摘要
Ⅰ期临床试验的主要目的是探索药物毒性最大耐受剂量MTD,而MTD估计的准确与否将影响之后的Ⅱ期和Ⅲ期临床试验研究的结果。抗肿瘤药物Ⅰ期试验的特点是直接对病人进行试验,且样本量较小,这对构造估计精确度高并具有安全性保障要求的统计设计方法提出了挑战。回顾三种常用的Ⅰ期试验设计方法有:3+3设计、CRM设计和mTPI设计。3+3设计是应用较为广泛的传统方法,后两者是当前常用的贝叶斯自适应试验设计方法。通过大量模拟研究对三种方法从最优分配、安全性和估计MTD精确性三方面给以全面考察,并结合中国实际得出mTPI设计是比较适合推荐的I期临床试验设计方法的结论。
The primary goal of phase Ⅰ clinical trial is to identify the maximum tolerated dose（MTD） of a new drug.The high uncertainty from the estimation of MTD may result in the selected doses for phase Ⅱ and phase Ⅲ studies yielding either insufficient efficacy or ineffectively defining the most suitable dosing regimen for approval.The anit-tumor drug studies conduct clinical trials on patients and ordinarily the corresponding sample size is small,which need the development of more accurate and safer statistical methods for phase Ⅰ studies.This paper introduces three influential design methods in phase Ⅰ.Among them,3＋3 is the traditional design,CRM and mTPI designs are Bayesian adaptive clinical trial designs.By conducting extensive simulations,we examine the operating characteristics of three approaches from the angles of optimal-assignment,safety control and precision of estimation of MTD.This paper ends with a discussion from the prospective of China’s current situation and concludes that the mTPI approach is a reliable and an ethical method and may easily to be implemented to the real clinical practice.
出处
《统计与信息论坛》
CSSCI
2013年第5期25-32,共8页
Journal of Statistics and Information
基金
西安财经学院校级科研项目<试验设计中样本量估计的贝叶斯方法研究与应用>(11XCK10)
国家自然科学基金项目<贝叶斯非劣效桥接试验设计框架与分析方法研究>(81002190)
