Validation of HER2 status in whole genome sequencing data of breast cancers with AI-driven, ploidy-corrected approach

The HER2 protein overexpression is one of the most significant biomarkers for breast cancer diagnostics, prediction, and prognostics. The availability of HER2-inhibitors in routine clinical practice directly translates into the diagnostic need for precise and robust marker identification. At the brink of the genomic era, multigene next-generation sequencing methodologies slowly take over the field of single-biomarker molecular and cytogenetic tests. However, copy number alterations such as amplification of the HER2-coding ERBB2 gene, are certainly harder to validate as an NGS biomarker than simple SNV mutations. They are characterized by several compound genomic factors i.a. structural heterogeneity, dependence on chromosome count and genomic context of ploidy. In our study, we tested the approach of using whole genome sequencing instead of NGS panels to robustly and accurately determine HER2 status in clinical setup. Based on the large dataset of 877 breast cancer patients genomes with curated clinical data and a machine learning approach for optimization of an unbiased diagnostic classifier, we provide a reliable algorithm of HER2 status assessment.