Association of high-grade serous ovarian cancer sample clonality with clinical outcomes

Keely Ulmer; Harvey Stephens; Sofia Gabrilovich; Andrew Polio; David Bender; Michael Goodheart; Jesus Gonzalez Bosquet

doi:10.1016/j.ygyno.2024.07.208

Objectives Genetic profiling of tumors has potential implications for treatment and survival. The presence of different cellular clones in high-grade serous ovarian cancer (HGSOC) tumors may alter their response to treatment and ultimately affect survival. Our objectives were to describe the clonality of HGSOC samples from different biopsy sites and assess the association of clonality with the immune contexture of the specimen. Lastly, we aimed to assess the association of clonality with clinical outcomes, including surgical outcomes, response to chemotherapy, and cancer overall survival. Methods This was a case-control study that used normal fallopian tubes (FT) (n = 12) as controls and HGSOC tissue samples (n = 112) as cases. RNA was extracted from the specimens, and the Illumina HiSeq platform was used to sequence it. Clonality, or sample cellular fraction, is calculated based on the variant allele frequency (VAF) and accounts for local copy number variation (CNV). Files resulting from the sequencing alignment, BAM files, were used to create VCF files. Then, the tool superFreq was used to determine genomic variation and allele frequency of the variants, as well as CNV calls. Clonality was performed by superFreq based on VAF hierarchical clustering while accounting for local copy number. The resulting clusters were required to be consistent with a phylogenetic tree: immediate subclones were not allowed to have a significantly higher summed clonality than that of the parental clone. Deconvolution of bulk RNAseq was done to identify immune cell proportions: B cells, M1 and M2 macrophages, monocytes, neutrophils, natural killer (NK) cells, non-regulatory CD4+ T cells, CD8+ T cells, Treg cells, and myeloid dendritic cells (DC). We performed univariate and multivariate regression analysis to assess the association of clonality and tumoral immune cell proportions. Results There were no differences in the number of subclones based on sample origin (P = 0.885). Normal fallopian tube clonality analysis, as a reference, had only germline cells, no clones. The only variable associated with biopsy origin was the fraction of all immune cells invading the sample: higher in ascites, then ovarian, and last omentum/peritoneum. Macrophages M2 tumor invasion was the only variable significantly associated with surgical outcomes in univariate and multivariate analysis: an increased fraction of M2 cells was associated with optimal surgery (OR < 1). The average fraction of germline cells and NK cells were associated independently with chemotherapy response: a higher fraction of NK cells was associated with a better response to chemotherapy (OR < 1); a higher fraction of germline cells was associated with a worse response (OR > 1). Subclone composition of HGSOC was not associated with cancer overall survival. Conclusions Samples with a higher average fraction of germline cells seemed to respond worse to chemotherapy, whereas a higher fraction of NK cells was associated with a better response to chemotherapy. Further analysis is needed to determine if this could be a potential tool in determining chemotherapy response before treatment initiation. All biopsy sites had a similar number of subclones, which may have implications for treatment and resistance in the future.

Association of high-grade serous ovarian cancer sample clonality with clinical outcomes

View Online

Abstract

Details

Metrics