Abstract
Background: Early diagnosis of lung cancer requires lung nodule biopsies, which can lead to severe complications. This study aimed to identify optimized panels of exosomal microRNAs (miRNAs) for non-invasive diagnosis of early-stage non-small cell lung cancer (NSCLC). Materials and Methods: This study comprised four phases: discovery, validation, optimization, and confirmation. In the discovery phase, next-generation sequencing profiled 2656 exosomal miRNAs in serum samples (n = 76) from patients with benign lung nodules and stage-specific NSCLC. The validation phase used qPCR to analyze selected miRNAs in serum samples (n = 75). The optimization phase employed a self-devised diagnostic platform, the "up-down ratio (UDR)," to identify miRNA panels. The confirmation phase involved miRNA-target gene interaction and enrichment analyses. Results: The discovery phase identified 15 candidate miRNAs, of which six were validated by qPCR: miR-1976, miR-150-5p, miR-301b-3p, miR-369-3p, miR-497-5p, and miR-610. UDR platform identified a panel of four miRNAs optimized for early detection of NSCLC with ROC over 0.93. Bioinformatics analysis revealed 20 target genes, with VEGFA, BCL2, and PTEN showing strong interactions with the miRNAs, particularly with miR-150-5p, miR-205-5p, miR-1976, miR-301b-3p, and miR-497-5p. Conclusions: This four-phase study suggests that exosomal miRNA panels have potential diagnostic value for early-stage lung cancer. The UDR platform enabled the selection of a four-miRNA panel (miR-150-5p, miR-301b-3p, miR-369-3p, and miR-497-5p), with bioinformatics analyses providing supportive evidence.