癌症统计学治愈的两种检验方法

Two Test Procedures or Statistical Cure in Cancer

本文介绍用于癌症统计学治愈评价的两种非参数方法:反射和检基检验,引入同源性生存率方差估计值将其加以改进,并以临床实例展示其计算步骤。这两种检验在无终检时还原为对应的经典中位数检验。前者的还原形式可转换为x^2检验,可能更易于被接受,获得较多临床应用机会。

This paper concerns two non-parametric tests: reflected and test-based,for assessing statis- tical cure in cancer. They are developed by in- troducing the homogenetic estimate for the vari- ance of survival rate. A clinical example illus- trates the methodology. The two tests reduce to their corresponding classical median tests when there is no censoring. The reflected test is probably easy to be accepted and applied in clinical practice as its reduced form can be con- verted to chi square test.