欢迎来到《四川大学学报(医学版)》
朱静, 徐维国, 肖欢, 等. 影像组学联合预测模型在预测非小细胞肺癌淋巴结转移中的建立和应用价值[J]. 四川大学学报(医学版), 2019, 50(3): 373-378.
引用本文: 朱静, 徐维国, 肖欢, 等. 影像组学联合预测模型在预测非小细胞肺癌淋巴结转移中的建立和应用价值[J]. 四川大学学报(医学版), 2019, 50(3): 373-378.
ZHU Jing, XU Wei-guo, XIAO Huan, et al. Application of a Radiomics Model for Preding Lymph Node Metastasis in Non-small Cell Lung Cancer[J]. Journal of Sichuan University (Medical Sciences), 2019, 50(3): 373-378.
Citation: ZHU Jing, XU Wei-guo, XIAO Huan, et al. Application of a Radiomics Model for Preding Lymph Node Metastasis in Non-small Cell Lung Cancer[J]. Journal of Sichuan University (Medical Sciences), 2019, 50(3): 373-378.

影像组学联合预测模型在预测非小细胞肺癌淋巴结转移中的建立和应用价值

Application of a Radiomics Model for Preding Lymph Node Metastasis in Non-small Cell Lung Cancer

  • 摘要:
      目的  开发并验证影像组学模型,用于预测非小细胞肺癌术前淋巴结转移风险。
      方法  2014年1月至2015年12月100例经临床病理确诊的非小细胞肺癌100例组成训练组,并用该数据建立影像组学预测模型。影像组学特征在平扫及增强CT上进行提取。Lasso-logistic模型用于数据降维、特征选择以及影像组学标记的建立。一致性系数(ICCs)用于评价观察者内部以及观察者之间的重复一致性。以一致性指数(C-index)评价影像组学标签对淋巴结转移的鉴别预测能力,并采用受试者工作特性(ROC)曲线下面积(AUC)展示。多因素logistic回归分析用于建立影像组学联合预测模型,该预测模型的参数包括影像组学标记和独立的临床危险因素。建立的影像组学模型由2016年1月至2017年12月连续纳入的100例非小细胞肺癌病例组成验证组进行验证。采用AUC评价该模型的鉴别预测效能,并用Delong检验进行模型间(联合预测模型与单纯使用22个影像组学标记的模型之间)的比较;用Hosmer-Lemeshow good of fit test(拟合优度检验)评价预测模型的校准度,其结果使用校正曲线表示,以比较模型预测的结果与实际淋巴结转移的一致性。
      结果  提取特征时,观察者内部和观察者间的一致性好,ICC均大于0.75。从300个影像组学特征中提取出22个,其组成的影像组学标记,对于鉴别预测淋巴结转移状态的AUC,训练组为0.781,验证组为0.776。建立的影像组学预测模型包含了影像组学标记和血清癌胚抗原(CEA)、细胞角蛋白19片段抗原(CYFRA21-1)、癌抗原125(CA125)水平。用此联合预测模型预测淋巴结转移状态,训练组的AUC为0.836,验证组的AUC为0.821,均高于训练组和验证组单纯使用22个影像组学标记的模型,差异有统计学意义(P<0.05)。影像组学联合预测模型在训练组和验证组中均有较好的校准度,与实际淋巴结转移一致性高。
      结论  本研究开发了一个包含了影像组学特征、临床危险因素的影像组学联合预测模型,该模型能够直观预测非小细胞肺癌患者术前的淋巴结转移风险。

     

    Abstract:
      Objective  To establish a radiomic model for predicting lymph node (LN) metastasis in patients with non-small cell lung cancer (NSCLC).
      Methods  The prediction model was developed using a training cohort comprising 100 patients with clinicopathologically confirmed NSCLC. Data were gathered from January 2014 to December 2015. Radiomic features of NSCLC were obtained from non-contrast and enhancement computed tomography (CT). Lasso-logistic regression models were established for data dimension reduction, feature selection, and radiomics signature building. Consistency coefficient (ICCs) was used to evaluate the consistency between observer interior and interobserver.The consistency index (C-index)is used to evalutate the prediction of lymph node metastasis by using the radiomics signature, shown with the area under the receiver operating characteristic curve (AUC).Multivariate logistic regression analyses were performed to develop the prediction model, considering radiomics signature and clinicopathologic risk factors. The radiomics model was validated in a validation cohort comprising 100 consecutive NSCLC patients from January 2016 to December 2017 in terms of its calibration and discrimination. AUC was used to evaluate the predictive effectiveness of the model, and Delong test was used to compare models. Hosmer-Lemeshow good of fit test was used to evaluate the calibration of prediction models.The results were represented by correction curves to compare the consistency between the predicted results of the model and the actual probability of LN metastasis.
      Results  The consistency between observer interior and interobserver was good, with ICC higher than 0.75.The radiomics signature, including 22 selected features, was associated with LN metastasis. AUC was 0.781 in training cohort and 0.776 in validation cohort. The individualized prediction model identified radiomics signature, neuron specific enolase (CEA), cytokeratin 19 fragment antigen 21-1 (CYFRA21-1), and carbohydrate antigen 125 (CA125) as independent predictors. The model showed good discrimination, with 0.836 AUC in the training cohort, and 0.821 AUC in the validation cohort. The model in both the training and validation cohorts had good calibration,which demonstrated high consistency with the actual LN metastasis.
      Conclusion  The radiomics model incorporating radiomics signature and clinical risk factors can be conveniently used to facilitate preoperative individualized prediction of LN metastasis in patients with NSCLC.

     

/

返回文章
返回