GLI1 gene alterations including fusions and amplifications compromise a subset of malignant mesenchymal tumors exhibiting characteristic monomorphic nested morphology and frequent S100 positivity, which mimic glomus tumors or well differentiated neuroendocrine tumors. We report four high-grade uterine endometrial stromal sarcomas (ESS) harboring GLI1 and MDM2/CDK4 co-amplifications with a median age of 51.5 years (range 43 ~ 72 years). Histologically, tumors showed a heterogenous morphology, including ovoid to spindle cells, showing nested/nodular arrangement (4/4). Myxoid background was observed at least partially in 4 tumors with prominent capillary networks. Mitoses index was 2 to 20/10 HPF (median 9.5/10 HPF). Immunochemically, tumors showed diffuse staining of CD10 (3/4) with frequently positive CyclinD1(2/4 tested) and mostly negative S100 protein (3/4). Next-generationsequencing (NGS) studies revealed GLI1 and MDM2/CDK4 co-amplification in all cases (4/4) and GLI1 fusion in 1 case (1/4), which were validated by fluorescence in situ hybridization (FISH) analysis. BCOR fusions were firstly identified with GLI1 and MDM2/CDK4 co-amplification in 2 cases (2/4). Copy number (CN) segmentation data showed GLI1 co-amplified cases present generally a single peak at the 12q13.3-15 locus. Follow-up (range:3 to 112 months; median 37.5 months) showed recurrence and/or metastasis in all cases (4/4), in which 1 patient developed lungs and liver metastasis. Relapse-free survival (RFS) analysis showed similar median RFS between GLI1 co-amplified HGESS and GLI1 non-amplified HGESS groups, which were shorter than LGESS group. Unusual clinicopathologic features of these HGESS with GLI1 and MDM2/CDK4 co-amplification mimicked other neoplasms, which caused significant diagnostic challenge and pitfalls. However, identification of GLI1 alterations in these tumors is beneficial for diagnosis and potential use of targeted GLI1 inhibitors.