Abstract PS1-11-20: Rna-based ESR1 and PGR gene expression provides a real-time readout of hormone receptor status in patients with breast cancer

Abstract Background: The hormone receptor (HR) status, traditionally determined by IHC-based expression levels of the estrogen receptor (ER; encoded by ESR1) and progesterone receptor (PR; encoded by PGR) in tumor tissues, is used for clinical decision-making in breast cancer. With the increase in IHC-based predictive biomarkers, there is a need for development of tissue conservation strategies. Further, receptor subtype switching occurs in ∼20% of cases during progression, potentially impacting treatment decisions made based on IHC at initial diagnosis and thus requiring repeat testing. Multiplex RNA gene expression could serve as a screening or surrogate for protein expression by IHC especially when tissue availability is limited. Herein, we examined a cohort of patients with breast cancer profiled with a targeted RNA sequencing assay to evaluate the accuracy of ESR1 and PGR expression when compared to ER/PR IHC expression and assess the feasibility of utilizing gene expression for HR subtype inference. Methods: The study cohort comprised 497 patients with breast cancer who received RNA-based profiling during routine clinical care, capturing the expression of 1,517 genes (FoundationOne®RNA; a laboratory developed test by Foundation Medicine, Inc.). IHC-based ER/PR annotation from pathology reports was extracted and compared to ESR1 and PGR RNA expression. Almost 25% of the samples were held out while the remaining were split into training and test sets (70:30). For the model, transcripts per million (TPM) cutoffs for ESR1/PGR that maximized accuracy in the training set were selected (ESR1 TPM cutoff: 517.3; PGR TPM cutoff: 50.3). Performance of the model was assessed in the test and the held-out sample sets using accuracy, sensitivity, specificity, and F1 score (range 0 to 1). Results: Hormone receptor IHC and gene expression were found to be highly correlated. ER+ cases by IHC had 54-fold higher ESR1 expression compared to ER- cases (median, IQR 6,969.7 [2,846-11,742.9] TPM in ER+ vs 128 [78-246.8] TPM in ER-; P<0.0001). Similarly, PR+ cases had 30-fold higher expression of PGR compared to PR- cases (median, IQR 242.6 [62.2-616.3] TPM in PR+ vs 7.5 [3.7-15.6] TPM in PR-; P<0.0001). In the training dataset, a cutoff of 517.3 TPM for ESR1 and 50.3 TPM for PGR maximized accuracy for predicting HR status. In the held-out set, applying a model with these cutoffs resulted in a sensitivity of 94% and specificity of 88% for detecting HR+ IHC status. Accuracy was 92% with an F1 score of 0.94, suggesting this model can achieve both high precision and high recall. Of the cases with low ER IHC staining (between 1-9%) (n=4), the model predicted two samples as HR+ and the other two samples as HR-. The median ESR1 TPM of the two samples predicted to be HR+, despite the low ER IHC staining, was 1,466.9 TPM whereas the TPM of the two samples predicted to be HR- was 66.4. Conclusion: ESR1 and PGR RNA expression is highly correlated with ER/PR IHC status, and gene expression-based thresholding from RNA expression profiling could potentially be utilized to infer real-time HR status with high accuracy. Utilizing a multiplex RNA expression methodology as a screen or surrogate for predictive biomarkers such as ER/PR will become increasingly important with more IHC-based predictive biomarkers on the horizon. Further validation will be required to assess its utility in clinical decision making for breast oncologists. Citation Format: L. Pusztai, S. D. Sisoudiya, R. B. Keller-Evans, J. J. Pao, J. S. Ross, V. Sethunath, R. S. Huang, A. B. Schrock, L. Poyser, L. Dennis, G. Frampton, E. M. Ebot, E. S. Sokol, S. Sivakumar, A. C. Annan. Rna-based ESR1 and PGR gene expression provides a real-time readout of hormone receptor status in patients with breast cancer [abstract]. In: Proceedings of the San Antonio Breast Cancer Symposium 2025; 2025 Dec 9-12; San Antonio, TX. Philadelphia (PA): AACR; Clin Cancer Res 2026;32(4 Suppl):Abstract nr PS1-11-20.