A Sensitive and Robust Plasma-Based DNA Methylation Panel for Early Detection of Target Gastrointestinal Cancers
Yanmiao Dai,Hui Li,Qianqian Wu,Jie Wang,Kai Wang,Sujuan Fei,Bing Pei,Lishuang Song,Guangxia Chen,Yong Ma,Chenjing Xia,Shangmin Xiong,Minxue Zheng,Ying Xue,Guodong Zhao,Hongwei Xu
DOI: https://doi.org/10.1016/j.neo.2023.100941
2023-01-01
Abstract:Background: Target gastrointestinal cancers (GICs), encompassing esophageal cancer (EC), gastric cancer (GC), and colorectal cancer (CRC), originate within a single readily accessible luminal organ system and are diagnosable using endoscopy. However, endoscopy is an invasive procedure with low compliance and no plasma-based DNA methylation assay for the early detection of GICs. Methods: Nine potential DNA methylation markers were identified and evaluated in tissue (n=60) and plasma (n=155) cohorts to select the most suitable markers. A training cohort (n=244) and a validation cohort (n=199), including GICs patients, benign tumors, gastrointestinal polyps, and controls, were enrolled to develop and validate a DNA methylation panel. An independent prospective cohort (n=158) was used to validate the panel's performance and compare it with blood protein tumor markers. Results: Six out of nine candidate methylation markers with excellent discrimination abilities in both tissue and plasma cohorts were selected for the DNA methylation panel. The panel demonstrated high AUC values of 0.937 (EC), 0.968 (GC), and 0.987 (CRC) in training cohort, and achieved AUC values of 0.921 (EC), 0.921 (GC), and 0.959 (CRC) in validation cohort. Notably, it achieved impressive AUC values of 0.971 and 0.843 for identifying stage I GICs in the training and validation cohorts, respectively. In the prospective cohort, the six-marker panel showed comparable AUC values to CEA, AFP, and CA19-9 (0.935, 0.769, 0.663, and 0.668, respectively). Conclusion: This study successfully developed and validated a novel, robust, sensitive, and specific plasma-based DNA methylation panel, offering a promising strategy for the early detection of GICs.