P-2: Risk Prediction of platinum-based chemotherapy and Radiotherapy Outcome

该项目的首要目标是确定宿主流行病学和遗传因素， 预测含铂化疗或联合胸部放疗的疗效和毒性， NSCLC患者。我们将构建一个1,200例NSCLC患者（III期和IV期- 接受一线铂类化疗或确定性胸部放疗）。这批人将是 研究了流行病学、临床和大量合理选择的生殖系多态性， 与临床结果相关，使我们能够构建临床疗效的预测风险模型， 毒性我们估计将有约600名患者接受铂类化疗， 接受含铂化疗加明确胸部放疗的III期患者。有 三个具体目标：1）我们将确定可预测铂类药物疗效和毒性的新型遗传基因座 所有1,200名患者接受化疗和放疗。我们将采用基于途径的基因分型， 分析方法，以评估频率约8,000单核苷酸多态性的基因参与相关的途径， 铂和辐射反应。我们将研究单个SNP的主效应、单倍型和 SNP在调节功效和毒性中的累积效应。我们的假设是特定的基因型 改变铂剂的代谢或作用或与基因毒性效应相关的 放疗可能影响患者对这些疗法的疗效和毒性。2)我们将应用 识别影响NSCLC的基因-基因和基因-环境相互作用的机器学习工具 结果。我们将开发算法，以确定不同的铂或放射治疗的亚组 功效或毒性。我们的假设是，治疗反应是由共同的，低的， 多态性，并且这些多态性相互作用和/或宿主 决定对治疗反应的因素。3)我们将构建生存的预测风险模型， 通过将临床和流行病学数据与该项目的遗传数据相结合， 来自专门用于这些队列的其他R 01研究的信息，例如一系列表型测定。我们 假设在标准临床和流行病学变量中加入遗传标记 将改善最终风险评估模型的生存和毒性预测。我们将 比较所有患者、仅接受化疗的患者和 采用联合治疗。该项目产生的风险模型可能允许临床医生识别 治疗开始前最有可能和最不可能获益或发生毒性的患者， 在为个体患者规划化疗和放疗方面具有巨大的临床益处。

The overarching goal of this project is to determine host epidemiologic and genetic factors that will be predictive of efficacy and toxicity of platinum-based chemotherapy or combined with thoracic radiotherapy in NSCLC patients. We will construct a well-characterized cohort of 1,200 NSCLC patients (Stages III and IV - receiving first-line platinum-based chemotherapy ¿ definitive thoracic radiotherapy). This cohort will then be studied for epidemiologic, clinical and a large number of rationally selected germline polymorphisms to correlate with the clinical outcome to allow us to construct predictive risk models for clinical efficacy and toxicity. We estimate there will be ~ 600 patients treated by platinum-based chemotherapy alone, and 600 Stage III patients receiving platinum-based chemotherapy plus definitive thoracic radiotherapy. There are three specific aims: 1) we will identify novel genetic loci that predict efficacy and toxicity to platinum-based chemotherapy and radiotherapy in all 1,200 patients. We will adopt a pathway-based genotyping and analyzing approach to evaluate frequencies .of about 8,000 SNPs in genes involved in pathways relevant to platinum and radiation response. We will examine individual SNP main effects, haplotypes, and the cumulative effect of SNPs in modulating efficacy and toxicity. Our hypothesis is that specific genotypes that alter the metabolism or action of platinum agents or relevant to the genotoxic effects of radiotherapy may impact the efficacy and toxicity of patients to these therapies. 2) we will apply machine-learning tools to identify gene-gene and gene-environment interactions influencing NSCLC outcome. We will develop algorithms to identify subgroups with differing platinum or radiotherapy treatment efficacy or toxicity. Our hypothesis is that therapeutic response is modulated by common, low penetrance polymorphisms, and that these polymorphisms interact with each other and/or host factors in determining response to therapy. 3) we will construct predictive risk models for survival and toxicity by integrating clinical and epidemiologic data with the genetic data from this project,and additional information from other R01 studies devoted to these cohorts such as a series of phenotypic assays. We hypothesize that the addition of genetic markers to the standard clinical and epidemiologic variables will improve the prediction of survival and toxicity of the final risk assessment models. We will compare the prediction accuracy among all patients, patients receiving chemotherapy alone, and patients treated by combined modality. The risk models resulting from this project may permit clinicians to identify patients before the start of therapy who are most and least likely to benefit or to develop toxicity and will have immense clinical benefit in terms of planning chemotherapy and radiotherapy for individual patients.