局部晚期非小细胞肺癌术后脑转移高危因素分析及风险模型建立  

Regression analysis and risk model establishment of high risk factors for postoperative brain metastases in locally advanced non-small cell lung cancer

作　　者：周开甲[1] 张鸣[1] 刘伯伟[1] 郑宗清[1] ZHOU Kaijia;ZHANG Ming;LIU Bowei;ZHENG Zongqing(Fujian Provincial Cancer Hospital/ the Affiliated Cancer Hospital of Fujian Medical University,Fuzhou,Fujian 350014,China)

机构地区：[1]福建省肿瘤医院/福建医科大学附属肿瘤医院

出　　处：《现代医药卫生》2019年第13期1947-1949,1954,共4页Journal of Modern Medicine & Health

摘　　要：目的通过多元回归分析建立局部晚期非小细胞肺癌术后脑转移相关高危因素的数学模型,为应用该模型预警肺癌脑转移提供理论支持。方法回顾性分析2011年6月至2013年6月106例在该院行外科手术的局部晚期非小细胞肺癌术后脑转移患者的临床特征、病理情况及相关分子标记物表达情况,通过多元回归分析其发生脑转移的高危因素,建立相关的数学模型。结果局部晚期非小细胞肺癌患者脑转移发生率与纵隔淋巴结转移数目、术前血清癌胚抗原(CEA)水平、肿瘤病理类型、EGFR基因突变与否、VGFR/Ki-67表达高低有关(P<0.05),纵隔淋巴结多发转移、术前血清CEA大于10μg/L、腺癌、EGFR基因突变、VGFR/Ki-67高表达为脑转移的高危因素(P<0.05),而脑转移的发生与年龄、性别、吸烟史、肿瘤分化程度及手术方式、P53表达情况无关(P>0.05)。局部晚期非小细胞肺癌术后脑转移相关因素的高危数学模型(多因素分析):logit(P)=9.198-0.721×病理类型-0.892×纵隔淋巴结阳性数-0.639×术前血清CEA水平-0.629×EGFR基因型-1.907×VGFR-1.041×Ki-67。P=0.48时,敏感度与特异度最高(敏感度80%,特异度78%)。即当P≥0.48为脑转移高危,P<0.48为脑转移低危。结论该数学模型对局部晚期非小细胞肺癌术后脑转移高危患者的预测准确性较好。通过对多个脑转移高危因素综合考虑,数字量化处理,能更加全面、准确地预测脑转移高危患者,可作为局部晚期非小细胞肺癌术后筛选脑转移高危患者的依据,为后续早期预防脑转移的治疗提供证据支持。ObjectiveTo establish a mathematical model for predicting brain metastasis in locally advanced non-small cell lung cancer(NSCLC)by multiple regression analysis,to provide theoretical support for the application of the model in the early warning of brain metastasis of lung cancer.MethodsThe clinical features,pathological changes and the expression of related molecular markers of 106 patients with brain metastases in locally advanced NSCLC after surgery from June 2011 to June 2013 in the hospital were retrospectively analyzed.The risk factors of brain metastasis were analyzed by multiple regression analysis,and the relevant mathematical model was established.ResultsThe incidence of brain metastasis in locally advanced NSCLC were related to the number of mediastinal lymph node metastasis,preoperative serum CEA level,tumor pathological type,EGFR gene mutation,VGFR/Ki-67 expression(P<0.05).The mediastinal multiple lymph node metastasis,preoperative serum CEA greater than 10μg/L,adenocarcinoma,EGFR gene mutation,VGFR/Ki-67 over expression were high risk factors for brain metastasis(P<0.05),but the occurrence of brain metastasis were not related to age,sex,smoking history,tumor differentiation degree and surgical method,P53 expression(P>0.05).Multivariate analysis:logit(P)was 9.198-0.721×pathological type-0.892×mediastinal lymph node positive number-0.639×preoperative serum CEA level-0.629×EGFR genotype-1.907×VGFR-1.041×Ki-67.When P=0.48 was used,sensitivity and specificity were the highest(sensitivity was 80%,specificity was 78%).That was to say,when P≥0.48 was high risk of brain metastases,P<0.48 was a low risk of brain metastases.ConclusionsThe model is accurate in predicting brain metastasis in patients with locally advanced NSCLC.Considering many high risk factors of brain metastasis,digital quantification can predict the patients at high risk of brain metastasis more comprehensively and accurately.It can be used as a basis for screening patients at high risk of brain metastasis after operation of locally

关 键 词：非小细胞肺癌 脑转移 高危因素 数学模型 

分 类 号：R734.2[医药卫生—肿瘤]