The goal of this study was to identify and characterize germline and somatic genetic variation in women with high grade serous ovarian cancers treated at a single cancer center and their associations with relapse-free and overall survival. We conducted paired tumor/normal targeted next-generation sequencing of 577 genes in pathways of DNA repair, response to DNA damage, cell-cycle regulation, programmed cell death, MAPK, and P13K/AKT/MTOR signaling in 114 tumors from 71 women. In addition, we investigated single copy number alterations (SCNAs) and loss of heterozygosity in a subset of 61 tumors. We validated our results with data from ovarian tumors in TCGA. Our results were similar to other studies of high-grade serous carcinomas. As an example, we found that approximately one third of the tumors harbored loss of function variants in homologous recombination genes and that >90% had <a href="https://www.ncbi.nlm.nih.gov/gene/7157">TP53</a> somatic mutations. Through GISTIC analysis of the OncoScan data on tumor DNA from 61 cases, we identified several regions of high copy number alterations including including <a href="https://www.ncbi.nlm.nih.gov/gene/4854">NOTCH3</a> and <a href="https://www.ncbi.nlm.nih.gov/gene/5296">PIK3R2</a> that were significantly associated with an increase in cancer recurrence and a reduction in overall survival. We identified few changes in mutation profiles where we had multiple tissues, either from multiple surgeries or from the primary site to metastatic sites within the same surgery. However, we did not investigate copy number alterations in multiple tissues from the same individuals, and so it may be that copy number alterations are more important. Data available through dbGAP will be the BAM files and the phenotypic data on the tumor samples, the age, race, ethnicity of the participants and their outcomes, and the Oncoscan OSCHP files.

Genomic Analyses of Germline and Somatic Variation in High-Grade Serous    Ovarian Cancer

Carcinoma, Ovarian Epithelial study with CNV Genotypes, Targeted-Capture data on dbGaP (71 participants)