Abstract:The genomic landscape of primary high-grade serous ovarian cancer (HGSC) is characterized by ubiquitous TP53 mutations, frequent disruption to the homologous recombination (HR) DNA repair pathway, CCNE1 amplification associated with primary treatment resistance, and extensive copy number aberrations. We recently described a cohort of 49 HGSC patients where we performed genomic and phenotypic profiling of multi-site tumors collected at primary cytoreductive surgery and at relapse, identifying 3 distinct evolutionary patterns and inter-tumor heterogeneity. The aim of this study was to further characterize the heterogeneity of genomic alterations in HGSC. Twenty-one patients undergoing upfront maximal effort cytoreductive surgery for advanced HGSC underwent mapping of their tumor locations and biopsies were collected from disseminated tumors. Tumor samples were also collected from ten patients who had a biopsy or surgery at relapse. DNA was extracted from frozen tumors (range 4-12, median=5 tumors per patient) and whole genome sequencing was performed. Analysis of mutations, structural variants, copy number and mutational signatures identified a somatic pathogenic TP53 mutation in all samples, and inter-tumor heterogeneity of somatic copy number profiles and mutational signature abundance. The prominent single base substitution mutational signatures are associated with defective HR and age. Whole genome duplication (WGD), defined as over 50% of the autosomal genome with a major copy number of at least 2, was detected in 15 cases (71%). However, in 5 of these cases the WGD was subclonal and only detected in a subset of the primary tumor sites. In the 8 cases with recurrent tumor samples, we observed that the WGD status varied between timepoints in 5 cases, e.g. in one case all primary tumor samples sequenced had WGD but the relapse sample did not. We have also observed that some cases have undergone more than one WGD. The relative abundance of the ovarian specific copy number mutational signature 4, which is associated with WGD, is significantly higher in samples with WGD than those without. Furthermore, there appears to be a positive association between the number of WGDs and the abundance of copy number signature 4. By contrast the abundance of copy number mutational signature 3, which is associated with BRCA1/2 related HR deficiency, is significantly lower in samples with WGD and is inversely related to the number of WGDs. These differences are also seen within the cases where only a subset of tumor sites have WGD. We also observe differences in the abundance of a number of the single base substitution and INDEL mutational signatures by WGD status. Our data is allowing us to understand the spatial heterogeneity of the genomic landscape of HGSC in primary disease and its evolution over time. We observe extensive inter-tumor heterogeneity of various features, including WGD, that highlights the limitations of performing single site genomic analysis for investigating HGSC biology. Citation Format: Nikki Burdett, Christina Fotopoulou, Ahwan Pandey, Marc Lorentzen, James Clark, Jonathan Krell, David D.L. Bowtell, Paula Cunnea, Elizabeth L Christie. Inter-tumor heterogeneity revealed by multi-site whole genome sequencing of high grade serous ovarian cancer [abstract]. In: Proceedings of the AACR Special Conference on Ovarian Cancer; 2023 Oct 5-7; Boston, Massachusetts. Philadelphia (PA): AACR; Cancer Res 2024;84(5 Suppl_2) nr A057.

Abstract A057: Inter-tumor heterogeneity revealed by multi-site whole genome sequencing of high grade serous ovarian cancer