Predictive circulating small extracellular vesicle miRNAs for immunochemotherapy response in ES-SCLC.

Abstract

e20146 Background: Extensive-stage small cell lung cancer (ES-SCLC) is a highly aggressive malignancy with limited treatment options. While immunochemotherapy constitutes standard first-line therapy, predictive biomarkers for treatment response remain undefined, impeding personalized therapeutic strategies. Methods: We investigated circulating small extracellular vesicle (sEV)-derived microRNAs (miRNAs) as non-invasive predictive biomarkers in treatment-naïve ES-SCLC patients receiving immunochemotherapy. Plasma samples from a training cohort (n = 33) and an independent prospective cohort (n = 5) were collected pre-treatment. sEVs were isolated, characterized, and subjected to small RNA sequencing to quantify miRNA expression. Results: Small RNA sequencing revealed 23 differentially expressed sEV miRNAs between responders (n = 19) and non-responders (n = 14). Machine learning refined these candidates into a predictive model. Recursive feature elimination (RFE) yielded an 11-sEV-miRNA signature. The top-performing model (Extra Trees Gini) incorporated 5 sEV miRNAs and 1 clinical feature, demonstrating high predictive accuracy in the training set (AUC = 0.855, sensitivity = 95%, specificity = 80%). Preliminary validation in the prospective cohort achieved 80% accuracy (4/5 correct classifications). Conclusions: We established a novel sEV-miRNA biomarker panel that robustly predicts immunochemotherapy response in ES-SCLC. High discriminatory performance and initial prospective validation underscore its clinical utility for guiding treatment decisions and optimizing outcomes by avoiding ineffective therapy in non-responders.