Background: Endometrial cancer (EC) is the most common gynecologic malignancy in the United States. Human epidermal growth factor 2 (HER2) overexpression or gene amplification occurs in 10-60% of EC. The addition of trastuzumab to carboplatin-paclitaxel chemotherapy improves survival in women with HER2-positive serous EC. However, criteria for assessment of HER2 status in EC are not standardized, and furthermore, intratumor heterogeneity may impact therapeutic efficacy. We aimed to identify factors associated with treatment response and survival outcomes in a retrospective cohort of ECs treated with anti-HER2 therapy. Methods: Patients with advanced/recurrent EC treated with trastuzumab or trastuzumab emtansine alone or in combination with chemotherapy from 2013-2021 at our institution were identified. Clinical and treatment information were retrieved from medical records. Central pathology review, HER2 immunohistochemistry (IHC) and fluorescence in-situ hybridization (FISH) were performed. HER2 IHC was scored using modified ASCO/CAP 2007 guidelines for breast cancer. HER2 expression heterogeneity was defined as a 2-degree difference in staining intensity in ≥5% of tumor cells or discordant IHC score on different specimens. HER2 amplification (tumor/normal fold change) was also estimated from targeted next-generation sequencing (NGS). Best response was assessed by RECIST v1.1 as complete response (CR), partial response (PR), stable disease (SD), or progressive disease (PD), with clinical benefit defined as CR/PR/SD. Results: Among 39 patients included, median age was 69. Histologic subtypes included serous (n = 18, 46%), carcinosarcoma (n = 7, 18%), high-grade EC with ambiguous features (n = 13, 33%) and low-grade endometrioid carcinoma (n = 1, 3%). HER2 IHC results were 0 or 1+ (n = 2, 5%), 2+ (n = 17, 46%), 3+ (n = 18, 49%), with HER2 heterogeneity in 15 cases. By FISH, the HER2/CEP17 ratio was ≥2.0 in 34 (92%) and < 2.0 in 3 (8%); HER2 copy number (CN) was ≥6.0 in 25 (68%) and < 6.0 in 12 (32%). HER2 copy number ≥6.0 was significantly correlated with HER2 amplification by NGS (p< 0.001). 18 (49%) patients had a co-existing somatic PIK3CA mutation. Amongst parameters assessed, only higher level of HER2 amplification by NGS was significantly associated with clinical benefit (p= 0.036). None of these features were associated with progression-free survival. Notably, 2 patients with IHC 0/1+ had PD and 1 patient with IHC 3+, but non-amplified by FISH, achieved PR. Conclusions: In our cohort of EC patients treated with anti-HER2 therapy, higher level of HER2 amplification was associated with clinical benefit. Further studies with larger sample sizes are needed to determine whether any additional characteristics are associated with treatment response and survival.