Through the wide adaptation of next-generation sequencing (NGS) technology within clinical practice, molecular profiling of the tumor has been the principal component of personalized treatment. In our study, we have generated a large collection of cancer genomes on East Asian epithelial ovarian carcinoma (EOC) patients and demonstrate the feasibility and utility of NGS platforms to explore the dynamic interrelations of major cancer driver alterations and their impacts on clinical prognosis and management. A total of 652 EOC patients have undergone clinical NGS panels to determine the prevalence of germline and somatic mutations. Notably, TP53 was the most frequently altered event (73%), followed by both BRCA1 and BRCA2 (22% each) and MYC (19%) through pan-EOC analysis. When analyzed based on individual histopathological levels, TP53 mutation was highly dominant in high-grade serous and mucinous histology, whereas mutations in PIK3CA and ARID1A were mostly observed in clear cell carcinoma, and KRAS, BRAF, and CDKN2A mutations were enriched in endometrioid, low-grade serous, and mucinous tumors, respectively. The network-based probabilistic model showed significant co-occurrences of TP53 with BRCA1 and ALK with BRCA2, NOTCH1, and ROS1, whereas mutual exclusivity of TP53 with KRAS and PIK3CA was evident. Furthermore, we utilized machine-learning algorithms to identify molecular correlates that conferred increased sensitivity to platinum and olaparib treatments including somatic mutations in BRCA1, ATM, and MYC. Conversely, patients with ALK mutation were considerably resistant to both treatment modalities. Collectively, our results demonstrate the clinical feasibility of prospective genetic sequencing to facilitate personalized treatment opportunities for patients with EOC.
Keywords: cancer genome; epithelial ovarian carcinoma; next-generation sequencing; olaparib; personalized medicine.
© 2022 UICC.