY, et al. Ability of 18F-FDG PET/CT radiomic features to differentiate EGFR mutation status in patients with lung adenocarcinoma[J]. Chin J Nucl Med Mol Imaging, 2021, 41(2): 65−70. DOI: 10.3760/cma.j.cn321828-20191108-00255.
    [18] Li XF, Yin GT, Zhang YF, et al. Predictive power of a radiomic signature based on 18F-FDG PET/CT images for EGFR mutational status in NSCLC[J/OL]. Front Oncol, 2019, 9: 1062[2021-07-11]. https://www.frontiersin.org/articles/10.3389/fonc.2019.01062/full. DOI: 10.3389/fonc.2019.01062.
    [19] 王子阳, 尹国涛, 李小凤, 等. 机器学习结合18F-FDG PET/CT影像组学特征对肺腺癌EGFR突变亚型的预测价值[J]. 中华核医学与分子影像杂志, 2021, 41(8): 479−485. DOI: 10.3760/cma.j.cn321828-20201105-00401.Wang ZY, Yin GT, Li XF, et al. Value of machine learning and 18F-FDG PET/CT radiomics features in lung adenocarcinoma EGFR mutation subtypes prediction[J]. Chin J Nucl Med Mol Imaging, 2021, 41(8): 479−485. DOI: 10.3760/cma.j.cn321828-20201105-00401.
    [20] Zhang M, Bao YM, Rui WW, et al. Performance of 18F-FDG PET/CT radiomics for predicting EGFR mutation status in patients with non-small cell lung cancer[J/OL]. Front Oncol, 2020, 10: 568857[2021-07-11]. https://www.frontiersin.org/articles/10.3389/fonc.2020.568857/full. DOI: 10.3389/fonc.2020.568857.
    [21] Liu QF, Sun DZ, Li N, et al. Predicting EGFR mutation subtypes in lung adenocarcinoma using 18F-FDG PET/CT radiomic features[J]. Transl Lung Cancer Res, 2020, 9(3): 549−562. DOI: 10.21037/tlcr.2020.04.17.
    [22] Yang B, Ji HS, Zhou CS, et al. 18F-fluorodeoxyglucose positron emission tomography/computed tomography-based radiomic features for prediction of epidermal growth factor receptor mutation status and prognosis in patients with lung adenocarcinoma[J]. Transl Lung Cancer Res, 2020, 9(3): 563−574. DOI: 10.21037/tlcr-19-592.
    [23] Koyasu S, Nishio M, Isoda H, et al. Usefulness of gradient tree boosting for predicting histological subtype and EGFR mutation status of non-small cell lung cancer on 18F FDG-PET/CT[J]. Ann Nucl Med, 2020, 34(1): 49−57. DOI: 10.1007/s12149-019-01414-0.
    [24] Shiri I, Maleki H, Hajianfar G, et al. Next-generation radiogenomics sequencing for prediction of EGFR and KRAS mutation status in NSCLC patients using multimodal imaging and machine learning algorithms[J]. Mol Imaging Biol, 2020, 22(4): 1132−1148. DOI: 10.1007/s11307-020-01487-8.
    [25] LeCun Y, Bengio Y, Hinton G. Deep learning[J]. Nature, 2015, 521(7553): 436−444. DOI: 10.1038/nature14539.
    [26] Apostolopoulos ID, Pintelas EG, Livieris IE, et al. Automatic classification of solitary pulmonary nodules in PET/CT imaging employing transfer learning techniques[J/OL]. Med Biol Eng Comput, 2021, 59(6): 1299−1310[2021-07-11]. https://link.springer.com/article/10.1007/s11517-021-02378-y. DOI: 10.1007/s11517-021-02378-y.
    [27] Han Y, Ma Y, Wu ZY, et al. Histologic subtype classification of non-small cell lung cancer using PET/CT images[J]. Eur J Nucl Med Mol Imaging, 2021, 48(2): 350−360. DOI: 10.1007/s00259-020-04771-5.
    [28] Jemaa S, Fredrickson J, Carano RAD, et al. Tumor segmentation and feature extraction from whole-body FDG-PET/CT using cascaded 2D and 3D convolutional neural networks[J]. J Digit Imaging, 2020, 33(4): 888−894. DOI: 10.1007/s10278-020-00341-1.
    [29] Park C, Na KJ, Choi H, et al. Tumor immune profiles noninvasively estimated by FDG PET with deep learning correlate with immunotherapy response in lung adenocarcinoma[J/OL]. Theranostics, 2020, 10(23): 10838−10848[2021-07-11]. https://pubmed.ncbi.nlm.nih.gov/32929383. DOI: 10.7150/thno.50283.
    [30] Mu W, Jiang L, Zhang JY, et al. Non-invasive decision support for NSCLC treatment using PET/CT radiomics[J/OL]. Nat Commun, 2020, 11(1): 5228[2021-07-11]. https://www.nature.com/articles/s41467-020-19116-x. DOI: 10.1038/s41467-020-19116-x.
    [31] Xiao ZY, Song Y, Wang K, et al. One-step radiosynthesis of 18F-IRS: a novel radiotracer targeting mutant EGFR in NSCLC for PET/CT imaging[J]. Bioorg Med Chem Lett, 2016, 26(24): 5985−5988. DOI: 10.1016/j.bmcl.2016.10.084.
  • 加载中
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

