Molecular and AI enabled prognostication in endometrial cancer a 2015 to 2024 bibliometric Atlas and critical review

BACKGROUND: Endometrial cancer (EC) is a leading gynecologic malignancy with increasing incidence and mortality, particularly in high-income countries. Traditional prognostic models based on clinicopathological features often fail to accurately stratify risk, especially in early-stage and low-risk cases. METHODS: A bibliometric analysis was performed on research articles published between 2015 and 2024, sourced from the Web of Science and Scopus databases. The analysis included publications on prognostic or risk models in EC and assessed key trends and themes within the field. RESULTS: The study reveals a significant shift toward molecular-based stratification, particularly with the Cancer Genome Atlas (TCGA) subtypes, and the incorporation of AI-driven models. Molecular markers, such as POLE-ultramutated and p53-abnormal tumors, alongside biomarkers like HE4 and L1CAM, offer improved prognostic accuracy. AI models, including radiomics and deep learning approaches, show promise in predicting disease recurrence and patient outcomes. CONCLUSION: Advances in molecular classification and AI have improved EC prognostication. However, challenges remain in prospective validation and broader clinical implementation. Future research should focus on multi-omics integration and international collaboration to improve the accuracy and applicability of EC prognostic models.