Immunohistochemical Assay Variability and Human Epidermal Growth Factor Receptor 2 (HER2)-Low Breast Cancer...

HER2-low breast cancer has gained clinical relevance owing to the efficacy of trastuzumab deruxtecan. However, variability in detecting low HER2 expression may affect patient eligibility. This nationwide study in the Netherlands evaluated the variability in HER2 immunohistochemistry (IHC) assays at low expression levels and compared this with the reported HER2-low rates across pathology laboratories.

All Dutch pathology laboratories (n = 36) were invited. Unstained slides containing dynamic range cell lines and IHC calibrators were distributed and processed using each laboratory's routine HER2 IHC protocol and centrally analyzed with artificial intelligence--based software. The limit of detection (LOD) was determined with peptide-coated microbeads. Retrospective HER2 data of primary breast cancer cases were extracted from the Dutch Nationwide Pathology Databank.

Overall, 29 out of 36 laboratories were included. The median HER2 expression of the dynamic range cell line was 0.0 (range, 0.31-1) in core 0, 0.4 (range, 0.62-2) in core 1+, and 4.2 (range, 0.65-1) in core 2+. LOD, evaluated in 20 laboratories, ranged from 17,000 to 41,000 (median, 27,000). It showed a significant correlation with HER2 expression intensity in core 1+ of the cell line (P = .002), with lower LODs associated with higher expression. However, there was no significant correlation between expression levels in core 1+ of the cell lines and reported HER2 0 or HER2 1+ rates in the Dutch Nationwide Pathology Databank (P = .412 and P = .852, respectively). Similarly, no correlation was found between LOD assays and the reported HER2-low rate (P = .910).

HER2 IHC assays show substantial interlaboratory variability at low expression levels. However, this variability does not translate into equivalent differences in reported HER2-low rates in routine clinical practice. Low HER2 expression in cell lines and a high LOD do not show a correlation with reported clinical expression levels. Standardization of analytical and interpretative practices, along with clinical validation of LOD thresholds, are essential.