基于MSCT灌注参数的非小细胞肺癌纵隔淋巴结转移的列线图模型的构建与验证  

Construction and validation of a nomogram model of mediastinal lymph node metastasis in non-small cell lung cancer based on MSCT perfusion parameters

作　　者：王东 徐凤琳[2] 于小涵 盖雪[1] 郝春晓 赵立群[1] WANG Dong;XU Fenglin;YU Xiaohan;GAI Xue;HAO Chunxiao;ZHAO Liqun(The First Affiliated Hospital of Qiqihar Medical College,Heilongjiang Qiqihar 161000,China;Pathological Center of Qiqihar Medical College,Heilongjiang Qiqihar 161000,China)

机构地区：[1]齐齐哈尔医学院附属第一医院,黑龙江齐齐哈尔161000 [2]齐齐哈尔医学院病理中心,黑龙江齐齐哈尔161000

出　　处：《现代肿瘤医学》2023年第22期4206-4211,共6页Journal of Modern Oncology

基　　金：黑龙江省齐齐哈尔市科技计划联合引导项目(编号:LSFGG-2022074)。

摘　　要：目的:探究基于多层螺旋CT(MSCT)灌注参数构建列线图模型预测非小细胞肺癌(NSCLC)纵隔淋巴结转移的预测价值。方法:纳入2010年1月至2020年1月我院380例NSCLC患者作为建模组,其中纵隔淋巴结转移患者159例,未转移患者221例;另于2020年2月至2023年2月多中心选取120例NSCLC患者作为验证组。比较两组临床资料及MSCT灌注参数,采用Logistic回归分析筛选NSCLC纵隔淋巴结转移的影响因素,构建列线图模型;通过校准曲线、受试者工作特征(ROC)曲线分析该模型的一致性与诊断效能,并进行外部验证。结果:Logistic回归分析结果显示,病理类型(OR=0.730,95%CI=0.567~0.941)、肿瘤直径(OR=12.195,95%CI=3.145~47.289)、肿瘤同侧纵隔淋巴结短径(OR=14.125,95%CI=2.857~69.832)、血清癌胚抗原(CEA)水平(OR=9.814,95%CI=2.415~39.882)、病灶区域血流量(BF)(OR=11.348,95%CI=3.074~41.892)、血容量(BV)(OR=11.456,95%CI=2.895~45.337)、平均通过时间(MTT)(OR=0.456,95%CI=0.274~0.759)、表面通透性(PS)(OR=10.475,95%CI=2.419~45.363)是NSCLC患者纵隔淋巴结转移的独立影响因素(P<0.05);对列线图预测模型进行内部验证与外部验证,模型的一致性指数(C-index)在建模组与验证组分别为0.921、0.918;校准曲线显示该列线图模型预测建模组与验证组NSCLC患者纵隔淋巴结转移风险均具有良好的区分度与精准度;应用ROC曲线分析该列线图模型预测建模组与验证组NSCLC患者纵隔淋巴结转移风险的效率,建模组曲线下面积(AUC)为0.956(95%CI=0.942~0.978),验证组AUC为0.950(95%CI=0.934~0.991)。结论:基于MSCT灌注参数构建NSCLC纵隔淋巴结转移的列线图模型区分度及校准度良好,可实现对NSCLC患者纵隔淋巴结转移的个体化评估。Objective:To investigate the predictive value of constructing a nomogram model to predict mediastinal lymph node metastasis in non-small cell lung cancer(NSCLC)based on multi-slice CT(MSCT)perfusion parameters.Methods:380 patients with NSCLC in our hospital from January 2010 to January 2020 were included as the modeling group,including 159 patients with mediastinal lymph node metastasis and 221 patients without metastasis.Another 120 patients with NSCLC were selected from February 2020 to February 2023 in a multicenter as the validation group.The clinical data and MSCT perfusion parameters of the two groups were compared,and Logistic regression analysis was used to screen the factors influencing mediastinal lymph node metastasis in NSCLC and construct a nomogram model.The consistency and diagnostic efficacy of this model were analyzed by calibration curves and receiver operating characteristic(ROC)curves,and external validation was performed.Results:Logistic regression analysis showed that the type of pathology(OR=0.730,95%CI=0.567~0.941),tumor diameter(OR=12.195,95%CI=3.145~47.289),short diameter of mediastinal lymph nodes ipsilateral to the tumor(OR=14.125,95%CI=2.857~69.832),serum carcinoembryonic antigen(CEA)(OR=9.814,95%CI=2.415~39.882),blood flow(BF)in the lesion area(OR=11.348,95%CI=3.074~41.892),blood volume(BV)(OR=11.456,95%CI=2.895~45.337),mean transit time(MTT)(OR=0.456,95%CI=0.274~0.759),and permeability surface(PS)(OR=10.475,95%CI=2.419~45.363)were independent factors influencing mediastinal lymph node metastasis in NSCLC patients(P<0.05).The C-index of the model was 0.921 and 0.918 in the modeling group and the validation group,respectively.The calibration curve showed that the nomogran model predicted the risk of mediastinal lymph node metastasis in the modeling group and the validation group with good discrimination and accuracy.The area under the curve(AUC)was 0.956(95%CI=0.942~0.978)in the modeling group and 0.950(95%CI=0.934~0.991)in the validation group,when the ROC curve was applied to analyz

关 键 词：非小细胞肺癌 纵隔淋巴结转移 MSCT灌注成像 定量参数 

分 类 号：R734.2[医药卫生—肿瘤]