Impact of HER2-ultralow heterogeneity and optimal threshold on trastuzumab deruxtecan (T-DXd) efficacy in metastatic breast cancer: A national multicenter cohort study (HEROIC).

1117 Background: Trastuzumab deruxtecan (T-DXd) has been approved for patients (pts) with HER2-ultralow metastatic breast cancer (MBC). HER2 discordance commonly occurs between primary and metastatic lesions within the same patient; however, its incidence remains unknown in the HER2-ultralow era. Additionally, there is still controversy about which specimen to use to determine HER2-ultralow status and optimal threshold to guide T-DXd therapy. Methods: This national, multicenter cohort study included MBC pts treated with T-DXd (5.4 mg/kg) with HER2 status available for both primary tumors and matched metastases between January 2020 and October 2024 (NCT06551220). HER2 status was determined according to the DB-06 protocol. Pts were divided into three cohorts based on HER2 discordance patterns: cohort 1 (HER2-positive/low/ultralow in both primary and metastases), cohort 2 (HER2-positive/low/ultralow in primary and HER2-null in metastases), and cohort 3 (HER2-null in primary and HER2-positive/low/ultralow in metastases). Endpoints included progression-free survival (PFS), overall survival (OS), objective response rate (ORR), disease control rate, and clinical benefit rate. Results: From 24 centers nationwide, 3546 pts met the criteria and were included. The incidence of HER2 discordance between primary and matched metastases has changed across eras of HER2-positivity definitions: HER2-positive era (9.8%, K = 0.78), HER2-low era (25.0%, K = 0.39), and HER2-ultralow era (20.2%, K = 0.16). Among T-DXd-treated pts (n = 1052), a higher response rate was observed in cohort 1 (ORR = 55.7%) and cohort 3 (ORR = 53.1%) compared to cohort 2 (ORR = 13.0%). ORR is positively correlated with HER2 expression if metastatic lesions are used as the examined tissue (positive 62.9%, low 49.8%, ultralow 47.0%, null 13.0%). However, the correlation between ORR and HER2 expression is not significant when primary lesions were examined (positive 57.8%, low 41.5%, ultralow 54.4%, null 53.1%). Additionally, cohort 1 (mPFS = 11.6 mo, mOS = 30.7 mo) and cohort 3 (mPFS = 10.9 mo, mOS = 18.4 mo) exhibited significantly superior PFS and OS compared to cohort 2 (mPFS = 6.1 mo, mOS = 12.3 mo). Faint incomplete membrane staining percentage ≥5% in metastatic lesion was the best threshold to distinguish PFS (HR = 0.54, P = 0.02; mPFS, 11.4 vs 8.6 mo) and ORR (OR = 4.00, P = 0.01; 60% vs 27%) among HER2-ultralow MBC treated with T-DXd. Conclusions: A high HER2-ultralow discordance rate was observed between primary tumors and matched metastases. HER2 status in metastatic specimens more accurately predicts T-DXd efficacy compared to primary specimens. A staining threshold of ≥5% tumor cells in metastatic lesions may optimize T-DXd treatment in HER2-ultralow MBC. Therefore, re-evaluating HER2 status in metastatic lesions is recommended for T-DXd treatment decision.