Abstract A60: Phylogenetic analyses reveal variable patterns of tumor evolution in HGSOC

Abstract

Abstract High-grade serous ovarian cancer (HGSOC) is the most common and difficult to treat ovarian cancer subtype. HGSOC is diagnosed typically at late stage when it has already metastasized to many tissues in peritoneal cavity. In order to gain understanding of ovarian cancer evolution and reasons for chemoresistance, patients’ sequencing and clinical data were extensively analyzed. We studied 124 cancerous samples from 30 HGSOC patients and sequenced whole genomes from multiple tissues collected at diagnosis (N=92), interval debulking surgery (N=17), and relapse (N=15), and in addition deep sequencing data (600 genes) from seven patients with longitudinal circulating tumor DNA (ctDNA). Subclones were identified using somatic mutations and copy number variants (CNV). Patients were divided into three primary therapy response groups (poor, intermediate, and good) using platinum free interval (PFI), primary therapy outcome, and residual tumor size after surgery. The poor response group consisted of patients having progressive disease. Phylogenetic analysis revealed clearly distinct evolutionary models in HGSOC. Briefly, patients can be first divided into two main groups based on genetic diversity between tissues at diagnosis and secondly into three subgroups by comparing them to patients’ relapse samples affected by chemotherapy. In our study 70% of patients fall into the model displaying high genetic diversity between tissues. Although genetic diversity did not correlate with clinical outcome, the number of subclones differed significantly between primary therapy response groups (ANOVA, p=0.013): good outcome patients had more subclones. Relapse or interval stage samples were available from ten patients. Four of them displayed static pattern, continuing the low genetic diversity detected at diagnosis. In this evolution model, subclonal composition is very similar in all the samples regardless of tissue or treatment stage. In the six other patients, we identified models with a) complete subclonal turnover (N=2), b) accumulation of novel mutations (N=2), or c) mixed pattern (N=2). Accumulation of novel mutations in the two patients (PFI 6.0 and 2.7 months) may result from the defective homologous recombination machinery identified by relapse-specific mutational signature (COSMIC signature SBS3). As a common feature of metastasis, evidence of polyclonal seeding was detected in almost all the patients. HGSOC tumors exhibit multiple distinct models of evolution. The most patients are highly genetically different. However, in a quarter of the patients tumors were quite static at diagnosis and half of them remained static even after chemotherapy despite obvious selection pressure. In the majority of cases, the largest number of changes were detected once influenced to chemotherapy. Because of inter-tissue heterogeneity and diverging nature of HGSOC on-time samples, for example through ctDNA, are needed to find optimal drug combinations for this moving target. Citation Format: Jaana Oikkonen, Nil Campamà Sanz, Antti Häkkinen, Yilin Li, Ingrid Schulman, Mikko Kivikoski, Kaisa Huhtinen, Sakari Hietanen, Seija E. Grénman, Johanna Hynninen, Rainer Lehtonen, Sampsa Hautaniemi. Phylogenetic analyses reveal variable patterns of tumor evolution in HGSOC [abstract]. In: Proceedings of the AACR Special Conference on Advances in Ovarian Cancer Research; 2019 Sep 13-16, 2019; Atlanta, GA. Philadelphia (PA): AACR; Clin Cancer Res 2020;26(13_Suppl):Abstract nr A60.