| Neuroblastoma is a pediatric cancer of the developing nervous system that most commonly affects young children and is often lethal. The etiology of this embryonal malignancy is poorly understood, and until very recently, little was known about the factors predisposing to metastatic transformation. Cumulative research in the area of tumor biology over the past two decades, however, has positioned the disease as a paradigm for the utility of molecular and biological features in guiding clinical practice. Indeed, somatically acquired genomic copy number aberrations (CNAs) are of fundamental importance for predicting tumor phenotype and are routinely used in treatment stratification. Despite the ability to classify patients into broad risk groups based on a small set of clinical and biological features, the survival for high-risk patients has reached a plateau at an unacceptable rate of 30-40% and the system remains imperfect. Genome-wide profiling of tumors at the molecular level will likely provide important refinements to the current classifications allowing for enhanced care, and ultimately individualized patient therapies. In this thesis, we explore the role of genomic copy number on all facets of tumorigenesis, and present statistical and analytical methods developed to achieve the biological goals of this work. We describe the first genome-wide association study of germline copy number variations (CNVs) in any human cancer, analyzing over 2,000 controls and over 1,200 neuroblastoma cases currently enrolled in clinical trials. Next, the role of somatically acquired CNAs in initiation and progression of disease is explored through two studies of CNAs in nearly 500 primary tumors. Heritable CNVs predisposing to neuroblastoma, candidate oncogenes and tumor suppressor genes, as well as potential refinements to the current risk stratification system are discussed. |
