Abstract 1032: Development of a plasma circRNA signature for the discrimination of malign lung nodules using the nCounter platform

Abstract Introduction: The evaluation of suspicious lung nodules detected by imaging techniques is currently being performed by invasive methods, such as surgical biopsy or fine-needle aspiration (FNA). Liquid biopsies could offer a minimally invasive alternative in this setting. Some circRNAs have recently been investigated as biomarkers for the early detection of lung cancer and other solid tumors. In this project, we are evaluating the levels of whole plasma circRNAs using the nCounter technology in order to develop a prognostic signature that can aid the clinician to differentiate benign vs. malign nodules. Methods: Plasma samples from patients with malign lung nodules (n=135) and controls (n=149), including patients with benign nodules (n=66/149) have been collected. RNA was purified using the QIAsymphony DSP Virus/Pathogen Midi Kit in a QIAsymphony robot (Qiagen). The nCounter Low RNA Input Amplification Kit (NanoString) was used to retrotranscribe and pre-amplify 4 μL of plasma RNA. Overnight hybridization and posterior nCounter processing (NanoString) steps were carried out following the manufacturer’s instructions. Differential expression and machine learning (ML) methods were performed for the development of a lung cancer signature. Results: Preliminary analysis of the 88 samples first included in the study shown a cluster of 6 circRNAs differentially expressed in lung cancer. Among these circRNAs, circSMAD2, circCOL11A1, circCHST15, and circFUT8 were upregulated, while circACP3 and circLYPLAL1 were found downregulated in plasma of lung cancer patients. In addition, a 64-circRNA signature selected by ML was able to differentiate NSCLC patients from controls, including those with benign nodules, The algorithm assigned signature scores to samples, which ranged from 0 to 1. Using a cut-off value of 0.5, preliminary signature was able to classify samples with an AUC ROC of 0.90 and accuracy of 78.41% (CI = 68.35% - 86.47%) with the final model. Conclusion: Using ML on nCounter analysis of 88 samples, we have generated a preliminary 64-circRNA signature in plasma predictive of early-stage NSCLC. Final results in the entire cohort of 294 samples will be presented at the meeting. Citation Format: Carlos Pedraz Valdunciel, Giovanna Casagrande, Elizabeth Martínez-Pérez, Ana Gimenez-Capitán, Joselyn Valarezo, Pablo Rubinstein, Andrés Aguilar-Hernández, Leticia Ferro-Leal, Cristina Marino-Buslje, Rui Reis, Rafael Rosell, Miguel Ángel Molina-Vila. Development of a plasma circRNA signature for the discrimination of malign lung nodules using the nCounter platform [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 1032.