Clinico-pathological and Genomic Features of HER2-low Early Breast Cancer (eBC). Results of Retrospective Analysis of Seven Adjuvant Trials by the Hellenic Cooperative Oncology Group (HeCOG)

Although breast cancer remains one of the most common and fatal cancer types among women worldwide, earlier diagnosis and improved therapeutic approaches have led to decreased mortality over last years. Indeed, the use of HER2-targeted drugs has improved clinical outcomes for patients with HER2 positive breast cancer, while, on the other hand, until recently, little progress has been made for HER2-negative (IHC score 0 and 1+) patients.

However, it seems that among these HER2-negative breast cancers, substantial heterogeneity exists regarding the expression of hormone receptors (HR) and HER2.

Biomarkers are critical for translating the biological heterogeneity of breast cancer into prognostically and therapeutically useful information. In recent years gene expression profiling and genomic analysis have shown utility in specific clinical scenarios, but immunohistochemistry (IHC) remains the cornerstone of biomarker testing in both early and advanced/metastatic disease.

In breast cancer, achieving efficacy with endocrine therapy or with HER2-blockade requires identifying patients whose tumors show significant survival dependency on the therapeutic target. This is achieved by using appropriate biomarkers cut-offs ensuring a favorable benefit-to-toxicity ratio for specific patients.

This paradigm has recently been challenged by a new class of HER2 drugs that use target expression not as a direct molecular lever but as a vehicle to deliver potent agents to cancer cells, breaking the straightforward link between the molecular target and the corresponding therapy. These drugs exhibit impressive activity at marker expression levels much lower than those required to effectively block HER2 signaling, shifting the diagnostic focus to the lower end of the staining spectrum, specifically distinguishing between HER2-zero and HER2-low expression.

Given the lack of a definitive molecular hallmark for cancers characterized by low HER2 expression, ongoing efforts aim to understand the biology of this heterogeneous group of tumors. This understanding is crucial to triage treatment, investigate resistance mechanisms, and inform potential combination strategies.

To enrich the available data, we perform a retrospective analysis of HER2-low patients in a large prospective cohort of early breast cancer patients from Greece enrolled in 7 randomized and observational clinical studies.

The study is structured as follows: From each patient with early breast cancer, at least two 2.5mm wide cores are punched from donated FFPE blocks and transferred onto low-throughput TMAs. The cores are selected to enrich for regions with higher tumor nucleated cellularity to facilitate NGS analyses. Each core is assessed for biomarker expression with biomarker cutoffs applied at the core level. A case is considered positive for a biomarker if any of its cores tested positive at the predefined cut-off value.

Cases showing any core with HER2 IHC 3+ and/or FISH amplification are excluded from this analysis. The remaining cases (n= 2751) are categorized into HER2-low and HER2-zero.

The study aims to investigate whether HER2-low and HER2-zero cases exhibit distinct characteristics, suggesting HER2-low may represent a unique biological subtype.

Given the importance of the hormone receptors and the known differential distribution of ER/PgR status between these two HER2 categories, after analyzing the entire population cohort, ER/PgR status will be treated as a confounding variable in the analysis (Luminal vs. TNBC) and the HER2 categories will be compared separately within ER/PR-pos (Luminal) and ER/PR-neg (TNBC) disease.

To find out whether intra-tumor discordance of HER-2 is suggestive of true biological heterogeneity we will also consider for the analysis the different HER2 IHC scores between the available cores for each case as well as the distribution of ERBB2 CN data (median of ERBB2 avg copies).