Prevalence and tissue concordance of BRCA2 copy number loss evaluated by single-cell, shallow whole genome sequencing of circulating tumor cells (CTCs) in castration-resistant prostate cancer (CRPC).

Abstract

5531 Background: Genomic studies have shown that up to 25% of prostate cancer tissue specimens harbor alterations in DNA Damage Repair (DDR) genes, which may sensitize the tumor to poly ADP-ribose polymerase inhibitors (PARPi). Trials evaluating PARPi in patients with DDR deficiencies have shown varied response rates and differences regarding which genomic alterations predict for sensitivity to these agents, with the majority of objective responses seen in BRCA2-altered tumors. These results highlight the need to develop biomarker assays which can predict benefit from PARPi therapy. Tissue and cell-free DNA (cfDNA) have been the most utilized sources of tumor material for analysis in this setting, but success rates of obtaining sufficient tumor for analysis from bone are low and detecting tumor-derived copy number variants (CNVs) in cfDNA is challenging. Circulating tumor cells (CTCs) represent an alternate source of genetic information, for which assays are available to isolate and sequence individual cells in a manner that eliminates background noise from stroma and healthy cells, while capturing inter-cellular heterogeneity. Methods: Blood samples, collected from 138 progressing metastatic CRPC patients within 30 days of a pre-treatment biopsy intended for sequencing using MSK-IMPACT, were sent to EPIC Sciences for CTC analysis. Detected CTCs underwent single cell, low pass whole genome sequencing. Prevalence and concordance of BRCA2 copy-loss, regardless of whether single copy or homozygous, was compared in matched tissue and CTC samples. Results: BRCA2 copy-loss was identified in 21% (23/108) and 50% (58/115) of successfully sequenced tissue and CTC samples, respectively. In the 58 patients with CTC-detected BRCA2 loss, BRCA2 loss was detected in 36% (220/565) of the sequenced CTCs, representing a median of 46% (range 4-100%) of CTCs found in each individual sample. When both sequencing assays were successful, BRCA2 loss was detected in CTCs in 84% (16/19) of the tissue-positive cases, whereas tissue sequencing detected BRCA2 loss in 35% (16/46) of CTC-positive cases. Conclusions: Data from this study supports the notion that single-cell CTC sequencing can detect BRCA2 copy-loss at a high frequency, including cases that were negative in tissue, while also characterizing inter-cellular heterogeneity. Further studies will investigate whether CTC BRCA2 copy-loss can predict the likelihood of response to PARPi.