A Simulation Model-based Framework to Support Oncology Guidelines and Practice

摘要 在这个前所未有的发现和大数据时代，个性化的癌症护理是复杂的。规模庞大且不断发展的 知识库要求临床医生和政策制定者综合不同的数据来设计新的试验、告知 临床指南、实践和政策。医学研究所和其他机构建议使用 在这些情况下进行模拟建模，以综合证据和支持临床护理。仿真建模 涉及到使用数学模型来结合各种证据来源来评估干预措施 可能会改变疾病的进展并影响结果。我的首要目标是用这个K99/R00奖 获得必要的技能和经验，成为独立的研究人员和使用 肿瘤学中的模拟建模。 此修订后的应用程序包括建模方法方面的培训，以填补我在知识和使用方面的空白 该培训旨在构建新的模拟模型，以解决基因表达谱(GEP)指导的护理 每年有近18万名美国女性被诊断出患有早期乳腺癌。培训目标为：目标K1。 应用数据发现和综合方面的训练来开发输入参数，以模拟化学药物的影响。 内分泌(与内分泌)治疗对癌症结果的影响，如复发、死亡率和化疗- 基于患者(例如，年龄、种族、共病)和肿瘤(例如，肿瘤大小、分级)的相关毒性 特征和GEP测试结果；目标K2。将训练应用于仿真建模以建立模型 组合来自目标K1的输入参数以预测癌症结果；以及目标K3。将培训应用于 不确定性量化以估计与要放置结果的模型化结果相关联的变异性 临床医生和指南制定者的背景。本研究的目的是：针对R1。执行模型验证以 演示模型使用未使用的独立数据源再现预测的能力 在目标K1中；对模型进行必要的修改；目标R2。使用经过验证的模型提供以下内容的摘要 支持发育的示范妇女群体化疗的利弊平衡 临床指南；以及目标R3。创建交互式Web界面，以提供有关 基于个体特征的化疗-内分泌(与内分泌)治疗对癌症预后的影响。 考虑到我在试点资金方面的良好记录，21 High Impact，我是唯一有资格获得这个奖项的人 出版物，坚实的量化研究基础，对研究事业的承诺，以及初步 与我的多学科指导团队一起进行建模研究。非同寻常的机构资源与 通过癌症干预和监测建模网络(CISNET)为此提供了理想的环境 申请。综合的研究和培训将使我成为一名独立的研究员 使用模拟建模来协助将快速发展的知识转化为肿瘤护理。

Abstract Personalized cancer care is complex in this unprecedented era of discovery and big data. The large and evolving knowledge base requires clinicians and policymakers to synthesize diverse data to design new trials, inform clinical guidelines, practice, and policy. The Institute of Medicine and others have recommended the use of simulation modeling in these situations to synthesize evidence and support clinical care. Simulation modeling involves the use of mathematical models to combine various sources of evidence to assess how interventions could alter the progression of a disease and affect outcomes. My overarching goal is to use this K99/R00 award to gain the necessary skills and experience to become an independent researcher and leader in the use of simulation modeling in oncology. This revised application includes training in modeling methods needed to fill gaps in my knowledge and use that training to build a new simulation model to address gene expression profile (GEP)-guided care for the nearly 180,000 US women annually diagnosed with early-stage breast cancer. The training aims are: Aim K1. Apply training in data discovery and synthesis to develop input parameters to model the effects of chemo- endocrine (vs. endocrine) therapy on cancer outcomes such as recurrence, mortality, and chemotherapy- related toxicity based on patient (e.g., age, race, comorbidity) and tumor (e.g., tumor size, grade) characteristics, and GEP test results; Aim K2. Apply training in simulation modeling to build a model combining input parameters from aim K1 to project cancer outcomes; and Aim K3. Apply training in uncertainty quantification to estimate the variability associated with modeled outcomes to place results in context for clinicians and guideline developers. The research aims are: Aim R1. Perform model validation to demonstrate the model's ability to reproduce predictions using an independent data source that was not used in aim K1; make necessary revisions to the model; Aim R2. Use the validated model to provide a summary of the balance of benefits and harms of chemotherapy in exemplar groups of women to support the development of clinical guidelines; and Aim R3. Create an interactive web-interface to provide model results on the effects of chemo-endocrine (vs. endocrine) therapy on cancer outcomes based on individual characteristics. I am uniquely qualified for this award given my strong track record of pilot funding, 21 high impact publications, a solid quantitative research foundation, commitment to a research career, and preliminary modeling research with my multidisciplinary mentoring team. The exceptional institutional resources coupled with the Cancer Intervention and Surveillance Modeling Network (CISNET) provide the ideal setting for this application. The integrated research and training will leave me poised to become an independent researcher using simulation modeling to assist in the translation of rapidly evolving knowledge into oncology care.