Abstract
<title>Abstract</title> <p> <bold>Background</bold> Circulating cell-free DNA (cfDNA) is a promising non-invasive biomarker for cancer diagnosis, and its value in differentiating benign and malignant pulmonary nodules (PNs) has attracted extensive attention. As a key subtype, long-fragment cfDNA concentration remains lacking in systematic research. <bold>Methods</bold> This study enrolled 154 participants, including healthy controls, patients with benign PNs, indeterminate pulmonary nodules (IPNs), and lung cancer. We detected plasma total and long-fragment cfDNA concentrations via qPCR, calculated cfDNA integrity (cfDI), and analyzed correlations between cfDNA indicators and clinical features. Binary logistic regression was applied to establish a multimodal diagnostic model combined with imaging indicators, and the ROC curve was used to evaluate diagnostic efficacy. <bold>Results</bold> Benign PN patients presented significantly higher total cfDNA and long-fragment cfDNA concentration than healthy control group. Age was weakly positively correlated with nodule diameter and total cfDNA concentration. The combined model of long-fragment cfDNA concentration, maximum nodule diameter, and Lung-RADS classification yielded an AUC of 0.911. The optimal cutoff range of long-fragment cfDNA was 1.27–3.79 ng/ml with a specificity of 90.43%. <bold>Conclusions</bold> Long-fragment cfDNA has good differential diagnostic value for PNs, and the multimodal model integrating cfDNA and imaging markers can markedly improve early lung cancer diagnostic accuracy, offering a reliable basis for clinical application. </p>