Methylation-Based Signatures for Gastroesophageal Tumor Classification

Area of Science:

• Oncology
• Genomics
• Molecular Biology

Background:

• Gastroesophageal junction (GEJ) tumors are frequently misclassified, leading to suboptimal treatment strategies.
• Misclassification rates for GEJ tumors can be as high as 45%, impacting treatment efficacy.
• Accurate classification is essential for tailoring effective therapeutic interventions.

Purpose of the Study:

• To leverage DNA methylation profiles for improved classification of gastroesophageal junction tumors.
• To identify robust methylation biomarkers for distinguishing GEJ tumors from esophageal and gastric cancers.
• To enhance diagnostic accuracy and treatment strategies for GEJ malignancies.

Main Methods:

• Utilized DNA methylation data from ~27,000 sites across four cohorts from Gene Expression Omnibus and The Cancer Genome Atlas.
• Employed the Multi-Survival Screening (MSS) algorithm to identify methylation biomarkers.
• Validated methylation signatures in independent discovery and validation datasets.

Main Results:

• Identified three distinct methylation signatures associated with protein binding, gene expression, and cellular component organization.
• Achieved high precision and recall rates (up to 99.2%) in the validation dataset.
• Discovered that methylation sites near transcription start sites (TSSs) are critical for tumorigenesis.

Conclusions:

• Presented the first methylation signatures with high predictive power for gastroesophageal tumors.
• These signatures offer a promising tool for improving the diagnosis of GEJ tumors.
• Enhanced classification can lead to more effective and personalized treatment plans for patients.