基于非小细胞肺癌原发灶影像组学的隐匿性淋巴结转移预测

Prediction of Occult Lymph Node Metastasis in NSCLC Based on Intratumoral Radiomics

目的 基于非小细胞肺癌(NSCLC)原发灶的影像组学特征构建预测模型,预测NSCLC隐匿性淋巴结转移。方法 搜集曾行胸部薄层对比增强CT(CE-CT)扫描,表现为单发周围型病灶,无淋巴结转移提示,后行根治术,病理结果为有/无淋巴结转移的NSCLC患者。搜集患者的肿瘤CE-CT特征、影像组学特征以及临床特征。采用PCC法降维,联合t检验与LASSO回归构建影像组学标签,采用二分类Logistic回归分析分别构建影像组学模型、CT联合临床模型、综合模型。结果 共纳入106例患者,比较各模型预测效能,综合模型为最优模型,在训练集及测试集的曲线下面积(AUC)分别为0.905、0.802。结论 综合影像组学特征及癌胚抗原(CEA)水平构建的综合模型,对NSCLC隐匿性淋巴结转移预测效能最佳。

Objective To construct a predictive model based on the radiomics features of non-small cell lung cancer(NSCLC) to predict occult lymph node metastasis.Methods Retrospective collection the patients underwent chest thin CE-CT scan,the images showed single peripheral lesions without lymph node metastasis,were treated by radical resection and those whose pathological results were NSCLC with or without lymph node metastasis.CT features,radiomics features and clinical features of the patients were collected.The dimension was reduced by PCC feature selection method,and the radiomics labels were constructed by combining t-test and LASSO regression.Logistic Regression method was used to construct the radiomics model,CT combined with clinical model,and comprehensive model,which were verified in the test set.Results A total of 106 patients were included.The comprehensive model was the optimal model.AUC of the training set and test set were 0.905 and 0.802,respectively.Conclusion The comprehensive model based on the combination of radiomic features and CEA level has the best predictive performance for occidental lymph node metastasis of NSCLC.