Clinical Correlative Studies of Neuroblastoma

描述（由申请人提供）：神经母细胞瘤是一种常见的儿童肿瘤，大约40%的患者在诊断时患有侵袭性转移性疾病（4期）。随着治疗的改进，生存率提高到35-40%。问题：在诊断时不可能预测哪些患者将成为长期无进展幸存者（PFS），哪些患者将死于疾病。假设：来自肿瘤RNA和DNA微阵列分析的分子“特征”将定义预后良好或不良的亚群。我们从MYCN基因非扩增肿瘤的表达谱中发现了55个基因特征，可识别79%和16%的PFS患者。总体目标：继续为这些患者开发治疗相关的基因组分类器。具体目的：1)确定RNA表达谱是否预测PFS。2)确定基于杂合性缺失和拷贝数异常的DNA标记是否能预测PFS，以及DNA和RNA标记结合是否能提高预测的准确性。研究设计：MYCN扩增和非扩增肿瘤将作为单独的组进行分析，因为它们在临床和生物学上是不同的。肿瘤可从儿童肿瘤组和其他合作者获得注释和临床随访。来自RNA和DNA微阵列的特征将用于建立有效的分子分类器，以预测PFS的可能性。RNA：最初，将比较人类外显子（HuEx）和标准HG 133微阵列的表达谱，以确定HuEx特征在预测PFS方面是否具有相似或更好的准确性。最佳平台将用于发现大约337个肿瘤的特征。临床适用的TaqMan(r)低密度阵列（TLDA）将被设计用于微阵列特征基因，并在相同的rna上进行测试，以验证其预测能力。最后，将对大约210个独立的外部肿瘤进行TLDA检测，以确认其有效性。DNA：来自最初用于RNA研究的相同样本的DNA以及配对的正常细胞将使用高密度500K SNP阵列进行测试，以确定DNA和DNA + RNA标记是否预测PFS。如果是这样，临床适用的TaqMan(r) DNA检测将根据SNP微阵列结果设计，并使用相同的DNA进行测试。最后，使用TaqMan对肿瘤的外部验证集进行测试。单独或与TLDA RNA检测确认其临床价值。总结：这是第一个也是目前唯一一个旨在确定临床定义的高危4期患者亚组的研究。使用基因组分类器预测长期PFS将有助于分配治疗和开发更有效的治疗方法。

DESCRIPTION (provided by applicant): Neuroblastoma is a common childhood tumor, and approximately 40% of patients have aggressive metastatic disease (stage 4) when diagnosed. With improved therapy, survival has increased to 35-40%. Problem: It is not possible at diagnosis to predict which patients will be long-term progression-free survivors (PFS) and which will succumb to disease. Hypothesis: Molecular "signatures" that are derived from microarray analyses of tumor RNA and DNA will define subgroups that have either excellent or poor outcomes. We discovered a 55 gene signature from expression profiles of MYCN gene non-amplified tumors that identifies patients with 79% and 16% PFS. Overall goal: Continue developing therapeutically relevant genomic classifiers for these patients. Specific aims: 1) Determine if RNA expression profiles predict PFS. 2) Determine if DNA signatures based upon loss of heterozygosity and copy number abnormalities predict PFS and if combining DNA and RNA signatures improves accuracy of prediction. Research Design: MYCN amplified and non-amplified tumors will be analyzed as separate groups because they are clinically and biologically distinct. Tumors are available from the Children's Oncology Group and other collaborators with annotation and clinical follow-up. Signatures derived from RNA and DNA microarrays will be used to build validated molecular classifiers that predict the likelihood of PFS. RNA: Initially, expression profiling with Human Exon (HuEx) and standard HG 133 microarrays will be compared to determine if HuEx signatures have similar or better accuracy in predicting PFS. The optimal platform will be used to discover signatures with approximately 337 tumors. Clinically applicable TaqMan(r) Low Density Arrays (TLDA) will be designed for microarray signature genes and tested on the same RNAs to validate their predictive ability. Finally, an independent external set of approximately 210 tumors will be tested with TLDA assays to confirm their validity. DNA: DNA from the same specimens originally used for RNA studies along with paired normal cells will be tested with high density 500K SNP arrays to determine if DNA and DNA + RNA signatures predict PFS. If so, clinically applicable TaqMan(r) DNA assays will be designed from SNP microarray results and tested using the same DNAs. Last, the external validation set of tumors will be tested with TaqMan. DNA assays to confirm their clinical value alone or with TLDA RNA assays. Summary: These are the first and currently only studies aimed at defining subgroups among clinically defined high-risk stage 4 patients. Prediction of long-term PFS using genomic classifiers will facilitate assignment of treatment and development of more effective therapy.