【作者】 王海鹏；

【导师】 邵永孚；

【作者基本信息】 中国协和医科大学 ， 肿瘤学， 2004， 博士

【摘要】 原发性肝癌分期方法的比较研究 目的 评估原发性肝癌五种分期方法的预后价值，包括三种临床分期方法：Okuda分期、CLIP(Cancer of The Liver Italian Program，CLIP)评分、LCSGJ(Liver Cancer Study Group of Japan，LCSGJ)第四版TNM(tumor-node-metastasis，TNM)分期及两种病理分期方法：UICC(International Union Against Cancer)／AJCC(American Joint Committee on Caner)第五版及第六版TNM分期，探讨其临床应用价值。 材料和方法 回顾性分析自1995年1月至1999年12月196例根治切除原发性肝癌的临床病理及随访资料。用log-rank检验进行生存率的单因素比较，用Cox回归进行多因素分析。用线性趋势卡方检验、似然比卡方检验、Harrell’s c-指数(concordant index，c-指数)评估各分期系统的同质性、判别力，用AIC(Akaike Information Criterion，AIC)寻找最佳的预后模型。统计分析用SPSS和S-PLUS完成。 结果 本组患者1、3、5年生存率分别为84.7％、64.1％、34.9％，中位生存时间41个月。单因素分析显示就诊时症状、白蛋白水平、肝硬化、门静脉高压、Child分级、肿瘤大体类型(单结节／多结节)、AFP(Alpha-fetoprotein，AFP)水平、门静脉瘤栓、肿瘤直径、微血管受侵、侵犯临近脏器、淋巴结转移对预后的影响有统计学显著意义；多因素分析显示，淋巴结转移、Child-Pugh分级、门静脉瘤栓、肿瘤大体类型、门静脉高压、微血管侵犯是预后的独立负相关因素。临床分期方法中，CLIP评分的同质性、单调性、判别力最佳。病理分期方法中，UICC／AJCC第六版TNM分期的同质性、单调性、判别力优于第五版UICC／AJCC TNM分期。CLIP评分和UICC／AJCC第六版TNM分期的统计学更多还原

【Abstract】 ObjectiveTo evaluate the prognostic value of five staging systems for HCC （hepatocellularcarcinoma）, including three clinical staging systems and two pathologic staging systems. The clinical staging systems evaluated were Okuda stage、CLIP （Cancer of The Liver Italian Program, CLIP） scoring system and the 4^th edition of LCSGJ （Liver Cancer Study Group of Japan, LCSGJ） TNM （tumor-node-metastasis, TNM） staging system. The pathologic staging systems compared were the 5th and 6th edition of UICC （International Union Against Cancer） /AJCC （American Joint Committee on Caner） TNM staging system.Materials and methodsClinicopathologic and follow-up data of 196 patients with curative resection of HCC from January 1995 to December 1999 were retrospectively studied. Univariate analysis of the prognostic significance of various clinicopathologic characteristics was performed using log rank test. Multivariate analysis was performed using Cox regression model. Linear trend chi-square test, likelihood ratio test and Harrel’s concordant index were used to compare the homogeneity, discriminatory capacity, and monotonicity of gradients. AIC （Akaike information criterion, AIC） was also calculated to find the most efficient model.Results1, 3, and 5-year survival rates were 84.7%、64.1% and 34.9%, respectively, and median survival time was 41 months. Survival was significantly worse in those with symptoms at presentation, lower level of albumin, cirrhosis, portal hypertension, higher Child score, multiple tumors, higher AFP （Alpha-fetoprotein AFP） level, tumor更多还原