Bentzen Søren M
Department of Human Oncology, University of Wisconsin School of Medicine and Public Health, Madison, WI 53792, USA.
Semin Radiat Oncol. 2008 Apr;18(2):75-88. doi: 10.1016/j.semradonc.2007.10.003.
Substantial research efforts into predictive radiation oncology have so far produced very little in terms of clinically applicable assays. This may change with the development of novel high-throughput assays that are of potential interest in a radiation oncology setting. However, it seems that much current research is opportunistic, driven by the available technologies rather than addressing pertinent clinical or biological questions. This review looks at the experience gained from the attempts to develop cellular radiobiology assays. The research process and, in particular, the need for rigorous validation of any promising assay in an independent dataset are stressed. Some common design problems are discussed using examples from radiation oncology. The statistical challenges and some of the key concepts in analyzing dense datasets from high-throughput assays are briefly reviewed. Single nucleotide polymorphisms, immunohistochemical markers, and DNA microarray gene signatures are used as examples of assays that show promise in radiation oncology applications. Some recent studies suggest a differential treatment response between tumor stem cells and other tumor cells. If this is a general pattern, then future predictive assays may have to be performed on stems cells rather than on unselected tumor cells. Advances in radiogenomics or radioproteomics will come from large collaborative research networks, collecting high-quality dosimetric and clinical outcome data and combining state-of-the-art laboratory techniques with appropriate biostatical methods.
到目前为止，对预测性放射肿瘤学的大量研究工作在临床适用检测方法方面成果甚微。随着新型高通量检测方法的发展，这种情况可能会改变，这些检测方法在放射肿瘤学环境中具有潜在的应用价值。然而，目前的许多研究似乎具有机会主义性质，是由现有技术驱动的，而不是针对相关的临床或生物学问题。这篇综述探讨了在开发细胞放射生物学检测方法过程中所获得的经验。强调了研究过程，特别是在独立数据集中对任何有前景的检测方法进行严格验证的必要性。利用放射肿瘤学的实例讨论了一些常见的设计问题。简要回顾了统计挑战以及分析高通量检测密集数据集时的一些关键概念。单核苷酸多态性、免疫组化标记物和DNA微阵列基因特征被用作在放射肿瘤学应用中显示出前景的检测方法的实例。最近的一些研究表明肿瘤干细胞和其他肿瘤细胞之间存在不同的治疗反应。如果这是一种普遍模式，那么未来的预测性检测可能必须在干细胞上进行，而不是在未选择的肿瘤细胞上进行。放射基因组学或放射蛋白质组学的进展将来自大型合作研究网络，收集高质量的剂量学和临床结果数据，并将最先进的实验室技术与适当的生物统计学方法相结合。
