A Multivariate Cell-Based Liquid Biopsy for Lung Nodule Risk Stratification: Analytical Validation and Early Clinical Evaluation

The iCAP is a tool for blood-based diagnostics that addresses the low signal-to-noise ratio of blood biomarkers by using cells as biosensors. The assay exposes patient serum to standardized cells in culture and classifies disease by AI analysis of gene expression readouts from the cells. This method simplifies the complexity of blood into a concise readout in a scalable cell-based assay. We developed the LC-iCAP as a rule-out test for nodule management in CT-based lung cancer screening. The assay achieved an AUC of 0.63 (95% CI 0.50-0.75) in blind temporal validation with retrospective samples. When integrated with CT data after validation, it demonstrated 90% sensitivity, 67% specificity and 95% NPV using an estimated 25% prevalence, significantly outperforming the Mayo Clinic model and potentially reducing unnecessary follow-up procedures. Analytical validation demonstrated LC-iCAP reproducibility, identified unwanted variation from long-term sample storage and indicated enrichment of hypoxia signaling in the differential assay readout. The goal of the experiment was to measure the gene expression response of human indicator cells to serum from patients with benign or malignant lung nodules and to find a differential pattern between the two conditions that can be used to predict presence of lung cancer in patients from their serum using machine learning tools. SAMPLES are total RNA extracted from cultured human indicator cell that have been exposed to either patient serum or controls. Control conditions included pooled malignant (pooled serum from multiple patient with malignant lung nodules), pooled benign (pooled serum from multiple patietns with benign lung nodules), DMOG chemical stimulant, or PBS. Global gene expression analysis of the samples was performed by RNAseq.