Abstract 1891: Identification of highly accurate early-stage specific NSCLC biomarkers from extracellular vesicles identifies immune response and metabolic biomarkers associated with cancer

Abstract Lung cancer is one of the leading causes of cancer-related deaths worldwide, with non-small cell lung cancer (NSCLC) accounting for approximately 85% of cases. Early-stage lung cancer is often asymptomatic, leading to delayed diagnosis. Challenges in early detection underscore the need for more reliable, non-invasive diagnostic biomarkers. Plasma extracellular vesicles (EVs) represent a promising target for tumor detection and monitoring as they provide an abundant, stable, an easily accessible source of biomarkers. The purpose of this study is to utilize EVs to identify novel cancer related biomarkers and translate these findings into a clinical/immunoassay. To identify novel biomarkers, we enriched EVs using size exclusion chromatography and a proprietary buffer system. TrueDiscovery™ Data-independent acquisition (DIA) mass spectrometry (MS) analysis was conducted on EVs from 74 NSCLC patients (IA/B (32), IIA/B (8); III (17); IV (17)) and 50 normal plasma samples. An in-house machine learning (ML) pipeline was developed to identify differentially expressed proteins and model candidate multiplexes to identify those with extremely high diagnostic accuracy (> 0.98). Additionally, candidate multiplex targets were tested by ELISA to assess mass spectrometry translation to a clinical platform. Comparing tumor vs normal samples (avg=2735 protein IDs) using our ML pipeline we identified 595 Pan Lung cancer differentially expressed proteins (absolute Log2FC>0.5; FDR adjusted p-value <0.01; AUC >0.5). When individual stages were compared to control plasma, a large number of differentially expressed proteins were also seen: Stage I – 549, Stage II – 120, Stage III – 355, and Stage IV – 373 proteins (absolute Log2FC>0.5; FDR adjusted p-value <0.01; AUC >0.5). Significant proteomic differences were also seen between early-stage and late-stage NSCLC, suggesting Stage specific biomarkers may be necessary for maximal diagnostic accuracy. Using the best individual biomarkers, we used Support Vector Machine to test all possible 3-Plexes and identify 3-Plexes with nearly perfect accuracy (>0.99) for each NSCLC Stage: Stage I (140 plexes), Stage II (92), Stage III (67), and Stage IV (50). Early-stage 3-Plexes consist largely of immune, inflammatory, and metabolic pathway proteins. Differential expression of key biomarkers was confirmed using ELISAs, and the performance of multiple highly accurate 3-Plexes is currently being assessed by ELISA on independent intended use case/control cohorts, for potential clinical development. The potential of these highly accurate 3-Plexes to detect even the smallest Stage Ia1 tumors suggests that multiplexed proteomics of EV-associated biomarkers may be an extremely sensitive and specific platform for detection of very small early-stage and recurrent tumors. Citation Format: Zack Opheim, Cheryl Bandoski, Poorva Mudgal, Valesca Anschau, Issa Isaac, Martin Mehnert, Alan Ezrin, Todd A. Hembrough. Identification of highly accurate early-stage specific NSCLC biomarkers from extracellular vesicles identifies immune response and metabolic biomarkers associated with 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 1891.