Abstract 5359: Multi-omics characterization of copy number variation in high-grade serous ovarian cancer

Abstract Ovarian cancer is one of the most lethal gynecologic malignancies, primarily due to its late-stage diagnosis and intrinsic tumor heterogeneity. Histologically, epithelial ovarian cancer (EOC) accounts for approximately 70% of cases and exhibits distinct molecular and clinical characteristics. High-grade serous ovarian cancer (HGSOC), the most aggressive form of EOC, stands out for its widespread genomic instability and poor prognosis. Copy number variations (CNVs) play a critical role in tumor heterogeneity, treatment resistance, and disease progression. While CNVs have been extensively studied using bulk sequencing approaches, their spatial heterogeneity within tumors remains underexplored. To investigate this, we integrated spatial transcriptomics (10x Genomics Visium Spatial) with whole-exome sequencing (WES) in 24 fresh-frozen HGSOC samples. This dual approach enabled us to validate CNV patterns inferred from spatial data while leveraging WES for a genome-wide view of CNVs. First, spatial transcriptomics data were processed using the Space Ranger pipeline (10x Genomics) to generate spatially resolved gene expression matrices. Tumor purity scores were estimated using the R package ESTIMATE, and were then used to inform CNV inference with the tool InferCNV. This analysis revealed varied CNVs across spatial domains, indicating intratumoral heterogeneity and the existence of subclones with distinct CNV profiles and spatial niches. We further resolved loss of heterozygosity (LOH) in these samples using the R package tLOH. Clonal LOH of chr 17p, encompassing the TP53 locus, was detected in several samples. Pseudotime and spatial trajectory analysis confirmed that chr17 LOH is an early event in the trajectory of CNV development. To validate the CNV and LOH patterns inferred from spatial transcriptomics, we performed WES on paired tumor and normal tissues from the same cohort. WES data were first aligned to the human reference genome GRCh38 using BWA, followed by CNV profiling with GATK CNV. Recurrent CNV patterns, including gains on chromosomes 1q, 3q, 6p, 8q, and 20q, and losses on chromosomes 4, 5q, 6q, 9p, 13q, 14q, and 17, were observed, consistent with previous reports. A comparison between spatially inferred CNV clusters and GATK CNV results showed strong concordance, further validating the robustness of spatial CNV inference. Interestingly, germline BRCA-mutated samples exhibited a higher frequency of CNVs than BRCA wild-type samples, suggesting a link between BRCA status and genomic instability. By placing CNVs in tissue context, our findings highlight the genetic and spatial heterogeneity of HGSOC, and enhance our understanding of the driving role of CNVs in tumor evolution and growth. This abstract was developed with the assistance of generative AI to revise and clarify the text. Citation Format: Yuxin Jin, Jing Qian, Michelle G. Webb, David W. Craig, Lynda Roman, John D. Carpten, Rania Bassiouni. Multi-omics characterization of copy number variation in high-grade serous ovarian cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 5359.