Abstract PS4-01-19: Isoform level RNA detection provides more detailed profiling of HER2 expression in breast cancer

Abstract Background: Accurate quantification of tumor HER2 expression is essential for guiding breast cancer therapy selection, as HER2 status directly determines treatment decisions and eligibility for targeted therapies. Current HER2 testing methods, primarily based on IHC and ISH, often lack the precision to confidently distinguish certain HER2 statuses, particularly in borderline cases, which can lead to under- or overtreatment. Developing a more accurate and reliable HER2 quantification method will enable safer and more effective treatment choices, ensuring patients receive the most appropriate therapy while minimizing unnecessary exposure to side effects. Ultimately, this will lead to better clinical outcomes and improved quality of care for breast cancer patients. Methods: We have developed a novel platform that leverages long-read sequencing to capture full-length mRNA and generate isoform-level expression profiles. By integrating tumor and blood transcriptome data from breast cancer patients, we characterized the expression of HER2 and 728 other cancer-related genes (based on COSMIC Cancer Gene Census and established ADC targets) to identify patterns that can be used to distinguish patients with different HER2 statuses. Patients: Twelve tumor biopsies and matched blood samples were collected from nine breast cancer patients with HER2 IHC scores of 0 or 1+ (ongoing study; numbers reported as of abstract submission). In addition, blood transcriptome profiles were analyzed from a separate cohort of 95 control patients with benign breast condition or no cancer on imaging. Results: In the 12 tumor samples and 729 genes analyzed, 29,804 non-reference transcript isoforms which had not been previously reported in the reference human genome annotations were identified in multiple patients. Of these, 16,533 (55.5%) were not found in any of 95 control samples, representing novel isoforms potentially unique to breast cancer tumors. For 551 genes, we identified non-reference isoforms shared between tumor and matched blood samples, including in HER2. Comparative analysis of tumor samples with different HER2 statuses (IHC 1+ vs. IHC 0) revealed significant differences in HER2 expression aligned with the IHC scores. The IHC 1+ group exhibited not only higher expression levels (median 79.5 vs. 37.5 parts per million; p = 0.015), but also greater transcript isoform diversity (median 38 vs. 9 transcripts per sample; p = 0.015) at the HER2 locus. On average, 23.3% of HER2 isoforms identified in HER2 IHC 1+ patients contained novel combinations of known splice donor/acceptor sites, as opposed to 11.5% in HER2 IHC 0 patients. Notably, some of these alternative splicing patterns were also detected in matched blood samples, offering valuable insight that could facilitate non-invasive diagnosis of patient HER2 status from blood. Conclusions: Our novel approach utilizing full-length RNA sequencing enables more comprehensive characterization and precise quantification of tumor-associated expression of HER2 and other cancer-related genes, providing new insights into transcript diversity and expression patterns that are critical for improving diagnostic accuracy, refining treatment selection, and ultimately enhancing patient outcomes. Citation Format: Y. Cheng, J. Bradley, K. Morris, I. Ivanova, M. Barnett, A. Séguret, O. Eve, A. Zyoud, G. Benitez, A. Robinson, A. Turnbull, J. Dixon, R. Hockett, H. Chuang, R. Kuo. Isoform level RNA detection provides more detailed profiling of HER2 expression in 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 PS4-01-19.