Genome-Wide DNA methylation profiling of tissue-plasma pairs identifies Novel cfDNA biomarkers for the diagnosis of gastric cancer.

A machine learning model based on eight cfDNA methylation markers achieves an AUC of 0.978 in training and 0.938 in testing for non-invasive gastric cancer diagnosis.

• Why it matters: Early detection of gastric cancer is vital for improving survival, but current methods are invasive and have limited sensitivity, creating a need for better non-invasive diagnostic tools.
• What they did: Researchers performed whole-genome bisulfite sequencing on tissue and plasma samples from 61 gastric cancer patients and 29 controls, identifying 32 stable DMRs, then used LASSO and RF algorithms to select eight core markers for model development.
• The result: The RF model demonstrated high accuracy (88.46%), sensitivity (94.44%), and specificity (75.00%), with high discriminatory power even after adjusting for age, supporting its potential as a non-invasive gastric cancer screening method.