Abstract A026: An alternative to HER2 IHC 0/1+/2+ status to predict which clinically HER2-negative patients will respond to anti-HER2 therapies: A rationale for the likely superiority of quantitative HER2 pathway RPPA measurements

Abstract

Abstract Immunohistochemistry (IHC) may not accurately measure HER2 levels and does not reflect HER2 activity/phosphorylation in HER2-negative patients in the HER2 LOW (IHC 1+/2+) and ULTRA-LOW (IHC 0) range [PMID: 35595825). Yet, accurate quantification of HER2 may be key for predicting response to antibody drug conjugates (ADC) such as fam-trastuzumab-deruxtecan-nxki (T-DxD). We showed previously that response to the receptor tyrosine kinase inhibitor (RTKI) neratinib and the ADC ado-trastuzumab emtansine (TDM-1) can be predicted by phospho(p) HER2 Y1248 and pEGFR Y1173, but not total HER2, in TN and HER2+ pts, respectively (PMID: 34741023; 32914002). We now utilize I-SPY2 TRIAL data to explore further relationships between HER2 IHC, Reverse Phase Protein Array (RPPA)-based quantitative HER2 measurement and a HER2 Activation Response Predictive Signature (HARPS) as candidate response predictors for HER2low-targeted agents. Tumors from 395 HER2-negative patients (HR+=200, TN=195) from 8 arms of the I-SPY2 TRIAL had HER2 IHC/FISH data, along with RPPA-based HER2, pHER2 Y1248 and pEGFR Y1173 data from Laser Capture Microdissection (LCM)-enriched tumor epithelium. Of these tumors, 37% (145/395) were HER2 IHC 0, 42% (168/395) were IHC 1+ and 21% (82/395) were IHC 2+. RPPA-based HARPS was categorized as HARPS+ (pHER2-high/pEGFR-high) or HARPS- (pHER2-low and/or pEGFR-low) using previously established combined cut points. Association between quantitative protein values and HER2 IHC was assessed using ANOVA/Tukey tests. Association between endpoint quartiles and IHC was assessed using chi-squared/posthoc tests. RPPA quantitative HER2 total protein levels were associated with HER2 IHC (Tukey p<1E-09), however, the median and ranges of HER2 expression highly overlapped between HER2 IHC 0, 1+ and 2+ (e.g. HER2=0.44 [0.27 - 0.6] in IHC 0 vs 0.69 [0.42-1.0] in IHC 2+). A significant number (34%) of HER2 IHC 0 tumors expressed HER2 above the median level in the HER2 LOW population. While pHER2 and HARPS status were not associated with HER2 IHC class, HARPS associated with total HER2 overall (chi-squared p=8E-05) and in TNBC (chi-squared p=2.4E-06). We found that 42% (81/195) of TNBC were HARPS+, as were 44% (40/89) of HER2 ULTRA-LOW and 40% (41/104) of HER2 LOW tumors. Our results reveal general disconcordance between IHC and RPPA quantitative HER2, with a third of HER2 IHC 0 tumors expressing HER2 and thus, potentially responsive to ADCs such as T-DxD or TDM1. The high proportion of HER2 ULTRA-LOW TNBC tumors classified HARPS+ by our CLIA-based assay identifies tumors predicted to respond to HER2-directed TKI therapies, representing a cohort with actionable HER2 biology missed by IHC. As HER2Low targeted therapies enter ISPY 2.2, we will test the hypothesis that quantitative HER2 is most predictive for HER2 ADCs, and HARPS is most predictive for RTKIs. Citation Format: Julia Wulfkuhle, Denise M. Wolf, Angela DeMichele, Christina Yau, Laura van ‘t Veer, Hope S. Rugo, Lajos Pusztai, Lamorna Brown-Swigart, Gillian L. Hirst, Rosa I. Gallagher, Amy L. Delson, Alexander D. Borowsky, Laura J. Esserman, Paula R. Pohlmann, Emanuel F. Petricoin. An alternative to HER2 IHC 0/1+/2+ status to predict which clinically HER2-negative patients will respond to anti-HER2 therapies: A rationale for the likely superiority of quantitative HER2 pathway RPPA measurements [abstract]. In: Proceedings of the AACR Special Conference in Cancer Research: Advances in Breast Cancer Research; 2023 Oct 19-22; San Diego, California. Philadelphia (PA): AACR; Cancer Res 2024;84(3 Suppl_1):Abstract nr A026.