Article Metrics

Article views(4529) PDF downloads(21) Cited by()

Related
Proportional views

The new progress of 18F-FDG PET/CT in predicting the mutation status of EGFR gene in non-small-cell lung cancer

    Corresponding author: Xiaonan Shao, scorey@sina.com
  • Department of Nuclear Medicine, the Third Affiliated Hospital of Soochow University, the First People′s Hospital of Changzhou, Changzhou 213003, China

Abstract: Lung cancer is the second most common cancer in the world, among which non-small-cell lung cancer (NSCLC) is the predominant type. Epidermal growth factor receptor (EGFR) plays an important role in the development of NSCLC and has become an important target in the treatment of NSCLC. Epidermal growth factor receptor-tyrosine kinase inhibitors have been widely used in the targeted therapy of NSCLC and have been shown to effectively prolong the survival of patients with EGFR mutations, and their efficacy and prognosis are closely related to those of EGFR gene mutations. 18F-fluorodeoxyglucose (FDG) PET/CT can non-invasively evaluate NSCLC and is of great significance in predicting EGFR gene mutation status in NSCLC. The authors review the recent advances in 18F-FDG PET/CT metabolic parameters and radiomics related to EGFR gene mutation status.

    HTML

  • 肺癌是全球范围内病死率最高的恶性肿瘤,其中非小细胞肺癌(non-small-cell lung cancer, NSCLC)约占80%[1]。随着肺癌个体化治疗的发展,识别与靶向治疗相关的基因突变状态[如表皮生长因子受体(epidermal growth factor receptor, EGFR)突变]已成为优化治疗策略的重要依据。相较于传统有创的活体组织病理学检查,18F-FDG PET/CT属于无创分子影像学技术,一次采集即可获得病灶的功能和解剖信息,《中华医学会肿瘤学分会肺癌临床诊疗指南(2021版)》已推荐其用于肺癌的诊断、分期及预后预测[2];近年来有研究结果显示,其也可有效预测EGFR基因的突变状态[3]。笔者就18F-FDG PET/CT常规代谢参数及影像组学在预测EGFR基因突变状态方面的研究进展作一综述。

  • 1.   18F-FDG PET/CT常规代谢参数与EGFR基因突变状态
    • 关于NSCLC EGFR基因突变状态与SUV的关系,目前尚无一致结论。有研究者认为EGFR野生组与EGFR突变组SUVmax的差异无统计学意义[4]。Kanmaz等[5]发现较高的SUVmax与肺腺癌EGFR基因突变显著相关。然而,更多的研究结果显示,EGFR基因突变可能与较小的SUVmax有关[6-8]。此外,郭虹霞等[6]发现原发灶较小的SUVmax及较短的肿瘤长径是肺腺癌患者EGFR基因突变的预测因子。一些研究者尝试将PET/CT常规代谢参数与血清肿瘤标志物结合来预测EGFR基因的突变状态。Gu等[7]证实较高的癌胚抗原水平和较小的SUVmax是EGFR基因突变的显著预测因子。另外,丁重阳等[8]发现,甲状腺转录因子1阳性和较小的SUVmax可在一定程度上预测EGFR基因突变。

