Methylation Patterns Partition Pancreatic Cancer into Distinct Prognostic Subtypes

Abstract

In the initiation and progression of pancreatic cancer, DNA methylation plays a critical role. The present study attempts to explore specific prognosis subtypes based on DNA methylation data and develop an epigenetic signature to predict the overall survival (OS) of patients with pancreatic cancer.147 samples were included in the training cohort, whereas the validation cohort included 226 samples. The 298 OS-related methylation sites in the training cohort were selected for consensus clustering, and the authors identified three subtypes with a significant difference in prognosis. Cluster1 was associated with poor OS, low-grade disease and high lymph node involvement. In addition, we identified 33 specific methylation sites in Cluster1. Subsequently, we developed a robust qualitative signature consisting of 14 methylation sites to individually predict OS in the training cohort, and the predictive accuracy of this model was confirmed in the validation cohort. Functional enrichment analysis showed that the selected genes in the model were mainly enriched in known cancer-related pathways. Patients were divided into high- and low-risk groups by the model, and a significant difference in OS was observed between these groups. Classification based on the modeling of a specific DNA methylation site can reveal the heterogeneity of pancreatic cancer and provide guidance for clinicians in predicting the prognosis of pancreatic cancer and providing personalized treatment.