影像组学评估进展期胰腺癌化疗敏感性及预后  

Radiomics evaluation of chemosensitivity and prognosis of advanced pancreatic cancer

作　　者：刘畅[1] 徐有莉 杨锐[1] 石喻[1] LIU Chang;XU You-li;YANG Rui;SHI Yu(Department of Radiology,Shengjing Hospital,China Medical University,Shenyang 110004,China)

机构地区：[1]中国医科大学附属盛京医院放射科,辽宁沈阳110004

出　　处：《中国临床医学影像杂志》2021年第2期120-124,共5页Journal of China Clinic Medical Imaging

基　　金：国家自然科学基金面上项目(81771802、81771893)。

摘　　要：目的:探讨影像组学对进展期胰腺癌(Advanced pancreatic cancer,APC)化疗敏感性和预后的评估价值。方法:回顾性分析经病理证实的114例患者的CT增强及形态学特征,采用AnalysisKit软件提取胰腺实质期APC纹理特征参数,采用R语言进行特征筛选,采用Logistic回归模型及Cox比例风险模型评价APC患者的化疗敏感性和预后生存期。结果:在预测化疗敏感性方面,影像组学特征的曲线下面积(Area under the curve,AUC)优于临床特征,AUC分别为0.914 (95%CI:0.862,0.965),0.728(95%CI:0.639,0.816),P=0.000 1。在预测预后生存期方面,联合影像组学及临床特征效果优于仅基于临床特征,建立的回归模型的C-index分别为0.692(95%CI:0.640,0.743),0.608(95%CI:0.551,0.665),P=0.003。结论:影像组学对评价APC的化疗敏感性及预后具有较好的临床价值。Objective:To explore the value of imaging histology in evaluating the chemosensitivity and prognosis of advanced pancreatic cancer(APC).Methods:CT enhancement and morphological features of 114 patients proved pathologically were analyzed retrospectively.APC texture feature parameters in pancreatic parenchymal phase were extracted by AnalysisKit software.Feature screening was performed by R language.Logistic regression model and Cox proportional hazards model were used to evaluate chemotherapy sensitivity and prognosis survival time of APC patients.Results:In predicting chemosensitivity,the area under the curve(AUC)of imaging features is superior to clinical features,with AUC of 0.914(95%CI:0.862,0.965),0.728(95%CI:0.639,0.816),P=0.0001 respectively.In predicting the survival time of prognosis,the combined imaging and clinical features are superior to those based on clinical features only.The C-index of the regression model established is0.692(95%CI:0.640,0.743),0.608(95%CI:0.551,0.665),P=0.003 respectively.Conclusion:Imaging histology has good clinical value in evaluating the chemosensitivity and prognosis of APC.

关 键 词：胰腺肿瘤 体层摄影术 螺旋计算机 

分 类 号：R735.9[医药卫生—肿瘤] R814.42[医药卫生—临床医学]