Background: Endometrial cancer (EC) is one of the most prevalent gynecological malignancies, ranking sixth in incidence among women worldwide. Although historically more frequent in developed countries, EC also represents a condition of increasing public health importance in Brazil, where it ranks sixth among cancers affecting women. The Brazilian National Cancer Institute (INCA) estimates approximately 7,840 new cases of uterine corpus cancer annually between 2023 and 2025, representing an increase of nearly 20% compared to the 2020–2022 period. By 2040, global EC incidence is expected to rise by more than 50% relative to 2021 levels, driven largely by population aging and increasing obesity rates, underscoring the urgent need for understanding and improving targeted management strategies. Aggressive tumors constitute a substantial proportion of EC cases with low prognosis and outcomes, presenting complex and heterogeneous biological profiles, yet their underlying molecular determinants remain poorly understood, particularly in admixed populations. Differences across racial groups have been reported, highlighting additional layers of biological heterogeneity in EC. 

Objectives and Methods: Given the high frequency of aggressive tumors and the underrepresentation of Brazilian patients in large genomic cohorts, comprehensive characterization of these cases is essential to enable the identification of key molecular pathways and biological processes. In this study, 146 tumor samples from patients diagnosed with  primary EC who received adjuvant therapy, including chemotherapy and radiotherapy, were included. Samples were collected at the National Tumor Bank between 2015 and 2020, comprising high-grade endometrioid carcinoma (27), serous carcinoma (40), carcinosarcoma (34) and clear cell carcinoma (45) subtypes with histological classification conducted by two pathologists. The RNA sequencing (RNA-seq) was prepared using the Ribo-Zero Plus kit (Illumina) and sequenced on the Illumina NovaSeq X Series platform at INCA. Raw sequencing data underwent preprocessing using nf-core pipelines, followed by downstream analyses.

Results: Quality control metrics confirmed adequate sample quality. Serous carcinoma showed the highest overall progression rate (62.5%), with 37.5% of patients developing recurrence and 25.0% presenting persistent disease. Similarly, carcinosarcoma exhibited a high burden of progression (50.0%), driven predominantly by persistent disease (32.4%), with recurrence occurring in 17.6% of cases. Mortality rates followed a similar distribution to disease progression across subtypes. Carcinosarcoma showed the highest mortality (85.3%), followed by serous carcinoma (72.5%). Higher prevalence of advanced-stage disease at the diagnosis (FIGO III–IV) were also observed in serous carcinoma (70%) and carcinosarcoma (75%). Grade 3 endometrioid tumors were more frequently diagnosed at early stages, partially explaining their improved outcomes. Unsupervised hierarchical clustering of RNA-seq data revealed distinct transcriptomic profiles between serous carcinoma and carcinosarcoma. Carcinosarcoma samples showed higher and more heterogeneous expression of genes related to extracellular matrix remodeling and stromal interaction, such as COL19A1, as well as immune-related transcripts including IGHG1 and IGHG3 in serous carcinoma. 

Conclusions: Overall, this cohort reveals pronounced heterogeneity in progression dynamics and survival across aggressive EC subtypes, highlighting serous carcinoma and carcinosarcoma as the subtypes with the greatest unmet clinical need and reinforcing the importance of subtype-specific prognostic studies and tailored therapeutic strategies.