Abstract

5614 Background: A new generation of novel HER2 targeting agents, including anti-body drug conjugates, has brought renewed focus to targeting ERBB2amp in EC. In this study, we sought to investigate ERBB2amp in EC including genomic biomarkers, ancestry, and real-world outcomes across diverse populations. Methods: Genomics cohort- Patients (pts) with a diagnosis of endometrial carcinoma or carcinosarcoma with tissue comprehensive genomic profiling by Foundation Medicine (FMI) were included. Predominant genetic ancestry was determined using a SNP-based approach using an algorithm trained on the 1000 Genomes data and classified as one of the five following ancestries: predominant African (AFR), European (EUR), admixed American (AMR), South Asian (SAS), or East Asian (EAS). Molecular subclassification was determined by MSI status and alterations in POLEand TP53. Outcomes cohort- This study used the nationwide (US-based) de-identified Flatiron Health-FMI EC clinico-genomic database, originating from approximately 280 US cancer clinics (~800 sites of care) with treatment received in the FH network between January 2011 and March 2022. Time to next treatment (TTNT) from first line therapy and real-world overall survival (rwOS) between groups were evaluated. Adjusted hazard ratios (aHR) from multivariable Cox proportional hazard were calculated to account for differences in the use of HER2 targeted therapy (tx). Results: 1,287/16,073 (8%) pts in the genomics cohort harbored ERBB2amp. ERBB2amp was more common in pts of AFR ancestry than EUR ancestry (12% vs 7%, p<0.001). Serous histology was more common in pts with than pts without ERBB2amp (47% vs 20%, p<0.001). Conversely endometrioid histology was less common in pts with ERRB2amp than pts without ERBB2amp (6% vs 26%, p<0.001). ERBB2amp pts were more likely to fall within the TP53mut molecular subtype (86% vs 46%) and less likely to be MSI-H (0.2% vs 19%) or no special molecular profile (7% vs 28%) (p<0.001). In the outcomes cohort, 20% of ERBB2amp pts received first line HER2 directed tx with or without chemotherapy. There was no difference in the frequency of HER2 tx use between ancestries. Pts with ERBB2amp (n=106) had worse rwOS (aHR 1.32, 95% CI 1.01-1.72, p=0.05) than pts without ERBB2amp (n= 1,169). There were no differences in TTNT (aHR 0.91, 95% CI 0.71-1.15, p=0.42). Specifically analyzing pts with AFR or EUR ancestry by ERBB2amp status, we observed no differences in TTNT or rwOS. Conclusions: The frequency of ERBB2amp was almost double that in pts of AFR ancestry compared to pts of EUR ancestry. We also observed significant differences in EC pts with ERBB2amp vs. no ERBB2amp with respect to histology, molecular classification, and rwOS. These differences should be considered when designing and enrolling patients in clinical trials using HER2 targeting agents.