Understanding and predicting relapse in acute myeloid leukemia

项目总结/摘要 复发仍然是白血病治愈的主要障碍。为什么有些病人会复发，而另一些人则治愈了？ 预测反应的能力，并在坦率复发发生之前用替代疗法进行干预， 潜在地优化治疗，并潜在地改变特定患者疾病的自然史。我们将 使用AML作为模型疾病，以证明可测量残留疾病（MRD）检测的功效 和基因表达特征来预测治疗反应。在这些研究中，我们将使用档案和 来自美国群体间研究的前瞻性样本，以便我们能够准确地将遗传变化与 临床特征和结果。在目标1中，我们将开发和测试一种新的测序方法， 测序，它可以在一百万个野生的背景中以一个突变等位基因的频率检测突变。 型等位基因在目标2中，我们将测试三个基因表达特征头对头预测短期和长期- 术语治疗反应，并发现反应组中不同的途径，这些途径可能是未来的治疗目标。 药物治疗在目标3中，我们将上述基因检测应用于同种异体移植环境，以预测 非清髓性移植后复发（因此，直接流产治疗）。我们的实验室 在CML和ALL中使用类似工具取得了相当大的成功，事实上，在这些疾病中的MRD检测 演变成临床试验结果的替代品。我们同样相信，在反洗钱中，这些目标最终将 使我们能够在治疗前预测结果，根据预测的可能途径选择治疗方法， 不良反应，然后监测MRD，如果MRD清除动力学 次优

PROJECT SUMMARY/ABSTRACT Relapse remains the major obstacle to cure in leukemia. Why do some patients relapse, while others are cured? The ability to predict response, and intervene with alternative therapies before frank relapse occurs, can potentially optimize therapy, and potentially change the natural history of a particular patient’s disease. We will use AML as a model disease to demonstrate the power of the detection of measurable residual disease (MRD) and gene expression signatures in predicting treatment response. For these studies we will use archival and prospective samples from U.S. Intergroup studies so that we can accurately associate genetic changes with clinical characteristics and outcomes. In Aim 1 we will develop and test a new sequencing method, duplex sequencing, which can detect a mutation at a frequency of one mutant allele in a background of a million wild type alleles. In Aim 2, we will test three gene expression signatures head to head in predicting short and long- term treatment response, and discover pathways different in the response groups that may be future targets of drug therapy. In Aim 3, we will apply the above genetic tests to the allogeneic transplant setting, to predict relapse following non-myeloablative transplantation (and hence, direct abortive therapy). Our lab has had considerable success in using similar tools in CML and ALL, and indeed, MRD detection in these diseases have evolved to a surrogate for outcome in clinical trials. We likewise believe that in AML these aims will eventually allow us to predict outcome prior to therapy, pick therapies based on likely pathways responsible for the predicted unfavorable response, then monitor MRD, changing therapy early if the kinetics of MRD clearance are suboptimal.