Combining Plasma miRNAs and Computed Tomography Features to Differentiate the Nature of Pulmonary Nodules

Abstract Objective The purpose of this study was to evaluate the diagnostic efficiency of combining plasma microRNAs (miRNAs) and computed tomography (CT) features in the diagnosis of pulmonary nodules. Methods Ninety-one pulmonary nodule patients who had undergone surgery were enrolled in our study from July 2016 to March 2018 at the Sun Yat-sen University Cancer Center. A prediction model was established by combining 3 miRNAs (miRNA-146a, -200b, and -7) and CT features to identify the pulmonary nodules of these patients. We evaluated the diagnostic performance of this prediction model for pulmonary nodules using the Receiver Operating Characteristic (ROC) curve. Results The expression levels of miRNA-146a, -200b, and -7 in early-stage non-small cell lung cancer (NSCLC) patients are significantly higher than those in benign nodule patients. We used these three miRNAs and CT features (pleural indentation and speculation) to establish a prediction model for early-stage NSCLC, with a sensitivity and specificity of 92.9%, 83.3% in the training set, respectively. For the validation process, with the sensitivity of 80.6% and the specificity of 79.2%. For ROC curve analyses, area under the curve (AUC) for tumor identification in the training stage and validation stage were 0.929 and 0.864, respectively. Conclusion Plasma miRNA-146a, miRNA-200b, and miRNA-7 may be potential biomarkers for the early diagnosis of lung cancer. Our prediction model can help to identify the nature of pulmonary nodules with a high diagnostic efficiency. Keywords：MicroRNA, Pulmonary nodules, Prediction model, Diagnosis, NSCLC