Designs for phase I trials of combinations of agents

描述（由申请人提供）：大多数用于肿瘤学I期试验设计的方法用于涉及单一细胞毒性药物的研究。这些研究的目的是估计“最大耐受剂量”，即在可接受的毒性水平下可以给予的最高剂量。这些方法的一个关键假设是剂量响应曲线的单调性。在这种情况下，剂量-反应曲线被称为遵循“简单顺序”，因为任何一对剂量的“剂量限制性毒性”（DLT）概率的顺序是已知的；施用更大剂量的药物可以预期在越来越多的患者中产生DLT。在I期试验中对药物组合进行测试变得越来越普遍。在这些研究中，与剂量组合相关的DLT概率通常遵循“偏序”，因为有对剂量组合的概率顺序是未知的。本建议使用贝叶斯方法，结合连续重新评估方法和顺序限制推理的特点，开发第一阶段试验的设计，其中概率遵循部分顺序。此外，我们将根据周期特异性毒性调整我们的方法。最后，我们将开发可上网的软件，以协助用户设计和执行部分订购的第一阶段试验。尽管我们的重点是第一阶段的联合试验，但我们开发的方法可以揭示第一阶段试验设计中的其他问题，包括有序群体的研究和癌症疫苗的试验。

DESCRIPTION (provided by applicant): The majority of methods for the design of Phase I trials for in oncology are intended for studies involving a single cytotoxic agent. The goal of these studies is to estimate the 'maximally tolerated dose', the highest dose that can be administered with an acceptable level of toxicity. A key assumption of these methods is the monotonicity of the dose- response curve. In this case, the dose-response curve is said to follow a 'simple order' because the ordering of the probabilities of a 'dose-limiting toxicity' (DLT) for any pair of doses is known; administration of greater doses of the agent can be expected to produce DLT's in increasing proportions of patients. It is becoming increasingly common for combinations of agents to be tested in phase I trials. In these studies, the probabilities of a DLT associated with the dose combinations often follow a 'partial order' in that there are pairs of dose combinations for which the ordering of the probabilities is not known. This proposal uses Bayesian methods, combining features of the continual reassessment method and order restricted inference to develop designs for phase I trials in which the probabilities follow a partial order. In addition, we will adapt our methods for cycle-specific toxicities. Finally, we will develop internet-accessible software to assist users in designing and carrying out partially ordered phase I trials. Even though our emphasis is on phase I trials of combinations, the methods we develop can shed light on other issues in phase I trial design, including the study of ordered groups and trials of cancer vaccines. PUBLIC HEALTH RELEVANCE: Dose-finding trials of combinations of agents are becoming increasingly common in cancer research. While there are many proposed methods for single agent trials, there are relatively few options for designing phase I trials of combinations. As in single agent trials, it is crucial to find a combination of doses that can be administered with an acceptable level of toxicity in order that these new therapies can be tested for efficacy. Without adequate statistical methods potentially effective combinations may be discarded as too toxic or get tested in subsequent studies at sub-optimal dose combinations. The overall goal of this proposal is to develop designs for phase I trials of combinations of agents.