Statistical theory and methods for cancer clinical trials

癌症是加拿大和世界各地的一个主要公共卫生问题。临床试验是开发治疗癌症或预防癌症发展的新药或新程序的关键组成部分。统计方法在癌症临床试验的设计和分析中起着关键作用。由于对癌症生物学的理解取得了快速进展，近年来开发了许多新的药物和治疗方法。新的统计方法需要使临床试验评估这些药物和程序更准确和有效。开发和评估这些统计方法在临床试验中的应用一直是我们研究计划的主要主题。我们计划继续这个研究项目，重点研究我们在实践中最近遇到的以下新问题：(1)配对生存数据的统计分析与应用于II期癌症临床试验的设计和分析；(2)评价肿瘤生物标志物效果的统计程序：(3)生存分析中高维模型的变量选择；(4)纵向生活质量和生存数据联合建模。预计本研究将为临床试验中具有事件时间终点的应用统计学家提供一些新的方法，并为这些方法奠定坚实的数学基础，从而更有效和可靠地确定癌症预防和治疗的新策略，从而降低癌症的发病率和死亡率。从这项研究中发展出来的方法也有可能应用于其他类型疾病的临床试验，如艾滋病毒、艾滋病和心血管疾病，以及一些其他类型的医学研究，如流行病学研究中的队列设计。它们也可以应用于其他科学和工程领域，如心理学和可靠性工程。拟议的研究还将为培养下一代研究人员提供极好的机会，这些研究人员致力于推进统计理论和方法，以及我们对疾病的理解和治疗方面的知识。

Cancer is a major public health problem in Canada and around the world. Clinical trial is a key component in the development of new drug or procedure to treat the cancer or prevent the development of the cancer. Statistical methods play a pivotal role in the design and analysis of cancer clinical trials. Because of rapid progress in the understanding of cancer biology, many new drugs and treatment procedures have been developed in the recent years. New statistical methods are required to make the clinical trials evaluating these drugs and procedures more accurate and efficient. Development and evaluation of these statistical methods with applications to the clinical trials has been primary topic of our research program. We plan to continue this research program with focus on the study of the following new problems encountered recently in our practices: (1) Statistical analysis for paired survival data with applications to design and analysis of phase II cancer clinical trial; (2) Statistical procedure for evaluating effect of cancer biomarkers: (3) Variable selection in high-dimensional models in survival analysis; and (4) Joint modeling of longitudinal quality of life and survival data. It is expected that this research will generate some new methods for applied statisticians working in clinical trials with time-to-event endpoints and lay solid mathematical foundations for these methods, which will result in more efficient and reliable identification of new strategies for the prevention and treatment of the cancer and, therefore, reduction of morbidity and mortality from cancer. Methods developed from this research have also potential applications to clinical trials of other types of diseases, such as HIV AIDs and cardiovascular diseases, and also to some other types of medical researches, such as cohort designs in epidemiological studies. They may also be applied to other areas of science and engineering, such as psychology and reliability engineering. The proposed research will also provide excellent opportunities to train next generation of researchers working in advancing statistical theory and methodology and also knowledge in our understanding and treatment of diseases.