靶向药物全球同步研发设计中的几个问题

发布时间：2018-01-09 14:31

本文关键词：靶向药物全球同步研发设计中的几个问题　出处：《南京医科大学》2017年硕士论文　论文类型：学位论文

【摘要】：目的:近年来靶向药物研发和药物全球同步研发项目越发受到关注。靶向药物的临床试验设计仍仅限于单个地域,如果要在新地域上市,需要进行一个新的确证性临床试验,用于评价药物的有效性和安全性,这必将会延长药物研发周期,浪费大量人力物力。本文提出了靶向药物的全球同步研发项目方法设计,并对该方法的统计学性能(检验效能、一类错误)进行评价。此外,本文提出的设计在LCT(local clinical trial)部分仅需要较少的TE(target ethnic)人群样本,第二部分给出了不同参数设置下LCT所需的样本量大小,以期为实际工作提供指导意见。方法:将全人群分为NTE(non-target ethnic)人群和TE人群,其中又包括靶点阳性人群和靶点阴性人群,MRCT(multi-regional clinical trial)和LCT可以同时进行也可以序贯进行,分别采用全随机设计或富集设计,最终的分析将人群归类为NTE人群和TE人群,并将各自的检验统计量进行加权合并。对于LCT部分样本量的估计,考虑终点指标为连续性或二分类指标,TE人群与NTE人群的效应比值和权重对样本量的影响,并进行比较。结果:观察不同靶点人群比例和LCT样本量对两种设计检验效能的影响,结果显示,当TE人群与NTE人群、靶点阳性与靶点阴性人群效应相等时,相同分配样本量情况下全随机设计与富集设计的检验效能最高,经过筛选的富集设计检验效能最低;当靶点阴性人群效应较低时,富集设计的分配样本量与全随机设计相等时,富集设计的检验效能较高,经过筛选的富集设计检验效能最低;靶点阳性人群比例增大,检验效能随之增大;两种设计的一类错误均较为稳定,不随权重及靶点比例变化而变化。当TE人群与NTE人群的效应比值较大时,本研究方法的样本量较为贴合实际,当效应比值较小,尤其是二分类资料,样本量异常大。另外,随着权重的增加,样本量变化率增加明显。结论:在靶向药物全球同步研发项目中,根据种族因素及靶向药物针对不同靶点人群的效应差异,当有足够的证据表明靶向药物对靶点阳性人群有更好的疗效且拥有可靠的靶点检测方法时,我们建议采用富集设计。当不能确定靶向药物对阴性人群是否有效或无可靠地靶点检测方法时,建议采用全随机设计。采用富集设计可以减少样本量,降低研究成本,但是也要考虑筛选过程花费,以及伦理学是否可行。
[Abstract]:Objective: in recent years, more and more attention has been paid to the research and development of targeted drugs and their global synchronization. The clinical trials of targeted drugs are still limited to a single region, if they are to be listed in new regions. A new corroborative clinical trial is needed to evaluate the efficacy and safety of drugs, which is bound to prolong the drug development cycle. Waste a lot of manpower and material resources. In this paper, we propose a global synchronous research and development project design of targeted drugs, and evaluate the statistical performance of the method (test effectiveness, a class of errors). The design proposed in this paper requires only a small number of TE(target ethnic population samples in the LCT(local clinical Trial part. In the second part, the sample size required by LCT with different parameters is given. Methods: the whole population was divided into NTE(non-target ethnic group and te group. These include target positive and target negative groups. MRCT(multi-regional clinical trialand LCT can be carried out simultaneously or sequentially, using full random design or enrichment design, respectively. The final analysis classifies the population as NTE population and te population, and combines their test statistics weighted. For the estimation of partial sample size of LCT, the endpoint index is considered as continuity or two-classification index. The effect ratio and weight of te population and NTE population on sample size were compared. Results: the effects of different target population ratio and LCT sample size on the effectiveness of two design tests were observed. When the effect of target positive and target negative was equal between te and NTE, the test efficiency of full random design and enrichment design was the highest when the sample size was the same. The efficiency of enrichment design was the lowest. When the target negative population effect is low and the sample size of enrichment design is equal to that of full random design, the test efficiency of enrichment design is higher and that of enrichment design selected is the lowest. The proportion of target positive population increased and the test efficiency increased. The two kinds of errors are stable and do not change with weight and target ratio. When the ratio of effect between te population and NTE population is large, the sample size of this method is more practical. When the effect ratio is small, especially the two-classification data, the sample size is unusually large. In addition, with the increase of weight, the sample size change rate increases obviously. Conclusion: in the target drug global synchronization research and development project. According to the racial factors and the effects of targeting drugs on different target populations, when there is sufficient evidence that targeted drugs have better efficacy on target positive people and have reliable target detection methods. We recommend the use of enrichment design. When it is not certain whether the targeted drug is effective for negative population or there is no reliable target detection method, we recommend the full random design. The enrichment design can reduce the sample size. Reduce the cost of research, but also consider the cost of the screening process and the feasibility of ethics.
【学位授予单位】：南京医科大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：R91

【参考文献】