论文部分内容阅读
目的构建肝癌新的分期系统。方法1999-2003年166例肝癌纳入研究,单因素和多因素分析预后影响因素。内部效度交叉程序(internal cross-validation procedure)将研究样本分为运算组(67%)和测试组(33%),重复10次。最佳模型在每个运算组计算,Cox模型计算每个变量相对风险度。测试组评估新分期的效度。结果最后模型由TNM、Ch ild-Pugh和AFP 3个协变量构成,分配线性记分(0/1/2)给每个协变量产生一个0~5分的联合分期记分。该分期较Okuda分期提供了更准确的信息并具有CLIP记分相似的预后鉴别。结论新的分期系统考虑肝功能和肿瘤特征因素,能够较准确预测肝癌预后,较适宜晚期肝癌。
Objective To construct a new staging system for liver cancer. Methods A total of 166 cases of HCC from 1999 to 2003 were included in the study. The prognostic factors were analyzed by univariate and multivariate analysis. Internal cross-validation procedure The study samples were divided into arithmetic group (67%) and test group (33%), repeated 10 times. The best model is calculated for each calculation group, and the Cox model calculates the relative risk for each variable. The test group assesses the validity of the new staging. Results The final model consisted of 3 covariates of TNM, Ch ild-Pugh and AFP. The distribution of linear scores (0/1/2) produced a joint score of 0-5 for each covariate. This stage provides more accurate information than the Okuda stage and has similar prognostic discrimination with CLIP score. Conclusion The new staging system, considering the liver function and tumor characteristics, can predict the prognosis of HCC more accurately and is more suitable for advanced HCC.