ObjectiveTo evaluate the clinical value of a combined diagnostic model integrating circulating cell-free DNA (cfDNA) methylation markers and CT imaging features for differentiating benign and malignant lung nodules and for early lung cancer detection. This study pioneers a two-step multi-omics modeling approach to construct a robust diagnostic model. MethodsA retrospective cohort of 140 patients (70 malignant and 70 benign, confirmed by postoperative pathology) with lung nodules who underwent surgical treatment at West China Hospital, Sichuan University, from January 2014 to December 2024 was included. Methylation profiles of 54 cfDNA regions were detected via targeted methylation sequencing. CT imaging features (e.g., nodule size, type, and signs) were extracted. A two-step modeling strategy was applied: ① imaging features were modeled directly using binary logistic regression, while methylation features were selected via LASSO regression before modeling; ② a combined model was constructed using the scores from both models. Model performance was evaluated using receiver operating characteristic (ROC) curves, with internal validation via Bootstrap (1000 iterations). ResultsAll patients were split into a training set (n=84) and a test set (n=56). In the test set, the combined model achieved an area under the ROC curve (AUC) of 0.86 [95% confidence interval (CI): 0.74~0.95], with both sensitivity and specificity reaching 82%. This outperformed the individual imaging model (AUC=0.74) and methylation model (AUC=0.82). ConclusionThe multi-omics combined diagnostic model significantly improved the ability to distinguish benign from malignant lung nodules, particularly for early-stage lesions like ground-glass opacities. Its non-invasive and high-sensitivity features provide a promising translational tool for lung cancer screening, with promising clinical application prospects.