      相较于作为单个像素值的SUVmax,代谢肿瘤体积(metabolic tumour volume,MTV)和总病灶糖酵解(total lesion glycolysis,TLG)更能反映肿瘤整体的葡萄糖代谢信息,可以克服因部分容积效应而产生的统计偏差。姜阳等[9]在一项回顾性研究中将NSCLC患者分为EGFR突变组及EGFR野生组,比较2组间的18F-FDG PET/CT代谢参数,结果显示,EGFR突变组的MTV与TLG均低于EGFR野生组;他们进一步将临床病理因素(如性别、吸烟史及血清肿瘤标志物等)与PET/CT代谢参数(如SUVmax、MTV和TLG等)结合进行多因素分析,结果显示,MTV是EGFR基因突变的独立预测因子。赵承勇等[10]则认为较低的TLG是EGFR基因突变的独立预测因子。

      产生上述分歧的主要原因可能是:(1)研究人群的异质性,入选患者的TNM分期和组织学类型可能对结果有影响,如只纳入了晚期肺癌患者[11],或只纳入了肺鳞癌患者[10],或只纳入了肺腺癌患者[5];此外,大部分研究未区分实性与亚实性病灶。(2)部分文献报道的样本量偏小,结果可能会存在偏倚。因此,综合上述研究结果,提示PET/CT常规代谢参数可能不是预测EGFR基因突变状态的可靠指标。

    2.   18F-FDG PET/CT影像组学与EGFR基因突变状态
    • 影像组学是近年来迅速发展的研究领域,涉及医学图像中定量指标(即影像组学特征)的提取、计算、选择、降维及数据处理[12-13]。随着影像组学的不断发展,对PET/CT影像组学特征与EGFR基因突变状态间关系的研究日益受到关注。Yip等[14]研究了348例NSCLC患者的18F-FDG PET影像组学特征与EGFR基因突变之间的关联,结果显示,8个PET特征与EGFR基因突变状态显著相关,其中灰度共生矩阵(gray level co-occurrence matrix,GLCM)参数InvDiffomor(体现肿瘤匀质性的PET特征)甚至可直接预测EGFR基因的突变状态。

      随着大数据时代的到来,更多的研究者用机器学习的方法开展影像组学研究,其中最常用的就是最小绝对收缩和选择算子算法,通过筛选大量存在多重共线性的PET/CT影像组学特征来预测EGFR基因的突变状态。Zhang等[15]选取5个PET特征及5个CT特征(如描述感兴趣区体积的球形程度、紧凑程度的形态特征与描述图像均匀性、异质性的纹理特征,GLCM等)建立的影像组学模型具有良好的预测效能,其验证集AUC为0.85。Jiang等[16]选取 13个PET特征(主要为形态特征及GLCM)和4个CT语义特征(分叶、毛刺、空泡、胸膜牵拉)建立预测模型,结果显示,该模型对EGFR基因突变状态具有良好的预测效能,其AUC达到了0.955。杨天红等[17]则从PET、CT、PET/CT图像参数中分别筛选出3、3、7个特征构成3种回归模型,其中由PET/CT特征(包括一阶特征、形态特征及纹理特征)建立的模型具有最高的预测效能(AUC=0.866)。Li等[18]提取7个PET特征(包括1个一阶特征及6个纹理特征)和2个CT一阶特征建立影像组学模型,当结合临床危险因素后,显著改善了该模型的预测性能(AUC从0.805提升到0.822)。上述研究所提取的影像组学特征中均包含GLCM,均取得了不错的预测效果,但尚未充分发挥机器学习在分类问题上的强大功能;且单纯使用最小绝对收缩和选择算子回归容易造成重要信息的丢失,从而导致训练模型的欠拟合[19]

      随机森林(random forest, RF)算法可以更方便、更直观地帮助研究者找到与预测目标相对应的最相关的影像组学特征,以此为基础建立机器学习模型来预测EGFR基因的突变状态。Zhang等[20]应用RF算法经过多步特征筛选后选择了4个CT纹理特征和2个PET一阶特征,分别构建了CT模型、PET模型及PET/CT联合模型,其中PET/CT联合模型预测EGFR基因突变的效能最高,AUC达到0.868,灵敏度为92.8%、特异度为66.3%、准确率为77.1%。Liu等[21]先剔除方差膨胀因子高于阈值的特征,随后进一步应用RF算法并引入logistic回归模型来筛选影像组学特征,最终应用极端梯度提升(eXtreme gradient boosting)机器学习算法来建立模型,其训练集AUC为0.93,而测试集AUC为0.87。在另一项回顾性研究中,Yang等[22]选择影像组学特征与临床病理因素相结合,以构建RF模型来识别EGFR基因的突变状态,结果显示,该方法同样具有良好的预测效能(验证组的AUC=0.71)。Koyasu等[23]对极端梯度提升机器学习算法和RF算法2种预测模型进行比较,结果表明,极端梯度提升模型具有更好的预测性能及泛化能力。上述研究结果均展现了不同机器学习算法强大的预测效能,尤其是极端梯度提升机器学习算法。

