Background: Trastuzumab deruxtecan (T-DXd) and sacituzumab govitecan (SG) are both approved antibody-drug conjugates (ADCs) for patients (pts) with HER2 low metastatic breast cancer (MBC). Biomarkers that predict response and/or resistance to sequential use of these therapies are needed.

Methods: In this cohort of patients (pts) with HER2 low MBC treated with both ADCs (in either sequence), we previously reported real-world efficacy and subgroup analyses by age, sites of disease, and use of intervening therapies. Here, we provide a descriptive analysis of next generation sequencing (NGS) results in a subset of pts (N=66) who underwent liquid and/or tissue NGS testing at any of the following time points: prior to ADC1, prior to ADC2, or after receipt of both ADCs. Genomic information captured included specific genomic mutations, microsatellite instability status (MSI), and tumor mutation burden (TMB). Commercial testing per routine clinical practice was performed by Guardant Health, Foundation Medicine, Tempus, Caris, and/or UCSF CLIA-approved testing. We performed Cox proportional hazard analysis to evaluate the relationship between patient and treatment characteristics and genomic expression with real-world overall survival (rwOS) from the start of ADC1.

Results: The study cohort included 74 pts from 4 sites, with 66/74 (89%) with available NGS data from at least one-time point [47 (71%) HR+/HER2 low and 19 (29%) HR-/HER2 low]. The median number of tissue NGS was 1 for both the HR+ (range: 0-4) and HR – (range: 0-3) cohorts. The median number of liquid NGS was 1 (range 0-3) and 0 (range 0-4) for the HR+ and HR- cohorts, respectively. The median total number of NGS (any combination of liquid/tissue) was 2 (range 1-5) and 1 (range 1-5) for the HR+ and HR- cohorts, respectively. Types of NGS testing and timepoints were as follows: Prior to ADC1 liquid 9, tissue 31, both 10, no NGS16; Prior to ADC2 liquid 9, tissue 13, both 1, no NGS 43; after both ADCs liquid 16, tissue 4, both 1, no NGS 45. In the entire cohort, 39 pts (59.1%) had a TP53 mutation; 29 pts (43.9%) had a PIK3CA mutation, and 13 pts (19.7%) had an ESR1 mutation. In the HR+ cohort, the most frequently detected mutations were PIK3CA in 26 (55.3%), TP53 in 24 (51.1%), and ESR1 in 13 (27.7%). In the HR- cohort, the most frequently detected mutations were TP53 in 15 (78.9%), PIK3CA and NF1 in 3 (15.8%). MSI testing demonstrated microsatellite stability (MSS) in all 66 patients. Median TMB was 4.70 (range: 1-14) for HR+ pts and 3.85 (range: 0.78-17.37) in HR- pts. In multivariate analysis evaluating the impact of HR status, age, time since MBC diagnosis, ADC order, de novo MBC, visceral disease, CNS disease, PIK3CA status, and TP53 status on rwOS from the start of ADC1, shorter time since MBC diagnosis [HR 0.96 (0.96-1.0, p=0.010] was the only factor associated with longer rwOS in a Cox proportional hazard analysis. Of note, the presence of TP53 mutations or PIK3CA mutations did not impact rwOS [HR 0.70 (0.36-1.36), p=0.293 and HR 1.74 (0.81-3.74), p=0.156, respectively].

Conclusion: In this cohort of pts with HER2low MBC who received sequential ADCs, most pts underwent NGS testing, and the genomic landscape of mutations was consistent with incidence in the reported literature. The presence of TP53 or PIK3CA mutations did not impact rwOS from the start of ADC1. Given the heterogeneity in the type of NGS testing and the timing related to ADC administration, correlations with outcomes are limited. This study highlights the need for prospective evaluation of NGS information to clarify mechanisms of response and resistance to ADCs.

VIEW THE PUBLICATION