Cox模型在乳腺癌临床病理学预后分析中的应用  

A clinicopathological analysis of prognosis in breast cancer by using Cox model

作　　者：郑唯强[1] 詹镕洲[1] 龚志锦[1] 李全华[1] 黄丽娟[1] 

机构地区：[1]第二军医大学基础医学部病理解剖学教研室,第二军医大学卫生勤务系卫生统计学教研室

出　　处：《第二军医大学学报》1995年第3期231-233,共3页Academic Journal of Second Military Medical University

基　　金：第二军医大学青年科学基金

摘　　要：应用半参数方法Ｃｏｘ模型并借助电子计算机，对８５例具有随访资料的乳腺癌患者的组织学分级、分型、临床分期、肿瘤大小、核仁组成区嗜银蛋白（ＡｓＮＯＲ）、核形状因子（ＦＦ）、年龄、ｐ５３癌基因蛋白、ＤＮＡ倍体、雌激素受体（ＥＲ）、核增殖抗原（ＰＣＮＡ）等１１个临床病理指标进行分析，结果经筛选进入Ｃｏｘ模型中与预后有关的指标为组织学分级、临床分期、ＦＦ、ＥＲ与ＰＣＮＡ，经统计学处理，该Ｃｏｘ模型的Ｓｃｏｒｅ值为４４．７４３，Ｐ＜０．０００１，具有非常显著意义。因而根据该Ｃｏｘ模型及有关统计学公式，可较客观而又正确地评估有关指标的相对危险度，并可由绘制的生存曲线图求得某一指标在一定时间内的生存率。In a following-up study of 85 breast cancer, the clinicopathological measurements including histological grade, type, clinical stage, tumor size, argynophilic nucleolar organizer regions(AgNOR) , nuclear form factor(FF) , age, p53 oncoprotein, DNA content, estrogen receptor(ER) , and proliferative cellular nuclear antigen(PCNA) were analyzed by the computer.The results showed that the elected markers relevant to the prognosis were histological grade, clinical stage, FF, ER and PCNA after Cox model analysis . It seemed to be a very significant Cox model(Score value was 44.743, df=5, P<0. 0001). So according to the Cox model and the statistics formula, we could objectively and rightly estimate the relative risk of the marker. In addition, we could easily get the survival rate of a certain marker during a certain period based on the worked survival curve picture.

关 键 词：COX模型 临床病理学 预后 乳腺癌 

分 类 号：R737.9[医药卫生—肿瘤]