Cancer Biol Med. 2020 Aug 15;17(3):652-663. doi: 10.20892/j.issn.2095-3941.2020.0063.
Objective: MicroRNA (miRNA), a short noncoding RNA, is claimed to be a potential blood-based biomarker. We aimed to identify and evaluate miRNAs as diagnostic biomarkers for non-small cell lung cancer (NSCLC). Methods: Profiles of 745 miRNAs were screened in the serum of 8 patients with NSCLC and 8 age- and sex-matched controls using TaqMan low-density arrays (TLDAs) and validated in 25 patients with NSCLC and 30 with other lung diseases (OLs) as well as in 19 healthy persons (HPs). The diagnostic performance of the candidate miRNAs was assessed in 117 cases of NSCLC and 113 OLs using quantitative real-time polymerase chain reaction (qRT-PCR). Differences in miRNA expression between patients with NSCLC and controls were assessed using the Mann-Whitney U test. The area under receiver operating characteristic (ROC) curve (AUC) was obtained based on the logistic regression model. Results: Ten miRNAs were found to be differentially expressed between patients with NSCLC and controls, including miR-769, miR-339-3p, miR-339-5p, miR-519a, miR-1238, miR-99a#, miR-134, miR-604, miR-539, and miR-342. The expression of miR-339-3p was significantly higher in patients with NSCLC than in those with OLs (P < 0.001) and HPs (P = 0.020). ROC analysis revealed an miR-339-3p expression AUC of 0.616 [95% confidence interval (CI): 0.561-0.702]. The diagnostic prediction was increased (AUC = 0.706, 95% CI: 0.649-0.779) in the model combining miR-339-3p expression and other known risk factors (i.e., age, smoking status, and drinking status). Conclusions: MiR-339-3p was significantly upregulated in patients with NSCLC compared with participants without cancer, suggesting a diagnostic prediction value for high-risk individuals. Therefore, miR-339-3p expression could be a potential blood-based biomarker for NSCLC.