Development and Evaluation of Accelerated Hazards Mixture Cure Model

描述(申请人提供)：摘要：随着癌症研究的进展，越来越多的癌症患者被治愈，他们不会因为癌症而复发或死亡。在给定患者的一系列危险因素的情况下，了解患者被治愈的机会变得越来越重要，或者在患者未被治愈的情况下，了解特定时间的存活率变得越来越重要。本项目的主要目的是建立一种新的、可行的混合治疗模型，用于估计治愈患者的比例和未治愈患者的生存概率。目前该领域的方法学假设治疗在试验的早期阶段是有效的，而新的模式可以允许治疗在初始时间不起作用，并在以后对未治愈的患者进行逐渐的e(R)效应。这一独特的特征使所提出的模型成为参与癌症流行病学研究的从业者或研究人员在对长期幸存者的经审查的生存数据进行建模时的重要替代模型。我们将开发一种半参数估计方法，将其与现有的模型和方法进行比较，通过仿真研究对所有模型进行评估，并将该模型应用于乳腺癌数据集。这项研究将为从业者分析治愈比例的癌症数据和评估风险e(R)Ets提供重要工具。该软件在R环境下的开发，使混合治疗模型在癌症流行病学研究中的应用成为可能。 公共卫生相关性：随着癌症研究的进步，越来越多的癌症患者被治愈。这个项目将探索一种新的模型来分析治愈患者的癌症生存数据。在R环境下的软件开发将使所提出的方法更容易应用于癌症流行病学研究。

DESCRIPTION (provided by applicant): Abstract : With the advances in cancer research, more and more cancer patients are cured, and they will not relapse or die due to the cancer. It becomes increasingly important to know a patient's chance of being cured given a set of risk factors of the patient, or the survival rate at a certain time if the patient is not cured. The primary aim of this project is to develop a new and 0exible mixture cure model for estimating the proportion of cured patients, and the survival probability of uncured patients. Current methodologies in this area assume that the treatment is e(R)ective at the early stage of trials, while the new model can allow the treatment to have no e(R)ect at the initial time and a gradual e(R)ect later on for uncured patients. This distinctive feature makes the proposed model an important alternative model for practitioners or researchers involved in cancer epidemiology studies in modeling censored survival data with long term survivors. We will develop a semiparametric estimation method, compare it with existing models and methods, evaluate all models by the simulation study, and apply the model to breast cancer data sets. The study will provide practitioners an important tool in analyzing the cancer data with cured fraction and evaluating the risk e(R)ects. The development of the software in the R environment will enable the possible use of the mixture cure model in the epidemiology cancer study easily. PUBLIC HEALTH RELEVANCE: Narrative With the advances in cancer research, more and more cancer patients are cured. This project will investigate a new model for analyzing the cancer survival data with cured patients. The software development in R environment will enable the proposed method used easily in epidemiology cancer study.