A framework for multi-indication evidence synthesis in oncology Health Technology Assessment

越来越多的药物被用于治疗不同类型的癌症。这些药物具有共同的作用机制，如阻断癌血管生长（抗血管生成药物，如贝伐单抗）。这些药物通常首先由监管机构许可用于特定适应症（用于特定治疗线中的特定癌症组织学），并且随着时间的推移，其许可证扩展到包括其他适应症。例如，贝伐单抗目前已被监管机构批准用于11种适应症，在获得特定适应症的监管许可证后，该药物将由卫生系统进行评估（临床和经济价值），以确定该药物是否可用于该适应症的患者。这就是所谓的卫生技术评估（HTA）。支持HTA的证据通常包括一项适应症特异性研究，该研究可以基于中间（替代）结局，如无进展生存期，而不是HTA机构要求的最终患者相关终点（如总生存期）。对单一适应症的关注意味着，即使存在大量关于药物其他适应症的证据，每次评估也会受到最终终点的高度证据不确定性的影响。我们建议探索是否可以通过跨适应症以及适应症内的研究来更好地利用证据。这种分析将在各种适应症之间共享信息，减少现有适应症的决策不确定性，并对药物在未来适应症中的价值提供更现实的预测。因此，该项目的总体目标是开发一个多适应症证据综合框架，以支持肿瘤学HTA。该框架将概述指定、进行和解释多指标证据综合所需的判断和分析，并将纳入一套工具。三个互补的方法学工作包将支持该框架的制定：WP1：在本工作包中，我们将确定贝伐珠单抗在11种适应症中的相对有效性证据，将在本工作方案中作为范例的真实的案例研究。为了支持多适应症分析，必须适当总结证据基础，供临床和政策决策者考虑。因此，我们将扩展现有方法，将证据映射到多适应症肿瘤学环境（跨适应症，具有多种结局，并显示证据的不成熟程度）。WP2：多适应症证据综合方法在本工作包中，我们将开发肿瘤学多适应症证据综合方法。我们将考虑在单个终点上跨适应症共享信息的模型，或者，在替代关系上跨适应症共享信息的模型。我们还将制定方法，明确方法所规定的适应症之间的共享水平，确保决策者和临床专家可以明确考虑基本假设的适当性。WP 3：支持政策框架的额外工作我们将进行模拟研究，从贝伐珠单抗案例研究推广到更广泛的多适应症药物，并支持框架的建议和指导原则，这些建议和指导原则涉及何时可能值得或不值得进行多适应症分析。我们还将开发用于正式诱导临床专家判断的方法，以支持多适应症分析。最后，我们将展示如何更充分地使用多适应症证据有助于卫生技术评估决策，通过评估的影响，整合多适应症证据合成使用贝伐单抗的案例研究的成本效益。

An increasing number of drugs are used to treat different types of cancer. These present a common mechanism of action, such as blocking cancer blood vessel growth (anti-angiogenics such as bevacizumab). These drugs are typically first licensed by regulatory authorities for use in a specific indication (for a particular cancer histology in a particular line of treatment), and over time their licences are extended to include additional indications. Bevacizumab, for example, is currently approved by regulators for use in 11 indications.Closely after the regulatory license is awarded for a particular indication, the drug is appraised (for clinical and economic value) by health systems to determine whether it is made available (funded) to patients in that indication. This is called Health Technology Assessment (HTA). The evidence supporting HTA typically comprises of an indication-specific study that can be powered on intermediate (surrogate) outcomes, such as progression-free survival, rather than the final patient-relevant endpoints required by HTA agencies (e.g. overall survival). The focus on single indications means that every appraisal is subject to a high level of evidential uncertainty on final endpoints, even where considerable evidence on the drug exists for other indications. We propose exploring whether better use of evidence can be made by looking across, as well as within, indications. Such analyses would share information across indications, reducing decision uncertainty across the existing indications and providing more realistic predictions of the value of the drug in future indications. The overall aim of this project is, therefore, to develop a framework for multi-indication evidence synthesis to support oncology HTA. The framework will outline the judgements and analyses required to specify, conduct, and interpret the synthesis of multi-indication evidence, and will incorporate a package of tools.Three complementary work packages of methodological work will support the development of the framework:WP1: Presenting and summarising evidence In this work package, we will identify the relative effectiveness evidence for bevacizumab across its 11 indications, the real case study that will be used as an exemplar in this programme of work. To support multi-indication analyses, it is important that the evidence base is summarised appropriately for consideration by clinical and policy decision makers. Therefore, we will extend existing methods for the mapping of evidence to the multi-indication oncology setting (across indications, with multiple outcomes, and displaying the level of immaturity of evidence). WP2: Methods for multi-indication evidence synthesisIn this work package, we will develop methods for multi-indication evidence synthesis in oncology. We will consider models for sharing of information across indications on a single endpoint and, alternatively, for sharing of information across indications on surrogate relationships. We will also develop methods for making explicit the level of sharing across indications imposed by the methods, ensuring that the appropriateness of the underlying assumptions can be explicitly considered by decision makers and clinical experts. WP3: Additional work to support a policy framework We will perform a simulation study to generalise from the bevacizumab case study to a broader set of multi-indication drugs and support the framework's recommendations and guiding principles relating to when multi-indication analyses may, or may not, be worthwhile. We will also develop methods for the formal elicitation of the judgment of clinical experts to support multi-indication analyses. Finally, we will demonstrate how the fuller use of multi-indication evidence can contribute to HTA decisions, by evaluating the impact on cost-effectiveness of integrating multi-indication evidence synthesis using the bevacizumab case study.