Dominant chromosomal abnormalities in breast cancer metastasis to CNS as compared with systemic metastasis demonstrated by liquid biopsy.

Abstract

1035 Background: Breast cancer (BrC) central nervous system (CNS) metastases are believed to be biologically and therapeutically distinct from systemic metastases. Detecting BrC metastasis to CNS using liquid biopsy (LBx) of the CSF is emerging as a reliable approach not only for confirming metastasis, but also for defining the molecular and biological characteristics of this metastasis. We compared the findings of LBx performed on CSF with those obtained by peripheral blood (PB) LBx. Methods: LBx was performed by next generation sequencing (NGS) of cell-free DNA and RNA (cfDNA/cfRNA) in PB plasma or CSF as well as RNA sequencing of CSF cell pellet. RNA sequencing was performed using a panel of almost 1600 genes and DNA sequencing using 302 genes. Results: Of the 339 tested CSF samples, 29 (8.5%) were negative for any mutation and 57 (16.8%) showed low level mutations characteristic of CHIP (clonal hematopoiesis of indeterminate potential). However, two samples showed only chromosomal gain or loss without mutations and 2 showed chromosomal abnormalities with CHIP but no cancer-related mutations. In contrast, 19 of 279 PB samples (6.8%) were negative for any abnormality and 33 (11.8%) had CHIP. One sample of the PB cases showed chromosomal abnormalities without cancer-related somatic mutations. Of the 253 positive CSF samples 207 (82.1%) showed chromosomal structural abnormalities while only 71 of the 227 positive PB samples (31.2%) showed chromosomal structural gain or loss (P <0.0001). The most common chromosomal abnormality was 1q gain, detected in 42.8% of CSF and in 18.5% of PB samples. 17p (TP53) deletion was detected in 25.4% of CSF and in 6.6% of PB samples. 10q (PTEN) deletion was detected in 9.4% of CSF and 3.1% of PB samples. ERBB2 gene amplification was detected in 6.3% of CSF cases and in 1.3% of PB cases. In CSF positive samples, the most mutated genes were TP53, KMT2C, PIK3CA, DNMT3A, CDH1, NF1, PTEN and ESR1, detected in 37.3%, 29%, 27.8%, 10.3%, 8.7%, 8.3%, 7.9%, and 7.5%, respectively. These genes were detected in positive PB samples at 51%, 35.2%, 42.7%, 81.1%, 11.8%, 8.8%, 9.6%, and 20.2%, respectively. Variant allele frequency in all mutations was significantly higher in CSF (P=0.0001) than PB. Germline mutations in BRCA1/2 were detected in 5.1% of CSF and in 4.8% of PB. Ten patients had their PB and CSF tested concurrently. Two were positive by CSF and negative by PB and one was positive by PB but negative by CSF. The rest had concordant results. Conclusions: BrC with CNS involvement represents a distinct subgroup whose biological characteristics may best be defined using LBx to evaluate chromosomal abnormalities and gene gain/loss. As such, optimal characterization and qualification minimal residual disease in patients with BrC with CNS involvement warrants integration of CSF LBx-defined genetic abnormalities as it may permit earlier and targeted interventions.