The Prognosis Prediction Model for Endometrial Cancer Based on DNA Methylation Signature

PMCID: PMC12174473

PMID: 40528483

DOI: 10.1002/cnr2.70218

Journal: Cancer reports (Hoboken, N.J.)

Publication Date: 2025-6-17

Authors: Ran R, Wang M, Miao J

Key Points

• Identified 11 key genes (including POU4F3, MAL, MAGI2) that play critical roles in endometrial cancer progression through epigenetic dysregulation
• Recurrence model AUC values: 0.671 (1-year), 0.708 (3-year), 0.689 (5-year), demonstrating strong predictive capabilities
• High-risk patients may benefit more from immune checkpoint inhibitors and targeted therapies like cetuximab

Summary

This comprehensive study investigated DNA methylation-related genes as potential prognostic and predictive biomarkers in endometrial cancer (EC). By analyzing 544 EC cases from the TCGA database, researchers developed sophisticated risk models using LASSO Cox regression to identify key methylation-related gene signatures that can predict cancer recurrence and patient outcomes.

The research identified 25 methylation-related genes with significant clinical implications, ultimately constructing two predictive models: a recurrence prediction model and a prognostic model. These models demonstrated robust performance, with Area Under the Curve (AUC) values ranging from 0.671 to 0.731 across different time points. Notably, the models outperformed traditional clinical staging in predicting patient prognosis, revealing complex interactions between epigenetic modifications and tumor biology.