Novel DNA methylation markers for early detection of gastric cardia adenocarcinoma and esophageal squamous cell carcinoma
Zhiyuan Fan,Jiajie Hao,Feifan He,Hao Jiang,Jinwu Wang,Minjuan Li,Xinqing Li,Ru Chen,Wenqiang Wei
DOI: https://doi.org/10.1007/s11427-024-2642-8
2024-08-30
Abstract:Gastric cardia adenocarcinoma (GCA) and esophageal squamous cell carcinoma (ESCC) present significant health challenges in China, often diagnosed at advanced stages with poor prognoses. However, effective biomarkers for early detection remain elusive. This study aimed to integrate methylome and transcriptome data to identify DNA methylation markers for the early detection of GCA and ESCC. In the discovery stage, we conducted Infinium MethylationEPIC array analysis on 36 paired GCA and non-tumor adjacent tissues (NAT), identifying differentially methylated CpG sites (DMCs) between GCA/ESCC and NAT through combined analyses of in-house and publicly available data. In the validation stage, targeted pyrosequencing and quantitative real-time RT-PCR were performed on paired tumor and NAT samples from 50 GCA and 50 ESCC patients. In the application stage, an independent set of 438 samples, including GCA, ESCC, high- and low-grade dysplasia (HGD/LGD), and normal controls, was tested for selected DMCs using pyrosequencing. Our analysis validated three GCA-specific, two ESCC-specific, and one tumor-shared DMCs, exhibiting significant hypermethylation and decreased expression of target genes in tumor samples compared with NAT. Leveraging these DMCs, we developed a GCA-specific 4-marker panel (cg27284428, cg11798358, cg07880787, and cg00585116) with an area under the receiver operating characteristic curve (AUC) of 0.917, effectively distinguishing between cardia HGD/GCA patients and cardia LGD/normal controls. Similarly, an ESCC-specific 3-marker panel (cg14633892, cg04415798, and cg00585116) achieved an AUC of 0.865 in distinguishing esophageal HGD/ESCC cases. Furthermore, integrating cg00585116, age, and alcohol consumption yielded a tumor-shared logistic model with good discrimination for two cancer/HGD (AUC, 0.767; 95% confidence interval, 0.720-0.813). The mean AUC of the model after 5-fold cross-validation was 0.764. In summary, our study identifies novel DNA methylation markers capable of accurately distinguishing GCA/ESCC and HGD from LGD and normal controls. These findings offer promising prospects for targeted DNA methylation assays in future minimally invasive cancer screening methods.