Liquid biopsy in pancreatic ductal adenocarcinoma from diagnostic accuracy to clinical utility

Abstract

Pancreatic ductal adenocarcinoma (PDAC) is often diagnosed after the opportunity for curative treatment has passed. Blood-based biomarkers offer opportunities for earlier detection and serial molecular assessment, but their value depends on the intended clinical use. This nonsystematic narrative review critically examines circulating tumor DNA (ctDNA), circulating tumor cells, extracellular vesicles, circulating RNAs, proteins and multimodal panels, including models developed using artificial intelligence. We distinguish population screening, high-risk surveillance, diagnostic evaluation, molecular profiling, treatment monitoring and postoperative recurrence assessment. High discrimination in selected case-control cohorts does not establish performance in asymptomatic populations or clinical benefit. Differences in disease stage, comparator groups, sample handling and validation design preclude reliable ranking of biomarker classes using published area-under-the-curve values. ctDNA has a more established role in selected molecular-profiling situations than in early detection; postoperative detection and longitudinal changes provide prognostic information, but do not by themselves justify treatment modification. Multimodal assays require comparison with clinical assessment and CA19–9 at prespecified thresholds, external validation, calibration and evaluation of downstream harms and costs. This review presents an intended-use framework and a summary of representative prospective studies. Limitations include purposive, non-exhaustive evidence selection and the absence of formal risk-of-bias or certainty grading. Clinical translation requires reproducible assays and evidence that biomarker-informed decisions improve outcomes. Liquid biopsy should therefore be evaluated within defined clinical pathways alongside imaging, pathology and established molecular testing.