An Exosome-based and Machine-learning-powered Liquid Biopsy for Pancreatic Cancer Early-detection and Disease Monitoring

Pancreatic ductal adenocarcinoma (PDAC) is a highly lethal malignancy characterized by an asymptomatic early phase, late diagnosis, and poor survival, particularly in individuals who develop disease outside the context of early-stage detection. Early detection strategies are currently limited to imaging-based surveillance (MRI and endoscopic ultrasound) in selected high-risk populations-including individuals with hereditary or familial pancreatic cancer and those with mucinous pancreatic cystic lesions-but these approaches are invasive, costly, and suboptimal in sensitivity.

The aim of this study is to evaluate circulating cell-free and exosome-bound microRNAs as non-invasive biomarkers of PDAC risk and disease biology, with the hypothesis that combined microRNA profiling can improve molecular risk stratification and potentially anticipate disease progression compared with standard imaging-based surveillance alone. The study population will include adult men and women (≥18 years) at increased risk for PDAC due to familial or hereditary predisposition or mucinous pancreatic cysts, with generally stable health status and existing clinical follow-up; no vulnerable populations are specifically targeted.

No medicinal products or medical devices are administered in this study. The procedure under investigation consists exclusively of laboratory-based measurement of microRNA expression from previously collected plasma samples, using validated exosome isolation and quantification techniques, with no impact on clinical management. Available preliminary and published data demonstrate that combined cell-free and exosomal microRNA signatures can detect early-stage PDAC with high accuracy and show dynamic behavior during disease progression and treatment, supporting their relevance for risk stratification and disease monitoring. Clinical, demographic, and laboratory data will be retrieved.