Abstract 4569: MiRNA signatures in liquid biopsy for early detection of non-small cell lung cancer

Background: Liquid biopsy-based biomarkers present a promising approach for non-invasive risk assessment and early detection of lung cancer. Aim: To identify a miRNA signature in sputum and plasma samples as a non-invasive biomarker for early detection of non-small cell lung cancer (NSCLC). Methods: A case-control study was conducted within the Barretos Cancer Hospital Screening Program, utilizing a low-dose CT mobile unit. Sputum and plasma samples were collected from high-risk controls (NLST and/or PLCO criteria; n=61) and NSCLC patients (n=62), matched for age, sex, and smoking history. MiRNA expression was analyzed using the nCounter HumanV3 miRNA panel (Nanostring™). Counts were normalized by the top stable miRNAs, followed by differential expression analysis (p<0.05). MiRNA selection was conducted using Boruta and LASSO. Dataset was partitioned into training and validation sets (85/15 split). Machine learning models (ML) were conducted (Azure) and performance was assessed using AUC (Table 1). Results: In plasma, 68 miRNAs were identified as differentially expressed, and 8 miRNAs were retained following filtering. Among the all-tested models (Table 1), the best ML (Voting Ensemble) for the 8-miRNA plasma signature showed a high accuracy in both training (80.3%, AUC = 0.931) and validation set (73.3%, AUC = 0.892). In sputum, 46 miRNAs were identified as differentially expressed, and 7 miRNAs were retained. The best ML (SVD,Logistic Regression) for the 7-miRNA sputum signature showed a high accuracy in both training (81.42%, AUC = 0.943) and validation set (78.57%, AUC = 0.916). The two defined signatures exhibited no overlap in miRNAs between fluids. Conclusion: This study identified two distinct miRNA fluid-specific signatures in sputum and plasma as potential non-invasive biomarkers for the early detection of lung cancer, underscoring their potential to advance precision-based screening and improve clinical decision-making in lung nodules management. Citation Format: Alessandro Pascon Filho, Giovana M. Stanfoca Casagrande, Rodrigo Sampaio Chiarantano, Fabiana de Lima Vazquez, Rui Manuel Reis M. Reis, Leticia Ferro Leal. MiRNA signatures in liquid biopsy for early detection of non-small cell lung cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 4569.