      尽管几种机器学习算法可单独或组合用于影像组学分析中的特征选择和分类,但由于各种机器学习的性能已被证明依赖于应用或数据类型,因此理论上不存在“一刀切”的方法。王子阳等[19]使用了K最近邻、支持向量机及Adaboost(一种迭代算法)这3种分类器分别对CT、PET和PET/CT融合组学特征进行学习并建立模型,结果显示,其在预测EGFR基因突变状态方面有一定价值(AUC为0.67~0.74)。Shiri等[24]使用6种特征选择方法和12种分类器构建了72种交叉组合的机器学习框架,通过选择最优组合成功预测了患者的EGFR基因突变状态。这种搭建机器学习框架的研究方法,可能是未来PET/CT影像组学研究的趋势。

      深度学习是一种更深层次的机器学习方法,其灵感来自大脑的神经元网络[25]。而深度学习算法中最常用的是多层前馈神经网络——卷积神经网络,其可接受三维图像的输入,可以在学习高分辨图像特征的同时以有监督的方法进行端到端的训练[21]。在PET/CT肺部肿瘤成像中,卷积神经网络的具体应用包括但不限于对病变的检测和分类[26-27]、自动图像分割[28]、对免疫治疗反应的预测[29]、影像组学特征的提取及建模[30]等。

      深度学习在PET/CT中已有广泛应用,但其联合PET/CT预测NSCLC EGFR基因突变状态的研究仍较少,原因可能是缺少高质量的大数据集。Mu等[30]进行了一项包含681例NSCLC患者的多中心研究,使用PET/CT图像的深度学习模型来预测患者EGFR基因的突变状态,在训练、验证和独立测试队列中的AUC分别为0.86、0.83和0.81。此外,迁移学习也可以解决样本量有限的问题[26],其是一种将来自其他领域图像上预先训练好的网络调优到一个新的数据集上的方法,或许能为PET/CT影像组学结合深度学习提供技术基础。

    3.   小结与展望
    • 既往的研究者认为,常规的PET/CT参数(如SUVmax、MTV、TLG等)与EGFR基因突变存在一定关系,但仍有争议[4-11]。且由于常规参数过于简单,不足以与基因组学、代谢组学或蛋白质组学的数据结合使用。而来自PET/CT影像组学的信息,通过结构分析提取大量数据后,与基因、代谢或蛋白质组学信息结合使用,具有更高的预测效能。

      目前EGFR基因突变状态预测相关的PET/CT影像组学研究仍然存在局限性。首先,现有研究多为回顾性单中心研究,由于成像方案的非标准化,采集和重建参数各异,结果的可重复性较低;其次,大多数研究的样本量偏少,模型的泛化能力有待进一步检验,需要多中心、大样本、前瞻性队列研究来验证;最后,PET/CT显像基因组仍然是一个新兴的领域,期待更完善的组学软件及深度学习方法的引入。

      综上,PET/CT预测NSCLC的EGFR基因突变状态是一个很有前景的领域。除常规显像剂18F-FDG外,靶向EGFR的分子探针(如18F-MPG、18F-IRS等)也显示出较好的检测EGFR基因突变的能力[30-31]。近年来也有研究者尝试应用PET/CT影像组学区分外显子19 del和外显子21 L858R突变,并取得了一定成果[19-22]。PET/CT影像组学结合多模态影像信息、建立深度学习模型,联合基因组、蛋白质组等多组学信息,将会为NSCLC患者的诊断、治疗决策的制定以及预后预测提供更多、更有价值的信息。

      利益冲突 所有作者声明无利益冲突

      作者贡献声明 高建雄负责文献的检索、综述的撰写;牛荣、邵小南、王跃涛负责命题的提出、综述的审阅与修改

Reference (31)

Catalog

/

DownLoad:  Full-Size Img  PowerPoint
Return
Return