HER2-targeted immunoconjugates for breast cancer: Ancestry and dose adjustment for thrombocytopenia.

Abstract

1048 Background: Thrombocytopenia (TCP) is a common toxicity of HER2-targeted agents, trastuzumab emtansine (TDM1) and trastuzumab deruxtecan (TDXd). A high incidence of this toxicity was observed in pivotal trials that led to approval of these agents. We investigated whether Asian ancestry increases risk of dose adjustment for TCP on TDM1/TDXd and its impact on dosing in the real-world setting. Methods: Females with HER2+ breast cancer who initiated TDM1/TDXd between 1/16/17-10/26/21 were identified by retrospective review. Primary endpoint was number of chemotherapy cycles until adjustment of TDM1/TDXd dose for TCP; competing endpoints included discontinuation for other toxicity, disease progression, and completion of treatment. Individuals who were switched from TDM1 to TDXd contributed an additional observation post-switch. Recurrent events analysis evaluated Asian ancestry (p<0.05) using a proportional means model, with robust sandwich estimate recognizing correlation between repeated observations per individual. Covariates (age, metastatic disease, drug, prior adjustment) were retained if they improved the model. Results: The study excluded individuals who declined to self-identify (n=23), self-identified as other than Asian or White (n=28), and/or dissented to research (n=24). The study included n=181 individuals (mean age 55.1+12.8 years), of whom n=48 (26.5%) identified as Asian and n=124 (68.5%) had metastatic disease. Overall, 33 individuals received TDXd exclusively, leaving 148 (81.8%) individuals who received TDM1, including 45 individuals who later switched to TDXd after development of TCP while on TDM1 (n=9) or other toxicity (n=36) on TDM1. For n=226 observations (total 2551 cycles), the endpoint was dose adjustment for TCP (n=32), discontinuation for other toxicity or disease progression (n=112), completion of treatment (n=27), or censoring (n=55). Taking into account history of switching drug for TCP, Asian ancestry was associated with increased risk of dose adjustment for TCP (Table). Neither age, metastatic disease, nor specific drug improved the model (data not shown). Conclusions: Among individuals taking TDM1 and/or TDXd for HER2+ breast cancer, we observed that those with Asian ancestry are at greater risk of dose reduction for TCP than their non-Asian counterparts. Upon confirmation in additional individuals with HER2+ cancers of the breast and other sites, this heightened susceptibility to TCP among Asian individuals should be further investigated to elucidate the underlying mechanism and optimize clinical guidelines for prevention and management. [Table: see text]