Abstract
OBJECTIVE: To investigate the value of circulating tumor DNA (ctDNA) methylation for early detection and prognostic monitoring of lung cancer. METHODS: A retrospective analysis was conducted on the DNA methylation test results of 150 lung cancer patients and 100 patients with benign lung lesions enrolled from January 2021 to December 2023. Quantitative methylation fluorescence analysis and bisulfite sequencing were used to measure ctDNA gene methylation, with ROC curves assessing early-stage diagnostic value. Lung cancer patients were followed for 1 year, then grouped by survival status. Cox regression identified poor prognosis factors, and ROC curves evaluated ctDNA methylation's prognostic value. An external cohort of 80 lung cancer patients from May 2022 to December 2023 validated the model using ROC and calibration curves. RESULTS: The positivity rates of SHOX2, CDO1, and SOX17 were significantly higher in lung cancer patients than those in controls (all P<0.05). The combined diagnostic model of these genes had a higher AUC than single-marker tests (Z = 2.363, 2.157, 2.061, all P<0.05). Compared with the favorable prognosis group, the unfavorable prognosis group had a higher proportion of stage III-IV tumors and higher positivity rates of SHOX2, CDO1, SOX17, and HOXA7 (all P<0.05). SHOX2, CDO1 and SOX17 were identified as independent poor-prognosis risk factors (all P<0.05), and their combined prognostic assessment outperformed single-marker tests (Z = 3.316, 2.394, 2.696, all P<0.05). Kaplan-Meier analysis showed that patients with negative methylation of SHOX2, CDO1, and SOX17 had longer survival (Log-rank χ (2) = 6.273, 4.524, 4.364, P<0.001). The model showed good predictive performance (AUC = 0.773), and external validation confirmed its accuracy (AUC = 0.682). CONCLUSIONS: Abnormal methylation of SHOX2, CDO1, and SOX17 is prevalent in lung cancer, potentially serving as biomarkers for early diagnosis and prognostic monitoring.