基于基线CT预测进展期胃癌新辅助化疗反应的Logistic回归模型研究  

Development of Logistic regression model based on baseline CT to predict neoadjuvant chemotherapy response in advanced gastric cancer

作　　者：吴岢珊 李柯颖 皈燕[3] 陈天武 WU Ke-shan;LI Ke-ying;GUI Yan;CHEN Tian-wu(Department of Radiology Affiliated Hospital of North Sichuan Medical College,Nanchong 637000,Sichuan;Department of Radiology,Jinshan Hospital Affiliated of Fudan University,Shanghai 210508;Department of Oncology,Affiliated Hospital of North Sichuan Medical College,Nanchong 637000,Sichuan;Department of Radiology,the Second Affiliated Hospital of Chongqing Medical University,Chongqing 400010,China)

机构地区：[1]川北医学院附属医院放射科,四川南充637000 [2]复旦大学附属金山医院放射科,上海210508 [3]川北医学院附属医院肿瘤科,四川南充637000 [4]重庆医科大学附属第二医院放射科,重庆400010

出　　处：《川北医学院学报》2025年第2期156-159,共4页Journal of North Sichuan Medical College

基　　金：国家自然科学基金项目(82271959)。

摘　　要：目的:构建基于基线CT预测进展期胃癌新辅助化疗反应的Logistic回归模型。方法:收集156例进展期胃癌患者接受新辅助化疗前后的CT影像资料及临床资料,按3∶1的比例随机分为训练集(n=117)与验证集(n=39)。单因素分析训练集治疗前相关指标与胃癌新辅助化疗反应性的关系;多因素Logistic回归分析影响预测胃癌新辅助化疗反应性的独立因素,并构建Logistic回归模型;受试者工作特征(ROC)曲线下面积(AUC)评估模型的预测效能,并在验证集中通过Kappa检验予以验证。结果:单因素分析显示,训练集患者治疗前cT分期、淋巴结转移及原发肿瘤体积(GTV)在有反应和无反应患者中比较,差异有统计学意义(P<0.05)。多因素Logistic回归分析结果显示,cT分期、cN分期、GTV是影响治疗反应性的独立影响因素(P<0.05)。ROC曲线分析显示,基于独立影响因素构建的Logistic回归模型在训练集中预测治疗反应性的AUC为0.724。Kappa检验显示,基于独立影响因素构建的Logistic回归模型在验证集中预测效能表现良好(Kappa=0.623)。结论:基于治疗前cT分期、淋巴结转移、GTV构建的Logistic回归模型对预测进展期胃癌新辅助化疗反应性有较大价值。Objective:To construct a Logistic regression model based on baseline CT to predict neoadjuvant chemotherapy response in advanced gastric cancer.Methods:Pretherapeutic and posttreatment CT imaging data and clinical data of 156 patients with advanced gastric cancer receiving neoadjuvant chemotherapy were retrospectively collected.The collected cases were randomly assigned into the training cohort(n=117) and the validation cohort(n=39) at a ratio of 3∶1.In the training cohort, the univariate analyse were performed to explore the relationship between relevant pretherapeutic indicators and the response of gastric cancer, and the indicators with statistical difference were included in a multivariate Logistic regression to determine the independent predictors.Subsequently, a Logistic regression model was constructed based on above independent predictors.Predictive performance of the model was evaluated by the receiver operating characteristic curve(ROC) and area under the ROC curve(AUC).In the validation cohort, the prediction efficiency of the model was verified by Kappa test.Results:In the training cohort, the univariate analysis showed statistically significant difference in cT stage, cN stage, and gross tumor volume(GTV) between patients with and without response(P<0.05).Multivariate analyses showed that cT stage, cN stage, and GTV were independent influencing factors of the response(P<0.05).ROC showed that the AUC of the Logistic regression model based on independent predictors to predict the treatment response was 0.724.In the validation cohort, the predictive model also performed well(Kappa=0.623).Conclusion:The Logistic regression model developed based on the cT staging, cN stage, and GTV is of great value in predicting the response of advanced gastric cancer after neoadjuvant chemotherapy.

关 键 词：胃癌 新辅助化疗 治疗反应性 预测 体层摄影术 X线计算机 

分 类 号：R735.2[医药卫生—肿瘤]