Lung cancer has the highest global incidence and mortality among malignancies. While surgery is the primary curative treatment for early-stage patients, approximately 25-35% are ineligible due to comorbidities. For these patients, Co-ablation system therapy is a key minimally invasive option. However, even after imaging indicates complete tumor removal (tumor-free status), minimal residual disease (MRD) may persist, leading to recurrence. Current mainstream MRD detection relies on identifying mutations in circulating tumor DNA (ctDNA). This approach faces challenges like high mutational heterogeneity and the frequent need for tumor tissue sequencing (Tumor-informed). In contrast, DNA methylation markers offer advantages: they do not require prior knowledge of tumor mutations, appear early in tumorigenesis, are tissue-specific, and allow sensitive detection via multiple consistent CpG sites. Recent studies confirm that ctDNA methylation-based MRD detection can predict recurrence in lung and colorectal cancers earlier than imaging and effectively stratify patient risk. This study aims to investigate the role of SHOX2 and PTGER4 genes in plasma, for monitoring MRD and evaluating therapeutic efficacy in lung cancer patients after Co-ablation system therapy. The study will enroll non-small cell lung cancer (NSCLC) patients undergoing Co-ablation system therapy. Peripheral blood will be collected at multiple timepoints (pre-treatment up to 24 months post-treatment) for ctDNA methylation analysis. The correlation between methylation levels and radiological findings will be assessed. The predictive power for recurrence will be evaluated using ROC curves. Patients will be stratified into high-risk and low-risk groups based on methylation status. Kaplan-Meier survival analysis and Cox regression models will compare recurrence-free survival between groups and evaluate the independent predictive value of SHOX2/PTGER4 methylation for recurrence risk, providing a scientific basis for personalized treatment decisions and recurrence prediction.