Key points are not available for this paper at this time.
Abstract Background: Renewed interest in clinical relevance of low-HER2 protein in breast cancer management prompted a retrospective investigation of a unique de-identified Database of quantified biomarkers associated with patient outcomes. Currently, IHC of ER, PR and HER2 proteins provides semi-quantitative results, at times complicated by variation in methods and interpretation (cf. ASCO/CAP Guidelines). Our goal was to ascertain relationships of these clinical analytes to predict cancer relapse when quantified for protein products and cognate gene expression. Methods: Database contained 2756 quantified HER2 results by ELISA (Oncogene Sciences) or by EIA (Triton Diagnostics) of which 198 patients had associated clinical outcomes. No definition of HER2-low has been accepted for IHC, although 50% of biopsies may be categorized as HER2-low (cf. ASCP CE Programs). Microarray results obtained for ~ 22,000 genes using RNA purified from LCM-procured carcinoma cells of 247 primary breast biopsies were analyzed. ESR1, PGR and ERBB2 gene expression levels in 276 biopsies had been validated by RTqPCR. To assess distribution and dispersion of HER2 protein, cumulative relative frequency distribution analyses were employed, focusing on cut-offs within interquartile range. Permutations and combinations of the 3 biomarkers based on IHC definitions (cf. ASCO/CAP Guidelines) were examined. Tertiles and quartiles of data sets were used to explore association between biomarker levels and clinical outcomes. Box plots depicted protein and gene expression levels, Progression-Free Survival (PFS), or Overall Survival (OS) across ER/PR/HER2 categories and compared differences using a Kruskal-Wallis ANOVA Rank Test. Cox regressions were used to contrast hazard ratios among the ER/PR/HER2 groups. Kaplan-Meier plots visualized survival rates among groups and a log-rank test detected statistical discrepancies. We investigated differences between groups for categorical variables using chi-squared tests (Fisher's exact test for non-normally distributed data), while for ordinal variables, ANOVA tests were used (Kruskal-Wallis for non-normally distributed data). Categorical variables were summarized as count (%), and ordinal variables were summarized as median IQR. Analyses were performed using R version 4.2.2 (Innocent and Trusting). Results: Neither quantified ER or PR alone nor ER/PR collectively were related to HER2 oncoprotein levels. Influence of menopausal status on biopsy biomarker profiles was investigated. With a median cutoff for HER2 protein, patients with ER+/PR+/HER2+ (TPBC) exhibited increased OS, whereas those with TNBC had decreased OS, of 8 combinations possible. Using tertiles and quartiles, TPBC was exhibited by 34-37% of specimens compared to TNBC, which represented 9.1-12.7% of biopsies when biomarker proteins were quantified. After qPCR data were stratified by ER/PR/HER2 status with a 50% quantile cut for HER2, 39% were TPBC and 10% were TNBC. Using microarray data with the same cutoffs, only 14% were TPBC and 17% were TNBC. Once TPBC were identified by either microarray or by qPCR, increased ESR1, PGR and ERBB2 expression collectively was associated with increased PFS and OS of patients. In general, biopsies that were TNBC or other combinations of biomarker’s gene expression correlated with poorer prognosis and overall survival. Conclusions: Quantifying levels of ER, PR and HER2 proteins as well as of ESR1, PGR and ERBB2 expression were correlated with prediction of clinical outcomes of TNBC, TPBC and patients with other biomarker subsets. Results support quantitative assessment of these biomarkers in biopsies to assess utility of the HER2-low type in personalizing breast cancer management and prediction of risk of relapse. Citation Format: James Wittliff, Michael Daniels. Enhanced Assessment of Clinical Behavior of TNBC and TPBC using Quantified Levels of HER2, Estrogen and Progestin Receptor Proteins and Expression of Their Cognate Genes abstract. In: Proceedings of the 2023 San Antonio Breast Cancer Symposium; 2023 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2024;84(9 Suppl):Abstract nr PO4-03-08.
Wittliff et al. (Thu,) studied this question.