DNA Methylation Panels for the Differentiation of Lung and Gastric Adenocarcinomas from Other Common Primary Adenocarcinomas.

: Gastric and lung adenocarcinomas are among the most common adenocarcinomas worldwide. Our research aimed to validate methylation biomarkers that differentiate gastric and lung adenocarcinomas from hepatocellular carcinoma, cholangiocarcinoma, colorectal carcinoma, pancreatic adenocarcinoma and paired healthy tissues. : The study analyzed 178 formalin-fixed, paraffin-embedded tissue samples, including 14 gastric adenocarcinomas, 15 lung adenocarcinomas, 15 hepatocellular carcinomas, 15 cholangiocarcinomas, 15 colorectal carcinomas, 15 pancreatic adenocarcinomas and their paired healthy tissues. Methylation status was determined experimentally by methylation-sensitive high resolution melting. The diagnostic panels were validated on bioinformatics datasets from The Cancer Genome Atlas and Gene Expression Omnibus, comprising 1981 and 773 samples, respectively. Sensitivity, specificity, diagnostic accuracy and predictive values for each cancer type were calculated for the experimental, Gene Expression Omnibus and The Cancer Genome Atlas datasets. : The gastric cancer-specific panel showed a sensitivity of 78.6-83.9%, a specificity of 86.6-94.6% and a diagnostic accuracy of 89.9-96.1% to differentiate between all tumors, and a sensitivity of 78.6-83.9%, a specificity of 89.2-96.4% and a diagnostic accuracy of 88-96.1% to differentiate between all tumors and healthy tissues. The lung adenocarcinoma-specific panel showed a sensitivity of 61.1-93.3%, a specificity of 70.3-90.8% and a diagnostic accuracy of 74.2-90.6% to differentiate between all tumors, and a sensitivity of 61.1-93.3%, a specificity of 77.9-93.4% and a diagnostic accuracy of 79.2% to 93.1% to differentiate between all tumors and healthy tissues. : This study demonstrates the potential of using diagnostic methylation panels to differentiate gastric and lung adenocarcinomas from other common adenocarcinomas and paired healthy tissues.