Circulating tumor DNA profiling of patients with metastatic luminal HER2-low and HER2-0 breast cancer via next-generation sequencing.

Abstract

e13090 Background: Despite substantial progress in the treatment of hormone receptor-positive/HER2-negative (HR+/HER2-) metastatic breast cancer, a significant proportion of patients will progress early on endocrine-based treatments. In patients with HER2-low expressing disease (i.e. HER2 1+ or 2+ without ERBB2 amplification), the antibody-drug conjugate trastuzumab deruxtecan (T-DXd) has substantially improved outcomes. Despite extensive research, potential biological differences between HER2-low and HER2-0 tumors remain poorly defined to date. Here we aimed to non-invasively establish molecular profiles of HER2-low and HER2-0 disease from circulating tumor DNA (ctDNA) in plasma in a large number of patients early in the course of metastatic disease, to infer potential differences and optimize treatment guidance. Methods: We analyzed 226 plasma samples from 185 metastatic breast cancer patients using an untargeted aneuploidy screening (mFAST-SeqS) to estimate tumor fractions (TFx), and a 77-gene panel (AVENIO ctDNA Expanded). Molecular profiles, the number of somatic variants, TFx, and variant allele frequencies (VAF), were compared between HER2-low and HER2-0, as well as between samples taken before 1 st line and 2 nd line treatment. For a subset of samples, we performed high-coverage whole-genome sequencing (WGS) for nucleosome positioning to identify potential differences in transcription factor (TF) activity between the groups. Results: Out of 226 samples, 130 were collected at baseline, 77 before 2 nd line treatment, and 19 before 3 rd or 4 th line treatment. 147 patients were classified as HR+/HER2-low, 76 as HR+/HER2-0 and 3 had an unknown HER2 status. Significantly higher TFx were detected in HER2-low vs. HER2-0 samples (median 2.89 vs. 1.85, P = 0.0036), but not in 2 nd vs. 1 st line samples. Somatic variants were detected in 199 samples. HER2-low and HER2-0 samples did not significantly differ in the number of somatic variants, or their VAFs. Both presented with a similar frequency of alterations in PIK3CA , TP53 and ESR1 genes. However, samples taken before 2 nd line treatment had significantly higher VAFs and the number of detected variants, and ESR1 alterations were more frequently detected compared to samples taken at baseline. Since distinctive genomic fingerprints are lacking between HER2-low and HER2-0 tumors, we are currently analyzing differential TF activity via LBFextract in 22 high-coverage WGS datasets that may better stratify the two cohorts. Conclusions: Consistent with previous reports, our study indicates that HER2-low and HER2-0 tumors share similar frequencies of common gene alterations. Besides significant differences in plasma TFx, only minimal variations in other mutational or molecular features were detected. Additional analyses may help identify differential impact of transcriptional drivers between the groups.