Predictive risk models for the diagnosis and cancer progression of pancreatic cysts

Abstract

The widespread use of cross-sectional imaging has led to a marked increase in the incidental detection of pancreatic cystic lesions (PCLs). Among these, mucin-producing cystic neoplasms—particularly intraductal papillary mucinous neoplasms (IPMNs) and mucinous cystic neoplasms (MCNs)—are of major clinical relevance due to their malignant potential. This review critically evaluates current predictive models for risk stratification of mucinous PCLs, with emphasis on imaging features, cyst fluid biomarkers, molecular alterations, and emerging artificial intelligence–based approaches. While guideline-based morphological criteria remain central to clinical decision-making, integrated models incorporating radiomics, molecular profiling, and multivariate nomograms may improve discrimination for high-grade dysplasia and invasive carcinoma. However, external prospective validation and integration into routine clinical workflows remain limited, representing key barriers to widespread adoption.