Recent Advances in Endoscopic Ultrasound (EUS) for Pancreatic Cystic Lesions

Abstract Pancreatic cystic lesions (PCLs), increasingly being detected via advanced imaging, pose diagnostic and management challenges due to their varying malignant potential. This review explores recent advances in endoscopic ultrasound (EUS) for PCL evaluation. PCLs are classified as neoplastic (e.g., intraductal papillary mucinous neoplasms [IPMNs], mucinous cystic neoplasms [MCNs]) or nonneoplastic (e.g., serous cystic neoplasms), with IPMNs and MCNs carrying higher malignancy risks (16–60% and 10–17%, respectively). Conventional EUS offers high-resolution imaging, outperforming computed tomography/magnetic resonance imaging in detecting high-risk features, though operator dependency limits reproducibility. Contrast-enhanced EUS enhances vascularity assessment, achieving 97% sensitivity for identifying high-grade dysplasia. Detective flow imaging EUS detects microvasculature without contrast, showing promise but requiring further validation. EUS-guided needle-based confocal laser endomicroscopy provides real-time histopathology, with 98% sensitivity for mucinous PCLs. EUS-guided sulfur hexafluoride pancreatography differentiates IPMNs with 96.6% accuracy. EUS-guided fine-needle aspiration and biopsy (FNB) improve diagnostic yield, with FNB offering 87% accuracy. Through-the-needle biopsy achieves 80 to 90% sensitivity for mucinous cysts, enhanced by molecular analysis (e.g., KRAS mutations). Artificial intelligence (AI) boosts EUS accuracy to 98.5% for cyst differentiation, reducing operator variability. Therapeutically, EUS-guided chemoablation and radiofrequency ablation offer minimally invasive options, with alcohol-free protocols improving safety (67% resolution). Challenges include complication risks and nonstandardized protocols and surveillance. Future directions involve AI integration, multiomics, and standardized protocols to optimize personalized PCL management, minimizing overtreatment while prioritizing high-risk lesions.